Latest from Tom's Hardware UK in Evga

Firmware

Like other board makers, EVGA's BIOS hasn’t changed much compared to the last one we looked at (Z690 generation). Once the system POSTs, you’re presented with four options (Setup, Default, Gamer Mode, and EVGA OC Robot) to access different functionality. After entering the Setup portion of the BIOS, where you can tweak settings, you see informative system information on top. There are headings for different sections below, with the rest of the screen taken up by options for each.

For the most part, there isn’t much digging around in the sub-menus to find the most frequently accessed items, but CPU and memory overclocking are in different sections. Gamer Mode slightly boosts the CPU, while the EVGA OC Robot finds a faster clock speed by raising clocks and playing with voltage during stress testing. Overall, thighs here are easy to read and navigate, with more options than most will ever use.

Tom's Hardware

Software

On the software side, EVGA’s Eleet X1 is a multi-functional monitoring and tweaking tool. For example, Eleet can overclock the CPU and Memory and monitor the system voltages, temperatures and fan speeds. Additionally, it offers RGB lighting control and several preset lighting modes, plus the ability to adjust each strip attached to the motherboard headers. The latest version of Eleet X1 (1.0.13.0) is easy to read and helpful. The only thing I feel is missing from the software is fan control.

Tom's Hardware

Test System / Comparison Products

We’ve updated our test system to Windows 11 64-bit OS with all updates applied. We kept the same Asus TUF RTX 3070 video card from our previous testing platforms but updated the driver. We also updated to F1 22 for our games and kept Far Cry 6. We use the latest non-beta motherboard BIOS available to the public unless otherwise noted. The hardware we used is as follows:

Test System Components
CPU	Intel Core i9-13900K
Memory	Kingston Fury Beast DDR5-6000 CL36 (KF560C36BBEAK2-32)
	GSkill Trident Z DDR5-5600 CL36 (F5-5600U3636C16GX2-TZ5RK)
GPU	Asus TUF RTX 3070
Cooling	Coolermaster MasterLiquid PL360 Flux
PSU	EVGA Supernova 850W P6
Software	Windows 11 64-bit (22H2)
Graphics Driver	Nvidia GeForce Driver 522.25
Sound	Integrated HD audio
Network	Integrated Networking (GbE or 2.5 GbE)

(Image credit: Tom's Hardware)

EVGA supplied our Supernova 850W P6 power supply (appropriately sized and more efficient than the 1.2KW monster we used previously) for our test systems, and G.Skill sent us a DDR5-5600 (F5-5600U3636C16GX2-TZ5RK) memory kit for testing.

EVGA

G.Skill

Benchmark Settings

Synthetic Benchmarks and Settings
Procyon	Version 2.1.459 64
	Office Suite (Office 365), Video Editing (Premiere Pro 22.6.2.2), Photo Editing (Photoshop 23.5.1, Lightroom Classic 11.5)
3DMark	Version 2.22.7359 64
	Firestrike Extreme and Time Spy Default Presets
Cinebench R23	Version RBBENCHMARK330542
	Open GL Benchmark - Single and Multi-threaded
Blender	Version 3.3.0
	Full benchmark (all 3 tests)
Application Tests and Settings
LAME MP3	Version SSE2_2019
	Mixed 271MB WAV to mp3: Command: -b 160 --nores (160Kb/s)
HandBrake CLI	Version: 1.2.2
	Sintel Open Movie Project: 4.19GB 4K mkv to x264 (light AVX) and x265 (heavy AVX)
Corona 1.4	Version 1.4
	Custom benchmark
7-Zip	Version 21.03-beta
	Integrated benchmark (Command Line)
Game Tests and Settings
Far Cry 6	Ultra Preset - 1920 x 1080, HD Textures ON
F1 2022	Ultra Preset - 1920 x 1080, Ultra High (default) Bahrain (Clear/Dry), FPS Counter ON

MORE: Best Motherboards

Benchmark Results

Our standard benchmarks and power tests are performed using the CPU’s stock frequencies (including any default boost/turbo), with all power-saving features enabled. We set optimized defaults in the BIOS and the memory by enabling the XMP profile. For this baseline testing, the Windows power scheme is set to Balanced (default) so the PC idles appropriately.

Synthetic Benchmarks

Synthetics provide a great way to determine how a board runs, as identical settings should produce similar performance results. Turbo boost wattage and advanced memory timings are places where motherboard makers can still optimize for either stability or performance, though, and those settings can impact some testing.

Tom's Hardware

Across our synthetic benchmarks, the Z790 Dark performed well overall. It was a bit slower than average in Cinebench and POV–Ray multi-threaded, but it did well in the Procyon tests (Office and Video/Photo editing) and exceedingly well in RAM bandwidth. Like with any modern board and CPU, if you keep temperatures in order/not throttling, you’ll get better performance.

Timed Applications

Tom's Hardware

In the timed applications, the Z790 Dark was above average in Corona, average in LAME, and both Handbrake tests. There’s nothing to worry about in these applications!

3D Games and 3DMark

Tom's Hardware

Starting with the launch of AMD’s Ryzen 7000 platform, we’ve updated one of our games, F1 21 to F1 22, while keeping Far Cry 6. We run the games at 1920x1080 resolution using the Ultra preset (details listed above). As the resolution goes up, the CPU tends to have less impact. The goal with these settings is to determine if there are differences in performance at the most commonly used (and CPU/system bound) resolution with settings most people use or strive for (Ultra). We expect the difference between boards in these tests to be minor, with most falling within the margin of error differences. We’ve also added a minimum FPS value, which can affect your gameplay and immersion experience.

To be succinct, the Z790 Dark is a competent gaming board. The 3DMark results were around the average, while the gaming tests showed our board up toward the top on average frames per second and average on minimums. You’d be hard-pressed to notice a difference without looking at a frame counter.

Overclocking

When overclocking, we aim to increase the power and add stress to the VRMs. We do so by increasing the clock speed and voltage until we’re at the thermal threshold for the processor during stress testing. However, where these CPUs are allowed to run with ‘boost’ speeds, those speeds are closer to the limit than ever before.

With our i9-13900K, the approach was different. We had to lower the voltage from what we recorded during stress tests to overclock our chip. We increased the clock speeds of the “P” and “E” cores by 100 MHz and 200 MHz, respectively, over the turbo boost and limited by our cooling. We ended up with 5.6 GHz P core and 4.5 GHz E core clocks using about 1.34V (from DMM). Temperatures peak just under the throttling point with this configuration. And as we’ve noted elsewhere, this is the intended performance for Intel’s flagship CPU.

(Image credit: Tom's Hardware)

Overclocking on the Z790 Dark on ambient cooling was as easy as you’d expect. Like with past motherboards, we set 1.35V in the BIOS, and on the Dark, load voltage in Windows is close enough to what was set in the BIOS not to enable LLC (for our meager overclock). As expected, the VRMs didn’t flinch (remember we tend to use a bit LESS power overclocking to our clock speeds), and the board ran stably throughout testing.

Using the latest BIOS (v1.13 at this time), it was smooth sailing on the memory side, too—all three memory kits, up to DDR5-7200, worked by simply enabling the XMP profile. As always, your mileage may vary, as reaching these speeds depends on the memory kit, BIOS, and processor. Still, most combos should be able to reach DDR5-7200 speeds on the Intel platform.

Power Consumption / VRM Temperatures

(Image credit: Tom's Hardware)

We used AIDA64’s System Stability Test with Stress CPU, FPU, Cache and Memory enabled for power testing, using the peak power consumption value. The wattage reading is from the wall via a Kill-A-Watt meter to capture the entire PC (minus the monitor). The only variable that changes is the motherboard; all other parts remain the same. Please note we moved to use only the stock power use/VRM temperature charts, as this section aims to ensure the power delivery can handle the chip even when overclocked. Since we’re using less power to get more clocks, those datasets are more novel than useful. We’re also temperature limited on the processor, so adding more power isn’t possible without increased throttling.

Idle power consumption on the premium Z790 Dark sat around 73W and peaked at 402W under load. The idle and load values more or less match up with other high-end boards. There’s nothing out of line on power use, but the Dark does use more juice than the average (as is expected with a high-end motherboard).

Tom's Hardware

VRM temperatures on our board peaked at just under 49 degrees Celsius, a great result for a 400W load (at the wall). The temperatures are well within the specifications of the high-end 105A SPS MOSFETS, so there’s nothing to worry about here, no matter what kind of overclocking you’re doing.

Bottom Line

For those looking to get the most out of their processor and memory through overclocking, the Z790 Dark Kingpin is one of the best options available, especially if your overclocking adventures involve sub-ambient cooling methods. That said, it’s more than competent as a daily driver too. Hardware-wise, the only thing missing is a PCIe 5.0-capable M.2 socket. The board has overbuilt power delivery, three M.2 sockets, eight SATA ports, ample USB count on the rear IO (including a 20 Gbps Type-C port), dual NICs (2.5 and 10 GbE), and a premium (though last-gen) audio solution. Performance across our testing places this board as average to above average, depending on what test you’re looking at. It gamed well, showed promise in the Procyon test suite, and had the most memory bandwidth among its peers.

Regarding competition, you can group this with overclocking boards like the Gigabyte Z790 Tachyon ($599.99) and the Asus ROG Strix Z790 Apex ($699.99). You can also put it up against the flagships, and outside of a 40 Gbps port and PCIe 5.0 M.2 socket(s), it compares well specs-wise and is cheaper than some of those. Compared to its kind, the Z790 Dark holds up well on specs but does cost more at $799.99 (assuming you can find the Tachyon or Apex in stock). All of these boards will hold up to sub-ambient/extreme overclocking as they are built for it. Of these, only the Apex has a PCIe 5.0 M.2 socket. Otherwise, it’s PCIe 4.0 x4 sockets (of which the Apex has the most at five). Outside of that, the specs are similar enough among them. It’s going to come down to price and appearance.

Ultimately, we love what the EVGA Z790 Dark Kingpin has to offer. While the price is higher than its peers, it’s currently the only one you can find in stock anywhere close to MSRP. Availability aside, it has a unique appearance with the rotated socket area and black-on-gold theme that stands out inside your chassis. The horizontal ports and sockets around the edges, along with the notch, are there to help clean up the wires, something the Asus doesn’t have. In all, the Z790 Dark is a well-rounded but expensive option in the premium mid-range/high-end and overclocking spaces. It’s not a great value among Z790 motherboards, but it is one of the more unique, capable, and overbuilt of the bunch.

MORE: Best Motherboards

How to Overclock Your Graphics Card

Jarred Walton — Tue, 25 Apr 2023 16:12:43 +0000

Overclocking your graphics card can improve its performance by 5% to as much as 15% (or more), depending on the particular card model. Of course, as with any overclocking, care should be taken to avoid running your GPU at unsafe settings. But if you have one of the best graphics cards and you're looking to eke out a bit more performance, we'll discuss how we go about overclocking and determining "safe" settings.

Before overclocking, it's useful to get a baseline measurement of how your graphics card performs. You want to be able to see how much faster your PC runs after tuning, after all — if performance doesn't improve, you'll want to know. Check our guide on how to test graphics card performance, and you'll also want to check your graphics card temperatures and clock speeds.

Once you have your baseline stock performance data, it's time to start overclocking. There are a variety of utilities, but we're going to focus on MSI Afterburner, which is one of the most popular and commonly used GPU overclocking tools. The same basic process can be used with other utilities, including EVGA Precision X1, Asus GPU Tweak, and other graphics card vendor utilities. It's also possible to overclock AMD and Nvidia GPUs using the built-in utility in AMD's drivers or Nvidia's GeForce Experience, which we'll cover below, but we'll start with the universal approach to overclocking your GPU.

Please note that overclocking very much depends on the "silicon lottery." Just because one person with a "HappyGPUs RTX 4090 X-Factor-Wow" model card (that's a made up name, if you couldn't tell) managed a 20% core overclock and a 25% memory overclock doesn't mean that every RTX 4090 card will hit those same clocks, or even that the same model will reach similar clocks. Case in point: Even though all the RTX 4090 cards we've seen so far appear to use Micron 21Gbps GDDR6X memory ("D8BZC" labeled chips), we've seen maximum stable memory overclocks range from 22.6Gbps to as high as 25Gbps.

Overclocking With MSI Afterburner

(Image credit: MSI)

Here's the main interface for MSI Afterburner, using the v3 skin. It shows sliders for core voltage, power limit, core clock, memory clock, and fan speed. By default, Afterburner locks out voltage adjustments. If you want to play with that option, go into the settings and check the box for "Unlock voltage control." You can try setting it to reference design, standard MSI, extended MSI, or third party — some cards will only allow voltage adjustments with the correct mode, and some might not allow it at all.

One option for overclocking is to use the automatic OC scanner tool. Press Ctrl+F, then click the OC Scanner button in the top-right corner and let it do its thing. This is supposed to make overclocking easier, but as with most auto-tuning utilities, your mileage may vary. We've had it crash, we've had it generate unstable overclocks, and sometimes it doesn't even match a quick and dirty +150 MHz (or whatever) manual overclock. It doesn't do anything for memory clocks either, so you'll likely end up going the manual route. The OC scanner can take 20 minutes to run, and most of the time, you can get a quick and easy overclock much faster on your own.

My approach is to fire up a graphically intense game or benchmark that will run in a window, then I run the game at 1920x1080 or 2560x1440 (depending on the GPU I'm using) and get to a spot where nothing is going to kill me. From there, I switch over to Afterburner and start trying some adjustments. Note that you'll want a game or test that doesn't stop rendering when you leave focus (so if it shows the menu when you Alt+Tab, that's not going to work).

There are four primary points of interest when overclocking a graphics card: GPU core clock, memory clock, GPU voltage, and fan speed. (In some cases, you can also adjust memory voltage.) I start by finding the maximum "stable" core clock, then I find the maximum "stable" memory clock, and then I try to find a blend of the two that results in optimal performance.

Note that "stable" is in quotation marks because even if the initial testing seems to work fine, there will often be exceptions where some games or applications crash while others work. When that happens, you'll have to go back to tune the settings and see if you can find some that allow the game to run properly. As you can probably guess, the more demanding the game (i.e., if it uses ray tracing or other advanced features), the more likely it is to require additional fine tuning and lower clocks.

(Image credit: CD Projekt Red)

For this tutorial, we'll use Cyberpunk 2077, running in a window at 2560x1440 with RT-Ultra settings and DLSS Quality mode, to find a reasonable overclock for a GeForce RTX 3090 Founders Edition card. That ticks all the boxes for a demanding game that will likely push the GPU to its limits, and it also has a built-in benchmark that will let us see what happens to performance after tuning.

We started with a GPU core overclock. After loading the game and applying the appropriate settings, we maxed out the power limit and set the fan speed to 80% to ensure the GPU and memory stayed cool. Everything was fine so far, and we applied a 100 MHz offset to the GPU core. That worked okay, which is almost always the case — if you can't get at least a 100 MHz core overclock, you probably shouldn't worry about going any further with your particular graphics card.

With that initial performance boost in play, we then tried a 150 MHz core overclock… and Cyberpunk 2077 almost immediately crashed. We tried maxing out the core voltage after restarting the game, and it still crashed, which means a 150 MHz offset won't work. We split the difference and dropped to 125 MHz, which also crashed. It appeared reasonably stable after maxing out the core voltage at +100 but still crashed after several minutes.

Ultimately, +100 MHz on the GPU cores appears to be as far as we're likely to get with this particular card, though do note that the increased power limit means we're running quite a bit faster than at stock — GPU clocks hovered around 1755 MHz at stock, compared to 1905 MHz with our overclock. A big chunk of that comes thanks to the increased power limit, as even with a 0 MHz GPU overclock, we saw average clocks of 1875 MHz in our testing.

Next, we worked on a memory overclock. We dropped the GPU core clock back to +0 MHz but left the power limit at the maximum (114% on this card). There's usually a fair amount of headroom, so we started with a +750 MHz offset on the GDDR6X memory. That gives an effective memory clock of 21Gbps, and after several minutes of running around in the game, we figured it was working fine. +1000 MHz (21.5Gbps) also appeared stable, as did +1250 MHz (22Gbps), but +1500MHz immediately crashed our PC and forced a system restart. Did we mention overclocking can be a trial and error process?

After additional tinkering and testing, we combined the GPU and VRAM overclocks to end up with +100 MHz on the GPU core and +1000 MHz on the GDDR6X memory. Note that Nvidia's RTX 30-series GPUs have "error detect and retry" (EDR) on the memory, which can sometimes obscure borderline unstable overclocks, and we figured 21.5Gbps was sufficiently fast.

GPU Overclocking Performance

What did the above do for performance? Here's how the stock, GPU core only, VRAM only, and GPU plus VRAM overclocks fared at 4K RT-Ultra with DLSS Quality in the built-in Cyberpunk 2077 benchmark. We also tested a second game, Borderlands 3, at 4K and Badass settings with the same four GPU settings.

Tom's Hardware

We'll forgive you if that feels underwhelming. Ultimately, our maximum GPU overclock improved performance on the RTX 3090 by 8.6% in Borderlands 3 and 8.5% in Cyberpunk 2077. Most of those gains came from the GPU core overclock, which improved performance by 7.3% in Borderlands 3 and 7.5% in Cyberpunk 2077. Meanwhile, just overclocking the VRAM improved performance by 2.1% in Borderlands 3 and 6.0% in Cyberpunk 2077.

Your actual gains can vary a lot, depending on your graphics card. For example, the RTX 3090 isn't known for being memory bandwidth starved in most cases, but the reference model definitely runs into power limits at stock. That's why the core overclock mattered more than the VRAM overclock. Still, the results we've shown above aren't atypical for graphics card overclocking, and we usually see about a 5–10% improvement in performance.

AMD and Nvidia "Built-In" Overclocking

(Image credit: Tom's Hardware)

You can use other utilities to overclock your GPU, but if you don't want to bother with MSI Afterburner, you should probably just use the built-in tools provided by AMD and Nvidia. Nvidia provides an auto-tuning tool and you can specify voltage and power limits, a temperature target, and a fan speed target. After that, you just leave it up to the drivers, which will overclock the GPU cores — but not the VRAM, so you're potentially missing out on a decent chunk of untapped performance.

(Image credit: Tom's Hardware)

AMD's overclocking tools built into Radeon Settings are far more useful. You can do everything you need, specifying voltages, clock speeds, fan speeds, memory speeds, and more. You can also ask the drivers to overclock the VRAM or the GPU — but not both. Alternatively, it can auto-undervolt the GPU, which reduces performance a bit but also tends to lower temperatures and power use. Basically, you don't need a separate GPU overclocking utility, though we still generally find MSI Afterburner to be more familiar and easier to use.

One big benefit of AMD's overclocking tools is that you can specify per-game OC profiles. Maybe one game isn't as intense and you can push clocks higher, or maybe you have a game that just needs the extra oomph to provide a smoother gameplay experience. It can be as complex or as simple as you want to make it, though ultimately, it will often still come down to trial and error.

Bottom Line: GPU Overclocking

(Image credit: Jean Michel "Wizerty" Tisserand)

Should you overclock your graphics card? We could be cheeky and say that if you need to ask, you should probably just leave well enough alone. The reality is that AMD and Nvidia — and their graphics card partners — tend to bin the chips and push performance nearly to the limit. Overclocking also increases the power use and often the temperature of your graphics card, which can cause it to wear out sooner. But perhaps even worse than that is the fact that overclocking usually only improves performance by 5–10%.

That might be noticeable, to some people, without running a bunch of benchmarks. In practice, there's usually not enough headroom that it's worth the hassle. You can give the auto-tuning options in the AMD and Nvidia drivers a shot, and that will probably get you at least half of what you'd get with manual tuning. Plus, it's a simple flip of a switch rather than something that requires a lot of trial and error.

As someone who has done a lot of overclocking over the years, I can say that one of the worst feelings is when you're in the middle of a game and it suddenly crashes due to a formerly "stable" overclock. You're then in the unenviable position of wondering if there's a bug in the game, if your overclock pushed things too far, or if there's something else wrong with your PC. I've lost countless hours troubleshooting my PCs, and these days I just don't find GPU overclocking to be particularly important. Applying upscaling of some form (DLSS, FSR2, XeSS, or even just a spatial upscaling filter) can provide a much larger performance improvement without increasing power use or affecting system stability.

You don't have to agree with me, however; perhaps your particular hardware will deliver a bigger boost in performance after tuning. Just keep an eye on your GPU and VRAM temperatures, and don't forget to give your PC a regular cleaning to keep it running smoothly. Also, don't go into GPU overclocking expecting a 20% improvement or more in performance, because those days are far in the past — unless you use liquid nitrogen.

MORE: Best Graphics Cards
MORE: GPU Benchmarks and Hierarchy
MORE: All Graphics Content

Homebrew Intel Arc OC Tool Released by Legendary Overclocker

Zhiye Liu — Tue, 10 Jan 2023 20:48:35 +0000

Overclocking legend Peter "Shamino" Tan, who currently works for Asus, has released Arc OC Tool, a homemade program to overclock Intel's Arc Alchemist graphics cards. Unfortunately, neither Intel nor Asus endorses this utility, so it's provided as is, and you should only use it at your own risk.

Among the best graphics cards, Intel Arc doesn't play nice with popular third-party overclocking software, such as MSI Afterburner, Asus GPU Tweak III, or EVGA Precision X1. Being a tenderfoot to the graphics card game, it's puzzling that Intel didn't make its graphics cards work with external overclocking applications or at least reach out to the respective parties to usher in support. Until now, Arc owners have been stuck with Intel's Arc Control software to overclock their graphics cards. Arc Control isn't a bad tool, but it certainly has limitations.

Tan built the Arc OC Tool around the Intel Graphics Control Library (IGCL), so the software comes with all the overclocking functions available as of January 9. Although he works for Asus, the Arc OC Tool works with any Arc graphics card and motherboard regardless of the vendor. The program is available for download from overclocking expert Pieter-Jan Plaisier's SkatterBencher website.

(Image credit: Tom's Hardware using various apps and utilities)

Arc OC Tool supports offset and static (locked) overclocking. The former taps into the graphics card’s voltage-frequency curve, whereas the latter enables consumers to dial in a specific voltage and frequency. For now, Intel’s own Arc Control software only allows offset overclocking, so the static feature is exclusive to Tan’s homebrew software.

Or that’s the theory. In limited testing with the Arc A770 Limited Edition, while running FurMark (above), we attempted to force a 2300 MHz clock speed. It didn’t take, and neither did 2,350 MHz. The card happily continued running at 2,200 MHz for the most part, with occasional blips to 2,250 or even 2,300 MHz. But it was doing that before we tried the Arc OC Tool.

Also, it’s worth noting the warning listed on the website:

“IMPORTANT! Be careful when setting the voltage! On my A380 I have to set 1.00000 for the default voltage, then 0.99999 for 10mV less and 1.00001 for 10mV more. If you’re not careful, it’s possible to set >2V as I demonstrated before.”

That... sounds bad. The software isn’t using the right values if a 0.01mV change represents a 10mV change. Given that warning, we didn’t want to poke around at voltage changes much, though we successfully locked up our PC by trying a 200 MHz and 0.01V voltage offset. Proceed with caution, in other words.

Unlike Arc Control, the Arc OC Tool doesn’t come with all the eye candy or the fancy sliders. Instead, Arc OC Tool aims to provide a simple user interface with the essential options for overclocking. So take the Arc OC Tool for a spin if you’re looking for a no-nonsense Arc overclocking application and don’t fancy having more bloatware on your system — not that Intel’s drivers won’t install Arc Control, regardless.

FSP Hydro G Pro 1000W ATX v3.0 Power Supply Review

Aris Mpitziopoulos — Tue, 13 Dec 2022 16:31:05 +0000

The FSP Hydro G Pro 1000W is among the few Gold PSUs in this capacity category featuring ATX v3.0 and PCIe 5.0 compatibility. Its performance is decent, but it isn't high enough to allow it to earn a place in our best PSUs article. Competitors like the Corsair RM1000x and the EVGA 1000 G7 achieve notably higher overall performance scores.

The Hydro G Pro with 1000W max power belongs to the new generation of PSUs, featuring ATX v3.0 and PCIe 5.0 compatibility. There aren't many choices with native 12VHPWR connectors in the 1000W Gold category, and FSP was among the first to introduce a related product.

Significant players, including Corsair, EVGA, and Seasonic, haven't released something on the market, yet. Thermaltake, with the Toughpower GF3 1000W, and Silverstone, with the DA1000R Gold, are two other choices in this category, meeting the ATX v3.0 spec for PSUs with 12VHPWR connectors.

Tom's Hardware

With 150mm depth, the Hydro G Pro 1000 has compact dimensions. Nonetheless, FSP has the smallest, in dimensions, 1000W Gold platform, which is used in the EVGA 1000 G7, with only 130mm depth! Compact PSUs are easier to install, especially in smaller chassis, but the over-populated PCBs don't help airflow, hence noise output is usually increased.

Moreover, it is not possible to install large enough cooling fans, with 135mm or 140mm diameter, in small PSUs, and smaller fans have to spin at higher speeds to offer the same airflow, producing more noise. The Hydro G Pro uses a 120mm, high-quality, fan, featuring a fluid dynamic bearing.

Tom's Hardware

Specifications of FSP Hydro G Pro 1000W

Manufacturer (OEM)	FSP
Max. DC Output	1000W
Efficiency	80 PLUS Gold, Cybenetics Gold (87-89%)
Noise	Cybenetics A- (25-30 dB[A])
Modular	✓ (fully)
Intel C6/C7 Power State Support	✓
Operating Temperature (Continuous Full Load)	0 - 50°C
Over Voltage Protection	✓
Under Voltage Protection	✓
Over Power Protection	✓
Over Current (+12V) Protection	✓
Over Temperature Protection	✓
Short Circuit Protection	✓
Surge Protection	✓
Inrush Current Protection	✓
Fan Failure Protection	✗
No Load Operation	✓
Cooling	120mm Fluid Dynamic Bearing Fan (MGA12012XF-O25)
Semi-Passive Operation	✓ (selectable)
Dimensions (W x H x D)	150 x 85 x 150mm
Weight	1.72 kg (3.79 lb)
Form Factor	ATX12V v3.0, EPS 2.92
Alternative Low Power Mode (ALPM) compatible	✓
Warranty	10 Years

Power Specifications of FSP Hydro G Pro 1000W

Rail		3.3V	5V	12V	5VSB	-12V
Max. Power	Amps	20	20	83.33	2.5	0.3
	Watts		120	1000	12.5	3.6
Total Max. Power (W)			1000

Cables & Connectors of FSP Hydro G Pro 1000W

Description	Cable Count	Connector Count (Total)	Gauge	In Cable Capacitors
ATX connector 20+4 pin (600mm)	1	1	18-22AWG	No
4+4 pin EPS12V (700mm)	2	2	18AWG	No
6+2 pin PCIe (650mm+150mm)	2	4	18AWG	No
6+2 pin PCIe (500mm+150mm)	1	2	18AWG	No
12+4 pin PCIe 600W (700mm)	1	1	16-24AWG	No
SATA (500mm+150mm+150mm+150mm)	2	8	18AWG	No
SATA (500mm+150mm) / 4-pin Molex (+150mm+100mm)	2	4 / 4	18AWG	No
SATA (500mm+150mm) / 4-pin Molex (+150mm) / FDD (+150mm)	1	2 / 1 / 1	18-22AWG	No
AC Power Cord (1350mm) - C13 coupler	1	1	18AWG	-

The cables are long, and the amount of connectors is satisfactory. The single 12VHPWR connector can deliver up to 600W. According to Intel's test plan, a 1000W ATX v3.0 PSU should be equipped with a 450W 12VHPWR connector, but most brands don't pay attention to this and use 600W connectors to make sure that the NVIDIA RTX 4090 graphics cards can go all the way up to 600W. Not all 4090s have such high maximum power limits, though.

Tom's Hardware

Component Analysis of FSP Hydro G Pro 1000W

We strongly encourage you to have a look at our PSUs 101 article, which provides valuable information about PSUs and their operation, allowing you to better understand the components we're about to discuss.

`General Data`	-
Manufacturer (OEM)	FSP
PCB Type	Double Sided
`Primary Side`	-
Transient Filter	4x Y caps, 2x X caps, 2x CM chokes, 1x MOV
Inrush Protection	NTC Thermistor SCK-056 (5 Ohm) & Relay
Bridge Rectifier(s)	2x HY GBJ2506 (600V, 25A @ 100°C)
APFC MOSFETs	2x Infineon IPA60R120P7 (600V, 16A @ 100°C, Rds(on): 0.12Ohm)
APFC Boost Diode	1x CREE C3D08060A (600V, 8A @ 150°C)
Bulk Cap(s)	1x Nippon Chemi-Con (450V, 680uF, 3,000h @ 105°C, KHS)
Main Switchers	2x Magnachip MMFT60R115PC (600V, 20.9A @ 100°C, Rds(on): 0.115Ohm)
APFC Controller	Infineon ICE2PCS02G
Resonant Controller	Champion CM6901T2X
Topology	Primary side: APFC, Half-Bridge & LLC converter Secondary side: Synchronous Rectification & DC-DC converters
`Secondary Side`	-
+12V MOSFETs	6x Infineon BSC014N04LSI (40V, 123A @ 100°C, Rds(on): 1.45mOhm)
5V & 3.3V	DC-DC Converters: 6x NEC 2SK3062-ZJ (60V, 70A, Rds(on): 8.5mOhm) PWM Controller(s): ANPEC APW7159C
Filtering Capacitors	Electrolytic: 4x Nippon Chemi-Con (2-5,000h @ 105°C, KZE), 2x Rubycon (3-6,000h @ 105°C, YXG), 2x Rubycon (2-10,000h @ 105°C, YXF), 1x Rubycon (6-10,000h @ 105°C, ZLH), 1x TK (105°C) Polymer: 29x Nippon Chemi-Con, 1x NIC
Supervisor IC	Weltrend WT7527RA (OCP, OVP, UVP, SCP,PG)
Fan Controller	APW9010
Fan Model	Protechnic Electric MGA12012XF-O25 (120mm, 12V, 0.52A, Fluid Dynamic Bearing)
`5VSB Circuit`	-
Rectifier	1x CET CEF04N7G (700V, 4A, Rds(on): 3.3Ohm) & 1x PFC P15L50SP SBR (50V, 15A)
Standby PWM Controller	97CL2N13

Tom's Hardware

The small PCB is equipped with equally small heat sinks. Except for the transient filter, the other circuits have enough space between electronic parts to allow for decent airflow. It is good to see the filtering caps on the secondary side being on the clear, without anything blocking airflow to them.

Typically, for an FSP platform, we find some potentiometers on a vertical PCB, which might look tempting to adjust, but you should not do it! The build quality is high, and all parts that FSP used are of high quality, too. FSP is among the few OEMs using bulk caps rated for 3,000h @ 105°C. All other brands use 2,000h @ 105°C bulk caps.

Tom's Hardware

The transient/EMI filter...

Tom's Hardware

The pair of bridge rectifiers can handle up to 50A.

Tom's Hardware

The APFC converter has two Infineon FETs and a single CREE boost diode. The PFC controller is an Infineon ICE2PCS02G IC.

Tom's Hardware

The main FETs are by Magnachip and arranged into a half-bridge topology. An LLC resonant converter is also used to boost efficiency. The resonant controller is a Champion CM6901T2X.

Tom's Hardware

Six Infineon FETs regulate the 12V rail. The same rail feeds two DC-DC converters, which generate the minor rails (5V and 3.3V).

Tom's Hardware

The filtering caps are provided by Japanese brands and are of good quality. Besides electrolytic caps

Tom's Hardware

The standby PWM controller is an 97CL2N13 IC.

Tom's Hardware

Many filtering caps are installed on the modular board.

(Image credit: Tom's Hardware)

The main supervisor IC is a Weltrend WT7527RA.

Tom's Hardware

Soldering quality is good.

Tom's Hardware

The cooling fan is a Protechnic Electric MGA12012XF-O25.

MORE: Best Power Supplies

To learn more about our PSU tests and methodology, please check out How We Test Power Supply Units.

Corsair RM1000x (2021)

EVGA 1000 G7

SilverStone SX1000 Platinum

Primary Rails And 5VSB Load Regulation

Tight load regulation also, among other factors, improves the system’s stability, especially under overclocked conditions and, at the same time, it applies less stress to the DC-DC converters that many system components utilize.

Tom's Hardware

Load regulation is not so tight. We want to see it within 1% at 12V and ideally below 0.8%.

Hold-Up Time

Put simply; hold-up time is the amount of time that the system can continue to run without shutting down or rebooting during a power interruption.

Tom's Hardware

The hold-up time is long and the power ok signal is accurate.

Inrush Current

Inrush current, or switch-on surge, refers to the maximum, instantaneous input current drawn by an electrical device when it is first turned on. A large enough inrush current can cause circuit breakers and fuses to trip. It can also damage switches, relays, and bridge rectifiers. As a result, the lower the inrush current of a PSU right as it is turned on, the better.

Tom's Hardware

Inrush current is high with voltage inputs we tried, 115V and 230V. A larger (higher resistance) NTC thermistor would help here.

Leakage Current

In layman's terms, leakage current is the unwanted transfer of energy from one circuit to another. In power supplies, it is the current flowing from the primary side to the ground or the chassis, which in the majority of cases is connected to the ground. For measuring leakage current, we use a GW Instek GPT-9904 electrical safety tester instrument.

The leakage current test is conducted at 110% of the DUT's rated voltage input (so for a 230-240V device, we should conduct the test with 253-264V input). The maximum acceptable limit of a leakage current is 3.5 mA and it is defined by the IEC-60950-1 regulation, ensuring that the current is low and will not harm any person coming in contact with the power supply's chassis.

(Image credit: Tom's Hardware)

Leakage current is low.

10-110% Load Tests

These tests reveal the PSU's load regulation and efficiency levels under high ambient temperatures. They also show how the fan speed profile behaves under increased operating temperatures.

Test	12V	5V	3.3V	5VSB	DC/AC (Watts)	Efficiency	Fan Speed (RPM)	PSU Noise (dB[A])	Temps (In/Out)	PF/AC Volts
10%	6.387A	1.977A	1.967A	0.974A	99.985	86.87%	0	<6.0	45.28°C	0.981
	12.273V	5.058V	3.355V	5.133V	115.1				40.96°C	115.16V
20%	13.786A	2.968A	2.954A	1.172A	199.927	90.053%	0	<6.0	46.2°C	0.996
	12.262V	5.054V	3.351V	5.119V	222.007				41.46°C	115.12V
30%	21.541A	3.465A	3.45A	1.371A	299.968	90.867%	0	<6.0	47.36°C	0.995
	12.252V	5.051V	3.348V	5.106V	330.115				42.01°C	115.1V
40%	29.264A	3.963A	3.946A	1.571A	399.469	90.913%	0	<6.0	48.52°C	0.995
	12.243V	5.047V	3.345V	5.093V	439.392				42.77°C	115.06V
50%	36.677A	4.957A	4.938A	1.772A	499.168	90.573%	1033	24.6	43.18°C	0.995
	12.233V	5.043V	3.341V	5.079V	551.124				49.19°C	115.04V
60%	44.181A	5.958A	5.935A	1.976A	599.697	90.113%	1035	24.6	43.49°C	0.995
	12.220V	5.036V	3.336V	5.062V	665.492				50.23°C	115V
70%	51.626A	6.957A	6.933A	2.179A	699.419	89.454%	1076	25.7	43.78°C	0.994
	12.209V	5.031V	3.332V	5.047V	781.873				51.25°C	114.97V
80%	59.156A	7.959A	7.929A	2.283A	799.425	88.671%	1509	35.8	44.23°C	0.993
	12.197V	5.027V	3.328V	5.036V	901.563				52.25°C	114.93V
90%	67.026A	8.46A	8.419A	2.387A	899.187	87.796%	1882	43.3	45.2°C	0.992
	12.185V	5.023V	3.324V	5.026V	1024.191				54.29°C	114.89V
100%	74.977A	8.965A	8.94A	2.492A	999.906	86.809%	2276	47.5	45.83°C	0.991
	12.174V	5.018V	3.321V	5.016V	1151.846				55.85°C	114.85V
110%	82.555A	9.971A	10.037A	2.495A	1099.721	85.486%	2687	51.0	46.5°C	0.99
	12.161V	5.014V	3.316V	5.009V	1286.444				57.42°C	114.81V
CL1	0.114A	14.318A	14.278A	0A	121.272	82.989%	0	<6.0	48.26°C	0.988
	12.264V	5.042V	3.34V	5.156V	146.129				42.74°C	115.14V
CL2	0.112A	19.811A	0A	0A	101.383	81.743%	939	21.6	43.89°C	0.982
	12.270V	5.048V	3.349V	5.164V	124.028				50.99°C	115.15V
CL3	0.112A	0A	19.707A	0A	67.364	77.267%	0	<6.0	52.64°C	0.971
	12.260V	5.061V	3.348V	5.16V	87.185				44.56°C	115.16V
CL4	82.057A	0A	0A	0A	999.805	87.427%	2019	44.4	45.29°C	0.991
	12.184V	5.031V	3.331V	5.118V	1143.59				55.21°C	114.86V

The PSU delivers full and 110% load without any issues at high temperatures, but the fan's noise goes through the roof.

20-80W Load Tests

In the following tests, we measure the PSU's efficiency at loads significantly lower than 10% of its maximum capacity (the lowest load the 80 PLUS standard measures). This is important for representing when a PC is idle with power-saving features turned on.

Test	12V	5V	3.3V	5VSB	DC/AC (Watts)	Efficiency	Fan Speed (RPM)	PSU Noise (dB[A])	Temps (In/Out)	PF/AC Volts
20W	1.206A	0.494A	0.491A	0.194A	19.99	66.898%	0	<6.0	40.19°C	0.884
	12.312V	5.062V	3.359V	5.166V	29.883				37.07°C	115.18V
40W	2.662A	0.691A	0.688A	0.29A	39.989	77.998%	0	<6.0	40.82°C	0.942
	12.279V	5.061V	3.358V	5.162V	51.269				37.46°C	115.17V
60W	4.118A	0.889A	0.884A	0.388A	59.987	82.921%	0	<6.0	41.82°C	0.962
	12.268V	5.061V	3.357V	5.157V	72.342				38.06°C	115.17V
80W	5.569A	1.087A	1.081A	0.485A	79.935	85.718%	0	<6.0	44.08°C	0.973
	12.266V	5.06V	3.356V	5.153V	93.254				40.1°C	115.16V

The fan doesn't spin at low lower than 80W loads, even at high operating temperatures.

2% or 10W Load Test

From July 2020, the ATX spec requires 70% and higher efficiency with 115V input. The applied load is only 10W for PSUs with 500W and lower capacities, while for stronger units, we dial 2% of their max-rated capacity.

12V	5V	3.3V	5VSB	DC/AC (Watts)	Efficiency	Fan Speed (RPM)	PSU Noise (dB[A])	Temps (In/Out)	PF/AC Volts
1.471A	0.255A	0.255A	0.044A	20.456	67.969%	0	<6.0	27.56°C	0.883
	12.293V	5.056V	3.356V	5.165V	30.098			22.66°C	115.17V

Ideally we want to see above 70% efficiency in this test.

Efficiency & Power Factor

Next, we plotted a chart showing the PSU's efficiency at low loads and loads from 10 to 110% of its maximum rated capacity. The higher a PSU’s efficiency, the less energy goes wasted, leading to a reduced carbon footprint and lower electricity bills. The same goes for Power Factor.

Tom's Hardware

Efficiency is good with normal loads, but pretty low at light loads.

5VSB Efficiency

Test #	5VSB	DC/AC (Watts)	Efficiency	PF/AC Volts
1	0.1A	0.516W	71.484%	0.069
	5.162V	0.722W		115.16V
2	0.25A	1.288W	78.707%	0.147
	5.156V	1.637W		115.16V
3	0.55A	2.829W	80.065%	0.263
	5.145V	3.533W		115.16V
4	1A	5.128W	80.323%	0.352
	5.128V	6.384W		115.17V
5	1.5A	7.666W	80.013%	0.408
	5.111V	9.58W		115.16V
6	2.499A	12.681W	79.346%	0.462
	5.074V	15.982W		115.16V

Tom's Hardware

The 5VSB rail is efficient.

Power Consumption In Idle And Standby

Mode	12V	5V	3.3V	5VSB	Watts	PF/AC Volts
Idle	12.465V	5.052V	3.354V	5.162V	8.343	0.508
						115.17V
Standby					0.071	0.007
						115.17V

Tom's Hardware

Vampire power is low with 115V but much higher than 0.1W with 230V input.

Fan RPM, Delta Temperature, And Output Noise

All results are obtained between an ambient temperature of 37 to 47 degrees Celsius (98.6 to 116.6 degrees Fahrenheit).

(Image credit: Tom's Hardware)

The fan's speed increases linearly to the load and it allows the fan to spin at high speeds, at high operating temperatures, to cope with the thermal loads.

The following results were obtained at 30 to 32 degrees Celsius (86 to 89.6 degrees Fahrenheit) ambient temperature.

(Image credit: Tom's Hardware)

The PSU's passive mode lasts long enough at normal operating temperatures, close to 30 degrees Celsius. Noise exceeds 30 dBA with more than 770W loads and over 35 dBA with 840W and higher loads. In no case the output noise exceeds 40 dBA under normal operating temperatures. Hence the overall noise output remains low.

MORE: Best Power Supplies

Protection Features

Check out our PSUs 101 article to learn more about PSU protection features.

OCP (Cold @ 21°C)	12V: 105.2A (126.25%), 12.141V 5V: 24.2A (121%), 5.033V 3.3V: 25A (125%), 3.335V 5VSB: 3.8A (152%), 5.021V
OCP (Hot @ 41°C)	12V: 104.8A (125.77%), 12.152V 5V: 22.1A (110.5%), 5.044V 3.3V: 22.5A (112.5%), 3.343V 5VSB: 3.9A (156%), 5.021V
OPP (Cold @ 27°C)	1277W (127.7%)
OPP (Hot @ 44°C)	1272.06W (127.21%)
OTP	✓ (97°C @ secondary side)
SCP	12V to Earth: ✓ 5V to Earth: ✓ 3.3V to Earth: ✓ 5VSB to Earth: ✓ -12V to Earth: ✓
PWR_OK	Proper operation
NLO	✓
SIP	Surge: MOV Inrush: NTC Thermistor & Bypass relay

The OCP triggering points are correctly set on all rails, and the same goes for OPP. The remaining protection features are present and work well.

DC Power Sequencing

According to Intel’s most recent Power Supply Design Guide (revision 1.4), the +12V and 5V outputs must be equal to or greater than the 3.3V rail at all times. Unfortunately, Intel doesn't mention why it is so important to always keep the 3.3V rail's voltage lower than the levels of the other two outputs.

Tom's Hardware

No problems here since the 3.3V rail is always lower than the other two.

Cross Load Tests

To generate the following charts, we set our loaders to auto mode through custom-made software before trying more than 25,000 possible load combinations with the +12V, 5V, and 3.3V rails. The deviations in each of the charts below are calculated by taking the nominal values of the rails (12V, 5V, and 3.3V) as point zero. The ambient temperature during testing was between 30 to 32 degrees Celsius (86 to 89.6 degrees Fahrenheit).

Load Regulation Charts

Tom's Hardware

Efficiency Graph

(Image credit: Tom's Hardware)

Ripple Graphs

The lower the power supply's ripple, the more stable the system will be and less stress will also be applied to its components.

Tom's Hardware

Infrared Images

We apply a half-load for 10 minutes with the PSU's top cover and cooling fan removed before taking photos with a modified Fluke Ti480 PRO camera able to deliver an IR resolution of 640x480 (307,200 pixels).

Tom's Hardware

The temperatures inside the PSU are low, with the filtering caps on the secondary side being among the coolest parts, which is highly beneficiary for their health.

MORE: Best Power Supplies

Advanced Transient Response Tests

For details about our transient response testing, please click here.

In the real world, power supplies are always working with loads that change. It's of immense importance, then, for the PSU to keep its rails within the ATX specification's defined ranges. The smaller the deviations, the more stable your PC will be with less stress applied to its components.

We should note that the ATX spec requires capacitive loading during the transient rests, but in our methodology, we also choose to apply a worst case scenario with no additional capacitance on the rails.

Advanced Transient Response at 20% – 20ms

Voltage	Before	After	Change	Pass/Fail
12V	12.245V	12.062V	1.50%	Pass
5V	5.044V	4.900V	2.85%	Pass
3.3V	3.346V	3.218V	3.84%	Pass
5VSB	5.106V	5.069V	0.72%	Pass

Advanced Transient Response at 20% – 10ms

Voltage	Before	After	Change	Pass/Fail
12V	12.248V	12.079V	1.37%	Pass
5V	5.046V	4.903V	2.84%	Pass
3.3V	3.347V	3.219V	3.82%	Pass
5VSB	5.107V	5.071V	0.71%	Pass

Advanced Transient Response at 20% – 1ms

Voltage	Before	After	Change	Pass/Fail
12V	12.253V	12.069V	1.50%	Pass
5V	5.049V	4.907V	2.82%	Pass
3.3V	3.349V	3.215V	4.01%	Pass
5VSB	5.109V	5.062V	0.92%	Pass

Advanced Transient Response at 50% – 20ms

Voltage	Before	After	Change	Pass/Fail
12V	12.212V	12.121V	0.74%	Pass
5V	5.032V	4.890V	2.82%	Pass
3.3V	3.336V	3.198V	4.13%	Pass
5VSB	5.067V	5.020V	0.93%	Pass

Advanced Transient Response at 50% – 10ms

Voltage	Before	After	Change	Pass/Fail
12V	12.215V	12.124V	0.74%	Pass
5V	5.034V	4.890V	2.85%	Pass
3.3V	3.337V	3.205V	3.96%	Pass
5VSB	5.069V	5.033V	0.70%	Pass

Advanced Transient Response at 50% – 1ms

Voltage	Before	After	Change	Pass/Fail
12V	12.218V	12.137V	0.66%	Pass
5V	5.037V	4.896V	2.80%	Pass
3.3V	3.338V	3.201V	4.10%	Pass
5VSB	5.070V	5.011V	1.17%	Pass

Tom's Hardware

Transient response is good enough at 12V but mediocre at 5V and 3.3V. Nevertheless, the 12V rail is the most important.

ATX 3.0 Transient Response Tests

The following table shows the load that we applied.

Duty Cycle	Time for Power Excursion (Te)	Time Constant (Tc)	Power @ Te	Power @ Tc
5%	100μs	1900μs	2000W	917.7W
8%	1ms	11.5ms	1800W	897.3W
12.5%	10ms	70ms	1600W	881.6W
25%	100ms	300ms	1200W	923.8W

Tom's Hardware

The PSU successfully passed all ATX 3.0 transient response tests for units equipped with 12VHPWR connectors.

Turn-On Transient Tests

In the next set of tests, we measure the PSU's response in simpler transient load scenarios—during its power-on phase. Ideally, we don't want to see any voltage overshoots or spikes since those put a lot of stress on the DC-DC converters of installed components.

Tom's Hardware

The spike at 5VSB looks nasty but it is below the 5.5V limit allowed by the ATX spec.

Power Supply Timing Tests

From year 2020, the PSU's Power-on time (T1) has to be lower than 150ms and the PWR_OK delay (T3) from 100 to 150ms, to be compatible with the Alternative Sleep Mode.

PSU Timings Table
T1 (Power-on time) & T3 (PWR_OK delay)
Load	T1	T3
20%	49ms	131ms
100%	52ms	130ms

Tom's Hardware

The PWR_OK delay is within the 100-150ms region, so the PSU supports the alternative sleep mode recommended by the ATX spec.

Ripple Measurements

Ripple represents the AC fluctuations (periodic) and noise (random) found in the PSU's DC rails. This phenomenon significantly decreases the capacitors' lifespan because it causes them to run hotter. A 10-degree Celsius increase can cut into a cap's useful life by 50%. Ripple also plays an important role in overall system stability, especially when overclocking is involved.

The ripple limits, according to the ATX specification, are 120mV (+12V) and 50mV (5V, 3.3V, and 5VSB).

Test	12V	5V	3.3V	5VSB	Pass/Fail
10% Load	16.9 mV	8.4 mV	7.8 mV	9.6 mV	Pass
20% Load	28.0 mV	13.6 mV	15.8 mV	43.1 mV	Pass
30% Load	24.4 mV	12.8 mV	13.1 mV	36.3 mV	Pass
40% Load	14.1 mV	8.3 mV	8.1 mV	9.3 mV	Pass
50% Load	29.4 mV	12.9 mV	17.4 mV	39.0 mV	Pass
60% Load	14.9 mV	9.0 mV	10.6 mV	11.0 mV	Pass
70% Load	32.2 mV	15.0 mV	19.0 mV	38.2 mV	Pass
80% Load	15.9 mV	10.8 mV	14.0 mV	11.7 mV	Pass
90% Load	27.1 mV	16.0 mV	19.9 mV	36.3 mV	Pass
100% Load	23.0 mV	11.9 mV	16.2 mV	14.2 mV	Pass
110% Load	23.8 mV	11.9 mV	17.1 mV	14.5 mV	Pass
Crossload 1	18.8 mV	14.8 mV	14.8 mV	10.5 mV	Pass
Crossload 2	16.5 mV	12.1 mV	10.6 mV	9.8 mV	Pass
Crossload 3	24.4 mV	12.2 mV	15.6 mV	37.0 mV	Pass
Crossload 4	23.1 mV	11.2 mV	15.1 mV	13.9 mV	Pass

Tom's Hardware

Ripple suppression is good on all rails.

Ripple At Full Load

Tom's Hardware

Ripple At 110% Load

Tom's Hardware

Ripple At Cross-Load 1

Tom's Hardware

FSP Hydro G Pro 1000W

Tom's Hardware

Ripple At Cross-Load 4

Tom's Hardware

EMC Pre-Compliance Testing – Average & Quasi-Peak EMI Detector Results

Electromagnetic Compatibility (EMC) is the ability of a device to operate properly in its environment without disrupting the proper operation of other nearby devices.

Electromagnetic Interference (EMI) stands for the electromagnetic energy a device emits, and it can cause problems in other nearby devices if too high. For example, it can cause increased static noise in your headphones or/and speakers.

΅We use TekBox's EMCview to conduct our EMC pre-compliance testing.

(Image credit: Tom's Hardware)

Three spurs go over the limit with the average EMI detector, and two with the peak detector.

MORE: Best Power Supplies

Performance Rating

(Image credit: Tom's Hardware)

Overall performance is decent, but not competitive. The difference with the Corsair RM1000x and the EVGA 1000 G7 is large.

Noise Rating

The graph below depicts the cooling fan's average noise over the PSU's operating range, with an ambient temperature between 30 to 32 degrees Celsius (86 to 89.6 degrees Fahrenheit).

(Image credit: Tom's Hardware)

Under normal operating temperatures, the average noise output is low.

Efficiency Rating

The following graph shows the PSU's average efficiency throughout its operating range with an ambient temperature close to 30 degrees Celsius.

(Image credit: Tom's Hardware)

The average efficiency would be higher if the platform performed better with light loads. Still, the difference with most similar capacity Gold units is not high. The Seasonic and EVGA units, with the latter using an FSP platform, perform like Platinum units and not Gold ones.

Power Factor Rating

The following graphs show the PSU's average power factor reading throughout its operating range with an ambient temperature close to 30 degrees Celsius and 115V/230V voltage input.

Tom's Hardware

The APFC converter performs well.

MORE: Best Power Supplies

The FSP Hydro G Pro 1000W, SilverStone DA1000R, and Thermaltake Toughpower GF3 1000W are among the few Gold units with this capacity, offering ATX v3.0 and PCIe 5.0 compatibility. This will change soon once Corsair, Seasonic, Cooler Master, and other significant brands release their new products.

You might not need an ATX v3.0 PSU yet, for any of the existing GPUs, and the upcoming AMD GPUs won't even require 12VHPWR connectors. Still, for graphics cards needing 12+4 pin connectors, it is highly preferable to avoid using adapters. Adapters increase resistance. Thus they are not suitable for high-power applications.

(Image credit: Tom's Hardware)

The Hydro G Pro 1000 is a decent offering in the corresponding category. Its primary assets are the native 600W 12VHPWR connector, the ATX v3.0 compatibility, and the low noise output under normal operating temperatures. Nevertheless, its overall performance is not high enough to worry the competition. Units like the Corsair RM1000x, and the EVGA 1000 G7, which uses an FSP platform, perform better. The 1000 G7 is noisy, though, because of its small dimensions.

MORE: Best Power Supplies

Disclaimer: Aris Mpitziopoulos is Tom's Hardware's PSU reviewer. He is also the Chief Testing Engineer of Cybenetics and developed the Cybenetics certification methodologies apart from his role on Tom's Hardware. Neither Tom's Hardware nor its parent company, Future PLC, are financially involved with Cybenetics. Aris does not perform the actual certifications for Cybenetics.

EVGA Auctions Unreleased RTX 4090 GPU For Charity, $6,200 And Rising

Zhiye Liu — Sun, 11 Dec 2022 18:57:58 +0000

After failed attempts to auction off the brand’s GeForce RTX 4090 on eBay, EVGA has taken to its forums to carry out the auction. After the company’s abrupt exit from the GPU market, the GeForce RTX 4090 FTW3 prototype, or “Next Gen Graphics Card” as EVGA calls it due to licensing issues, will make a fine collector’s item.

EVGA previously took to eBay to auction off the graphics card. It had fetched up to $13,200 before eBay removed the listing. The company has restarted the auction on the EVGA forums, where the current bid, at the time of writing, is $6,200. However, it still has a long way to go before it matches the last bid on eBay.

All the proceeds from EVGA’s auction will go to St. Jude Children’s Research Hospital. The auction finalizes on December 16th at 11:59 AM PT, so interested parties have more than enough time to submit a bid. However, the brand is auctioning the graphics card as it is and will not provide a warranty or accept returns.

Although it’s a preproduction sample, EVGA’s GeForce RTX 4090 FTW3 is a working graphics card. YouTuber JayzTwoCents has already taken one of the samples for a spin. Since EVGA ended its partnership with Nvidia, the graphics card uses a qualification AD102 die and relies on EVGA’s home-brewed firmware. The prototype features a 2,520 MHz boost clock, the same as the Founders Edition. Therefore, its performance should be in the same alley as the GeForce RTX 4090 Founders Edition.

EVGA was one of the most popular Nvidia partners and was very good at making graphics cards. So it’s a shame that the company departed from the graphics card market. Many enthusiasts would love to see EVGA hook up with another significant chipmaker like AMD or Intel. However, the company has said it is concentrating on its other hardware products, which we suspect potentially offer higher profit margins.

If you’re an EVGA fan, it’s your chance to own what is likely the last graphics card that the company manufactured. It won’t come cheap, but it’s for a good cause.

Thermaltake Toughpower GF3 850W ATX v3.0 Power Supply Review

Aris Mpitziopoulos — Wed, 16 Nov 2022 12:00:39 +0000

The Thermaltake GF3 with 850W max power offers good performance, but not high enough for it to make our list of the best PSUs. Nonetheless, it has high build quality and is one of the first 850W PSUs offering ATX 3.0 and PCIe 5.0 compatibility. The Corsair RM850x and the EVGA 850 G7, which achieve notably higher performance, are not ATX 3.0 ready and do not have 12VHPWR connectors.

Thermaltake recently introduced the GF3 line, which consists of ATX 3.0 and PCIe 5.0 ready PSUs. In this review, we will look at the model with 850W capacity, equipped with a 300W 12VHPWR connector, on paper at least, because in reality, the connector's sense pins are set for a max load of 600W. Channel Well Technology provides the platforms for all GF3 models but the 1350W and 1650W ones.

Tom's Hardware

The GF3 850 is rated Gold by 80 Plus and Cybenetics and has a Cybenetics A- noise rating. The fully modular cable design is a must in this price range ($120-130) and it uses a fluid dynamic bearing fan to ensure the absence of problems during the extended, ten-year warranty. Lastly, its dimensions are normal, with a 160 mm depth.

Specifications

Manufacturer (OEM)	CWT
Max. DC Output	850W
Efficiency	80 PLUS Gold, Cybenetics Gold (87-89%)
Noise	Cybenetics A- (25-30 dB[A])
Modular	✓ (fully)
Intel C6/C7 Power State Support	✓
Operating Temperature (Continuous Full Load)	0 - 40°C
Over Voltage Protection	✓
Under Voltage Protection	✓
Over Power Protection	✓
Over Current (+12V) Protection	✓
Over Temperature Protection	✓
Short Circuit Protection	✓
Surge Protection	✓
Inrush Current Protection	✓
Fan Failure Protection	✗
No Load Operation	✓
Cooling	135mm Fluid Dynamic Bearing Fan (HA13525H12SF-Z)
Semi-Passive Operation	✓(selectable)
Dimensions (W x H x D)	150 x 85 x 160mm
Weight	1.66 kg (3.66 lb)
Form Factor	ATX12V v3.0, EPS 2.92
Alternative Low Power Mode (ALPM) compatible	✓
Warranty	10 Years

Power Specifications

Rail		3.3V	5V	12V	5VSB	-12V
Max. Power	Amps	22	22	70.8	3	0.3
	Watts		120	849.6	15	3.6
Total Max. Power (W)			850

Cables & Connectors

Description	Cable Count	Connector Count (Total)	Gauge	In Cable Capacitors
ATX connector 20+4 pin (600mm)	1	1	16AWG	No
4+4 pin EPS12V (700mm)	1	1	16AWG	No
8 pin EPS12V (700mm)	1	1	16AWG	No
6+2 pin PCIe (500mm+150mm)	2	4	16-18AWG	No
12+4 pin PCIe (610mm) (300W)	1	1	16-24AWG	No
SATA (500mm+150mm+150mm+150mm)	3	12	18AWG	No
4-pin Molex (500mm+150mm+150mm+150mm)	1	4	18AWG	No
FDD Adapter (100mm)	1	1	22AWG	No

There are enough cables and connectors to cover the unit's maximum power efficiently. The single 12VHPWR connector makes the difference compared to the existing 850W units. This type of connector is new, but soon enough, every PSU with over 450W capacity will have one.

Tom's Hardware

Component Analysis

`General Data`	-
Manufacturer (OEM)	CWT
Platform	CSZ
PCB Type	Double Sided
`Primary Side`	-
Transient Filter	4x Y caps, 2x X caps, 2x CM chokes, 1x MOV
Inrush Protection	1x NTC Thermistor SCK-075 (7 Ohm) & Relay
Bridge Rectifier(s)	2x Yangjie Electronic GBU1506 (600V, 15A @ 100°C)
APFC MOSFETs	2x STMicroelectronics STF33N60M2 (600V, 16A @ 100°C, Rds(on): 0.125Ohm)
APFC Boost Diode	1x On Semiconductor FFSP0865A (650V, 8A @ 155°C)
Bulk Cap(s)	1x Rubycon (420V, 680uF, 2,000h @ 105°C, MXE)
Main Switchers	2x On Semiconductor
APFC Controller	Champion CM6500UNX & CM03X
Resonant Controller	Champion CU6901VAC
Topology	Primary side: APFC, Half-Bridge & LLC converter Secondary side: Synchronous Rectification & DC-DC converters
`Secondary Side`	-
+12V MOSFETs	6x International Rectifier IRFH7004PbF (40V, 164A @ 100°C, Rds(on): 1.4mOhm)
5V & 3.3V	DC-DC Converters: 2x UBIQ QN3107M6N (30V, 70A @ 100°C, Rds(on): 2.6mOhm) & 2x UBIQ QM3054M6 (30V, 61A @ 100°C, Rds(on): 4.8mOhm) PWM Controller(s): uPI-Semi uP3861P
Filtering Capacitors	Electrolytic: 3x Nichicon (2-5,000h @ 105°C, HD), 4x Nichicon (4-10,000h @ 105°C, HE), 1x Rubycon (2-10,000h @ 105°C, YXF), 1x Nippon Chemi-Con (4-10,000h @ 105°C, KY), 1x Nippon Chemi-Con (4-10,000h @ 105°C, KYA) Polymer: 10x Elite, 6x APAQ, 8x CapXon, 4x NIC
Supervisor IC	Weltrend WT7502R
Fan Controller	Microchip PIC16F1503
Fan Model	Hong Hua HA13525H12SF-Z (135mm, 12V, 0.5A, Fluid Dynamic Bearing Fan)
`5VSB Circuit`	-
Rectifier	1x PS1045L SBR (45V, 10A)
Standby PWM Controller	On-Bright OB2365T

Tom's Hardware

CWT provides the platform, and its code name is CSZ. This is not a brand-new platform, but rather a modified design to offer the desired ATX 3.0 compatibility and a 12VHPWR connector. The parts that CWT used, after TT's instructions, are of high quality. The PCB is large enough to allow for good airflow, but typically CWT used small heat sinks, with the secondary side without any. With the proper fan speed profile, this won't affect the platform's reliability.

Tom's Hardware

The transient/EMI filter is complete and does a fair job. Surge protection is offered through an MOV, and inrush protection is handled by an NTC thermistor and bypass relay combo.

Tom's Hardware

The pair of bridge rectifiers can rectify up to 30A.

Tom's Hardware

The APFC converter uses two STMicroelectronics FETs and an On Semiconductor boost diode. The bulk cap is by Rubycon, and its capacity is barely enough to allow for more than 17 ms hold-up time.

Tom's Hardware

The main FETs are installed in a half-bridge topology. An LLC resonant converter is also used for higher efficiency.

Tom's Hardware

The 12V FETs are installed on a vertical board next to the main transformer for lower power losses. The minor rails are generated through a pair of DC-DC converters.

Tom's Hardware

The electrolytic filtering caps are of high quality. Most polymer caps are by lower-cost brands, but these caps are highly tolerant to harsh conditions.

Tom's Hardware

The standby PWM controller is an On-Bright OB2365T.

Tom's Hardware

We find several polymer caps on the modular PCB, for ripple filtering purposes.

(Image credit: Tom's Hardware)

The main supervisor IC is a Weltrend WT7502R. It is installed on the same board with a Microchip PIC16F1503, which controls the fan's speed.

Tom's Hardware

Soldering quality is typical CWT, that is good.

Tom's Hardware

The cooling fan is by Hong Hua and it uses a fluid dynamic bearing, so it should work for a nice, long time.

MORE: Best Power Supplies

To learn more about our PSU tests and methodology, please check out How We Test Power Supply Units.

Corsair RM850x (2021)

Seasonic FOCUS GX-850

ASUS ROG STRIX 850G

Primary Rails And 5VSB Load Regulation

The following charts show the main rails' voltage values recorded between a range of 40W up to the PSU's maximum specified load, along with the deviation (in percent). Tight regulation is an important consideration every time we review a power supply because it facilitates constant voltage levels despite varying loads. Tight load regulation also, among other factors, improves the system’s stability, especially under overclocked conditions and, at the same time, it applies less stress to the DC-DC converters that many system components utilize.

Tom's Hardware

The 12V rail starts low and increases its voltage with more than 60W load. Since we consider the minimum and maximum voltage deviations on all rails, sudden voltage increases affect load regulation.

Hold-Up Time

Put simply, hold-up time is the amount of time that the system can continue to run without shutting down or rebooting during a power interruption.

Tom's Hardware

The hold-up time is longer than 17 ms, but the power ok signal's hold-up time is lower than 16ms, which is the minimum that the ATX spec requires.

Inrush Current

Tom's Hardware

Inrush current is low with 115V and on the high side with 230V.

Leakage Current

(Image credit: Tom's Hardware)

Leakage current is low.

10-110% Load Tests

These tests reveal the PSU's load regulation and efficiency levels under high ambient temperatures. They also show how the fan speed profile behaves under increased operating temperatures.

Test	12V	5V	3.3V	5VSB	DC/AC (Watts)	Efficiency	Fan Speed (RPM)	PSU Noise (dB[A])	Temps (In/Out)	PF/AC Volts
10%	5.244A	1.99A	2.01A	0.989A	85.015	84.401%	0	<6.0	44.77°C	0.979
	12.094V	5.025V	3.284V	5.056V	100.724				40.41°C	114.91V
20%	11.498A	2.987A	3.019A	1.189A	169.975	88.519%	0	<6.0	45.58°C	0.99
	12.095V	5.022V	3.28V	5.047V	192.023				40.87°C	114.89V
30%	18.096A	3.487A	3.525A	1.368A	254.994	90.163%	0	<6.0	46.89°C	0.994
	12.100V	5.02V	3.277V	5.119V	282.819				41.57°C	114.87V
40%	24.744A	3.987A	4.032A	1.564A	340.096	90.256%	418	7.8	41.84°C	0.991
	12.079V	5.017V	3.274V	5.115V	376.812				47.89°C	114.85V
50%	31.025A	4.986A	5.045A	1.762A	425.107	89.92%	418	7.8	42.18°C	0.991
	12.074V	5.015V	3.271V	5.108V	472.757				48.63°C	114.83V
60%	37.269A	5.985A	6.059A	1.961A	509.623	89.419%	624	17	42.96°C	0.992
	12.069V	5.013V	3.268V	5.1V	569.925				49.99°C	114.8V
70%	43.582A	6.986A	7.075A	2.161A	594.932	88.797%	829	26.4	43.15°C	0.993
	12.065V	5.011V	3.266V	5.092V	669.985				50.71°C	114.79V
80%	49.903A	7.987A	8.091A	2.262A	679.78	88.103%	1021	32.8	43.87°C	0.994
	12.061V	5.01V	3.263V	5.085V	771.582				52.06°C	114.75V
90%	56.622A	8.489A	8.588A	2.364A	765.213	87.34%	1394	41.6	44.68°C	0.995
	12.056V	5.008V	3.26V	5.078V	876.131				54.03°C	114.74V
100%	63.069A	8.992A	9.117A	2.963A	850.026	86.454%	1684	47.1	45.98°C	0.995
	12.055V	5.006V	3.258V	5.063V	983.212				55.99°C	114.71V
110%	69.421A	9.995A	10.231A	2.966A	934.602	85.484%	1928	50.1	46.91°C	0.996
	12.047V	5.004V	3.255V	5.058V	1093.306				57.82°C	114.68V
CL1	0.116A	14.412A	14.57A	0A	121.33	82.161%	442	8.3	42.08°C	0.988
	12.107V	5.011V	3.274V	5.061V	147.672				48.53°C	114.89V
CL2	0.116A	21.944A	0A	0A	111.428	80.589%	445	8.3	43.01°C	0.987
	12.109V	5.014V	3.289V	5.068V	138.267				50.09°C	114.9V
CL3	0.116A	0A	22.225A	0A	73.991	75.34%	424	7.9	44.43°C	0.978
	12.112V	5.03V	3.266V	5.063V	98.214				52.47°C	114.91V
CL4	70.572A	0A	0A	0A	849.791	87.157%	1537	44.2	45.78°C	0.995
	12.042V	5.03V	3.271V	5.122V	975.014				55.73°C	114.7V

The PSU delivers 110% of its max-rated capacity at high operating temperatures and for prolonged periods without any problems. Noise output is high, though, exceeding 50 dBA.

20-80W Load Tests

In the following tests, we measure the PSU's efficiency at loads significantly lower than 10% of its maximum capacity (the lowest load the 80 Plus standard measures). This is important for representing when a PC is idle with power-saving features turned on.

Test	12V	5V	3.3V	5VSB	DC/AC (Watts)	Efficiency	Fan Speed (RPM)	PSU Noise (dB[A])	Temps (In/Out)	PF/AC Volts
20W	1.242A	0.496A	0.501A	0.197A	20.01	76.407%	0	<6.0	40.12°C	0.834
	11.961V	5.042V	3.294V	5.075V	26.191				37.01°C	114.93V
40W	2.732A	0.694A	0.701A	0.296A	40.01	82.884%	0	<6.0	41.07°C	0.938
	11.967V	5.041V	3.293V	5.072V	48.272				37.61°C	114.93V
60W	4.220A	0.894A	0.904A	0.395A	60.011	85.000%	0	<6.0	42.32°C	0.964
	11.974V	5.031V	3.287V	5.069V	70.601				38.55°C	114.92V
80W	5.653A	1.094A	1.105A	0.494A	79.974	84.591%	0	<6.0	43.07°C	0.978
	12.090V	5.028V	3.285V	5.066V	94.546				39.11°C	114.92V

Efficiency is high with light loads.

2% or 10W Load Test

12V	5V	3.3V	5VSB	DC/AC (Watts)	Efficiency	Fan Speed (RPM)	PSU Noise (dB[A])	Temps (In/Out)	PF/AC Volts
1.224A	0.25A	0.25A	0.052A	16.968	74.57%	0	<6.0	26.87°C	0.806
11.940V	5.041V	3.292V	5.079V	22.754				26.81°C	114.94V

The PSU easily breaks the 70% mark with a 2% load.

Efficiency & Power Factor

Tom's Hardware

Efficiency with normal loads is low. The situation turns around with light and super-light loads.

5VSB Efficiency

Test #	5VSB	DC/AC (Watts)	Efficiency	PF/AC Volts
1	0.1A	0.508W	75.819%	0.065
	5.08V	0.67W		114.93V
2	0.25A	1.27W	78.397%	0.147
	5.078V	1.62W		114.93V
3	0.55A	2.791W	79.556%	0.268
	5.073V	3.508W		114.93V
4	1A	5.066W	79.426%	0.36
	5.065V	6.378W		114.93V
5	1.5A	7.585W	79.592%	0.421
	5.056V	9.53W		114.93V
6	3.001A	15.092W	78.111%	0.497
	5.03V	19.321W		114.92V

Tom's Hardware

The 5VSB rail achieves high enough efficiency.

Power Consumption In Idle And Standby

Mode	12V	5V	3.3V	5VSB	Watts	PF/AC Volts
Idle	11.936V	5.041V	3.292V	5.082V	2.855	0.208
						114.93V
Standby					0.019	0.002
						114.93V

Tom's Hardware

Vampire power is low.

Fan RPM, Delta Temperature, And Output Noise

All results are obtained between an ambient temperature of 37 to 47 degrees Celsius (98.6 to 116.6 degrees Fahrenheit).

(Image credit: Tom's Hardware)

The fan's speed increases linearly to the load, even at increased operating temperatures, but at high speeds, the fan is noisy, reaching 50 dBA. Larger heat sinks would help keep noise output lower under harsh conditions.

The following results were obtained at 30 to 32 degrees Celsius (86 to 89.6 degrees Fahrenheit) ambient temperature.

(Image credit: Tom's Hardware)

At normal operating temperatures, close to 30 degrees Celsius, and with the load pattern that we used, where we fully load the minor rails from the start, the semi-passive operation doesn't last long. Nevertheless, the fan's noise is minimal at up to 460W. Noise exceeds 30 dBA with more than 720W load, and the 40 dBA mark is passed with 825W.

MORE: Best Power Supplies

Protection Features

Check out our PSUs 101 article to learn more about PSU protection features.

OCP (Cold @ 27°C)	12V: 88.6A (125.15%), 11.978V 5V: 30.5A (138.64%), 4.997V 3.3V: 30.9A (140.45%), 3.256V 5VSB: 4.7A (156.67%), 5V
OCP (Hot @ 43°C)	12V: 88.8A (125.44%), 12.025V 5V: 30.4A (138.18%), 4.999V 3.3V: 30.7A (139.55%), 3.251V 5VSB: 4.7A (156.67%), 4.995V
OPP (Cold @ 28°C)	1061.34W (124.92%)
OPP (Hot @ 43°C)	1067.91W (125.7%)
OTP	✓ (131°C @ 12V Heat Sink)
SCP	12V to Earth: ✓ 5V to Earth: ✓ 3.3V to Earth: ✓ 5VSB to Earth: ✓ -12V to Earth: ✓
PWR_OK	Accurate but lower than 16ms
NLO	✓
SIP	Surge: MOV Inrush: NTC Thermistor & Bypass relay

OCP is correctly set at 12V, and the same goes for OPP. However, this is not the case for the minor rails, where OCP is way higher than 130%. Moreover, the power ok signal's hold-up time is slightly lower than 16ms, which the ATX sets as the minimum level.

DC Power Sequencing

Tom's Hardware

No problems here, as the 3.3V rail is always lower than the other two.

Cross Load Tests

Load Regulation Charts

Tom's Hardware

Efficiency Graph

(Image credit: Tom's Hardware)

Ripple Graphs

The lower the power supply's ripple, the more stable the system will be and less stress will also be applied to its components.

Tom's Hardware

Infrared Images

Tom's Hardware

As expected, the board holding the 12V FETs is the hottest part, given the lack of proper heat sinks.

MORE: Best Power Supplies

Advanced Transient Response Tests

For details about our transient response testing, please click here.

Advanced Transient Response at 20% – 20ms

Voltage	Before	After	Change	Pass/Fail
12V	12.061V	11.876V	1.54%	Pass
5V	5.014V	4.916V	1.95%	Pass
3.3V	3.276V	3.165V	3.38%	Pass
5VSB	5.049V	4.986V	1.25%	Pass

Advanced Transient Response at 20% – 10ms

Voltage	Before	After	Change	Pass/Fail
12V	12.060V	11.903V	1.30%	Pass
5V	5.014V	4.922V	1.83%	Pass
3.3V	3.276V	3.167V	3.32%	Pass
5VSB	5.049V	4.989V	1.19%	Pass

Advanced Transient Response at 20% – 1ms

Voltage	Before	After	Change	Pass/Fail
12V	12.062V	11.911V	1.24%	Pass
5V	5.014V	4.915V	1.98%	Pass
3.3V	3.276V	3.170V	3.25%	Pass
5VSB	5.049V	4.986V	1.26%	Pass

Advanced Transient Response at 50% – 20ms

Voltage	Before	After	Change	Pass/Fail
12V	12.027V	11.902V	1.04%	Pass
5V	5.007V	4.913V	1.88%	Pass
3.3V	3.267V	3.153V	3.50%	Pass
5VSB	5.108V	5.073V	0.68%	Pass

Advanced Transient Response at 50% – 10ms

Voltage	Before	After	Change	Pass/Fail
12V	12.028V	11.909V	0.99%	Pass
5V	5.008V	4.912V	1.92%	Pass
3.3V	3.267V	3.152V	3.53%	Pass
5VSB	5.108V	5.068V	0.78%	Pass

Advanced Transient Response at 50% – 1ms

Voltage	Before	After	Change	Pass/Fail
12V	12.029V	11.947V	0.68%	Pass
5V	5.007V	4.917V	1.79%	Pass
3.3V	3.267V	3.149V	3.61%	Pass
5VSB	5.108V	5.072V	0.70%	Pass

Tom's Hardware

Transient response is average at 12V but pretty tight on the other rails. Still, I want to see over 3.2V at 3.3V in all cases.

ATX 3.0 Transient Response Tests

The following table shows the load that we applied.

Duty Cycle	Time for Power Excursion (Te)	Time Constant (Tc)	Power @ Te	Power @ Tc
5%	100μs	1900μs	1700W	780.0W
8%	1ms	11.5ms	1530W	762.7W
12.5%	10ms	70ms	1360W	749.3W
25%	100ms	300ms	1020W	785.2W

Tom's Hardware

The PSU successfully passed all ATX 3.0 transient response tests for units equipped with 12VHPWR connectors.

Turn-On Transient Tests

Tom's Hardware

The turn-on transient response is good overall. We noticed a slight drop during the last test, which won't cause any issues.

Power Supply Timing Tests

There are several signals generated by the power supply, which need to be within specified, by the ATX spec, ranges. If they are not, there can be compatibility issues with other system parts, especially mainboards. From year 2020, the PSU's Power-on time (T1) has to be lower than 150 ms and the PWR_OK delay (T3) from 100 to 150 ms, to be compatible with the Alternative Sleep Mode.

PSU Timings Table
T1 (Power-on time) & T3 (PWR_OK delay)
Load	T1	T3
20%	62ms	126ms
100%	63ms	126ms

Tom's Hardware

The PWR_OK delay is within the 100-150ms region, so the PSU supports the alternative sleep mode recommended by the ATX spec.

Ripple Measurements

The ripple limits, according to the ATX specification, are 120mV (+12V) and 50mV (5V, 3.3V, and 5VSB).

Test	12V	5V	3.3V	5VSB	Pass/Fail
10% Load	4.9 mV	8.4 mV	6.2 mV	5.0 mV	Pass
20% Load	5.6 mV	8.2 mV	6.1 mV	5.2 mV	Pass
30% Load	17.8 mV	7.5 mV	6.4 mV	6.6 mV	Pass
40% Load	14.3 mV	7.6 mV	6.5 mV	6.8 mV	Pass
50% Load	13.5 mV	8.0 mV	7.0 mV	5.8 mV	Pass
60% Load	12.8 mV	8.1 mV	7.3 mV	6.5 mV	Pass
70% Load	12.6 mV	10.1 mV	12.9 mV	7.0 mV	Pass
80% Load	12.4 mV	11.9 mV	14.4 mV	7.4 mV	Pass
90% Load	12.4 mV	9.7 mV	9.4 mV	8.0 mV	Pass
100% Load	19.0 mV	11.1 mV	10.8 mV	8.3 mV	Pass
110% Load	19.9 mV	11.4 mV	11.0 mV	9.0 mV	Pass
Crossload 1	13.6 mV	10.4 mV	8.9 mV	9.5 mV	Pass
Crossload 2	13.4 mV	12.6 mV	7.2 mV	10.1 mV	Pass
Crossload 3	6.0 mV	6.2 mV	10.5 mV	8.9 mV	Pass
Crossload 4	16.7 mV	9.4 mV	8.8 mV	10.5 mV	Pass

Tom's Hardware

Ripple suppression is great on all rails.

Ripple At Full Load

Tom's Hardware

Ripple At 110% Load

Tom's Hardware

Ripple At Cross-Load 1

Tom's Hardware

Ripple At Cross-Load 4

Tom's Hardware

EMC Pre-Compliance Testing – Average & Quasi-Peak EMI Detector Results

Electromagnetic Compatibility (EMC) is the ability of a device to operate properly in its environment without disrupting the proper operation of other nearby devices.

΅We use TekBox's EMCview to conduct our EMC pre-compliance testing.

(Image credit: Tom's Hardware)

There is a single spur going over the limits, with both the average and peak EMI detectors.

MORE: Best Power Supplies

Performance Rating

(Image credit: Tom's Hardware)

The GF3 850 achieves good performance, but the Corsair RM850x and the EVGA 850 G7 are far ahead.

Noise Rating

The graph below depicts the cooling fan's average noise over the PSU's operating range, with an ambient temperature between 30 to 32 degrees Celsius (86 to 89.6 degrees Fahrenheit).

(Image credit: Tom's Hardware)

Under normal operating temperatures, the average noise output is below 30 dBA. With lower noise at high loads, it could be close to 25 dBA.

Efficiency Rating

The following graph shows the PSU's average efficiency throughout its operating range with an ambient temperature close to 30 degrees Celsius.

(Image credit: Tom's Hardware)

Average efficiency should be higher, above 88%, at least.

Power Factor Rating

The following graphs show the PSU's average power factor reading throughout its operating range with an ambient temperature close to 30 degrees Celsius and 115V/230V voltage input.

Tom's Hardware

The APFC converter registers good performance.

MORE: Best Power Supplies

The Thermaltake ToughPower GF3 850 achieves good overall performance, but that is not high enough to dethrone the Corsair RM850x and the EVGA 850 G7.

But the GF3 850 hides an ace up its sleeve; it is ATX 3.0 and PCIe 5.0 ready. Nvidia's RTX 4000 series might not require ATX 3.0 PSUs and 12VHPWR connectors to operate, but it is good to be as future-proof as possible because PSUs can outlive several GPU generations.

(Image credit: Tom's Hardware)

From now on, we will see more and more ATX 3.0 compatible PSUs featuring 12VHPWR connectors. Such PSUs are tolerant to dreadful power spikes, reaching up to 200% of the PSU's max power. This is an overkill feature for the current generation GPUs, at least, but you cannot know what the future will bring. We strongly advise you to invest in an ATX 3.0 PSU. Still, if you find a good bargain on a previous-generation PSU, you could go for it, considering that the RTX 4000 graphics cards can operate with the provided adapter cables. The use of adapters is not ideal since they allow for higher voltage drops, especially under high loads, but this is the most affordable way to retain compatibility with PSUs that don't have 12VHPWR connectors.

MORE: Best Power Supplies

MSI MEG Ai1000P PCIe 5 ATX v3.0 Power Supply Review

Aris Mpitziopoulos — Tue, 15 Nov 2022 15:00:09 +0000

The MSI MEG Ai1000P PCIe5 achieves top performance, and the fact that it is ATX 3.0 and PCIe 5.0 ready brings it to the newest standards. That combination earns it a place in our best PSUs article. It achieves almost similar performance to the Seasonic Prime Platinum 1000 and the EVGA SuperNOVA 1000 P6, but none of those units have 12VHPWR connectors, nor are they ATX 3.0 compatible.

Besides ATX 3.0 and PCIe 5.0 compatibility, the MSI MEG Ai1000P PCIe 5 also has an interesting (we could say appealing) exterior design. Thanks to the semi-digital platform that the PSU uses, provided by Channel Well Technology, there is compatibility with the MSI Gaming Intelligence application, through which you can adjust several settings, like the fan's speed and toggle on/off the single +12V mode while monitoring the PSU's vital functions and power delivery.

Tom's Hardware

The MEG Ai1000P PCIe5 has a fully modular cable design and compact enough dimensions, given its capacity, measuring 160mm in depth. MSI could go with a larger fan than the 120mm it used, to keep the noise output low. Nonetheless, the fan's build quality is high, and the product's warranty is extended at ten years to match the competition's offerings.

Specifications

Manufacturer (OEM)	CWT
Max. DC Output	1000W
Efficiency	80 PLUS Platinum, Cybenetics Platinum (89-91%)
Noise	Cybenetics Standard++ (30-35 dB[A])
Modular	✓ (fully)
Intel C6/C7 Power State Support	✓
Operating Temperature (Continuous Full Load)	0 - 50°C
Over Voltage Protection	✓
Under Voltage Protection	✓
Over Power Protection	✓
Over Current (+12V) Protection	✓
Over Temperature Protection	✓
Short Circuit Protection	✓
Surge Protection	✓
Inrush Current Protection	✓
Fan Failure Protection	✗
No Load Operation	✓
Cooling	120mm Hydraulic Bearing Fan (PLA12025S12H)
Semi-Passive Operation	✓
Dimensions (W x H x D)	150 x 85 x 160mm
Weight	1.99 kg (4.39 lb)
Form Factor	ATX12V v3.0, EPS 2.92
Alternative Low Power Mode (ALPM) compatible	✓
Warranty	10 Years

Power Specifications

Rail		3.3V	5V	12V	5VSB	-12V
Max. Power	Amps	22	22	83.33	3	0.3
	Watts		120	1000	15	3.6
Total Max. Power (W)			1000

Cables & Connectors

Description	Cable Count	Connector Count (Total)	Gauge	In Cable Capacitors
ATX connector 20+4 pin (600mm)	1	1	16-20AWG	No
4+4 pin EPS12V (710mm)	2	2	18AWG	No
6+2 pin PCIe (600mm)	6	6	16AWG	No
12+4 pin PCIe (610mm)	1	1	16-24AWG	No
2x 6+2 pin PCIe (610mm)	1	2	18AWG	No
SATA (500mm+150mm+150mm+150mm)	4	16	18AWG	No
4 pin Molex (500mm+150mm+150mm+150mm) / FDD (+150mm)	1	4 / 1	18-20AWG	No
USB Cable Mini to Type A (610mm)	1	1	24-28AWG	No
USB Mini to Motherboard Header Cable (590mm)	1	1	24-28AWG	No

The PSU is equipped with connectors, including two EPS and six PCIe, all in dedicated cables. It also has two more PCIe on a single cable and a 12VHPWR connector able to handle up to 600W. According to the ATX spec, a 1000W PSU should have up to 450W 12VHPWR connector,

Tom's Hardware

Component Analysis

`General Data`	-
Manufacturer (OEM)	CWT
Platform	CTT
PCB Type	Double Sided
`Primary Side`	-
Transient Filter	6x Y caps, 2x X caps, 2x CM chokes, 1x MOV
Inrush Protection	NTC Thermistor & Relay
Bridge Rectifier(s)	2x GBJ2506 (600V, 25A @ 100°C)
APFC MOSFETs	2x Infineon IPA60R099P6 (600V, 24A @ 100°C, Rds(on): 0.099Ohm) & 1x SYNC Power SPN5003 (to reduce no load consumption)
APFC Boost Diode	2x Infineon IDH08G65C6 (650V, 8A @ 145°C)
Bulk Cap(s)	1x Nichicon (400V, 680uF, 2,000h @ 105°C, GL) & 1x Nichicon (400V, 560uF, 2,000h @ 105°C, GG)
Main Switchers	4x Alpha & Omega AOTF29S50 (500V, 18A @ 100°C, Rds(on): 0.15Ohm)
IC Drivers	2x Novosense NSi6602 & 1x Infineon 2EDN752x
Digital APFC Controller	Texas Instruments UCD3138A
Digital Resonant Controller	Texas Instruments UCD3138A
Topology	Primary side: Semi-Digital, Interleaved PFC, Full-Bridge & LLC converter Secondary side: Synchronous Rectification & DC-DC converters
`Secondary Side`	-
+12V MOSFETs	6x Infineon BSC014N06NS (60V, 152A @ 100°C, Rds(on): 1.45mOhm)
5V & 3.3V	DC-DC Converters: 6x FETs PWM Controller(s): no info
Filtering Capacitors	Electrolytic: 5x Nippon Chemi-Con (105°C, W), 1x Nichicon (2-5,000h @ 105°C, HD), 4x Nippon Chemi-Con (4-10,000h @ 105°C, KY), 1x Nippon Chemi-Con (2-5,000h @ 105°C, KZE), 3x Rubycon (4-10,000h @ 105°C, YXJ), 3x Rubycon (4-10,000h @ 105°C, YXF) Polymer: 11x Nippon Chemi-Con, 15x FPCAP, 5x NIC
Supervisor IC	Weltrend WT7502R (OVP, UVP, SCP, PG)
MCU & Fan Controller	Microchip PIC32MM0064GPM036
Fan Model	MSI (Powerlogic) PLA12025S12H-4 (120mm, 12V, 0.36A, Hydro-Dynamic Bearing Fan)
`5VSB Circuit`	-
Rectifier	1x PS1045L SBR (45V, 10A)
Standby PWM Controller	On Bright OB2365T

Tom's Hardware

The platform is made by CWT and uses a semi-digital design with the code name CTT. This is a modern design where digital controllers handle the primary side and a part of the secondary side. Thanks to the digital controllers, the PSU can "talk" to the system so users, besides monitoring its vital functions, can also make changes to the fan speed control and toggle on/off the multi-rail mode for the 12V rail. CWT uses high-end parts everywhere, and our only objection is the small diameter fan, which is of extremely high quality, but the chassis could accommodate a 140 mm fan.

Tom's Hardware

The transient/EMI filter has all necessary parts, including an MOV for protection against voltage surges. Moreover, there is an NTC thermistor and bypass relay combo for suppressing high inrush currents.

Tom's Hardware

The pair of bridge rectifiers can handle up to 50A.

Tom's Hardware

The APFC circuit has two Infineon FETs and two boost diodes because it uses an interleaved design where two converters work in parallel. The bulk caps are by Nichicon, and their combined capacity is high, allowing for more than 17ms hold-up time easily.

Tom's Hardware

The four main FETs are installed into a full-bridge topology. An LLC resonant converter is also used for lowering energy losses.

Tom's Hardware

This is the board holding one of the Texas Instruments digital controllers.

Tom's Hardware

Six Infineon FETs regulate the 12V rail. Two VRMs generate the minor rails.

Tom's Hardware

All filtering caps are of high quality, electrolytic and polymer ones.

Tom's Hardware

The standby PWM controller is an On Bright OB2365T.

Tom's Hardware

Besides bus bars and shunt resistors, we also find several filtering caps at the face of the modular board.

(Image credit: Tom's Hardware)

The main supervisor IC is a Weltrend WT7502R.

Tom's Hardware

Soldering quality is not the best we have seen by CWT, probably because this was an early sample.

Tom's Hardware

The Powerlogic fan is of high quality, using a hydro-dynamic bearing. MSI could use a larger diameter fan, though, for lower noise output.

MORE: Best Power Supplies

To learn more about our PSU tests and methodology, please check out How We Test Power Supply Units.

Super Flower Leadex V Platinum PRO 1000W

Corsair RM1000x (2021)

EVGA SuperNOVA 1000 G6

Primary Rails And 5VSB Load Regulation

The following charts show the main rails' voltage values recorded between a range of 40 watts up to the PSU's maximum specified load, along with the deviation (in percent). Tight regulation is an important consideration every time we review a power supply because it facilitates constant voltage levels despite varying loads. Tight load regulation also, among other factors, improves the system’s stability, especially under overclocked conditions and, at the same time, it applies less stress to the DC-DC converters that many system components utilize.

Tom's Hardware

Load regulation is within 1% on all major rails, so it is pretty tight. It could be tighter at 12V, though.

Hold-Up Time

Put simply; hold-up time is the amount of time that the system can continue to run without shutting down or rebooting during a power interruption.

Tom's Hardware

The hold-up time is long and the power ok signal is accurate and longer than 16 ms.

Inrush Current

Tom's Hardware

Inrush current is low with 230V input, but pretty high with 115V.

Leakage Current

(Image credit: Tom's Hardware)

Leakage current is low.

10-110% Load Tests

These tests reveal the PSU's load regulation and efficiency levels under high ambient temperatures. They also show how the fan speed profile behaves under increased operating temperatures.

Test	12V	5V	3.3V	5VSB	DC/AC (Watts)	Efficiency	Fan Speed (RPM)	PSU Noise (dB[A])	Temps (In/Out)	PF/AC Volts
10%	6.530A	1.974A	1.979A	0.99A	99.994	85.609%	0	<6.0	44.53°C	0.979
	12.006V	5.066V	3.334V	5.053V	116.82				40.19°C	115.13V
20%	14.083A	2.963A	2.97A	1.189A	199.944	90.96%	0	<6.0	45.41°C	0.993
	12.003V	5.062V	3.333V	5.048V	219.779				40.58°C	115.1V
30%	22.029A	3.459A	3.465A	1.388A	299.991	91.918%	0	<6.0	46.59°C	0.996
	11.982V	5.06V	3.333V	5.044V	326.286				41.47°C	115.08V
40%	29.922A	3.955A	3.96A	1.588A	399.623	92.367%	0	<6.0	47.24°C	0.997
	11.978V	5.057V	3.333V	5.039V	432.642				41.62°C	115.05V
50%	37.485A	4.947A	4.951A	1.788A	499.353	92.349%	0	<6.0	48.83°C	0.997
	11.974V	5.054V	3.332V	5.034V	540.71				42.83°C	115.02V
60%	45.213A	5.941A	5.944A	1.989A	599.904	91.849%	786	13.3	43.66°C	0.997
	11.946V	5.051V	3.331V	5.029V	653.12				50.24°C	114.99V
70%	52.835A	6.936A	6.938A	2.19A	699.632	91.291%	1146	25.4	44.13°C	0.997
	11.934V	5.047V	3.33V	5.023V	766.378				51.14°C	114.97V
80%	60.507A	7.934A	7.932A	2.292A	799.673	90.681%	1493	34.7	44.48°C	0.999
	11.929V	5.043V	3.328V	5.019V	881.889				52.58°C	114.96V
90%	68.587A	8.434A	8.416A	2.393A	899.47	90.043%	1792	39.4	44.85°C	0.999
	11.911V	5.039V	3.326V	5.014V	998.961				54.17°C	114.92V
100%	76.422A	8.938A	8.93A	3A	999.511	89.529%	2079	43.2	45.62°C	0.998
	11.905V	5.035V	3.325V	5V	1117.686				55.69°C	114.9V
110%	84.195A	9.939A	10.017A	3.002A	1100.128	88.743%	2326	47.0	46.78°C	0.998
	11.899V	5.031V	3.324V	4.998V	1239.721				57.71°C	114.87V
CL1	0.117A	14.26A	14.322A	0A	121.295	83.348%	698	8.7	42.95°C	0.985
	11.994V	5.063V	3.33V	5.086V	145.539				48.88°C	115.13V
CL2	0.116A	21.715A	0A	0A	111.386	81.875%	1285	29.4	43.56°C	0.983
	12.007V	5.065V	3.333V	5.109V	136.12				50.64°C	115.13V
CL3	0.116A	0A	21.794A	0A	73.986	74.395%	1370	31.7	44.31°C	0.837
	12.022V	5.067V	3.331V	5.059V	99.443				52.49°C	115.15V
CL4	83.927A	0A	0A	0A	1000.076	90.096%	2058	42.8	45.97°C	0.998
	11.916V	5.041V	3.33V	5.046V	1112.815				55.92°C	114.9V

The PSU delivered 1,100 wats of load at close to 47 degrees Celsius without any issues, but you should not over-stress it if you want it to outlive the extended warranty. The cooling fan rotated at full speed only during the overload test.

20-80W Load Tests

Test	12V	5V	3.3V	5VSB	DC/AC (Watts)	Efficiency	Fan Speed (RPM)	PSU Noise (dB[A])	Temps (In/Out)	PF/AC Volts
20W	1.236A	0.493A	0.495A	0.197A	19.996	73.387%	0	<6.0	40.23°C	0.687
	12.019V	5.068V	3.335V	5.067V	27.316				37.09°C	115.15V
40W	2.723A	0.691A	0.693A	0.296A	39.994	75.323%	0	<6.0	41.24°C	0.791
	12.003V	5.067V	3.334V	5.064V	53.123				37.79°C	115.15V
60W	4.210A	0.888A	0.89A	0.395A	59.992	79.229%	0	<6.0	41.74°C	0.818
	12.002V	5.067V	3.335V	5.063V	75.588				38.05°C	115.14V
80W	5.682A	1.085A	1.088A	0.494A	79.942	81.888%	0	<6.0	43.29°C	0.836
	12.021V	5.067V	3.335V	5.062V	97.633				39.31°C	115.14V

Efficiency under light loads should be higher. We would like to see three out of four tests with over 80% efficiency.

2% or 10W Load Test

12V	5V	3.3V	5VSB	DC/AC (Watts)	Efficiency	Fan Speed (RPM)	PSU Noise (dB[A])	Temps (In/Out)	PF/AC Volts
1.472A	0.255A	0.255A	0.053A	20.115	73.301%	0	<6.0	25.77°C	0.678
	12.032V	5.06V	3.326V	5.061V	27.451			23.78°C	115.15V

With super-light loads, the platform delivers high efficiency.

Efficiency & Power Factor

Tom's Hardware

Efficiency is high enough with normal and super-light loads, but not so high with light loads. Moreover, the APFC converter needs tuning for higher PF readings with 230V input.

5VSB Efficiency

Test #	5VSB	DC/AC (Watts)	Efficiency	PF/AC Volts
1	0.1A	0.511W	76.183%	0.053
	5.108V	0.671W		115.16V
2	0.25A	1.276W	78.775%	0.123
	5.104V	1.62W		115.16V
3	0.55A	2.804W	79.183%	0.235
	5.097V	3.541W		115.15V
4	1A	5.087W	79.344%	0.339
	5.086V	6.411W		115.15V
5	1.5A	7.612W	79.897%	0.398
	5.074V	9.529W		115.16V
6	3A	15.111W	78.062%	0.482
	5.037V	19.357W		115.15V

Tom's Hardware

The 5VSB rail is not efficient and this is a great shame!

Power Consumption In Idle And Standby

Mode	12V	5V	3.3V	5VSB	Watts	PF/AC Volts
Idle	12.043V	5.056V	3.323V	5.057V	4.15	0.252
						115.16V
Standby					0.012	0.001
						115.16V

Tom's Hardware

Vampire power is low.

Fan RPM, Delta Temperature, And Output Noise

All results are obtained between an ambient temperature of 37 to 47 degrees Celsius (98.6 to 116.6 degrees Fahrenheit).

(Image credit: Tom's Hardware)

The fan speed profile is not aggressive, but the 120mm fan has to spin at high speeds under harsh conditions, to handle the thermal loads.

The following results were obtained at 30 to 32 degrees Celsius (86 to 89.6 degrees Fahrenheit) ambient temperature.

(Image credit: Tom's Hardware)

At normal operating temperatures, close to 30 degrees Celsius, the PSU is silent with up to 500W loads. With more than 650W at it enters the 30-35 dBA zone and the 40 dBA mark is passed with 860W and higher loads. As we already stated, with a larger fan, noise output could be lower at high loads.

MORE: Best Power Supplies

Protection Features

Check out our PSUs 101 article to learn more about PSU protection features.

OCP (Cold @ 26°C)	12V: 102.2A (122.64%), 11.927V 5V: 29A (131.82%), 5.046V 3.3V: 28.4A (129.09%), 3.311V 5VSB: 4.8A (160%), 4.993V
OCP (Hot @ 44°C)	12V: 101.4A (121.68%), 11.906V 5V: 28.8A (130.91%), 5.057V 3.3V: 27.4A (124.55%), 3.324V 5VSB: 4.7A (156.67%), 4.989V
OPP (Cold @ 28°C)	1223.42W (122.34%)
OPP (Hot @ 43°C)	1223.42W (122.34%)
OTP	✓ (168°C @ 12V Heat Sink)
SCP	12V to Earth: ✓ 5V to Earth: ✓ 3.3V to Earth: ✓ 5VSB to Earth: ✓ -12V to Earth: ✓
PWR_OK	Proper operation
NLO	✓
SIP	Surge: MOV Inrush: NTC Thermistor & Bypass relay

The OCP triggering points are correctly set at 12V and the same goes for OPP. It is a pleasant surprise to find proper OCP triggering points on the minor rails, too.

DC Power Sequencing

Tom's Hardware

No problems here since the 3.3V rail is always lower than the other two.

Cross Load Tests

Load Regulation Charts

Tom's Hardware

Efficiency Graph

(Image credit: Tom's Hardware)

Ripple Graphs

The lower the power supply's ripple, the more stable the system will be and less stress will also be applied to its components.

Tom's Hardware

Infrared Images

Tom's Hardware

The 12V board exceeded 100 degrees Celsius without active cooling and a half load for ten minutes. The FETs on this board can handle the heat, but up to a point, which is why active cooling is required at increased loads.

MORE: Best Power Supplies

Advanced Transient Response Tests

For details about our transient response testing, please click here.

Advanced Transient Response at 20% – 20ms

Voltage	Before	After	Change	Pass/Fail
12V	12.007V	11.930V	0.64%	Pass
5V	5.063V	4.997V	1.30%	Pass
3.3V	3.334V	3.228V	3.17%	Pass
5VSB	5.049V	4.990V	1.17%	Pass

Advanced Transient Response at 20% – 10ms

Voltage	Before	After	Change	Pass/Fail
12V	12.007V	11.898V	0.90%	Pass
5V	5.062V	4.984V	1.55%	Pass
3.3V	3.334V	3.229V	3.16%	Pass
5VSB	5.048V	4.966V	1.62%	Pass

Advanced Transient Response at 20% – 1ms

Voltage	Before	After	Change	Pass/Fail
12V	12.009V	11.909V	0.83%	Pass
5V	5.062V	4.996V	1.31%	Pass
3.3V	3.333V	3.225V	3.24%	Pass
5VSB	5.048V	4.997V	1.01%	Pass

Advanced Transient Response at 50% – 20ms

Voltage	Before	After	Change	Pass/Fail
12V	11.983V	11.909V	0.62%	Pass
5V	5.051V	4.982V	1.37%	Pass
3.3V	3.330V	3.219V	3.34%	Pass
5VSB	5.031V	4.973V	1.14%	Pass

Advanced Transient Response at 50% – 10ms

Voltage	Before	After	Change	Pass/Fail
12V	11.984V	11.890V	0.78%	Pass
5V	5.050V	4.968V	1.62%	Pass
3.3V	3.329V	3.219V	3.29%	Pass
5VSB	5.030V	4.977V	1.06%	Pass

Advanced Transient Response at 50% – 1ms

Voltage	Before	After	Change	Pass/Fail
12V	11.985V	11.906V	0.66%	Pass
5V	5.051V	4.987V	1.26%	Pass
3.3V	3.329V	3.214V	3.45%	Pass
5VSB	5.031V	4.978V	1.06%	Pass

Tom's Hardware

Transient response is great on all rails!

Advanced Transient Response Tests (ATX 3.0 & 12VHPWR)

The Intel ATX v3.0 spec, besides the 12VHPWR connector, also introduced new requirements for transient response testing. You can read more on our ATX v3.0 analysis. So we have to conduct extra transient response tests in the new generation PSUs to check if they meet the ATX v3.0 requirements.

Tom's Hardware

The PSU doesn't have a problem with the tougher ATX 3.0 transient response tests for units with 12VHPWR connectors. All rails are within limits in all tests.

Turn-On Transient Tests

Tom's Hardware

Everything went fine in the turn-on transient response tests.

Power Supply Timing Tests

There are several signals generated by the power supply, which need to be within specified, by the ATX spec, ranges. If they are not, there can be compatibility issues with other system parts, especially mainboards. From the year 2020, the PSU's Power-on time (T1) has to be lower than 150 ms and the PWR_OK delay (T3) from 100 to 150 ms, to be compatible with the Alternative Sleep Mode.

PSU Timings Table
T1 (Power-on time) & T3 (PWR_OK delay)
Load	T1	T3
20%	68ms	132ms
100%	74ms	127ms

Tom's Hardware

The PWR_OK delay is within the 100-150ms region, so the PSU supports the alternative sleep mode recommended by the ATX spec.

Ripple Measurements

The ripple limits, according to the ATX specification, are 120mV (+12V) and 50mV (5V, 3.3V, and 5VSB).

Test	12V	5V	3.3V	5VSB	Pass/Fail
10% Load	16.1 mV	5.4 mV	5.7 mV	13.0 mV	Pass
20% Load	13.7 mV	5.0 mV	5.0 mV	12.4 mV	Pass
30% Load	17.4 mV	5.5 mV	7.4 mV	12.6 mV	Pass
40% Load	18.2 mV	6.3 mV	6.5 mV	13.5 mV	Pass
50% Load	19.7 mV	6.0 mV	6.3 mV	13.6 mV	Pass

Tom's Hardware

Ripple suppression is good on all rails.

Ripple At Full Load

Tom's Hardware

Ripple At 110% Load

Tom's Hardware

Ripple At Cross-Load 1

Tom's Hardware

Ripple At Cross-Load 4

Tom's Hardware

EMC Pre-Compliance Testing – Average & Quasi-Peak EMI Detector Results

Electromagnetic Compatibility (EMC) is the ability of a device to operate properly in its environment without disrupting the proper operation of other nearby devices.

΅We use TekBox's EMCview to conduct our EMC pre-compliance testing.

(Image credit: Tom's Hardware)

There are some spikes with the average EMI detector, but everything is fine with the peak EMI detector.

MORE: Best Power Supplies

Performance Rating

(Image credit: Tom's Hardware)

The flagship MSI PSU achieves top performance, meeting the Seasonic 1000 Prime Platinum and the EVGA 1000 P6 eye-to-eye.

Noise Rating

The graph below depicts the cooling fan's average noise over the PSU's operating range, with an ambient temperature between 30 to 32 degrees Celsius (86 to 89.6 degrees Fahrenheit).

(Image credit: Tom's Hardware)

The average noise output is increased, because the fan has to spin at high speeds at higher loads.

Efficiency Rating

The following graph shows the PSU's average efficiency throughout its operating range with an ambient temperature close to 30 degrees Celsius.

(Image credit: Tom's Hardware)

The average efficiency is high enough.

Power Factor Rating

The following graphs show the PSU's average power factor reading throughout its operating range with an ambient temperature close to 30 degrees Celsius and 115V/230V voltage input.

Tom's Hardware

The APFC converter could be tuned for higher PF readings with 230V input.

MORE: Best Power Supplies

MSI has a strong entry in the high-end category, with the MEG Ai1000P PCIe 5, which as its code name implies comes with a 12VHPWR connector. This is a true ATX 3.0-ready PSU, able to keep up with the demands of the new generation GPUs and most likely with the upcoming ones, too, since the RTX 4090 doesn't require an ATX 3.0 or PCIe 5.0 ready PSU to operate correctly. It just needs a powerful unit, especially if you plan to push its power limit up to 600 watts.

(Image credit: Tom's Hardware)

The MSI MEG Ai1000P PCIe 5 is a fine PSU, topping our performance charts. Its main competitors are the Seasonic Prime 1000 Platinum and the EVGA 1000 P6, none of which are ATX 3.0 and PCIe 5.0 ready.

Even if you plan on getting an Nvidia GeForce RTX 4090, you don’t need an ATX 3.0, PCIe 5.0 ready PSU. Older but strong PSUs will deliver the necessary power, though you'll have to deal with unwieldy (and possibly dangerous) adapters. So if you're spending well over $1,000 on a new high-end graphics card, you may want to invest in one of these new power supplies. Even if you don't plan on buying a high-end GPU in the near future, if you're buying a new PSU now that you plan to use for several years, it's at least worth considering buying one that supports the latest standard.

MORE: Best Power Supplies

EVGA's Axed RTX 4090 FTW3 Gets Shown Off on YouTube

editors@tomshardware.com (Aaron Klotz) — Tue, 08 Nov 2022 01:37:37 +0000

YouTuber JayzTwoCents just released a video showcasing what would have been EVGA's RTX 4090 FTW3 graphics card, featuring an early prototype model that was delivered to him by EVGA directly. The card's existence perfectly demonstrates how abrupt EVGA's decision to leave the GPU market was, with its engineers working on future RTX 4090 models at the very moment EVGA broke ties with Nvidia.

To quickly recap, EVGA left the GPU market for good two months ago, which was announced by JayzTwoCents and Gamers Nexus on YouTube. The company reportedly left due to an unsatisfactory partnership with Nvidia -- and the move was abrupt, with only a couple of EVGA's higher-up leaders knowing about the decision before the news officially went public via the YouTubers.

Jay says he was actually shown a prototype RTX 4090 just like the one he has now, during his and Steve's (Gamers Nexus) talk with EVGA. This was shown to both of them to demonstrate (again) how truly abrupt and serious EVGA's decision was to leave the GPU market.

The card itself is a very early prototype model with the GeForce RTX 4090 badging replaced with "Next gen Graphics" due to licensing issues that would occur if the card was shown in public since EVGA no longer is in partnership with Nvidia.

The card shown has the FTW3 badge on it, so we know that this would have been one of EVGA's flagship 4090 variants. The card itself is surprisingly compact for an RTX 4090 partner card, with three massive fans on the front, and an elegant (for a high-end GPU) matte black-on-silver aesthetic.

(Image credit: YouTube - JayzTwoCents)

Compared to the RTX 3090 and 3090 Ti FTW3 models, this prototype FTW3 features a noticeably bigger form factor and substantially larger fans. From a visual perspective, the silver accents on the top and bottom of the shroud are a significant departure from the previous generation's pure black finish. Overall, the card is a bit more subdued, rather than a pure aggressive gaming theme like we're used to seeing on other cards.

According to Jay, the card's structural rigidity was particularly impressive to see, with the rear I/O backplate screwed directly into the GPU shroud to prevent GPU sagging. Jay demonstrated this by putting the GPU into a case and the card showed zero sag, despite the card's roughly four PCIe slot-wide thickness.

The last thing to mention about the card's design is the 16-pin power connector, which is at the rear of the card, forcing the power cable to the rear as well, instead of the front. FThis design choice will make the card look a lot cleaner when installed, compared to other 4090 models. However, users would have needed additional clearance at the front of their chassis, to accommodate the extra cable length.

Surprisingly the card is functional, with Jay demoing their running Unigine Heaven in a system. But according to Jay, the card is apparently running a "home made" vBIOS since this is, again, a very early prototype card. As a result, the card does not feature a higher-than 100% power limit in GPU overclocking software, so GPU core overclocking is limited on the card.

Nonetheless, Jay overclocked the card, and was surprised by the GDDR6X memory quality, which hit a 1900MHz positive offset. According to Jay, this memory frequency is one of the highest he's ever seen, only matched by MSI's RTX 4090 Suprim X AIB partner card. It's a shame the power limit is locked, or else he likely could have used this card to hit some incredibly high benchmark scores.

While it's sad to see EVGA leave the GPU market for good, at least this card gives us one last chance to see what EVGA was working on, right up until the very end.

Post-EVGA, Kingpin Seems Open to New Hardware Partnerships

Mark Tyson — Wed, 21 Sep 2022 16:23:41 +0000

Last week, EVGA announced it would be leaving the graphics card business — which came as a huge surprise and left everyone with a lot of questions about the brand's future. There's especially been a lot of discussion, on social media and in online forums, about what EVGA's decision means for the future of the premium Kingpin sub-brand, which was aimed at hardcore enthusiasts and overclockers.

Vince "Kingpin" Lucido has made a statement about the EVGA bombshell, and has tacitly indicated he is open to offers.

(Image credit: EVGA, Kingpin)

In a Facebook post to friends and fans, Lucido first thanked friends and colleagues in the industry before moving on to share some love for fans of EVGA Kingpin products. Lucido also made a statement regarding the future of Kingpin products:

“If the KP hardware is meant to continue on in one way or another, I'm sure that it will [smiley face],” wrote Lucido.

Reading between the lines, this suggests Lucido is interested in new hardware partnerships. If Kingpin hardware does continue, especially in the GPU sphere, there are some obvious contenders for collaboration. The big three — Asus, Gigabyte, and MSI — are all heavily involved in overclocking competitions and produce hardware tuned for overclocking feats. It's easy to imagine any of these brands partnering with Kingpin Cooling to create GPUs, motherboards, and maybe more. Of course, we might also see one of the lighter-weight contenders looking to use Kingpin's brand to raise its profile.

(Image credit: EVGA, Kingpin)

Digesting Friday’s Announcement

You may have been waiting for an official statement from EVGA about its decision to leave the GPU market, but the company didn't put out a press release. However, EVGA's Global Product Management Director Jacob Freeman posted on the official forums on Friday. Freeman provided the following concise summary:

EVGA will not carry the next generation graphics cards.
EVGA will continue to support the existing current generation products.
EVGA will continue to provide the current generation products.

The above bullet points are direct quotes — Freeman also emphasized that EVGA was very thankful to the “great community” that supported its graphics card products over the years.

If you're wondering why EVGA made its big decision so close to the GeForce RTX 40 series launch, we think it was probably thanks to an explosive mix of manufacturing economics combined with the weight of EVGA’s customer service commitments (extended warranty, step-up program, etc.) eating too deep into meager profit margins.

EVGA's Manufacturing, Consumer Policies Minimized GPU Profits, Report

editors@tomshardware.com (Aaron Klotz) — Mon, 19 Sep 2022 20:10:11 +0000

EVGA shocked the gamer and enthusiast world last week when it announced its departure from the graphics card industry entirely, due to conflicts with Nvidia as its "tyrannical" partner. Today, Igor of Igor's Lab published his thoughts on EVGA's departure, deeming most of EVGA's problems self-inflicted.

According to Igor, EVGA as an add-in-noard (AIB) manufacturer, operates very differently compared to Nvidia's other AIB partners. EVGA relies completely on third parties to create the circuit boards and the coolers, with engineering being the only part of the process EVGA covers directly.

As a result, EVGA's GPU margins are exceptionally low for an AIB partner, since much of its resources have to be fed back to the third parties responsible for manufacturing the actual graphics cards. Igor asked several competitors about margins, and found worst-case scenarios — which includes EVGA's strategy — account for around 5% in margin profits.

This is a significant difference compared to the 10% profit margins allotted to other AIB partners who do all the manufacturing in-house, allowing these companies to be a lot more efficient to gain those higher margins.

To make matters worse, EVGA is also operating on a volumetric loss compared to its other AIB partners, with far fewer GPU shipments overall. This is probably related to EVGA's sales coverage of primarily America (and Europe), compared to its AIB competitors who manufacture and ship GPUs worldwide. According to Igor, shipment volume is a big deal when you're only making 5% to 10% profit margins.

At the same time, EVGA has also tried to stand out by offering longer warranty periods, and a step-up program, both things no other competitor in the GPU industry offers. While this strategy does provide EVGA with a stellar track record for customer satisfaction, it is a "suicidal strategy" according to EVGA's competitors. One anonymous source at a competitor told Igor, "If it were profitable, we would have done it long ago."

Weighing Up Nvidia's Blame

(Image credit: JPR)

There is no denying that Nvidia has strict guidelines for its AIB partners, including what to do or not do with each graphics card design. Nvidia also goes the extra mile to compete with its AIB partners directly with its Founders Edition GPU models. Recent stats from JPR indicates Nvidia's gross margins have grown fairly steadily since 2005, while the margins of its AIB partners have fallen since 2000.

We don't doubt EVGA's professed reasons for leaving Nvidia are true, including Nvidia withholding MSRP information until GPUs are announced on stage, and forcing AIB partners to set GPU prices to specific categories on specific models. But, it is interesting to see this new information come to light, indicating EVGA has one of the lowest profit margins compared to the rest of Nvidia's AIB partners.

It's clear we don't know the full details of the situation. And with Nvidia and its partners seemingly uninterested or unable to go on the record and be candid about sales, profits and other stats, we may never know much more than we do today. But it would not be surprising to hear EVGA left the GPU market due to issues originating from Nvidia directly, as well as from self-inflicted financial problems combined.

Still, take all this with a grain of salt. Igor is a respected, well-connected individual in the PC industry, but we may still learn more, given we're just a few days past EVGA's bombshell announcement.

EVGA Abandons the GPU Market, Reportedly Citing Conflicts With Nvidia

Andrew E. Freedman — Fri, 16 Sep 2022 20:17:43 +0000

EVGA, often considered Nvidia's top add-in-board partner, is making a drastic shift. It's done doing business with Nvidia and will stop making GPUs altogether. That's according to YouTube channels JayZTwoCents and GamersNexus, which both sat down with EVGA CEO Andrew Han to discuss his frustrations with Nvidia as a partner and reasoning for making the decision. Jon Peddie Research also published a brief article on the matter.

"We are not going to be on [Nvidia CEO] Jensen [Huang]'s lap on stage, so I don't want people to speculate what's going on [when we're not there]," GamersNexus Steve Burke quotes Han as saying. "EVGA has decided not to carry the next gen."

When asked for comment, EVGA product management director Jacob Freeman pointed Tom's Hardware to its forums, where the company offered the following message:

"EVGA is committed to our customers and will continue to offer sales and support on the current lineup. Also, EVGA would like to say thank you to our great community for the many years of support and enthusiasm for EVGA graphics cards." EVGA also confirmed that it will not carry the next generation of GPUs, that it will support current generation products, and that it will "continue to provide the current generation products," though it didn't specify for how long.

It is notable that Han met with two of the biggest tech YouTubers and had the videos embargoed, so there may be more to the story.

"We’ve had a great partnership with EVGA over the years and will continue to support them on our current generation of products," Bryan Del Rizzo, director of global public relations for GeForce at Nvidia told Tom's Hardware. "We wish Andrew and our friends at EVGA all the best."

EVGA will reportedly continue the existing RTX 30-series product line until it runs out of stock. The company won't be moving to partner with AMD or Intel, either, with plans to instead focus on other products for the foreseeable future (EVGA already makes power supplies, coolers and motherboards, for instance), but also reportedly won't expand into new product categories.

The company has reportedly withheld some stock for the market for the purposes of replacing cards in warranty, which should protect current customers, at least for a while.

Han told JayZTwoCents and GamersNexus that EVGA won't be selling the business. GamersNexus states that Nvidia's top brass were first notified in April, while Peddie reports that Nvidia was informed in June. Apparently, EVGA did make some RTX 40-series cards that made it as far as the engineering sample stage, however.

It's unclear how EVGA will keep its employees together. It's not a huge company, and while the GamersNexus video suggests Han said he wants to take care of employees, it's unclear what some of these engineers will have to do. The company laid off some employees in Taiwan earlier in the summer.

Both videos suggest that EVGA feels that Nvidia has stifled it, and suggest they don't find out details like MSRP, costs to buy GPU chips and more until Nvidia announces them, often publicly on stage. There is also talk of price ceilings as well as not being able to customize cards by changing the specs, like adding more memory.

It's also an issue because Nvidia is competing against partners with its own Founders Edition cards, both videos say, because Nvidia, as the supplier and manufacturer, doesn't have to worry about profit margins as much.

EVGA has built its reputation on delivering some of the best Nvidia GPU AIB cards. In fact, the VGA stands for video graphics adapter, which is odd for a company that no longer makes them. According to Jon Peddie Research, EVGA's shift here will "put a big hole in Nvidia's North American sales of AIBs," suggesting that EVGA had 40% of the market, as well as fans of the brand in western Europe. Peddie notes that as GPUs require more and more electricity, EVGA may be able to sell individuals and OEMs power supplies to make up for some of the profits.

Updated September 16, 4:23 p.m. ET with the statement from EVGA's forums.

EVGA SuperNOVA 1000 G7 Power Supply Review

Aris Mpitziopoulos — Thu, 15 Sep 2022 08:00:15 +0000

The EVGA SuperNOVA 1000 G7 is a high-performance power supply, challenging some of the highest capacity models on our best PSUs,list. FSP delivered an excellent platform to EVGA with tiny dimensions and high build quality. The strongest competitor is the Corsair RM1000x (2021), which delivers the same performance levels with notably lower noise output.

After numerous EVGA G series revisions, we have reached G7, with FSP providing the platforms. The G7 units come in four flavors, ranging from 650 W to 1,000 W. All have super compact dimensions, with only 130mm depth, the same as the SFX-L form factor. The dimensions of the 1000 G7 unit, which I will evaluate in this review, are 150 (W) x 85 (H) x 130 mm (L). In comparison the SilverStone SX1000 SFX-L measures 125 x 65 x 130 mm. This makes the 1000 G7 the smallest 1,000 W ATX PSU available on the market. One might wonder why vendors keep making smaller power supplies when very specific standards define the dimensions, but as long as it doesn't affect noise output and performance, I don't mind.

There was no room for a larger fan, so FSP had to install 120 mm fans. Thanks to their fluid dynamic bearings, they shouldn't have a problem with the prolonged, 10-year warranty. EVGA was among the first to offer such long warranty periods, along with Corsair and Seasonic.

Tom's Hardware

The product's box is small, although it contains a powerful 1,000 W PSU. The protection inside is good, with foam spacers covering the PSU's edges.

Tom's Hardware

Once you take the PSU out of the box, you will also notice some LED indicators, which depict the load level of the PSU.

Tom's Hardware

Specifications

Manufacturer (OEM)	FSP
Max. DC Output	1,000 W
Efficiency	80 PLUS Gold, Cybenetics Platinum (89-91%)
Noise	Cybenetics Standard++ (30-35 dB[A])
Modular	✓(Fully)
Intel C6/C7 Power State Support	✓
Operating Temperature (Continuous Full Load)	0 - 50°C
Over Voltage Protection	✓
Under Voltage Protection	✓
Over Power Protection	✓
Over Current (+12V) Protection	✓
Over Temperature Protection	✓
Short Circuit Protection	✓
Surge Protection	✓
Inrush Current Protection	✓
Fan Failure Protection	✗
No Load Operation	✓
Cooling	120mm Fluid Dynamic Bearing Fan (MGA12012XF-O25)
Semi-Passive Operation	✓(selectable)
Dimensions (W x H x D)	150 x 85 x 130mm
Weight	1.72 kg (3.79 lb)
Form Factor	ATX12V v2.52, EPS 2.92
Warranty	10 Years

Power Specifications

Rail		3.3V	5V	12V	5VSB	-12V
Max. Power	Amps	24	24	83.3	3	0.5
	Watts		120	1000	15	6
Total Max. Power (W)			1000

Cables and Connectors

Description	Cable Count	Connector Count (Total)	Gauge	In Cable Capacitors
ATX connector 20+4 pin (600mm)	1	1	18-22AWG	Yes
4+4 pin EPS12V (700mm)	2	2	18AWG	No
6+2 pin PCIe (700mm+150mm)	3	6	18AWG	No
6+2 pin PCIe (700mm)	2	2	18AWG	No
SATA (550mm+100mm+100mm)	4	12	18AWG	No
4-pin Molex (550mm+100mm+100mm+100mm)	1	4	18AWG	No
FDD Adapter (100mm)	1	1	22AWG	No
AC Power Cord (1390mm) - C13 coupler	1	1	16AWG	-

Plenty of long cables and connectors are provided, including two EPS and eight PCIe connectors. You won't find any of the new 12+4 pin PCIe connectors, though, since this PSU is not ATX 3.0 or PCIe 5.0 ready. Moreover, I would like to see a longer distance between the peripheral connectors.

Tom's Hardware

(Image credit: Tom's Hardware)

Only the ATX connector has extra caps, which are for better ripple filtering.

Component Analysis

`General Data`	-
Manufacturer (OEM)	FSP
PCB Type	Double Sided
`Primary Side`	-
Transient Filter	4x Y caps, 2x X caps, 2x CM chokes, 1x MOV
Inrush Protection	NTC Thermistor SCK-056 (5 Ohm) & Relay
Bridge Rectifier(s)	2x
APFC MOSFETs	3x
APFC Boost Diode	1x
Bulk Cap(s)	2x Nippon Chemi-Con (420V, 470uF each or 940uF combined, 2,000h @ 105°C, KHE)
Main Switchers	2x Infineon IPP60R120P7 (600V, 16A @ 100°C, Rds(on): 0.12Ohm)
IC Driver	1x Novosense NSi6602
APFC Controller	Infineon ICE2PCS02
Resonant Controller	Champion CM6901T2X
Topology	Primary side: APFC, Half-Bridge & LLC converter Secondary side: Synchronous Rectification & DC-DC converters
`Secondary Side`	-
+12V MOSFETs	no info
5V & 3.3V	DC-DC Converters: 6x Infineon BSC0901NS (30V, 94A @ 100°C, Rds(on): 1.9mOhm) PWM Controller(s): ANPEC APW7159C
Filtering Capacitors	Electrolytic: 5x Rubycon (3-6,000h @ 105°C, YXG), 3x Rubycon (1-5,000h @ 105°C, ZL) Polymer: 18x Nippon Chemi-Con, 6x NIC
Supervisor IC	Weltrend WT7527RA (OCP, OVP, UVP, SCP, PG)
Fan Controller	Microchip PICF15324
Fan Model	Protechnic Electric MGA12012XF-O25 (120mm, 12V, 0.52A, Fluid Dynamic Bearing Fan)
`5VSB Circuit`	-
Rectifier	1x NIKO-SEM P1006BD (60V, 42A @ 100°C, Rds(on): 10mOhm) FET
Standby PWM Controller	Power Integrations INN2603K

Tom's Hardware

FSP makes this tiny platform. As expected, because of the high capacity, it is overpopulated, and identifying all parts without possibly destroying the PSU was pretty hard. Such PCBs always give us trouble during the part analysis process.

The design is modern with a half-bridge topology and an LLC resonant converter on the primary side. We also find a synchronous rectification scheme on the secondary side for the 12V rail and DC-DC converters for the minor rails.

Tom's Hardware

The transient/EMI filter has all the necessary parts, so I expect low EMI emissions. Moreover, it is equipped with an MOV, which handles voltage surges coming from the mains grid, and I also found an NTC thermistor and relay combo for suppressing high inrush currents.

(Image credit: Tom's Hardware)

The pair of bridge rectifiers is sandwiched between two heat sinks.

Tom's Hardware

The APFC converter uses three FETs and a single boost diode. The PFC controller is installed on a vertical board, an Infineon ICE2PCS02. The same board also hosts an operational amplifier (op-amp).

Tom's Hardware

Two Infineon IPP60R120P7 are the main FETs, configured in a half-bridge topology. Their driver IC is a Novosense NSi6602, and the LLC resonant converter is the usual aspect, a Champion CM6901T2X IC. Parts of the LLC resonant converter are installed on a daughter board because there was no room on the main PCB.

Tom's Hardware

The FETs that regulate the 12V rail are installed on a board next to the main transformer to minimize energy losses. Another vertical board hosts the DC-DC converters that handle the minor rails.

Tom's Hardware

The filtering caps are high quality. On the secondary side, all electrolytic caps are by Rubycon and the polymer ones by Chemi-Con and NIC.

Tom's Hardware

The standby PWM controller is by Power Integrations, and a NIKO-SEM P1006BD FET is the rectifier on the secondary side of the 5VSB circuit.

Tom's Hardware

Many Chemi-Con polymer caps are installed on the modular PCB for ripple filtering purposes.

(Image credit: Tom's Hardware)

The main supervisor IC is a Weltrend WT7527RA.

Tom's Hardware

Soldering quality is good.

Tom's Hardware

The fan is of high quality, and the fluid dynamic bearing will help it outlive the extended warranty if you don't have it operating under high operating temperatures all the time (> 40 degrees Celsius).

MORE: Best Power Supplies

To learn more about our PSU tests and methodology, please check out How We Test Power Supply Units.

Corsair RM1000x (2021)

EVGA SuperNOVA 1000 P6

Seasonic FOCUS GX-1000

Primary Rails And 5VSB Load Regulation

Tom's Hardware

Load regulation is tight enough on all rails.

Hold-Up Time

Put simply; hold-up time is the amount of time that the system can continue to run without shutting down or rebooting during a power interruption.

Tom's Hardware

The hold-up time is long and the power ok signal is accurate.

Inrush Current

Tom's Hardware

Inrush current is high with both voltage inputs that we tried.

Leakage Current

(Image credit: Tom's Hardware)

Leakage current is low.

10-110% Load Tests

These tests reveal the PSU's load regulation and efficiency levels under high ambient temperatures. They also show how the fan speed profile behaves under increased operating temperatures.

Test	12V	5V	3.3V	5VSB	DC/AC (Watts)	Efficiency	Fan Speed (RPM)	PSU Noise (dB[A])	Temps (In/Out)	PF/AC Volts
10%	6.374A	1.936A	1.958A	0.98A	99.987	88.011%	0	<6.0	44.64°C	0.971
	12.298V	5.165V	3.371V	5.103V	113.606				40.33°C	115.14V
20%	13.754A	2.907A	2.939A	1.178A	199.926	91.314%	0	<6.0	45.75°C	0.991
	12.290V	5.16V	3.368V	5.094V	218.946				40.95°C	115.11V
30%	21.487A	3.394A	3.43A	1.377A	299.965	92.273%	0	<6.0	46.35°C	0.994
	12.282V	5.157V	3.367V	5.085V	325.085				41.23°C	115.08V
40%	29.184A	3.88A	3.922A	1.576A	399.492	92.427%	0	<6.0	47.75°C	0.995
	12.277V	5.154V	3.366V	5.075V	432.222				42.12°C	115.05V
50%	36.559A	4.853A	4.903A	1.777A	499.227	92.164%	0	<6.0	49.13°C	0.995
	12.274V	5.152V	3.365V	5.066V	541.676				43.02°C	115.02V
60%	44.092A	5.831A	5.895A	1.982A	599.759	91.601%	1236	30.2	43.16°C	0.996
	12.246V	5.146V	3.359V	5.045V	654.748				49.87°C	114.98V
70%	51.496A	6.808A	6.882A	2.184A	699.465	91.089%	1405	34.7	43.43°C	0.995
	12.241V	5.142V	3.357V	5.034V	767.891				50.58°C	114.95V
80%	59.004A	7.787A	7.868A	2.288A	799.488	90.442%	1717	39.8	43.73°C	0.995
	12.229V	5.138V	3.354V	5.024V	883.981				51.78°C	114.91V
90%	66.840A	8.276A	8.351A	2.392A	899.243	89.713%	2040	44.7	44.56°C	0.994
	12.221V	5.134V	3.352V	5.014V	1002.344				53.66°C	114.87V
100%	74.494A	8.769A	8.864A	3.005A	999.212	88.953%	2343	47.5	46.24°C	0.994
	12.210V	5.131V	3.35V	4.99V	1123.303				56.37°C	114.84V
110%	82.078A	9.752A	9.946A	3.009A	1099.856	87.97%	2683	50.8	46.55°C	0.994
	12.203V	5.127V	3.347V	4.983V	1250.246				57.43°C	114.8V
CL1	0.112A	14.013A	14.163A	0A	121.278	85.802%	0	<6.0	47.97°C	0.98
	12.308V	5.152V	3.367V	5.119V	141.341				42.83°C	115.13V
CL2	0.112A	23.299A	0A	0A	121.381	84.245%	0	<6.0	49.51°C	0.981
	12.308V	5.151V	3.368V	5.124V	144.075				43.12°C	115.13V
CL3	0.112A	0A	23.45A	0A	80.57	79.28%	0	<6.0	51.83°C	0.967
	12.299V	5.164V	3.377V	5.114V	101.638				44.64°C	115.14V
CL4	81.835A	0A	0A	0A	999.846	89.444%	2277	47.5	45.19°C	0.994
	12.218V	5.144V	3.358V	5.073V	1117.842				55.16°C	114.85V

The PSU doesn't have a problem operating under high temperatures for prolonged periods, but you should expect high noise output.

20-80W Load Tests

Test	12V	5V	3.3V	5VSB	DC/AC (Watts)	Efficiency	Fan Speed (RPM)	PSU Noise (dB[A])	Temps (In/Out)	PF/AC Volts
20W	1.207A	0.484A	0.489A	0.195A	19.99	69.134%	0	<6.0	40.23°C	0.788
	12.299V	5.168V	3.372V	5.123V	28.913				37.15°C	115.17V
40W	2.657A	0.677A	0.685A	0.293A	39.99	80.216%	0	<6.0	41.24°C	0.9
	12.299V	5.167V	3.372V	5.12V	49.851				37.87°C	115.16V
60W	4.108A	0.871A	0.881A	0.391A	59.989	84.662%	0	<6.0	41.61°C	0.941
	12.298V	5.167V	3.372V	5.117V	70.856				37.86°C	115.15V
80W	5.554A	1.064A	1.076A	0.489A	79.935	86.967%	0	<6.0	43.76°C	0.96
	12.298V	5.166V	3.372V	5.114V	91.915				39.79°C	115.15V

The unit achieves high efficiency under light loads, with minimal noise output because the fan doesn't spin.

2% or 10W Load Test

12V	5V	3.3V	5VSB	DC/AC (Watts)	Efficiency	Fan Speed (RPM)	PSU Noise (dB[A])	Temps (In/Out)	PF/AC Volts
1.472A	0.255A	0.255A	0.053A	20.545	70.164%	0	<6.0	35.04°C	0.792
	12.298V	5.167V	3.371V	5.128V	29.282			23.65°C	115.16V

The PSU achieves over 70% efficiency with a 2% load, as the ATX spec recommends.

Efficiency & Power Factor

Tom's Hardware

This is a highly efficient platform. There is room for improvement, though, at super-light loads.

5VSB Efficiency

Test #	5VSB	DC/AC (Watts)	Efficiency	PF/AC Volts
1	0.1A	0.512W	76.602%	0.064
	5.124V	0.668W		115.16V
2	0.25A	1.28W	80.835%	0.142
	5.123V	1.584W		115.16V
3	0.55A	2.813W	81.626%	0.26
	5.115V	3.446W		115.16V
4	1A	5.107W	79.616%	0.357
	5.107V	6.415W		115.16V
5	1.5A	7.644W	79.448%	0.412
	5.096V	9.622W		115.16V
6	2.999A	15.16W	78.969%	0.477
	5.055V	19.198W		115.16V

Tom's Hardware

The 5VSB rail is efficient.

Power Consumption In Idle And Standby

Mode	12V	5V	3.3V	5VSB	Watts	PF/AC Volts
Idle	12.278V	5.162V	3.363V	5.122V	7.955	0.42
						115.16V
Standby					0.058	0.006
						115.16V

Tom's Hardware

We would like to see below 0.1W vampire power with 230V input.

Fan RPM, Delta Temperature, And Output Noise

All results are obtained between an ambient temperature of 37 to 47 degrees Celsius (98.6 to 116.6 degrees Fahrenheit).

(Image credit: Tom's Hardware)

The fan's speed increases linearly as the load and the operating temperatures increase. Given the small and overpopulated PCB and the high Wattage, this is the best FSP could do, also considering the ten-year warranty.

The following results were obtained at 30 to 32 degrees Celsius (86 to 89.6 degrees Fahrenheit) ambient temperature.

(Image credit: Tom's Hardware)

At normal operating temperatures, close to 30 degrees Celsius, the PSU's passive mode lasts for long, and the load on the minor rails doesn't seem to affect it. The 30 dBA mark is passed with 700 W and, with above 850 W, noise exceeds 40 dBA.

MORE: Best Power Supplies

Protection Features

Check out our PSUs 101 article to learn more about PSU protection features.

OCP (Cold @ 25°C)	12V: 99A (118.84%), 12.170V 5V: 31A (129.17%), 5.133V 3.3V: 30.1A (125.42%), 3.354V 5VSB: 4.2A (140%), 5.022V
OCP (Hot @ 41°C)	12V: 97.2A (116.67%), 12.195V 5V: 29.2A (121.67%), 5.140V 3.3V: 28.5A (118.75%), 3.366V 5VSB: 4.3A (143.33%), 5.023V
OPP (Cold @ 24°C)	1291.64W (129.16%)
OPP (Hot @ 44°C)	1258.75W (125.88%)
OTP	✓ (156°C @ 12V Secondary Side)
SCP	12V to Earth: ✓ 5V to Earth: ✓ 3.3V to Earth: ✓ 5VSB to Earth: ✓ -12V to Earth: ✓
PWR_OK	Proper Operation
NLO	✓
SIP	Surge: MOV Inrush: NTC Thermistor & Bypass Relay

OCP and OPP are set correctly. Lastly, the other protection features are present and work properly.

DC Power Sequencing

Tom's Hardware

No problems here since the 3.3V rail is always lower than the other two.

Cross Load Tests

Load Regulation Charts

Tom's Hardware

Efficiency Graph

(Image credit: Tom's Hardware)

Ripple Graphs

The lower the power supply's ripple, the more stable the system will be and less stress will also be applied to its components.

Tom's Hardware

Infrared Images

Tom's Hardware

The main transformer and the vertical board holding the 12V FETs are the hottest parts. All in all, the temperatures on all parts are kept in control during this test.

MORE: Best Power Supplies

Advanced Transient Response Tests

For details about our transient response testing, please click here.

Advanced Transient Response at 20% – 20ms

Voltage	Before	After	Change	Pass/Fail
12V	12.275V	12.182V	0.75%	Pass
5V	5.158V	5.056V	1.99%	Pass
3.3V	3.366V	3.255V	3.30%	Pass
5VSB	5.085V	5.063V	0.44%	Pass

Advanced Transient Response at 20% – 10ms

Voltage	Before	After	Change	Pass/Fail
12V	12.275V	12.186V	0.73%	Pass
5V	5.158V	5.056V	1.98%	Pass
3.3V	3.366V	3.255V	3.29%	Pass
5VSB	5.086V	5.066V	0.38%	Pass

Advanced Transient Response at 20% – 1ms

Voltage	Before	After	Change	Pass/Fail
12V	12.278V	12.206V	0.59%	Pass
5V	5.158V	5.060V	1.89%	Pass
3.3V	3.366V	3.251V	3.42%	Pass
5VSB	5.086V	5.070V	0.32%	Pass

Advanced Transient Response at 50% – 20ms

Voltage	Before	After	Change	Pass/Fail
12V	12.250V	12.179V	0.58%	Pass
5V	5.147V	5.048V	1.93%	Pass
3.3V	3.359V	3.246V	3.37%	Pass
5VSB	5.052V	5.033V	0.37%	Pass

Advanced Transient Response at 50% – 10ms

Voltage	Before	After	Change	Pass/Fail
12V	12.250V	12.171V	0.64%	Pass
5V	5.147V	5.045V	1.99%	Pass
3.3V	3.359V	3.248V	3.31%	Pass
5VSB	5.052V	5.034V	0.37%	Pass

Advanced Transient Response at 50% – 1ms

Voltage	Before	After	Change	Pass/Fail
12V	12.253V	12.178V	0.61%	Pass
5V	5.148V	5.048V	1.94%	Pass
3.3V	3.360V	3.244V	3.46%	Pass
5VSB	5.054V	5.032V	0.43%	Pass

Tom's Hardware

Transient response is good, overall, especially at 12V where it matters the most.

Turn-On Transient Tests

Tom's Hardware

We didn't spot any issues during the turn-on transient tests.

Power Supply Timing Tests

There are several signals generated by the power supply, which need to be within specified, by the ATX spec, ranges. If they are not, there can be compatibility issues with other system parts, especially mainboards. From year 2020, the PSU's Power-on time (T1) has to be lower than 150ms and the PWR_OK delay (T3) from 100 to 150ms, to be compatible with the Alternative Sleep Mode.

PSU Timings Table
T1 (Power-on time) & T3 (PWR_OK delay)
Load	T1	T3
20%	47ms	129ms
100%	43ms	129ms

Tom's Hardware

The PWR_OK delay is within the 100-150ms region, so the PSU supports the alternative sleep mode recommended by the ATX spec.

Ripple Measurements

The ripple limits, according to the ATX specification, are 120mV (+12V) and 50mV (5V, 3.3V, and 5VSB).

Test	12V	5V	3.3V	5VSB	Pass/Fail
10% Load	5.9 mV	4.7 mV	4.9 mV	8.4 mV	Pass
20% Load	6.6 mV	5.5 mV	6.0 mV	22.3 mV	Pass
30% Load	7.0 mV	5.5 mV	6.5 mV	12.9 mV	Pass
40% Load	7.5 mV	5.9 mV	6.3 mV	12.2 mV	Pass
50% Load	8.3 mV	7.3 mV	7.2 mV	23.4 mV	Pass

Tom's Hardware

Ripple suppression is good on all rails.

Ripple At Full Load

Tom's Hardware

Ripple At 110% Load

Tom's Hardware

Ripple At Cross-Load 1

Tom's Hardware

Ripple At Cross-Load 4

Tom's Hardware

EMC Pre-Compliance Testing – Average & Quasi-Peak EMI Detector Results

Electromagnetic Compatibility (EMC) is the ability of a device to operate properly in its environment without disrupting the proper operation of other nearby devices.

΅We use TekBox's EMCview to conduct our EMC pre-compliance testing.

(Image credit: Tom's Hardware)

Several spurs exceed the limits with the average EMI detector, but everything is fine with the peak detector.

MORE: Best Power Supplies

Performance Rating

(Image credit: Tom's Hardware)

Overall performance is high.

Noise Rating

The graph below depicts the cooling fan's average noise over the PSU's operating range, with an ambient temperature between 30 to 32 degrees Celsius (86 to 89.6 degrees Fahrenheit).

(Image credit: Tom's Hardware)

With a larger PCB and fan, this platform could have a lower overall noise output.

Efficiency Rating

The following graph shows the PSU's average efficiency throughout its operating range with an ambient temperature close to 30 degrees Celsius.

(Image credit: Tom's Hardware)

The 1000 G7 is highly efficient and this is why is rated as Platinum in the Cybenetics scale.

Power Factor Rating

The following graphs show the PSU's average power factor reading throughout its operating range with an ambient temperature close to 30 degrees Celsius and 115V/230V voltage input.

Tom's Hardware

The APFC converter has decent performance.

MORE: Best Power Supplies

EVGA is near the top spot on our performance charts with the 1000 G7 unit, and this is not an easy feat, considering the mighty Corsair RM1000x that this unit has to face off against. Besides tiny dimensions, the 1000 G7 offers high efficiency, tight enough load regulation, good ripple suppression, and excellent transient response at 12V, which is the most important rail. The build quality is also high, and FSP used top-notch parts, including Rubycon and Chemi-Con caps, along with an FDB fan from a respected manufacturer.

(Image credit: Tom's Hardware)

The LED load indicators on the PSU's side are a gimmick, though one that may still come in handy to some. Moreover, PSU downsizing has reached its limits with this product, which is in the SFX-L form factor's territory with only 130 mm depth. Anything less than that will require a smaller fan, and this is not an option if you want to keep noise output low enough. Despite the tiny dimensions, the EVGA 1000 G7 meets the Corsair RM1000x in overall performance, but it loses the fight in noise output because of the smaller fan and the overpopulated PCB. Still, it isn't noisy under normal operating conditions.

MORE: Best Power Supplies

EVGA SuperNOVA 750 GA Power Supply Review

Aris Mpitziopoulos — Tue, 13 Sep 2022 21:35:40 +0000

The EVGA SuperNOVA 750 GA doesn't outperform the majority of competing offerings in the 750W category, including those among the best PSUs. Nonetheless, at $80, it is a good bargain, given its high build quality and decent overall performance. Moreover, the ten-year warranty shows EVGA's faith in its reliability.

There are multiple SuperNOVA power supply lineups from EVGA. The GA one range consists of four members with capacities ranging from 550W to 850W. We saw this one had positive reviews on Amazon and other resellers, so we chose to get and test one.

The 750W member of the SuperNOVA GA line will be our test subject. This unit has Gold efficiency in the 80 Plus standard and Silver in the Cybenetics scale. It also has a Cybenetics Standard++ noise rating, meaning that you won't need earplugs to use it. Besides the fully modular cable design and compact dimensions, the double ball-bearing fan is another interesting feature. DBB fans, in general, are preferred to FDB ones under harsh conditions because they are more tolerant to increased operating temperatures.

Tom's Hardware

The OEM behind the SuperNOVA 750 GA is Andyson, a manufacturer we don't often see. This is a shame because we have seen some terrific platforms from this OEM in the past, including the Aerocool Project 7 PSUs.

Tom's Hardware

Specifications

Manufacturer (OEM)	Andyson
Max. DC Output	750W
Efficiency	80 PLUS Gold, Cybenetics Silver (85-87%)
Noise	Cybenetics S++ (30-35 dB[A])
Modular	✓ (Fully)
Intel C6/C7 Power State Support	✓
Operating Temperature (Continuous Full Load)	0 - 50°C
Over Voltage Protection	✓
Under Voltage Protection	✓
Over Power Protection	✓
Over Current (+12V) Protection	✓
Over Temperature Protection	✓
Short Circuit Protection	✓
Surge Protection	✓
Inrush Current Protection	✓
Fan Failure Protection	✗
No Load Operation	✓
Cooling	135mm Double Ball Bearing Fan (RL4Z B1352512EH)
Semi-Passive Operation	✓ (Selectable)
Dimensions (W x H x D)	150 x 85 x 150mm
Weight	1.54 kg (3.4 lb)
Form Factor	ATX12V v2.52, EPS 2.92
Warranty	10 Years

Power Specifications

Rail		3.3V	5V	12V	5VSB	-12V
Max. Power	Amps	20	20	62.5	3	0.3
	Watts		120	750	15	3.6
Total Max. Power (W)			750

Cables & Connectors

Description	Cable Count	Connector Count (Total)	Gauge	In Cable Capacitors
ATX connector 20+4 pin (590mm)	1	1	16-18AWG	No
4+4 pin EPS12V (700mm)	2	2	18-22AWG	Yes
6+2 pin PCIe (700mm+150mm)	2	4	18-22AWG	Yes
6+2 pin PCIe (700mm)	2	2	18-22AWG	Yes
SATA (560mm+105mm+105mm)	3	9	18AWG	No
4-pin Molex (550mm+100mm+100mm+100mm)	1	4	18AWG	No
FDD Adapter (100mm)	1	1	22AWG	No
AC Power Cord (1400mm) - C13 coupler	1	1	18AWG	-

There are two EPS connectors on dedicated cables and six PCIe on four cables. The number of peripheral cables is adequate, but the distance between the peripheral connectors is short, at just 100 mm. Lastly, the in-cable caps in the ATX, EPS, and PCIe cables might cause trouble during cable management.

Tom's Hardware

Component Analysis

`General Data`	-
Manufacturer (OEM)	Andyson
PCB Type	Single Sided
`Primary Side`	-
Transient Filter	4x Y caps, 2x X caps, 2x CM chokes, 1x MOV
Inrush Protection	NTC Thermistor SCK205R0 (5 Ohm) & Relay
Bridge Rectifier(s)	2x
APFC MOSFETs	2x Infineon IPP60R099P7 (600V, 20A @ 100°C, Rds(on): 0.099Ohm)
APFC Boost Diode	1x CREE C3D08060A (600V, 8A @ 150°C)
Bulk Cap(s)	1x Nichicon (400V, 560uF, 2,000h @ 105°C, GG)
Main Switchers	4x MagnaChip MDF18N50 (500V, 11A @ 100°C, Rds(on): 0.27Ohm)
APFC Controller	Champion CM6500UNX
Resonant Controller	Champion CM6901T6X
Topology	Primary side: APFC, Full-Bridge & LLC converter Secondary side: Synchronous Rectification & DC-DC converters
`Secondary Side`	-
+12V MOSFETs	4x Infineon IRFB7440PbF (40V, 122A @ 100°C, Rds(on): 2.5mOhm)
5V & 3.3V	DC-DC Converters: 6x Nexperia PSMN4R0-30YLD (30V, 67A @ 100°C, Rds(on): 4mOhm) PWM Controller(s): ANPEC APW7159C
Filtering Capacitors	Electrolytic: 2x Nichicon (2-5,000h @ 105°C, HD), 4x Rubycon (4-10,000h @ 105°C, YXJ), 4x Rubycon (6-10,000h @ 105°C, ZLH), 1x Nichicon (4-10,000h @ 105°C, HE), 1x Nichicon (5-6,000h @ 105°C, HV) Polymer: 16x FPCAP
Supervisor IC	IN1S429I-DCG
Fan Model	Globe Fan RL4Z B1352512EH (135mm, 12V, 0.5A, Double Ball Bearing Fan)
`5VSB Circuit`	-
Rectifier	1x 100R25C SBR & 1x IPS ITU02N70A FET
Standby PWM Controller	ATK AT6002H
`-12V Circuit`	-
Rectifier	1x STMicroelectronis L7912CV (-12V, 3mA)

Tom's Hardware

The platform is by Andyson. The PCB is small and overpopulated. The heat sinks are large enough and the build quality is high. Finding good Japanese caps in an 80-dollar PSU and a double ball-bearing fan is a pleasant surprise.

Tom's Hardware

The transient filter has all necessary parts, including an MOV and an NTC thermistor. The latter is supported by a relay, which bypasses it once the PSU starts, to allow for higher efficiency and increased protection from inrush currents since the NTC cools down, increasing its resistance.

(Image credit: Tom's Hardware)

The pair of bridge rectifiers are sandwiched between two heat sinks.

Tom's Hardware

The APFC converter uses two Infineon FETs and a single CREE boost diode. The APFC controller is a Champion CM6500UNX IC. The bulk cap is by Nichicon and is of good quality, but its capacity is not high enough to allow for a longer than 17ms hold-up time.

Tom's Hardware

The main FETs are four MagnaChip MDF18N50, installed in a full-bridge topology. These FETs are not as good in terms of quality as the FETs used in the APFC converters, but thanks to the full-bridge topology, the stress applied to them is lower. An LLC resonant is also used to provide a notable efficiency boost. The resonant controller is a Champion CM6901T6X.

Tom's Hardware

Four Infineon FETs regulate the 12V rail. Two DC-DC converters regulate the minor rails. They use in total six FETs, and the common PWM controller is by Anpec.

Tom's Hardware

Japanese manufacturers provide the filtering caps. Besides electrolytic caps, a large number of polymer ones are also used.

Tom's Hardware

The standby PWM controller is an ATK AT6002H. On the primary side, the 5VSB rail uses a FET, and on the secondary side, we find an SBR.

Tom's Hardware

Several electrolytic and polymer caps are installed on the face of the modular PCB.

(Image credit: Tom's Hardware)

The main supervisor IC is an IN1S429I-DCG.

(Image credit: Tom's Hardware)

The -12V rail uses an STMicroelectronis L7912CV. It is nice to see a regulating IC for this rail since it provides enhanced protection.

Tom's Hardware

Soldering quality is good.

Tom's Hardware

The cooling fan uses double ball-bearings and it is of high quality.

MORE: Best Power Supplies

To learn more about our PSU tests and methodology, please check out How We Test Power Supply Units.

EVGA SuperNOVA 750 GA

Corsair RM750

Corsair CX750M

Primary Rails and 5VSB Load Regulation

The following charts show the main rails' voltage values recorded between a range of 40W up to the PSU's maximum specified load, along with the deviation (in percent). Tight regulation is an important consideration every time we review a power supply because it facilitates constant voltage levels despite varying loads. Tight load regulation, among other factors, improves the system’s stability, especially under overclocked conditions and, at the same time, it applies less stress to the DC-DC converters that many system components utilize.

Tom's Hardware

Load regulation is tight on all rails but 5VSB.

Hold-Up Time

Put simply; hold-up time is the amount of time that the system can continue to run without shutting down or rebooting during a power interruption.

Tom's Hardware

The hold-up time is notably shorter than 17 ms, which is what the ATX spec requires.

Inrush Current

Tom's Hardware

With 115V input the inrush current is low, but this is not the case with 230V.

Leakage Current

(Image credit: Tom's Hardware)

Leakage current is low.

10-110% Load Tests

These tests reveal the PSU's load regulation and efficiency levels under high ambient temperatures. They also show how the fan speed profile behaves under increased operating temperatures.

Test	12V	5V	3.3V	5VSB	DC/AC (Watts)	Efficiency	Fan Speed (RPM)	PSU Noise (dB[A])	Temps (In/Out)	PF/AC Volts
10%	4.383A	1.992A	1.971A	0.979A	75.008	86.748%	0	<6.0	44.81°C	0.964
	12.184V	5.023V	3.348V	5.108V	86.466				40.5°C	115.19V
20%	9.778A	2.989A	2.959A	1.179A	149.975	90.305%	0	<6.0	46.13°C	0.968
	12.178V	5.019V	3.347V	5.092V	166.075				41.13°C	115.18V
30%	15.522A	3.489A	3.452A	1.379A	224.987	91.233%	0	<6.0	47.07°C	0.973
	12.172V	5.017V	3.346V	5.078V	246.608				41.49°C	115.17V
40%	21.273A	3.992A	3.952A	1.58A	300.078	90.73%	1134	35.2	41.72°C	0.97
	12.169V	5.011V	3.34V	5.064V	330.737				47.82°C	115.17V
50%	26.652A	4.996A	4.946A	1.783A	374.726	90.58%	1065	32.9	41.91°C	0.97
	12.164V	5.006V	3.337V	5.049V	413.697				48.71°C	115.17V
60%	32.060A	5.999A	5.938A	1.987A	449.642	90.193%	1096	33.7	42.25°C	0.969
	12.159V	5.002V	3.335V	5.033V	498.531				49.38°C	115.16V
70%	37.477A	7.005A	6.933A	2.193A	524.608	89.596%	1187	37.0	43.02°C	0.968
	12.154V	4.998V	3.333V	5.017V	585.525				50.89°C	115.16V
80%	42.968A	8.004A	7.923A	2.299A	599.793	88.905%	1319	39.1	43.48°C	0.969
	12.146V	4.996V	3.332V	5.003V	674.648				51.86°C	115.16V
90%	48.796A	8.515A	8.407A	2.405A	674.867	88.273%	1365	39.8	44.08°C	0.97
	12.140V	4.992V	3.33V	4.99V	764.524				53.12°C	115.15V
100%	54.428A	9.022A	8.923A	3.024A	750.104	87.387%	1428	41.1	44.66°C	0.971
	12.133V	4.989V	3.329V	4.962V	858.374				54.45°C	115.15V
110%	59.942A	10.031A	10.007A	3.033A	825.137	86.289%	1501	42.5	46.55°C	0.973
	12.125V	4.985V	3.327V	4.946V	956.254				57.31°C	115.15V
CL1	0.115A	14.434A	14.279A	0A	121.317	83.648%	0	<6.0	48.48°C	0.969
	12.160V	5.003V	3.34V	5.101V	145.032				42.31°C	115.19V
CL2	0.115A	19.998A	0A	0A	101.401	82.588%	0	<6.0	50.08°C	0.967
	12.169V	5.001V	3.353V	5.12V	122.779				43.02°C	115.19V
CL3	0.115A	0A	19.729A	0A	67.398	76.927%	0	<6.0	52.25°C	0.963
	12.166V	5.027V	3.345V	5.108V	87.612				44.05°C	115.19V
CL4	61.752A	0A	0.001A	0A	749.857	87.987%	1440	41.4	45.09°C	0.972
	12.143V	5.011V	3.343V	5.064V	852.237				55.17°C	115.14V

The 50 C temperature looks to be true. I pushed the PSU tough at close to 47 C with 110% of its max power, and it didn't sweat.

20-80W Load Tests

Test	12V	5V	3.3V	5VSB	DC/AC (Watts)	Efficiency	Fan Speed (RPM)	PSU Noise (dB[A])	Temps (In/Out)	PF/AC Volts
20W	1.218A	0.497A	0.492A	0.195A	20.007	71.025%	0	<6.0	39.98°C	0.878
	12.190V	5.028V	3.35V	5.137V	28.169				36.86°C	115.19V
40W	2.682A	0.697A	0.69A	0.292A	40.005	81.568%	0	<6.0	40.92°C	0.941
	12.189V	5.025V	3.349V	5.132V	49.045				37.61°C	115.19V
60W	4.146A	0.896A	0.887A	0.39A	60.004	85.735%	0	<6.0	42.09°C	0.956
	12.187V	5.024V	3.349V	5.127V	69.988				38.36°C	115.19V
80W	5.608A	1.095A	1.084A	0.488A	79.971	87.814%	0	<6.0	43.77°C	0.964
	12.185V	5.024V	3.349V	5.122V	91.068				39.68°C	115.19V

With light loads the PSU's fan doesn't spin, even at elevated temperatures.

2% or 10W Load Test

12V	5V	3.3V	5VSB	DC/AC (Watts)	Efficiency	Fan Speed (RPM)	PSU Noise (dB[A])	Temps (In/Out)	PF/AC Volts
1.056A	0.245A	0.245A	0.051A	15.202	65.623%	0	\|6.0	27.51°C	0.847
	12.197V	5.027V	3.347V	5.143V	23.165			27.12°C	115.18V

It would be nice to see above 70% efficiency with a 2% load.

Efficiency and Power Factor

Tom's Hardware

Efficiency is high with normal and light loads, while there is much room for improvement with a 2% load. The PF readings also need improvement.

5VSB Efficiency

Test #	5VSB	DC/AC (Watts)	Efficiency	PF/AC Volts
1	0.1A	0.515W	77.514%	0.096
	5.144V	0.664W		115.11V
2	0.25A	1.285W	83.168%	0.197
	5.138V	1.545W		115.09V
3	0.55A	2.821W	84.842%	0.316
	5.128V	3.325W		115.09V
4	1A	5.111W	85.157%	0.395
	5.11V	6.002W		115.09V
5	1.5A	7.64W	84.959%	0.437
	5.092V	8.993W		115.09V
6	3A	15.118W	82.466%	0.488
	5.039V	18.332W		115.1V

Tom's Hardware

The 5VSB rail is highly efficient.

Power Consumption In Idle and Standby

Mode	12V	5V	3.3V	5VSB	Watts	PF/AC Volts
Idle	12.204V	5.021V	3.339V	5.147V	7.343	0.496
						115.18V
Standby					0.077	0.011
						115.18V

Tom's Hardware

Vampire power is high with 230V input.

Fan RPM, Delta Temperature, and Output Noise

All results are obtained between an ambient temperature of 37 to 47 degrees Celsius (98.6 to 116.6 degrees Fahrenheit).

(Image credit: Tom's Hardware)

The fan speed profile is not aggressive under tough conditions.

The following results were obtained at 30 to 32 degrees Celsius (86 to 89.6 degrees Fahrenheit) ambient temperature.

(Image credit: Tom's Hardware)

At normal operating temperatures, close to 30 degrees Celsius, the PSU's passive mode lasts up to 300W with a minimal load on the minor rails. The more you push the minor rails, the faster you activate the fan. With a 450W load, noise exceeds 30 dBA, while with more than 580W, noise goes over 35 dBA.

MORE: Best Power Supplies

Protection Features

Check out our PSUs 101 article to learn more about PSU protection features.

OCP (Cold @ 22.5°C)	12V: 100.02A (160%), 12.131V 5V: 25.4A (127%), 4.969V 3.3V: 30.2A (151%), 3.296V 5VSB: 6.8A (226.7%), 4.857V
OCP (Hot @ 36°C)	12V: 100.39A (160.62%), 12.135V 5V: 25.3A (126.5%), 4.971V 3.3V: 29.9A (149.5%), 3.299V 5VSB: 6.8A (226.7%), 4.822V
OPP (Cold @ 21°C)	1213.38W (160.03%)
OPP (Hot @ 38°C)	1218.27W (160.62%)
OTP	✓ (140°C @ 12V Heat Sink)
SCP	12V to Earth: ✓ 5V to Earth: ✓ 3.3V to Earth: ✓ 5VSB to Earth: ✓ -12V to Earth: ✓
PWR_OK	Accurate but lower than 16ms
NLO	✓
SIP	Surge: MOV Inrush: NTC Thermistor

The OCP triggering point at 12V is set super high, and the same goes for OPP. Andyson and EVGA did this to cope with power spikes. This is not ideal, though, because such high triggering points practically render the PSU's protection features useless. On the contrary, the over-temperature protection is reasonably set.

DC Power Sequencing

Tom's Hardware

No problems here since the 3.3V rail is always lower than the other two.

Cross Load Tests

Load Regulation Charts

Tom's Hardware

Efficiency Graph

(Image credit: Tom's Hardware)

Ripple Graphs

The lower the power supply's ripple, the more stable the system will be and less stress will also be applied to its components.

Tom's Hardware

Infrared Images

Tom's Hardware

We didn't notice alarmingly high temperatures at the PSU's internals, and given the quality of the parts, the fan speed profile could be more relaxed. Perhaps EVGA and Andyson wanted to stay on the safe side because of the ten-year warranty.

MORE: Best Power Supplies

Advanced Transient Response Tests

For details about our transient response testing, please click here.

Advanced Transient Response at 20% – 20ms

Voltage	Before	After	Change	Pass/Fail
12V	12.178V	11.991V	1.53%	Pass
5V	5.017V	4.896V	2.41%	Pass
3.3V	3.343V	3.196V	4.41%	Pass
5VSB	5.092V	5.037V	1.09%	Pass

Advanced Transient Response at 20% – 10ms

Voltage	Before	After	Change	Pass/Fail
12V	12.179V	12.061V	0.97%	Pass
5V	5.016V	4.902V	2.28%	Pass
3.3V	3.342V	3.192V	4.50%	Pass
5VSB	5.092V	5.046V	0.90%	Pass

Advanced Transient Response at 20% – 1ms

Voltage	Before	After	Change	Pass/Fail
12V	12.179V	12.079V	0.82%	Pass
5V	5.016V	4.899V	2.33%	Pass
3.3V	3.342V	3.192V	4.50%	Pass
5VSB	5.092V	5.017V	1.48%	Pass

Advanced Transient Response at 50% – 20ms

Voltage	Before	After	Change	Pass/Fail
12V	12.164V	12.048V	0.95%	Pass
5V	5.007V	4.889V	2.37%	Pass
3.3V	3.337V	3.181V	4.67%	Pass
5VSB	5.048V	4.977V	1.41%	Pass

Advanced Transient Response at 50% – 10ms

Voltage	Before	After	Change	Pass/Fail
12V	12.165V	12.008V	1.29%	Pass
5V	5.006V	4.883V	2.46%	Pass
3.3V	3.337V	3.184V	4.58%	Pass
5VSB	5.048V	4.999V	0.97%	Pass

Advanced Transient Response at 50% – 1ms

Voltage	Before	After	Change	Pass/Fail
12V	12.164V	12.030V	1.10%	Pass
5V	5.006V	4.886V	2.39%	Pass
3.3V	3.336V	3.210V	3.76%	Pass
5VSB	5.048V	4.989V	1.18%	Pass

Tom's Hardware

Transient response is good enough at 12V, but mediocre on the minor rails.

Turn-On Transient Tests

Tom's Hardware

There is a significant voltage step at 5VSB, which doesn't worry us. The 12V slopes look better.

Power Supply Timing Tests

PSU Timings Table
T1 (Power-on time) & T3 (PWR_OK delay)
Load	T1	T3
20%	38ms	133ms
100%	38ms	133ms

Tom's Hardware

The PWR_OK delay is within the 100-150ms region, so the PSU supports the alternative sleep mode recommended by the ATX spec.

Ripple Measurements

The ripple limits, according to the ATX specification, are 120mV (+12V) and 50mV (5V, 3.3V, and 5VSB).

Test	12V	5V	3.3V	5VSB	Pass/Fail
10% Load	9.1 mV	5.4 mV	4.7 mV	6.3 mV	Pass
20% Load	11.1 mV	6.2 mV	5.9 mV	6.9 mV	Pass
30% Load	12.2 mV	6.4 mV	6.8 mV	7.6 mV	Pass
40% Load	13.0 mV	7.7 mV	8.6 mV	8.8 mV	Pass
50% Load	13.9 mV	8.3 mV	9.9 mV	8.8 mV	Pass
60% Load	15.0 mV	8.4 mV	10.4 mV	8.3 mV	Pass
70% Load	16.1 mV	9.5 mV	11.7 mV	8.6 mV	Pass
80% Load	16.8 mV	9.9 mV	16.7 mV	9.4 mV	Pass
90% Load	17.5 mV	10.6 mV	17.9 mV	11.9 mV	Pass
100% Load	24.4 mV	11.8 mV	21.3 mV	16.5 mV	Pass
110% Load	25.4 mV	13.5 mV	22.2 mV	30.7 mV	Pass
Crossload 1	14.3 mV	10.3 mV	12.6 mV	10.1 mV	Pass
Crossload 2	11.2 mV	10.5 mV	5.0 mV	8.6 mV	Pass
Crossload 3	10.3 mV	6.1 mV	12.5 mV	8.0 mV	Pass
Crossload 4	23.6 mV	10.3 mV	16.3 mV	15.3 mV	Pass

Tom's Hardware

Ripple suppression is good but the competition is strong in this section.

Ripple At Full Load

Tom's Hardware

Ripple At 110% Load

Tom's Hardware

Ripple At Cross-Load 1

Tom's Hardware

Ripple At Cross-Load 4

Tom's Hardware

EMC Pre-Compliance Testing – Average & Quasi-Peak EMI Detector Results

Electromagnetic Compatibility (EMC) is the ability of a device to operate properly in its environment without disrupting the proper operation of other nearby devices.

΅We use TekBox's EMCview to conduct our EMC pre-compliance testing.

(Image credit: Tom's Hardware)

We didn't notice any high spurs with the average and peak EMI detectors.

MORE: Best Power Supplies

Performance Rating

(Image credit: Tom's Hardware)

The overall performance is not that far from MSI's offering, but other units like the Corsair RM750x and the EVGA 750 G6 are better.

Noise Rating

The graph below depicts the cooling fan's average noise over the PSU's operating range, with an ambient temperature between 30 to 32 degrees Celsius (86 to 89.6 degrees Fahrenheit).

(Image credit: Tom's Hardware)

MSI's offering might have about the same performance but achieves a significant win in the noise output section.

Efficiency Rating

The following graph shows the PSU's average efficiency throughout its operating range with an ambient temperature close to 30 degrees Celsius.

(Image credit: Tom's Hardware)

Average efficiency is close to the 750 G6.

Power Factor Rating

The following graphs show the PSU's average power factor reading throughout its operating range with an ambient temperature close to 30 degrees Celsius and 115V/230V voltage input.

Tom's Hardware

The APFC converter's performance is low, with both voltage inputs that we tried.

MORE: Best Power Supplies

At $80, the EVGA SuperNOVA 750 GA is a terrific deal. It achieves high enough overall performance, has tight load regulation and decent ripple suppression and isn't noisy under normal conditions. Moreover, it utilizes a fully modular cable design, and its build quality is solid. We found Japanese caps everywhere, along with a quality double ball-bearing fan, and this is why EVGA supports this platform with a hefty, ten-year warranty. The significant downsides are the lower-than-17 ms hold-up time and the not correctly set protection features.

(Image credit: Tom's Hardware)

To retain compatibility with the newer generation GPUs, which deliver high power spikes, EVGA looks to have asked Andyson to heavily modify the protection features of the 750 GA, rendering them useless. You know that something is way off when you need to apply more than 1200W of load to a 750W PSU to force it to shut down. The unit might have survived our tests, but its lifespan will be greatly reduced with repetitive overloads like the aforementioned. If you want to spend more, you could also take a look at the Corsair RM750x or the EVGA 750 G6.

MORE: Best Power Supplies

EVGA SuperNOVA 850 G7 Power Supply Review

Aris Mpitziopoulos — Thu, 08 Sep 2022 16:29:00 +0000

It's time for a new leadership power supply: The EVGA SuperNOVA 850 G7 outperforms the mighty Corsair RM850x (2021), earning a place in our list of the best PSUs. It might not have a 12+4 pin PCIe connector yet, but it offers top performance and super-compact dimensions. The 850W G7 and G6 units have the same price tags. You should go with the first if you are interested in pure performance or choose the second if you care more about noise output.

FSP is providing the platforms. The G2 and G3 lines were by Super Flower, G5 by FSP and G6 used the Seasonic Focus platform. The power density scores go through the roof with only 130 mm in length for all G7 units. So far, we haven't encountered such small ATX form-factor PSUs, especially in 1,000 W and 850 W capacities. There is no room to go any smaller than that, without sacrificing the 120 mm fan for a smaller one, which would lead to increased noise. If you need smaller PSUs, you should look at the SFX-L and SFX form factors.

Tom's Hardware

The exterior design looks nice with a light blue fan grille, fully modular cable design and compact dimensions.

Tom's Hardware

On one of the PSU's sides you will find five LED indicators, which depict the load level.

Tom's Hardware

Specifications

Manufacturer (OEM)	FSP
Max. DC Output	850W
Efficiency	80 PLUS Gold, Cybenetics Platinum (89-91%)
Noise	Cybenetics Standard++ (30-35 dB[A])
Modular	✓
Intel C6/C7 Power State Support	✓
Operating Temperature (Continuous Full Load)	0 - 50°C
Over Voltage Protection	✓
Under Voltage Protection	✓
Over Power Protection	✓
Over Current (+12V) Protection	✓
Over Temperature Protection	✓
Short Circuit Protection	✓
Surge Protection	✓
Inrush Current Protection	✓
Fan Failure Protection	✗
No Load Operation	✓
Cooling	120mm Fluid Dynamic Bearing Fan (MGA12012XF-O25)
Semi-Passive Operation	✓(selectable)
Dimensions (W x H x D)	150 x 85 x 130mm
Weight	1.72 kg (3.79 lb)
Form Factor	ATX12V v2.52, EPS 2.92
Warranty	10 Years

Power Specifications

Rail		3.3V	5V	12V	5VSB	-12V
Max. Power	Amps	24	24	70.8	3	0.5
	Watts		120	850	15	6
Total Max. Power (W)			850

Cables and Connectors

Description	Cable Count	Connector Count (Total)	Gauge	In Cable Capacitors
ATX connector 20+4 pin (600mm)	1	1	18-22AWG	Yes
4+4 pin EPS12V (700mm)	2	2	18AWG	No
6+2 pin PCIe (700mm+150mm)	2	4	18AWG	No
6+2 pin PCIe (700mm)	2	2	18AWG	No
SATA (550mm+100mm+100mm)	3	9	18AWG	No
4-pin Molex (550mm+100mm+100mm+100mm)	1	4	18AWG	No
FDD Adapter (100mm)	1	1	22AWG	No
AC Power Cord (1390mm) - C13 coupler	1	1	16AWG	-

All the cables are long, and the amount of connectors is sufficient. With two EPS and six PCIe connectors, the PSU won't have any problems delivering its full power. There are also plenty of peripheral connectors, but the distance between them is short.

Tom's Hardware

(Image credit: Tom's Hardware)

FSP used in-cable caps in the ATX cable for better ripple suppression, so don't expect this cable to be highly flexible.

Component Analysis

`General Data`	-
Manufacturer (OEM)	FSP
PCB Type	Double Sided
`Primary Side`	-
Transient Filter	4x Y caps, 2x X caps, 2x CM chokes, 1x MOV
Inrush Protection	NTC Thermistor SCK-056 (5 Ohm) & Relay
Bridge Rectifier(s)	2x
APFC MOSFETs	3x
APFC Boost Diode	1x
Bulk Cap(s)	1x Nippon Chemi-Con (420V, 470uF, 2,000h @ 105°C, KHE) & 1x TK (420V, 330uF, 2,000h @ 105°C, LGW)
Main Switchers	2x Infineon IPP60R120P7 (600V, 16A @ 100°C, Rds(on): 0.12Ohm)
IC Driver	1x Novosense NSi6602
APFC Controller	Infineon ICE2PCS02
Resonant Controller	Champion CM6901T2X
Topology	Primary side: APFC, Half-Bridge & LLC converter Secondary side: Synchronous Rectification & DC-DC converters
`Secondary Side`	-
+12V MOSFETs	no info
5V & 3.3V	DC-DC Converters: 6x Infineon BSC0901NS (30V, 94A @ 100°C, Rds(on): 1.9mOhm) PWM Controller(s): ANPEC APW7159C
Filtering Capacitors	Electrolytic: 5x Rubycon (3-6,000h @ 105°C, YXG) Polymer: 17x Nippon Chemi-Con, 6x NIC
Supervisor IC	Weltrend WT7527RA (OCP, OVP, UVP, SCP, PG)
Fan Controller	Microchip PICF15324
Fan Model	Protechnic Electric MGA12012XF-O25 (120mm, 12V, 0.52A, Fluid Dynamic Bearing Fan)
`5VSB Circuit`	-
Rectifier	1x NIKO-SEM P1006BD (60V, 42A @ 100°C, Rds(on): 10mOhm) FET
Standby PWM Controller	Power Integrations INN2603K

Tom's Hardware

The OEM behind this platform is FSP, and it looks nice! The only problem is that because of the overpopulated, tiny PCB, we had a tough time identifying all parts without heavy de-soldering, which could destroy the PSU, and we need it for future reference in case we have to re-test it. FSP used good parts, and the design is up to date, with a half-bridge topology and an LLC resonant converter on the primary side. The secondary side hosts a synchronous rectification scheme for the 12V rail and DC-DC converters for the minor rails.

Tom's Hardware

The EMI filter is complete so that EMI emissions won't be an issue, both incoming and outgoing. We also found an MOV for handling voltage surges and an NTC thermistor and bypass relay combo, for protection against high inrush currents.

(Image credit: Tom's Hardware)

The pair of bridge rectifiers is sandwiched between two heat sinks.

Tom's Hardware

The APFC converter uses three FETs and a single boost diode. The PFC controller is installed on a vertical board, an Infineon ICE2PCS02. The same board also hosts an operational amplifier (op-amp). The bulk caps are a weird mix of Chemi-Con and TK. Usually, identical bulk caps are used, but in this case, we find two different. A good reason for this can be the shortage of caps.

Tom's Hardware

Two Infineon IPP60R120P7 are the main FETs, configured in a half-bridge topology. Their driver IC is a Novosense NSi6602, and the LLC resonant converter is a Champion CM6901T2X IC. Precisely the same parts are used in the 1000 G7 unit.

Tom's Hardware

The FETs that regulate the 12V rail are installed on a board next to the main transformer to minimize energy losses. The DC-DC converters are installed on another vertical board.

Tom's Hardware

The mix of Rubycon electrolytic and Chemi-Con and NIC polymer caps, is the best we could ask for this PSU.

Tom's Hardware

The standby PWM controller is by Power Integrations, and a NIKO-SEM P1006BD FET is the rectifier on the secondary side of the 5VSB circuit. The same 5VSB circuit is used on the 1000 G7 model.

Tom's Hardware

Many Chemi-Con polymer caps are installed on the modular PCB for ripple filtering purposes.

(Image credit: Tom's Hardware)

The main supervisor IC is a Weltrend WT7527RA.

Tom's Hardware

Soldering quality is decent overall, but we noticed some bad spots, which don't affect the unit's performance, though.

Tom's Hardware

Protechnic electric is a force in fans, so we are pleased to see one of its products in this unit. The fluid dynamic bearing ensures the fan's longevity and low noise output under average speeds. Keep in mind, though, that this fan is strong, so it won't be quiet at high speeds.

MORE: Best Power Supplies

To learn more about our PSU tests and methodology, please check out How We Test Power Supply Units.

Corsair RM850x (2021)

EVGA SuperNOVA 850 G6

Seasonic FOCUS GX-850

Primary Rails And 5VSB Load Regulation

The following charts show the main rails' voltage values recorded between a range of 40W up to the PSU's maximum specified load, along with the deviation (in percent). Tight regulation is an important consideration every time we review a power supply because it facilitates constant voltage levels despite varying loads. Tight load regulation also, among other factors, improves the system’s stability, especially under overclocked conditions. At the same time, it applies less stress to the DC-DC converters that many system components utilize.

Tom's Hardware

Load regulation is satisfactory on all rails.

Hold-Up Time

Put simply; hold-up time is the amount of time that the system can continue to run without shutting down or rebooting during a power interruption.

Tom's Hardware

The hold-up time is long and the power ok signal is accurate. The "PWR OK Inactive to DC Loss Delay" signal could be shorter, though.

Inrush Current

Inrush current, or switch-on surge, refers to the maximum, instantaneous input current drawn by an electrical device when it is first turned on. A large enough inrush current can cause circuit breakers and fuses to trip. It can also damage switches, relays and bridge rectifiers. As a result, the lower the inrush current of a PSU right as it is turned on, the better.

Tom's Hardware

Inrush current is high with both voltage inputs, 115V and 230V, that we tried.

Leakage Current

(Image credit: Tom's Hardware)

Leakage current is low.

10-110% Load Tests

These tests reveal the PSU's load regulation and efficiency levels under high ambient temperatures. They also show how the fan speed profile behaves under increased operating temperatures.

Test	12V	5V	3.3V	5VSB	DC/AC (Watts)	Efficiency	Fan Speed (RPM)	PSU Noise (dB[A])	Temps (In/Out)	PF/AC Volts
10%	5.158A	1.943A	1.943A	0.987A	85.022	87.313%	0	<6.0	44.85°C	0.97
	12.295V	5.147V	3.397V	5.069V	97.377				40.44°C	115.13V
20%	11.321A	2.917A	2.918A	1.186A	170.001	90.757%	0	<6.0	45.93°C	0.992
	12.286V	5.143V	3.394V	5.059V	187.311				41.08°C	115.11V
30%	17.834A	3.404A	3.407A	1.388A	255.027	91.83%	0	<6.0	46.85°C	0.992
	12.278V	5.142V	3.391V	5.046V	277.72				41.68°C	115.08V
40%	24.361A	3.892A	3.896A	1.589A	340.136	92.113%	0	<6.0	47.71°C	0.994
	12.271V	5.14V	3.389V	5.035V	369.258				41.94°C	115.06V
50%	30.554A	4.867A	4.874A	1.792A	425.218	91.929%	0	<6.0	48.39°C	0.995
	12.264V	5.137V	3.386V	5.023V	462.547				42.33°C	115.04V
60%	36.705A	5.844A	5.853A	1.994A	509.705	91.551%	0	<6.0	49.23°C	0.995
	12.257V	5.135V	3.383V	5.016V	556.746				42.57°C	115.02V
70%	42.988A	6.823A	6.844A	2.201A	595.029	90.901%	1288	31.7	43.21°C	0.995
	12.233V	5.132V	3.376V	5V	654.592				50.36°C	114.99V
80%	49.238A	7.803A	7.831A	2.306A	679.873	90.366%	1564	36.7	43.76°C	0.994
	12.225V	5.128V	3.371V	4.989V	752.364				52.04°C	114.96V
90%	55.891A	8.295A	8.314A	2.411A	765.312	89.716%	1976	43.3	44.81°C	0.994
	12.216V	5.126V	3.368V	4.98V	853.044				54.01°C	114.94V
100%	62.293A	8.787A	8.828A	3.028A	850.115	88.986%	2332	47.7	46.06°C	0.993
	12.206V	5.124V	3.365V	4.955V	955.33				56.16°C	114.91V
110%	68.574A	9.769A	9.91A	3.033A	934.686	88.128%	2693	51.2	47.3°C	0.993
	12.197V	5.12V	3.36V	4.948V	1060.586				58.19°C	114.88V
CL1	0.114A	14.089A	14.093A	0A	121.339	85.678%	1248	30.2	42.57°C	0.984
	12.287V	5.127V	3.385V	5.085V	141.629				48.03°C	115.12V
CL2	0.114A	23.473A	0A	0A	121.446	84.379%	1231	30.2	43.97°C	0.985
	12.289V	5.114V	3.399V	5.091V	143.926				50.17°C	115.12V
CL3	0.114A	0A	23.426A	0A	80.597	78.766%	870	19.5	44.65°C	0.972
	12.280V	5.149V	3.381V	5.083V	102.323				52.73°C	115.13V
CL4	69.588A	0A	0A	0A	849.771	89.583%	2088	45.9	45.35°C	0.993
	12.212V	5.14V	3.378V	5.045V	948.581				55.31°C	114.91V

The PSU can handle harsh operating conditions, that is, increased loads and temperatures, but don't expect it to be quiet.

20-80W Load Tests

Test	12V	5V	3.3V	5VSB	DC/AC (Watts)	Efficiency	Fan Speed (RPM)	PSU Noise (dB[A])	Temps (In/Out)	PF/AC Volts
20W	1.209A	0.486A	0.485A	0.196A	20.021	70.679%	0	<6.0	40.04°C	0.806
	12.299V	5.149V	3.399V	5.094V	28.329				36.93°C	115.15V
40W	2.660A	0.68A	0.68A	0.295A	40.019	81.124%	0	<6.0	40.92°C	0.913
	12.298V	5.149V	3.399V	5.089V	49.331				37.59°C	115.14V
60W	4.110A	0.874A	0.874A	0.393A	60.018	85.315%	0	<6.0	42.84°C	0.95
	12.297V	5.149V	3.399V	5.085V	70.347				39.13°C	115.14V
80W	5.560A	1.068A	1.068A	0.492A	79.989	87.556%	0	<6.0	43.35°C	0.968
	12.296V	5.148V	3.398V	5.081V	91.357				39.39°C	115.13V

The unit achieves high efficiency under light loads, with minimal noise output because the fan doesn't spin.

2% or 10W Load Test

12V	5V	3.3V	5VSB	DC/AC (Watts)	Efficiency	Fan Speed (RPM)	PSU Noise (dB[A])	Temps (In/Out)	PF/AC Volts
1.226A	0.25A	0.25A	0.052A	17.474	68.51%	0	<6.0	37.99°C	0.778
	12.299V	5.15V	3.4V	5.099V	25.508			35.45°C	115.15V

The PSU exceeds 60% efficiency with a 2% load. It would be nice to measure above 70%, though, in this test scenario.

Efficiency & Power Factor

Tom's Hardware

The platform is highly efficient with normal loads, also achieving good results with light loads. The competition is tough at light and super-light loads, so FSP could improve the platform in these sectors.

5VSB Efficiency

Test #	5VSB	DC/AC (Watts)	Efficiency	PF/AC Volts
1	0.1A	0.51W	77.34%	0.065
	5.093V	0.659W		115.15V
2	0.25A	1.273W	80.723%	0.146
	5.091V	1.577W		115.15V
3	0.55A	2.796W	82.185%	0.262
	5.082V	3.402W		115.15V
4	1A	5.074W	80.724%	0.362
	5.073V	6.285W		115.14V
5	1.5A	7.596W	79.19%	0.414
	5.063V	9.591W		115.15V
6	3.001A	15.056W	78.672%	0.478
	5.018V	19.137W		115.14V

Tom's Hardware

The 5VSB rail is efficient.

Power Consumption In Idle And Standby

Mode	12V	5V	3.3V	5VSB	Watts	PF/AC Volts
Idle	12.302V	5.151V	3.402V	5.103V	7.16	0.411
						115.15V
Standby					0.059	0.006
						115.15V

Tom's Hardware

We would like to see below 0.1W vampire power with 230V input.

Fan RPM, Delta Temperature, And Output Noise

All results are obtained between an ambient temperature of 37 to 47 degrees Celsius (98.6 to 116.6 degrees Fahrenheit).

(Image credit: Tom's Hardware)

The fan's speed is not aggressive considering the test conditions, the tiny and overpopulated PCB, and the high power output.

The following results were obtained at 30 to 32 degrees Celsius (86 to 89.6 degrees Fahrenheit) ambient temperature.

(Image credit: Tom's Hardware)

Some spots in the PSU's noise map at normal operating temperatures indicate that the fan speed increased notably to cope with heat build-up. This is why it is preferable to have the fan spinning at all times rather than employing a semi-passive operation mode. With more than 690W, the PSU exceeds 30 dBA; with 20-30W more, it goes over 35 dBA. With a larger PCB, allowing for more airflow between parts, and a 135-140mm fan, things could be better in noise output.

MORE: Best Power Supplies

Protection Features

Check out our PSUs 101 article to learn more about PSU protection features.

OCP (Cold @ 31°C)	12V: 101.2A (142.97%), 12.160V 5V: 30.7A (127.92%), 5.101V 3.3V: 29.1A (121.25%), 3.367V 5VSB: 4.3A (143.33%), 4.978V
OCP (Hot @ 46°C)	12V: 98.2A (138.71%), 12.179V 5V: 28.3A (117.92%), 5.106V 3.3V: 29A (120.83%), 3.369V 5VSB: 4.3A (143.33%), 4.983V
OPP (Cold @ 33°C)	1230.95W (144.82%)
OPP (Hot @ 43°C)	1196.11W (140.72%)
OTP	✓ (142°C @ 12V Secondary Side)
SCP	12V to Earth: ✓ 5V to Earth: ✓ 3.3V to Earth: ✓ 5VSB to Earth: ✓ -12V to Earth: ✓
PWR_OK	Proper Operation
NLO	✓
SIP	Surge: MOV Inrush: NTC Thermistor & Bypass Relay

OCP at 12V and OPP are highly set, most likely to cope with power spikes. That is not the best way to do it since it makes the PSU's protection features less effective. On the minor rails, the OCP triggering points are correctly set.

DC Power Sequencing

Tom's Hardware

No problems here since the 3.3V rail is always lower than the other two.

Cross Load Tests

Load Regulation Charts

Tom's Hardware

Efficiency Graph

(Image credit: Tom's Hardware)

Ripple Graphs

The lower the power supply's ripple, the more stable the system will be and less stress will also be applied to its components.

Tom's Hardware

Infrared Images

Tom's Hardware

The board holding the 12V FETs is the spot reporting the highest temperatures. Still, we didn't notice alarmingly high temperatures, thanks to the highly efficient platform that minimizes energy losses.

MORE: Best Power Supplies

Advanced Transient Response Tests

For details about our transient response testing, please click here.

Advanced Transient Response at 20% – 20ms

Voltage	Before	After	Change	Pass/Fail
12V	12.280V	12.178V	0.83%	Pass
5V	5.143V	5.036V	2.08%	Pass
3.3V	3.392V	3.278V	3.38%	Pass
5VSB	5.056V	5.034V	0.43%	Pass

Advanced Transient Response at 20% – 10ms

Voltage	Before	After	Change	Pass/Fail
12V	12.282V	12.211V	0.57%	Pass
5V	5.144V	5.045V	1.93%	Pass
3.3V	3.393V	3.276V	3.45%	Pass
5VSB	5.057V	5.040V	0.34%	Pass

Advanced Transient Response at 20% – 1ms

Voltage	Before	After	Change	Pass/Fail
12V	12.282V	12.189V	0.75%	Pass
5V	5.144V	5.036V	2.10%	Pass
3.3V	3.393V	3.293V	2.95%	Pass
5VSB	5.057V	5.034V	0.45%	Pass

Advanced Transient Response at 50% – 20ms

Voltage	Before	After	Change	Pass/Fail
12V	12.254V	12.182V	0.58%	Pass
5V	5.136V	5.037V	1.93%	Pass
3.3V	3.382V	3.265V	3.45%	Pass
5VSB	5.025V	5.006V	0.38%	Pass

Advanced Transient Response at 50% – 10ms

Voltage	Before	After	Change	Pass/Fail
12V	12.254V	12.174V	0.65%	Pass
5V	5.137V	5.030V	2.09%	Pass
3.3V	3.382V	3.268V	3.38%	Pass
5VSB	5.024V	5.006V	0.36%	Pass

Advanced Transient Response at 50% – 1ms

Voltage	Before	After	Change	Pass/Fail
12V	12.255V	12.187V	0.55%	Pass
5V	5.137V	5.036V	1.96%	Pass
3.3V	3.383V	3.275V	3.20%	Pass
5VSB	5.025V	5.006V	0.38%	Pass

Tom's Hardware

Transient response is good at 12V, 3.3V and 5VSB and satisfactory at 5V.

Turn-On Transient Tests

Tom's Hardware

We didn't notice any notable voltage overshoots or voltage spikes, in these tests.

Power Supply Timing Tests

There are several signals generated by the power supply, which need to be within specified, by the ATX spec, ranges. If they are not, there can be compatibility issues with other system parts, especially mainboards. From year 2020, the PSU's Power-on time (T1) has to be lower than 150ms and the PWR_OK delay (T3) from 100 to 150ms, to be compatible with the Alternative Sleep Mode.

PSU Timings Table
T1 (Power-on time) & T3 (PWR_OK delay)
Load	T1	T3
20%	47ms	127ms
100%	42ms	131ms

Tom's Hardware

The PWR_OK delay is within the 100-150ms region, so the PSU supports the alternative sleep mode recommended by the ATX spec.

Ripple Measurements

The ripple limits, according to the ATX specification, are 120mV (+12V) and 50mV (5V, 3.3V, and 5VSB).

Test	12V	5V	3.3V	5VSB	Pass/Fail
10% Load	6.3 mV	4.9 mV	6.7 mV	8.5 mV	Pass
20% Load	7.5 mV	5.3 mV	6.7 mV	10.0 mV	Pass
30% Load	7.6 mV	5.8 mV	7.4 mV	9.9 mV	Pass
40% Load	8.0 mV	6.7 mV	7.4 mV	9.3 mV	Pass
50% Load	8.7 mV	6.7 mV	7.8 mV	10.3 mV	Pass
60% Load	9.3 mV	7.2 mV	8.8 mV	12.1 mV	Pass
70% Load	9.9 mV	7.3 mV	9.6 mV	13.3 mV	Pass
80% Load	10.8 mV	8.1 mV	9.7 mV	13.5 mV	Pass
90% Load	11.9 mV	9.9 mV	11.2 mV	13.9 mV	Pass
100% Load	18.6 mV	10.4 mV	11.9 mV	16.3 mV	Pass
110% Load	20.4 mV	14.0 mV	12.7 mV	15.6 mV	Pass
Crossload 1	10.2 mV	7.5 mV	8.6 mV	6.8 mV	Pass
Crossload 2	6.8 mV	6.3 mV	7.7 mV	5.5 mV	Pass
Crossload 3	6.9 mV	6.4 mV	8.2 mV	4.1 mV	Pass
Crossload 4	18.4 mV	9.8 mV	10.5 mV	6.6 mV	Pass

Tom's Hardware

Ripple suppression is good on all rails.

Ripple At Full Load

Tom's Hardware

Ripple At 110% Load

Tom's Hardware

Ripple At Cross-Load 1

Tom's Hardware

Ripple At Cross-Load 4

Tom's Hardware

EMC Pre-Compliance Testing – Average and Quasi-Peak EMI Detector Results

Electromagnetic Compatibility (EMC) is the ability of a device to operate properly in its environment without disrupting the proper operation of other nearby devices.

Electromagnetic Interference (EMI) is the electromagnetic energy a device emits, and it can cause problems in other nearby devices if too high. For example, it can cause increased static noise in your headphones or/and speakers.

΅We use TekBox's EMCview to conduct our EMC pre-compliance testing.

(Image credit: Tom's Hardware)

EMI emissions are low, with both the average and peak EMI detectors.

MORE: Best Power Supplies

Performance Rating

(Image credit: Tom's Hardware)

Overall performance tops the chart.

Noise Rating

The graph below depicts the cooling fan's average noise over the PSU's operating range, with an ambient temperature between 30 to 32 degrees Celsius (86 to 89.6 degrees Fahrenheit).

(Image credit: Tom's Hardware)

The EVGA 850 G7 is not noisy under average loads and operating temperatures, but it will get loud if you push it hard. This affects its overall noise output.

Efficiency Rating

The following graph shows the PSU's average efficiency throughout its operating range with an ambient temperature close to 30 degrees Celsius.

(Image credit: Tom's Hardware)

The 850 G7 is highly efficient.

Power Factor Rating

The following graphs show the PSU's average power factor reading throughout its operating range with an ambient temperature close to 30 degrees Celsius and 115V/230V voltage input.

Tom's Hardware

The APFC converter has good performance with 115V, while there is room for improvement with 230V input.

MORE: Best Power Supplies

The EVGA 850 G7 is one of the best 850W units the market today. The Corsair RM850x (2021) is close in performance and has an advantage in noise output, while the 850 G6 loses notably in performance but achieves a significant win in noise output. The performance FSP delivered out of such a compact platform is impressive. We can't stop thinking, though, about the improvement in noise output with a larger PCB and cooling fan. Downsizing high-capacity PSUs at that degree comes at a cost, of increased noise output, under harsh conditions.

(Image credit: Tom's Hardware)

Besides the tiny footprint, the top performance, the fully modular cable design and the high build quality, EVGA threw in LED load indicators on one of the PSU's sides. Someone can argue that it would be better if an external PCB hosted these LED indicators, which the user could install on top of their desk to be easily accessible. This sounds good, but it would also likely increase the cost. We also noticed is that FSP used two different bulk caps in this unit, with the second provided by a less known manufacturer. We would like to see both caps have the same quality.

The EVGA 850 G7 brings up memories of the legendary G2 units, which were among the best PSUs that EVGA ever offered. We are eager to see the upgraded G7 versions featuring the new 12+4 pin PCIe connectors and ATX 3.0 support.

MORE: Best Power Supplies

Intel Core i9-12900K
MSI Pro Z690-A WiFi DDR4
Corsair 2x16GB DDR4-3600 CL16
Crucial P5 Plus 2TB
Cooler Master MWE 1250 V2 Gold
Cooler Master PL360 Flux
Cooler Master HAF500
Windows 11 Pro 64-bit

Steep RTX 30-Series Discounts Suggest Impending 40-Series Launch

editors@tomshardware.com (Aaron Klotz) — Mon, 08 Aug 2022 17:57:44 +0000

Hot on the heels of Nvidia's devastating earnings report, the company appears to be desperately trying to sell its existing RTX 20- and 30-series GPUs as quickly as possible. Graphics card prices have plummeted this week, with the EVGA store and Newegg Shuffle leading the charge. This suggests that Nvidia is readying up its RTX 40-series launch for the near future, and its partners are looking to dump inventory.

At the EVGA store, many of Nvidia's RTX 30-series GPU models are on sale. Flagship AIB partner models such as the 3080 FTW3 Ultra Hydro Copper, and other FTW3 models, are now priced incredibly closely, often at the GPU's official MSRP. Several other models are priced well under MSRP, including the RTX 3090 Ti, 3090, and 3080 Ti. EVGA is discounting most of its 3090 tier cards as much as 42%.

Some of the discounts have changed over the past weekend, and some cards are now out of stock. Perhaps the most surprising discount comes from the RTX 2060 6GB. Prices for this GPU are now as low as $239.99, which is just $5 more than its smaller — and significantly slower, if you check our GPU benchmarks — brother, the GTX 1660 Super. In our testing, the 2060 was about 25% faster. It's also worth mentioning that the RTX 2060 12GB variant was selling for just $249 yesterday, though it's now at $319.

The Newegg Shuffle also had some great deals, with an Asus RTX 3080 Ti Tuf model discounted as low as $833, a Zotac RTX 3080 Ti Extreme model for $839, an RTX 3070 Tuf for $515, and an RX 6600 Dual for just $249. All prices include promo codes featured on each listing, though of course you had to get lucky to get the GPUs at those prices.

(Image credit: Newegg)

Nvidia's Preliminary Q2'23 Financial Losses

Nvidia's financial revenue stream for 2022 is not looking good. In a recent preliminary financial results report, the company announced a whopping 44% quarterly loss in gaming revenue. Total revenue was down from $8.1 billion to $6.7 billion, a 19% quarterly decline, which gaming revenue plummeted from $2.04 billion to 3.64 billion.

This impending loss of funds is the result of the cryptocurrency crash of 2022. That has resulted in significantly reduced demand for Nvidia's GPUs, right when supply started to improve, leaving the company with excess soon-to-be-outdated parts. Nvidia CEO Jensen Huang announced channel price adjustments for its gaming partners, to combat the sting of reduced GPU demand, which is likely what we're seeing from EVGA and others right now.

Nvidia still appears to have a 2022 release window for the RTX 40-series, probably sooner than later, which also accounts for the highly discounted 30-series prices. The RTX 40-series has been rumored for a Q4 of 2022, possibly as early as September, so it makes sense to clear out the existing inventory in preparation for those cards.

But don't expect cheap RTX 40-series GPU pricing out of the gate. Many gamers are likely still looking to upgrade, and the new Nvidia Ada architecture sounds very promising. Nvidia pre-ordered a lot of TSMC 5nm wafers at the height of the global GPU shortage in 2021, but those wafers could end up being allocated to data center and mobile parts in order to keep average selling prices up. However, even the launch of RTX 40-series apparently won't make up for the pending quarterly results.

EVGA Slashes $1,000 Off GeForce RTX 3090 Ti GPUs, Starting at $1,149

Zhiye Liu — Sun, 07 Aug 2022 18:09:53 +0000

In an unpredecented move, EVGA (via FrameChasers) has slashed $1,000 of the brand's custom GeForce RTX 3090 Ti models. The GA102-powered graphics cards is one of the best graphics cards right now, which debuted early this year at $1,999.

The GeForce RTX 3090 Ti FTW3 Gaming, which is the least expensive SKU on EVGA's lineup, now sells for $1,149, 47% cheaper than what the company used to sell it for. At $1,149, the GeForce RTX 3090 Ti FTW3 Ultra Gaming is also one of the most affordable custom GeForce RTX 3090 Ti models on the market. It's no push-over on the performance side, either. The graphics card flaunts a 1,890 MHz boost clock.

If you need something with a bit more firepower, the GeForce RTX 3090 Ti FTW3 Ultra Gaming retails for $1,199 on EVGA's store. It carries a $50 premium over the non-ultra model but offers a 30 MHz higher boost clock speed. The GeForce RTX 3090 Ti FTW3 Ultra Gaming is the second fastest model in the lineup, with a factory overclock that pushes the boost lock to 1,920 MHz.

EVGA GeForce RTX 3090 Ti

Graphics card	New Price	MSRP	Boost Clock (MHz)	Part Number
GeForce RTX 3090 Ti K\|NGP\|N Hybrid Gaming	$1,999	$2,499	1,950	24G-P5-4998-KT
GeForce RTX 3090 Ti FTW3 Ultra Hybrid Gaming	$1,499	$2,199	1,920	24G-P5-4988-KR
GeForce RTX 3090 Ti FTW3 Black Gaming	$1,399	$1,999	1,860	24G-P5-4981-KR
GeForce RTX 3090 Ti FTW3 Ultra Gaming	$1,199	$2,199	1,920	24G-P5-4985-KR
GeForce RTX 3090 Ti FTW3 Gaming	$1,149	$2,149	1,890	24G-P5-4983-KR

The GeForce RTX 3090 Ti FTW3 Ultra Gaming is also available with a hybrid cooling design. The GeForce RTX 3090 Ti FTW3 Ultra Hybrid Gaming goes for $1,499, $300 more expensive than the air-cooled counterpart. In addition, the graphics card arrives with a beefy 360mm AIO cooler for maximum heat dissipation.

Even the exclusive GeForce RTX 3090 Ti K|NGP|N Hybrid Gaming is one sale, but unfortunately, it's out of stock. The over-engineered graphics card, which caters to extreme overclockers, used to carry an eye-watering price tag of $2,499. EVGA has lowered the price to $1,999, which still costs more than your average high-performance gaming PC. For what it's worth, EVGA includes a "free" SuperNOVA 1600W P2 power supply.

EVGA's GeForce RTX 3090 Ti graphics cards come with Nvidia's latest "Face Your Demons" bundle that consists of Ghostwire: Tokyo, Doom Eternal, and Doom Eternal Year One Pass (Includes Doom Eternal: The Ancient Gods Part One & Part Two and Battlemode).

Graphics card pricing started to plummet a few months ago as Ethereum is nearing its transition to a proof-of-stake (PoS) blockchain. Cryptocurrency miners have started to offload their hardware on the second-hand market to cut losses, whereas scalpers have also lowered pricing on platforms like eBay. EVGA is probably attempting to get rid of Ampere inventory while the company still can. Pricing will most likely continue to improve in the upcoming months as Nvidia readies its GeForce RTX 40-series (Ada Lovelace) graphics cards and AMD answering the call with its RDNA 3 graphics cards. Let's not forget that Intel also has Arc Alchemist, which the chipmaker expects to bring to the market in the summer.

Nvidia GeForce GTX 1630 Review: Lobotomized Turing

Jarred Walton — Thu, 04 Aug 2022 19:33:59 +0000

The Nvidia GeForce GTX 1630 started popping up in rumors in leaks a few months back. Part of me thought, "Surely Nvidia won't release a new and pathetically slow Turing variant this late in the game." But the realist in me knew it was only a matter of time — the GT 1030 and GT 730 cards that started shipping again last year was the only evidence we needed.

Let's be blunt: The GTX 1630 isn't anywhere close to being one of the best graphics cards, and in fact it lands near the very bottom of our GPU benchmarks hierarchy. The only slower GPUs that we've tested are the GTX 1050, RX 560, RX 550, and the aforementioned GT 1030. None of those are worth your time or money either, but at least they're not being released in mid-2022.

The only real contender for the GTX 1630 is AMD's recently launched Radeon RX 6400, but this new Nvidia card actually makes the lackluster 6400 look good. Actually, the real competition — and the reason no one should give the GTX 1630 the time of day — comes from the existing GTX 1650 and GTX 1650 Super. The latter basically doubles the specs of the 1630, and pretty much doubles performance as well.

Colorful sent us this sample for review, but there doesn't seem to be an official price from Nvidia. What we can find online suggests that the soft MSRP has been set at $199, which is just silly. EVGA lists its own GTX 1630 for $199, or you can also buy the far superior GTX 1650 Super for $199. The 1630 also nominally replaces the GT 1030, which had a launch price of $79 — $70 for the faster GDDR5 variant. It feels as though the GTX 1630 was priced according to mid-2021, but it's now laughably expensive.

Here's how the specifications for Nvidia's old-timer Turing TU117 GPUs stack up, with the RX 6400 and RX 6500 XT for comparison.

GPU Specifications
Graphics Card	GTX 1630	RX 6500 XT	RX 6400	GTX 1650 Super	GTX 1650
Architecture	TU117	Navi 24	Navi 24	TU116	TU117
Process Technology	TSMC 12FFN	TSMC N6	TSMC N6	TSMC 12FFN	TSMC 12FFN
Transistors (Billion)	4.7	5.4	5.4	6.6	4.7
Die size (mm^2)	200	107	107	284	200
SMs / CUs	10	16	12	20	14
GPU Cores	512	1024	768	1280	896
RT/RA Cores	N/A	16	12	N/A	N/A
Boost Clock (MHz)	1775	2815	2815	1725	1665
VRAM Speed (Gbps)	12	18	16	12	8
VRAM (GB)	4	4	4	4	6
VRAM Bus Width	64	64	64	128	128
ROPs	32	32	32	48	32
TMUs	32	64	48	80	56
TFLOPS FP32 (Boost)	1.8	5.8	4.3	4.4	3
Bandwidth (GBps)	96	144	128	192	128
PCIe Link	Gen3 x16	Gen4 x4	Gen4 x4	Gen3 x16	Gen3 x16
TDP (watts)	75	107	53	100	75
Launch Date	Jun-22	Jan-22	Jan-22	Nov-19	Apr-19
Launch Price	$199	$199	$159	$159	$149
Online Price	$199	$168	$149	~$125 used	~$90 used

AMD's RX 6400 basically tied the GTX 1650, so there's little question it will easily beat the lower spec GTX 1630. About the only advantage that the GTX 1630 has is the presence of video encoding hardware — the TU117 has Pascal-era hardware, not the improved Turing encoder, but it's still better than nothing.

What's particularly odd with the GTX 1630 is that Nvidia has been shipping the same TU117 GPU in laptops as the MX450 (and more recently MX550) for a couple of years, though granted NVENC is disabled on those parts. Apparently, there were enough chips that couldn't reach the required 14 SMs for the MX450 or GTX 1650, and Nvidia and its partners figured a cut-down 10 SM variant might still sell on desktops — to the uninformed, anyway.

You basically get all the same features as a GTX 1650, just with less performance. You don't even necessarily get a lower power card, as the GTX 1630 models we've seen still come with a 6-pin power connector. That might be because these were less desirable chips with defects, or maybe it's because Nvidia tried to make up for the lack of GPU cores with slightly higher clocks. At least the RX 6400 can be found in half-height models and doesn't require additional power.

The 4GB GDDR6 comes clocked at 12Gbps, with a 64-bit interface. As noted already, that's exactly half of what the GTX 1650 Super provides. Also note that, unlike AMD's RDNA 2 GPUs, there's no Infinity Cache to make up the difference, though the 512 GPU cores are already going to be a limiting factor.

With most modern graphics card selling at close to MSRP, the GTX 1630 feels like far too little, far too late. Last year we saw GTX 1650 cards going for $300 or more, and a $200 GTX 1630 might have made some kind of warped sense. Today, you can get the RX 6500 XT, RX 6400, GTX 1650, and even GTX 1650 Super for $200 or less. Maybe this was supposed to be for big OEMs, so they could toss in a weak GPU and claim to still offer dedicated graphics, but that's ultimately just going to lead to disappointed customers.

MORE: Best Graphics Cards
MORE: GPU Benchmarks and Hierarchy
MORE: All Graphics Content

Tom's Hardware

Colorful markets its GTX 1630 card under the Battle Ax name — surely a weapon to strike fear into the hearts of your opponents! If a name was all that was required to con unsuspecting gamers, maybe this would work. But, sorry Colorful, a tough sounding name does nothing to help the underwhelming GPU inside the box.

It's at least a small card, though it still has dual fans. It weighs 539g and measures 219x115x36mm. That's larger than the PowerColor RX 6400, but still small enough to fit in most mini-ITX cases.

(Image credit: Tom's Hardware)

The heatsink is milled aluminum, and it's more than sufficient to keep the GPU and memory cool. Judging by temperatures and fan speeds, though, it's not a high quality design. Cooling 75W is basically child's play these days, and throwing two fans and a larger heatsink at the problem represents a brute force approach rather than an elegant design.

The card itself is very barebones, with no lighting or other extras. The red and black plastic cooling shroud is as exciting as it gets, outside of the box artwork. Video outputs consist of one each for DisplayPort 1.4, HDMI 2.0, and DL-DVI-D — the last is for anyone still holding on to an older LCD. It would have been nice to get four HDMI/DisplayPort connections instead, since that would have at least provided a potential use case for people looking to cheaply add support for extra monitors.

MORE: Best Graphics Cards
MORE: GPU Benchmarks and Hierarchy
MORE: All Graphics Content

(Image credit: Tom's Hardware)

TOM'S HARDWARE 2022 GPU TEST PC

Our 2022 GPU testbed consists of a Core i9-12900K processor, MSI Pro Z690-A DDR4 WiFi motherboard, and DDR4-3600 memory (with XMP enabled). We're running Windows 11 to ensure we get the most out of Alder Lake, and we're running the latest game patches and drivers for this testing — 516.93 for the GTX 1630.

Our gaming tests consist of a "standard" suite of eight games, and since there's no ray tracing hardware on the GTX 16-series GPUs, we can skip that portion of testing. We've tested at 1080p with medium and ultra settings, though as we'll see shortly there's little point in enabling maximum settings in more demanding games on such a slow graphics card.

MORE: Best Graphics Cards
MORE: GPU Benchmarks and Hierarchy
MORE: All Graphics Content

Tom's Hardware

The GTX 1630 really only makes sense at 1080p low to medium settings, maybe high settings in lighter games. You can see by the overall average of just 36.5 fps that it's not going to win any performance awards.

But it does manage to beat the GTX 1050, which launched with a price of $110 back in 2016, and of course the GT 1030. It also beats the RX 560 4GB, and by extension the RX 550 4GB. Again, those were generally sub-$100 cards when they launched. On the other hand, it fails to beat the GTX 1050 Ti, a $140 card from 2016 with similar video encoding hardware.

In the individual games, nearly all of the games we tested at 1080p medium remained playable, breaking 30 fps. Flight Simulator came up just a hair short at 28 fps, but it's still playable at that speed, while Total War: Warhammer 3 fell far short of the 30 fps mark and averaged just 21 fps.

Tom's Hardware

As you can guess from the 1080p medium results, bumping the quality up to 1080p ultra generally exceeds the capabilities of the GTX 1630. Average performance across the eight tested games was just 18 fps, and there wasn't a single game that broke the 30 fps mark.

Some of that might be due to only having 4GB VRAM, but the 64-bit memory interface and lack of bandwidth are also a problem. The GTX 1650 and 1650 Super also have 4GB VRAM, but they delivered mostly playable 1080p ultra results, outside of a few games.

But we're not suggesting anyone buy a 4GB card with the intent of playing games at ultra quality settings. More and more titles are pushing beyond 4GB, even at 1080p, and the days of such GPUs are limited., which is the big problem with the GTX 1630, as it's already near the bottom of the performance ladder and clearly doesn't have any legs left.

MORE: Best Graphics Cards
MORE: GPU Benchmarks and Hierarchy
MORE: All Graphics Content

(Image credit: Tom's Hardware)

For our power, temperature, clock speed, and fan speed testing, we use Powenetics testing hardware and software. We capture in-line GPU power consumption by collecting data while looping Metro Exodus — this time at 1080p medium instead of our normal 1440p ultra, due to the underpowered GPU. We also test with the FurMark stress test at 1600x900. Our power testing PC uses an open testbed, as that's required for all the extra wires and riser card, and it's the same Core i9-9900K that we've used for the past several years.

Tom's Hardware

Metro Exodus appears to have taxed the GTX 1630 in such a way as to limit maximum power draw to 55W or less. Some games will push higher power use, particularly at more reasonable settings (for example, Horizon Zero Dawn at its original quality setting broke the 60W mark). FurMark meanwhile was able to max out the card's power use at a steady 72–73W.

There's certainly room for overclocking, but the GPU and memory on the GTX 1630 will fall well short of the speeds that would be necessary to make it a worthwhile purchase. Again, GTX 1650 Super doubles all the important elements and currently costs the same $199 — less if you buy it used. An extra 100 MHz and 14 Gbps VRAM won't come anywhere near closing that gap.

Clock speeds on the GTX 1630 were in line with what we're used to seeing from Turing GPUs: about 100 MHz higher than the rated boost clock for gaming, and even FurMark beat Colorful's stated 1815MHz boost clock.

Temperatures peaked at around 67C in FurMark, while Metro only averaged 57C. Again, some games and settings will push the GTX 1630 a bit harder, but we've used the same test for consistency. Fan speeds also topped out at around 1800 RPM in FurMark, while lighter workloads dropped the fans below 1500 RPM.

We measured noise levels at 10cm using an SPL (sound pressure level) meter, aimed between two two fans. That helps to minimize the impact of the CPU cooling fans, and the noise floor (with the CPU fans running) of our test environment and equipment was <33 dB(A). Colorful's GTX 1630 does not have a 0db mode for the fans, so they're always spinning, and the idle noise was a bit higher at 35.0 dB(A). During our gaming tests, that increased to 39.9 dB(A) and 42% fan speed. Manually setting the fans to 75% resulted in 54.3 dB(A) of noise, though it's unlikely the card would hit such speeds during normal use.

MORE: Best Graphics Cards
MORE: GPU Benchmarks and Hierarchy
MORE: All Graphics Content

(Image credit: Tom's Hardware)

It's hard to find things to like about the GTX 1630. If it arrived last year, at this same price, it might have had a few takers in the midst of the global GPU shortages. Today? It's too slow and too expensive. It's frankly absurd that some places like EVGA have GTX 1650 and GTX 1650 Super cards that cost less than the GTX 1630.

Maybe the pricing will correct, but realistically it needs to be cut in half for the cards to make any sort of sense. As a replacement in the product stack for the aging GTX 1050 — which hasn't been manufactured for a few years now, as far as we can tell — a GTX 1630 with a similar $110 price point would have been okay. You get double the VRAM at least, which means some of the games that choke on 2GB cards could still run. Alternatively, Nvidia could have made this a GT 1030 replacement, priced below $100. It still wouldn't be fast, but at least it would have the ultra-budget pricing angle going for it.

With AMD's Radeon RX 6400 delivering significantly better performance for $50 less, the only real selling point for the GTX 1630 would be its media encoding capabilities. If that's your primary consideration, waiting for Intel's Arc A380 to ship in the US would be a better choice. Gaming performance might be questionable, but the AV1 media encoding capabilities look quite promising.

(Image credit: Tom's Hardware)

The Turing TU117 GPU was never a potent chip, but disabling half of the cores and memory channels makes a relatively weak GPU into a solution that can barely hope to compete with integrated graphics. Yes, it's faster than the GTX 1050, but it's also slower than the GTX 1050 Ti, and both of those can readily be picked up used for about $100 these days.

You don't even get extra features compared to the 1050 Ti that would warrant the upgrade — it's the same NVENC hardware, more or less, with the same amount of VRAM, same video outputs, and worse performance. Power efficiency isn't an advantage either, as the GTX 1050 Ti only used 58W in our gaming test, and both are nominally 75W cards. So what's the point?

The point is to make something out of nothing, as far as I can tell. Nvidia probably had some TU117 chips that had functional NVENC hardware but with some of the memory controllers and SMs that couldn't hit the GTX 1650 or MX450 requirements. Just like there are GTX 1650 cards made using TU106 silicon, but with all the ray tracing and tensor core hardware disabled, and with only 14 of the potential 36 SMs enabled, now we have a GTX 1630 made with chips where 50% of the cores and memory controllers apparently failed.

If Nvidia or its partners priced this appropriately, some people would inevitably find a use for it. That means that it needs to cost significantly less than the pre-existing GTX 1650 cards, since those are universally faster. Even Intel's Arc A380 will likely deliver superior performance, though it's a gamble on whether or not Intel will ever get the drivers to fully functional status. We'll hopefully have a card soon that we can put through its paces.

MORE: Best Graphics Cards
MORE: GPU Benchmarks and Hierarchy
MORE: All Graphics Content

EVGA E1 Chassis Frame Costs $1600

Mark Tyson — Wed, 27 Jul 2022 18:22:52 +0000

EVGA's open air PC case, the E1 Chassis Frame, is now ready to purchase. The company first took the wraps off its E1 design back in January. In brief, it is an open-air case made of carbon fiber, with the motherboard suspended via steel cables.

This kind of design and innovation doesn’t come cheap, but the starting price of $1,600 for the most basic kit (the chassis and a VGA vertical kit) is pretty breathtaking. EVGA isn’t scared of premium pricing — other EVGA E1 kits with more system components included are priced as high as $4,999.

(Image credit: EVGA)

It's worth noting that the E1 series is created with premium materials. EVGA says that the E1 structure is built entirely with 3K plain weave carbon fiber. This is, apparently, the same material used to construct high performance supercars, because it has an excellent combination of light weight (chassis weighs 2.76lbs) and strength. It looks pretty good too — you might mistake the frame for polished tubular steel from a distance, but the underlying woven carbon fiber becomes evident up-close.

(Image credit: EVGA)

There are three EVGA E1 kits on sale. In the intro we highlighted the ‘cheapest’ kit, dubbed the EVGA E1 KIT 2. EVGA’s fullest offering costs $4,999 and is called the EVGA E1 Bare Bones. Paying this much and still getting a barebones system is definitely...extraordinary. On top of the initial $4,999 outlay for the top-end kit, you'll also need to buy (or migrate) a CPU, memory, storage, and an operating system.

(Image credit: EVGA)

Potential purchasers of the $4,999 EVGA E1 can check out expansion options by pondering over our EVGA Z690 Dark KingPin E-ATX motherboard review from earlier this month. This well-rounded overclocking–focused motherboard should satisfy most enthusiasts as long as they don’t care about RGB (it’s Dark), and as long as three M.2 storage sockets are sufficient.

Another big-ticket component in the EVGA E1 Bare Bones model is the GeForce RTX 3090 Ti KingPin GPU. We have yet to get this premium overclocking GPU in the labs.

Those interested in the EVGA E1 kits can make their orders today, but even for the simplest and cheapest model there is said to be a build-to-order lag of three to four weeks. This delay is a little strange given these are all self-assembled products. EVGA is currently bundling a premium shipping case with the $4,999 kit; the company claims this bonus would normally cost $800.

Corsair CX750M (2021) Power Supply Review

Aris Mpitziopoulos — Wed, 20 Jul 2022 15:04:54 +0000

The budget 750W category is challenging, with several worthy options like the XPG Pylon 750, the Thermaltake Smart BM2 750, and the Cooler Master MWE Bronze 750. The Corsair CX750M stacks well against the competition, but it cannot lead the pack, so even if we had a budget 750W category in our best PSUs article, it wouldn't be our top choice, but an alternative one.

Corsair revamped its popular CX-M line in 2021, but so far, we didn't have the chance to review any of these units. Like the last line, the new one includes four members with capacities ranging from 450W to 750W. All use a semi-modular cable design, and although someone would expect that their platforms are inferior to the non-modular CX models, it is the other way around. The CX units use more modern platforms provided by CWT and Great Wall, while the CX-M units use a simple design by CWT to suppress production cost.

Tom's Hardware

The CX750M is the strongest of the line and with its two EPS and four PCIe connectors, it won't have a problem powering up a potent gaming system equipped with a power-hungry GPU. Its CWT platform scores Bronze in 80 PLUS and Silver in Cybenetics, and it also has a Standard++ rating by the latter. The semi-modular cable design allows for a lower production cost, while it only keeps the absolutely essential cables fixed, the ATX and one EPS. Finally, Corsair used a good rifle-bearing fan, since it provides a five-year warranty for this product.

Tom's Hardware

Specifications

Manufacturer (OEM)	CWT
Max. DC Output	750W
Efficiency	80 PLUS Bronze, Cybenetics Silver (85-87%)
Noise	Cybenetics Standard++ (30-35 dB[A])
Modular	✓ (Semi)
Intel C6/C7 Power State Support	✓
Operating Temperature (Continuous Full Load)	0 - 40°C
Over Voltage Protection	✓
Under Voltage Protection	✓
Over Power Protection	✓
Over Current (+12V) Protection	✓
Over Temperature Protection	✓
Short Circuit Protection	✓
Surge Protection	✓
Inrush Current Protection	✓
Fan Failure Protection	✗
No Load Operation	✓
Cooling	120mm Rifle Bearing Fan (HA1225H12F-Z)
Semi-Passive Operation	✗
Dimensions (W x H x D)	150 x 85 x 140mm
Weight	1.6 kg (3.53 lb)
Form Factor	ATX12V v2.52, EPS 2.92
Warranty	5 years

Power Specifications

Rail		3.3V	5V	12V	5VSB	-12V
Max. Power	Amps	20	20	62	3	0.3
	Watts		130	744	15	3.6
Total Max. Power (W)			750

Cables & Connectors

`Native Cables`
Description	Cable Count	Connector Count (Total)	Gauge	In Cable Caps
ATX Connector 20+4 pin (610mm)	1	1	18-22AWG	No
4+4 pin EPS12V (670mm)	1	1	18AWG	No
`Modular Cables`
4+4 pin EPS12V (650mm)	1	1	18AWG	No
6+2 pin PCIe (600mm+150mm)	2	4	16-18AWG	No
SATA (450mm+110mm+110mm+110mm)	2	8	18AWG	No
4-pin Molex (450mm+100mm+100mm) / FDD (+100mm)	1	3 / 1	18-22AWG	No
AC Power Cord (1400mm) - C13 coupler	1	1	18AWG	-

It is good to see only two cables fixed. Ideally, all should be modular, but with only two, essential, cables fixed there won't be any issues during the PSU's installation. On top of that, you will have lower resistance on the fixed cables, hence better load regulation, especially on the minor rails which mostly pass through the ATX cable.

Cable length is sufficient and as expected and there are no in-cable caps. The FDD connector is not required by the majority of users, so Corsair could replace it with a 4-pin Molex and increase their number to four. Lastly, the distance between 4-pin Molex is too short, at only 100mm.

Tom's Hardware

Component Analysis

`General Data`	-
Manufacturer (OEM)	CWT
PCB Type	Single-Sided
`Primary Side`	-
Transient Filter	4x Y caps, 2x X caps, 2x CM chokes, 1x MOV, 1x CAP200DG (Discharge IC)
Inrush Protection	NTC Thermistor SCK - 2R58 (2.5Ohm)
Bridge Rectifier(s)	1x GBU15L06 (800V, 10A @ 100°C)
APFC MOSFETs	2x Champion GP28S50 (500V, 28A, Rds(on): 0.125Ohm)
APFC Boost Diode	1x ON Semiconductor FFSP0665A (650V, 6A @ 153°C)
Bulk Cap(s)	1x Nichicon (400V, 390uF, 2,000h @ 105°C, GG)
Main Switchers	2x Champion GP23S60HX
PFC/PWM Combo Controller	Champion CM6800TX & Champion CM03X
Topology	Primary side: APFC, Double Forward Secondary side: Semi-Synchronous Rectification (12V) & DC-DC converters (5V & 3.3V)
`Secondary Side`	-
+12V	2x Advanced Power AP6N3R5P (60V, 80A @ 100°C, Rds(on): 3.58mOhm) FET & 2x PFC PFR40V60CT (60V, 40A @ 100°C) SBR
5V & 3.3V MOSFETs	2x UBIQ QM3054M6 (30V, 61A @ 100°C, Rds(on): 4.8mOhm) & 2x UBIQ QN3107M6N (30V, 70A @ 100°C, Rds(on): 2.6mOhm) PWM Controller: ANPEC APW7159C
Filtering Capacitors	Electrolytic: 11x Elite (2-5,000h @ 105°C, ED), 1x Nichicon (4-10,000h @ 105°C, HE), 2x Elite (4-10,000h @ 105°C, EY), 1x Elite (2-5,000h @ 105°C, EK), 1x Nippon Chemi-Con (4-10,000h @ 105°C, KY), 2x Elite (2,000h @ 105°C, PF) Polymer: 9x APAQ
Driver IC	Sync Power SP6019
Supervisor IC	INI1S429I - DCG (OVP, UVP, OCP, PG, SCP)
Fan Model	Hong Hua HA1225H12F-Z (120mm, 12V, 0.58A, Rifle Bearing Fan)
`5VSB Circuit`	-
Standby PWM Controller	Power Integrations TNY290PG

Tom's Hardware

Corsair used CWT's CSB-A platform, which utilizes a double forward topology on the primary side. On the secondary side, FETs and SBRs regulate the 12V rail, and two DC-DC converters handle the minor rails. It is a budget design with small heat sinks. Despite the small PCB, there is enough space between the parts for decent airflow. Lastly, the XPG Pylon 750 and the Thermaltake Smart BM2 750 units also use the same platform.

Tom's Hardware

The transient filter has all the necessary parts to do a fine job, including a discharge IC which provides a slight efficiency boost.

Tom's Hardware

There is an MOV to protect against voltage surges. An NTC thermistor lower inrush currents. It isn't supported by a relay, which usually is the case in budget designs.

(Image credit: Tom's Hardware)

The single bridge rectifiers can handle up to 10 Amps at higher than 100 degrees Celsius and up to 15 Amps at lower temps.

Tom's Hardware

The APFC converter uses two Champion FETs and a single boost diode. The bulk cap is by Nichicon, and it doesn't have enough capacity, to offer a longer than 17ms hold-up time.

Tom's Hardware

The main Champion FETs are installed into a double forward topology. The joint PFC/PWM controller is a Champion CM6800TX, which was highly popular several years ago. There is also an extra controller, a Champion CM03X, to lower vampire power.

Tom's Hardware

Two FETs and two SBRs regulate the 12V rail. This looks to be a semi-synchronous design. The minor rails use two DC-DC converters.

Tom's Hardware

The filtering caps are by Elite, a very good alternative to Japanese manufacturers, Nichicon and Chemi-Con. There are also nine APAQ polymer caps.

Tom's Hardware

The standby PWM controller is a TNY290PG.

(Image credit: Tom's Hardware)

The main supervisor IC is an INI1S429I - DCG.

Tom's Hardware

Soldering quality is decent.

Tom's Hardware

Hong Hua made a good name in the PSU market, so it is good to see the CX750M using a rifle bearing fan from this brand.

MORE: Best Power Supplies

To learn more about our PSU tests and methodology, please check out How We Test Power Supply Units.

Thermaltake Smart BM2 750W

ASUS TUF GAMING 750W

EVGA 750 BQ

Primary Rails And 5VSB Load Regulation

Tom's Hardware

Load regulation is tight at 5V, but not as tight on the other rails.

Hold-Up Time

Put simply; hold-up time is the amount of time that the system can continue to run without shutting down or rebooting during a power interruption.

Tom's Hardware

The hold-up time is dead low. The small bulk cap is the restrictive factor here.

Inrush Current

Tom's Hardware

Inrush current is high with 230V. This is not the case with 115V input.

Leakage Current

(Image credit: Tom's Hardware)

Leakage current is low.

10-110% Load Tests

These tests reveal the PSU's load regulation and efficiency levels under high ambient temperatures. They also show how the fan speed profile behaves under increased operating temperatures.

Test	12V	5V	3.3V	5VSB	DC/AC (Watts)	Efficiency	Fan Speed (RPM)	PSU Noise (dB[A])	Temps (In/Out)	PF/AC Volts
10%	4.396A	1.992A	1.985A	0.995A	74.987	84.682%	955	21.3	40.5°C	0.967
	12.145V	5.021V	3.324V	5.025V	88.551				45.65°C	115.13V
20%	9.812A	2.989A	2.983A	1.197A	149.899	88.511%	961	21.4	41.26°C	0.983
	12.127V	5.018V	3.318V	5.011V	169.356				46.66°C	115.13V
30%	15.594A	3.49A	3.485A	1.4A	224.888	89.239%	963	21.5	42.1°C	0.987
	12.110V	5.014V	3.314V	4.998V	252.006				47.91°C	115.12V
40%	21.399A	3.991A	3.989A	1.604A	299.965	89.127%	964	21.5	42.88°C	0.989
	12.093V	5.011V	3.309V	4.986V	336.559				49.01°C	115.12V
50%	26.812A	4.992A	4.995A	1.81A	374.226	88.478%	969	21.7	43.51°C	0.99
	12.074V	5.008V	3.303V	4.971V	422.961				50.39°C	115.12V
60%	32.298A	5.994A	6.004A	2A	449.06	87.607%	973	21.8	43.89°C	0.991
	12.055V	5.005V	3.298V	4.956V	512.586				51.41°C	115.11V
70%	37.804A	6.997A	7.017A	2.225A	524.076	86.399%	1253	29.6	44.02°C	0.992
	12.035V	5.003V	3.292V	4.941V	606.58				52.24°C	115.11V
80%	43.394A	8.001A	8.029A	2.332A	599.283	85.232%	1729	39.4	44.58°C	0.992
	12.016V	5V	3.286V	4.928V	703.117				53.67°C	115.11V
90%	49.339A	8.504A	8.528A	2.439A	674.307	83.908%	2104	46.7	44.66°C	0.993
	11.995V	4.996V	3.282V	4.917V	803.628				54.26°C	115.1V
100%	55.106A	9.01A	9.059A	3.063A	749.536	82.41%	2362	48.9	45.57°C	0.993
	11.974V	4.993V	3.277V	4.896V	909.522				55.66°C	115.09V
110%	60.749A	10.014A	10.171A	3.069A	824.561	80.803%	2358	48.5	46.75°C	0.994
	11.955V	4.992V	3.273V	4.886V	1020.459				57.68°C	115.09V
CL1	0.114A	15.542A	15.638A	0A	131.265	81.33%	985	22.2	43.16°C	0.983
	12.118V	5.038V	3.299V	5.009V	161.398				50.18°C	115.14V
CL2	0.113A	19.781A	0A	0A	101.375	81.919%	983	22.1	44.21°C	0.979
	12.135V	5.055V	3.312V	5.026V	123.75				52.43°C	115.14V
CL3	0.113A	0A	19.932A	0A	67.364	74.448%	962	21.5	44.82°C	0.968
	12.135V	5.016V	3.311V	5.025V	90.486				53.88°C	115.14V
CL4	62.516A	0A	0A	0A	749.364	83.173%	2377	50.3	45.63°C	0.993
	11.987V	4.992V	3.296V	4.972V	900.984				55.77°C	115.09V

This might be a budget PSU, but it doesn't have a problem operating at high temperatures. It is impressive to see this platform coping with full and 110% load at close to 47 degrees Celsius! Usually, we don't push budget units beyond 40 degrees Celsius.

20-80W Load Tests

Test	12V	5V	3.3V	5VSB	DC/AC (Watts)	Efficiency	Fan Speed (RPM)	PSU Noise (dB[A])	Temps (In/Out)	PF/AC Volts
20W	1.220A	0.498A	0.495A	0.198A	19.983	68.678%	946	21.0	37.21°C	0.902
	12.158V	5.023V	3.33V	5.048V	29.097				40.32°C	115.14V
40W	2.688A	0.697A	0.694A	0.297A	39.984	79.05%	947	21.0	37.35°C	0.948
	12.154V	5.022V	3.328V	5.044V	50.581				40.85°C	115.14V
60W	4.158A	0.896A	0.892A	0.397A	59.984	84.058%	950	21.1	38.6°C	0.963
	12.149V	5.021V	3.327V	5.04V	71.36				42.61°C	115.14V
80W	5.623A	1.095A	1.091A	0.496A	79.917	85.755%	953	21.2	39.69°C	0.969
	12.145V	5.02V	3.325V	5.035V	93.192				44.29°C	115.13V

At light loads the fan spins at low RPM, keeping noise output equally low.

2% or 10W Load Test

12V	5V	3.3V	5VSB	DC/AC (Watts)	Efficiency	Fan Speed (RPM)	PSU Noise (dB[A])	Temps (In/Out)	PF/AC Volts
1.047A	0.264A	0.264A	0.051A	15.195	62.928%	909	20.1	21.61°C	0.88
	12.158V	5.025V	3.333V	5.054V	24.147			24.1°C	115.14V

Efficiency at super light loads is high enough.

Efficiency & Power Factor

Tom's Hardware

Efficiency with normal levels is decent, while it could be a bit higher with super-light loads. The PF readings are high, as well.

5VSB Efficiency

Test #	5VSB	DC/AC (Watts)	Efficiency	PF/AC Volts
1	0.1A	0.505W	75.862%	0.074
	5.056V	0.666W		115.12V
2	0.25A	1.262W	78.943%	0.16
	5.054V	1.599W		115.12V
3	0.55A	2.775W	79.961%	0.27
	5.048V	3.47W		115.12V
4	1A	5.038W	80.498%	0.348
	5.039V	6.259W		115.12V
5	1.5A	7.542W	79.361%	0.393
	5.029V	9.503W		115.12V
6	2.999A	15W	78.064%	0.45
	5.002V	19.215W		115.12V

Tom's Hardware

The 5VSB rail is efficient.

Power Consumption In Idle And Standby

Mode	12V	5V	3.3V	5VSB	Watts	PF/AC Volts
Idle	12.162V	5.026V	3.336V	5.059V	4.673	0.413
						115.1V
Standby					0.049	0.005
						115.1V

Tom's Hardware

We would like to see below 0.1W vampire power with 230V input.

Fan RPM, Delta Temperature, And Output Noise

All results are obtained between an ambient temperature of 37 to 47 degrees Celsius (98.6 to 116.6 degrees Fahrenheit).

(Image credit: Tom's Hardware)

The fan speed profile is not aggressive, at high operating temperatures. Up to 450W loads, the fan rotates at low speeds, so noise output is minimal.

The following results were obtained at 30 to 32 degrees Celsius (86 to 89.6 degrees Fahrenheit) ambient temperature.

(Image credit: Tom's Hardware)

At normal operating temperatures, close to 30 degrees Celsius, the PSU is silent with up to 440-450W loads. If you push it with more than 580W, it will exceed 40 dBA, and for a brief part, it goes over 45 dBA.

MORE: Best Power Supplies

Protection Features

Check out our PSUs 101 article to learn more about PSU protection features.

OCP (Cold @ 23°C)	12V: 74.4A (120%), 11.957V 5V: 30.5A (152.5%), 5.055V 3.3V: 28.5A (142.5%), 3.309V 5VSB: 4.2A (140%), 4.980V
OCP (Hot @ 40°C)	12V: 74A (119.35%), 11.955V 5V: 29.5A (147.5%), 5.056V 3.3V: 27.5A (137.5%), 3.308V 5VSB: 4.2A (140%), 4.977V
OPP (Cold @ 22°C)	939.85W (125.31%)
OPP (Hot @ 39°C)	884.38W (118.87%)
OTP	✓ (185°C @ 12V Heat Sink)
SCP	12V to Earth: ✓ 5V to Earth: ✓ 3.3V to Earth: ✓ 5VSB to Earth: ✓ -12V to Earth: ✓
PWR_OK	Accurate but lower than 16ms
NLO	✓
SIP	Surge: MOV Inrush: NTC Thermistor

The OCP triggering points are set high on the minor rails but correctly set at 12V. The over power protection also works appropriately under high and low temperatures. Moreover, there is over temperature protection, which is essential to any PSU. Finally, it would be nice to see a bypass relay for the NTC thermistor, which would enhance inrush current protection, but most PSUs lack this relay in this price range.

DC Power Sequencing

Tom's Hardware

No problems here since the 3.3V rail is always lower than the other two.

Cross Load Tests

Load Regulation Charts

Tom's Hardware

Efficiency Graph

(Image credit: Tom's Hardware)

Ripple Graphs

The lower the power supply's ripple, the more stable the system will be and less stress will also be applied to its components.

Tom's Hardware

Infrared Images

Tom's Hardware

The secondary side gets hot, in the 12V rail's generation and filtering area. This PSU is not meant to operate without a fan, especially under the conditions we applied, so it is natural to see high temperatures in the parts that are stressed the most.

MORE: Best Power Supplies

Advanced Transient Response Tests

For details about our transient response testing, please click here.

Advanced Transient Response at 20% – 20ms

Voltage	Before	After	Change	Pass/Fail
12V	12.126V	11.927V	1.65%	Pass
5V	5.017V	4.879V	2.75%	Pass
3.3V	3.318V	3.152V	5.01%	Pass
5VSB	5.010V	4.938V	1.44%	Pass

Advanced Transient Response at 20% – 10ms

Voltage	Before	After	Change	Pass/Fail
12V	12.126V	11.969V	1.30%	Pass
5V	5.017V	4.903V	2.27%	Pass
3.3V	3.318V	3.145V	5.21%	Pass
5VSB	5.010V	4.956V	1.07%	Pass

Advanced Transient Response at 20% – 1ms

Voltage	Before	After	Change	Pass/Fail
12V	12.126V	11.963V	1.35%	Pass
5V	5.017V	4.896V	2.41%	Pass
3.3V	3.318V	3.145V	5.23%	Pass
5VSB	5.010V	4.938V	1.44%	Pass

Advanced Transient Response at 50% – 20ms

Voltage	Before	After	Change	Pass/Fail
12V	12.074V	11.905V	1.40%	Pass
5V	5.008V	4.893V	2.30%	Pass
3.3V	3.304V	3.127V	5.37%	Fail
5VSB	4.975V	4.907V	1.36%	Pass

Advanced Transient Response at 50% – 10ms

Voltage	Before	After	Change	Pass/Fail
12V	12.074V	11.896V	1.47%	Pass
5V	5.008V	4.865V	2.86%	Pass
3.3V	3.304V	3.129V	5.30%	Fail
5VSB	4.976V	4.908V	1.37%	Pass

Advanced Transient Response at 50% – 1ms

Voltage	Before	After	Change	Pass/Fail
12V	12.074V	11.916V	1.31%	Pass
5V	5.007V	4.889V	2.35%	Pass
3.3V	3.304V	3.121V	5.54%	Fail
5VSB	4.976V	4.918V	1.17%	Pass

Tom's Hardware

Transient response is good at 12V, decent at 5V and mediocre at 3.3V.

Turn-On Transient Tests

Tom's Hardware

There is a notable step at 5VSB, and the 12V slopes are not so smooth. Nonetheless, these results won't create any issues.

Power Supply Timing Tests

There are several signals generated by the power supply, which need to be within specified, by the ATX spec, ranges. If they are not, there can be compatibility issues with other system parts, especially mainboards. From year 2020, the PSU's Power-on time (T1) has to be lower than 150ms and the PWR_OK delay (T3) from 100 to 150ms, to be compatible with the Alternative Sleep Mode.

PSU Timings Table
T1 (Power-on time) & T3 (PWR_OK delay)
Load	T1	T3
20%	32ms	133ms
100%	39ms	133ms

Tom's Hardware

The PWR_OK delay is within the 100-150ms region, so the PSU supports the alternative sleep mode recommended by the ATX spec.

Ripple Measurements

The ripple limits, according to the ATX specification, are 120mV (+12V) and 50mV (5V, 3.3V, and 5VSB).

Test	12V	5V	3.3V	5VSB	Pass/Fail
10% Load	8.3 mV	8.5 mV	12.9 mV	10.9 mV	Pass
20% Load	9.4 mV	8.4 mV	15.2 mV	11.2 mV	Pass
30% Load	10.4 mV	9.6 mV	16.7 mV	11.7 mV	Pass
40% Load	11.9 mV	11.0 mV	17.7 mV	13.6 mV	Pass
50% Load	12.1 mV	11.6 mV	18.7 mV	13.9 mV	Pass
60% Load	14.7 mV	12.2 mV	20.2 mV	15.9 mV	Pass
70% Load	19.3 mV	11.6 mV	22.4 mV	18.7 mV	Pass
80% Load	24.3 mV	12.3 mV	24.8 mV	33.3 mV	Pass
90% Load	29.4 mV	13.6 mV	25.9 mV	39.4 mV	Pass
100% Load	38.3 mV	18.0 mV	30.8 mV	23.5 mV	Pass
110% Load	45.8 mV	20.9 mV	31.7 mV	25.3 mV	Pass
Crossload 1	23.4 mV	13.0 mV	22.8 mV	40.5 mV	Pass
Crossload 2	11.0 mV	11.3 mV	10.7 mV	24.8 mV	Pass
Crossload 3	16.9 mV	10.0 mV	21.8 mV	32.4 mV	Pass
Crossload 4	39.2 mV	15.6 mV	21.1 mV	14.0 mV	Pass

Tom's Hardware

Ripple suppression is decent. It would be nice, though, to see the same ripple levels with the previous CM750M unit.

Ripple At Full Load

Tom's Hardware

Ripple At 110% Load

Tom's Hardware

Ripple At Cross-Load 1

Tom's Hardware

Ripple At Cross-Load 4

Tom's Hardware

EMC Pre-Compliance Testing – Average & Quasi-Peak EMI Detector Results

Electromagnetic Compatibility (EMC) is the ability of a device to operate properly in its environment without disrupting the proper operation of other nearby devices.

΅We use TekBox's EMCview to conduct our EMC pre-compliance testing.

(Image credit: Tom's Hardware)

EMI emissions are low. The corresponding filter does a splendid job!

MORE: Best Power Supplies

Performance Rating

(Image credit: Tom's Hardware)

The PSU achieves good performance for a unit using a double-forward platform. The problem is that it costs the same with the EVGA 750 GA, which as you can see dominates the performance chart since it uses a more modern platform.

Noise Rating

The graph below depicts the cooling fan's average noise over the PSU's operating range, with an ambient temperature between 30 to 32 degrees Celsius (86 to 89.6 degrees Fahrenheit).

(Image credit: Tom's Hardware)

The average noise output could be lower. The increased noise at high loads notably affects the average noise output.

Efficiency Rating

The following graph shows the PSU's average efficiency throughout its operating range with an ambient temperature close to 30 degrees Celsius.

(Image credit: Tom's Hardware)

Average efficiency is high enough for this category's standards.

Power Factor Rating

The following graphs show the PSU's average power factor reading throughout its operating range with an ambient temperature close to 30 degrees Celsius and 115V/230V voltage input.

Tom's Hardware

The APFC converter does an excellent job!

MORE: Best Power Supplies

Corsair teamed up with CWT once more, and the result is good. The modern exterior design and the updated platform will allow the CX750M to retain its position in the market and compete with the other budget 750W PSUs. There are several good choices in this category, including the XPG Pylon 750, the Thermaltake Smart BM2 750, and the Cooler Master MWE Bronze 750, but the major competitor for all these units look is the EVGA SuperNOVA 750 GA, which, although it belongs to a higher category, has the same price tag.

The 750 GA offers higher efficiency, has a fully modular cable design, and is supported by a ten-year warranty, so having it at the same price, 80 dollars, as the CX750M creates enormous problems for the latter. To remain competitive, the CX750M and the other 750 Bronze and Silver units must have at least 20 dollars difference from the 750 GA. At the same price, it just doesn't make sense to go for a lower efficiency and performance unit.

(Image credit: Tom's Hardware)

The Corsair CX750M has good build quality and decent overall performance and won't be noisy under normal operating conditions. It is also supported by a long warranty, which the rifle bearing fan and the capacitors that Corsair used will outlive. Its main problem looks to be the relatively high price tag, although this can be countered with a Corsair discount code if you're lucky. It currently goes for 80 dollars on Newegg, and for this amount, someone can get the more capable EVGA 750 GA, so Corsair should do something about the price and lower it as close as possible to 60 dollars.

MORE: Best Power Supplies

RTX 4090 Allegedly 30% Faster in TimeSpy Extreme vs. LN2 RTX 3090 Ti

editors@tomshardware.com (Aaron Klotz) — Mon, 18 Jul 2022 17:39:13 +0000

According to our resident Twitter leaker @kopite7kimi, Nvidia's upcoming GeForce RTX 4090 cracked the 15,000 point barrier in 3DMark Time Spy Extreme by a long shot, with a record-breaking graphics score of 19,000 points in the famous benchmark. Furthermore, if Kopite's news is accurate, it puts the RTX 4090's performance barrier well ahead of anything available today, including RTX 3090 Ti GPUs chilled on liquid nitrogen.

For some perspective, the current reigning champion of the 3DMark TimeSpy Extreme benchmark is user "biso biso," with an LN2 cooled EVGA RTX 3090 Ti Kingpin Edition graphics card punching out a world recorded graphics result of 14,611 points.

RTX 4090, TSE >19000July 18, 2022

Again, if Kopites data is real, engineering samples of Nvidia's 4090's are already hitting TimeSpy Extreme scores 30% higher than the highest-overclocked RTX 3090 Ti's running today (not to mention normal RTX 3090 Ti's). As a result, we could see even higher scores once the RTX 4090 hits the market.

This data seems to confirm what we've already heard in the past: that Nvidia's RTX 40 series lineup will feature one of the most significant performance jumps we've seen in a single GPU generation from Nvidia.

Current rumors speculate that the top die for the 40 series, AD102, will pack 71% more CUDA cores and SMs than Ampere's top die GA102. In addition, it will reportedly feature similar or higher clock speeds compared to RTX 30 series, thanks to a more efficient TSMC 5nm process. We don't expect the rest of Nvidia's 40 series dies to pack the same core count increase, but they all are expected to pack a lot more cores if 71% is the ceiling for the 40 series.

Power consumption is also rumored to go up extensively, with flagship 40 series cards rumored to hit as much as 500 to 600 watts of power consumption.

As a result, high-performance gains should be expected to compensate for the crazy high power requirements for the 40 series. However, take this news with a grain of salt since we don't have full confirmation on this data. But it does have merit based on the current information that we have on Nvidia's next-generation GPUs.

EVGA Equips RTX 3090 Ti Kingpin Card with PowerLink 52U 12VHPWR Adaptor

ashilov@gmail.com (Anton Shilov) — Wed, 13 Jul 2022 13:00:18 +0000

EVGA's Kingpin (stylized as K|NGP|N) series graphics cards have always been a little special. They use a custom printed circuit board (PCB) with an enhanced voltage regulating module (VRM) along with a hybrid or liquid cooling system to ensure maximum performance. But with its GeForce RTX 3090 Ti, EVGA has gone one step further: it bundled a rather formidable 1600W power supply with the board. And now we see a special converter board for 12VHPWR.

PowerLink 52U for 3090 Ti KINGPIN 🤯 pic.twitter.com/Np7z1TDAPdJuly 12, 2022

EVGA's 1.6kW PSU lacks the 12VHPWR PCIe Gen5 auxiliary power connectors that EVGA's GeForce RTX 3090 Ti Kingpin graphics board uses, which is why EVGA had to create its special PowerLink 52U five 8-pin to two 16-pin (12VHPWR) power adapter exclusively for its flagship card (which will be among the best gaming graphics cards ever produced until the RTX refresh hits), Jacob Freeman, a spokesman for EVGA, has revealed in his Twitter.

The EVGA GeForce RTX 3090 Ti Kingpin uses two 12VHPWR PCIe Gen5 auxiliary power connectors that can deliver up to 1200W of power, an amount that the board can barely devour even when heavily overclocked. Nonetheless, to make its overkill VRM work properly, EVGA wants to feed both connectors with enough power. Apparently the five 8-pin PCIe auxiliary power plugs (that can deliver 750W) can do the job, albeit using a special adapter.

(Image credit: EVGA)

EVGA's PowerLink 52U will certainly simplify cable management for existing GeForce RTX 3090 Ti Kingpin owners since messing with cabled adapters is not particularly user friendly. What remains to be seen is whether the same adapter will be supplied with future GeForce RTX 40-series graphics cards (or sold separately) since they will use 12VHPWR PCIe Gen5 power connectors and will not be compatible with existing PSUs out-of-box.

Common logic says that since next-generation graphics cards will still be power hungry and will need new power connectors, there will be adapters for these boards to make them work with existing high-wattage power supply that will still be good for upcoming graphics adapters.

EVGA Z690 Dark KingPin Review: Overclocking-Focused Flagship

Joe Shields — Thu, 07 Jul 2022 12:07:33 +0000

Like it or not–and clearly only the motherboard makers like it–flagship boards these days often push up to or even above the $1,000 mark, all in the quest for the best features and performance. The EVGA Z690 Dark Kingpin (officially stylized K|NGP|N) hit the shelves priced at $849.99 and, as of today, is a more reasonable $699.99. Among flagship moboss, that price is comparitively reasonable when competitors cost between $900 and $1,200. The Z690 Dark has an updated appearance that carries over EVGA’s signature black styling, robust power delivery, horizontal connectors and a unique socket orientation that sets it apart from other boards.

Performance-wise, the Dark Kingpin lit up the charts as one of the fastest boards across our multi-threaded tests, and nearly did a clean sweep as the best-performing board in our Procyon Office tests. Single-threaded portions of our tests didn’t fare as well, but the board was still average across most of those tests. Gaming performance was also above average. Overall, this is a well-perofrming board out of the box, especially when pushed with heavy multi-threaded loads.

This board also has almost every high-end feature you can expect at the flagship level. You get two full-length PCIe 5.0 slots (and SLI support), 21-phase 90A MOSFETs, three M.2 sockets (though none are PCIe 5.0), eight SATA ports and a "legacy" U.2 port, a premium (though last-gen) audio codec with NU Audio software and dual 2.5 GbE ports along with integrated Wi-Fi 6E.

What sets the Kingpin apart from the other luxury boards is all the overclocking-focused functionality. Where most focus on style and shoehorning in extra features , the Kingpin has a slew of switches, buttons and headers on board (to connect the ProbeIt tool) to help push your Alder Lake system to the limit. Before we get into the details and see if this board has a chance to make our best motherboards list, here are its detailed specifications, direct from EVGA.

Specifications: EVGA Z690 Dark Kingpin

Socket	LGA1700
Chipset	Z690
Form Factor	E-ATX
Voltage Regulator	21 Phase (18+2+1 90A SPS MOSFETs for Vcore)
Video Ports	✗
USB Ports	(1) USB 3.2 Gen 2x2 Type-C (20 Gbps)
	(4) USB 3.2 Gen 2 (10 Gbps)
	(2) USB 2.0 (480 Mbps)
Network Jacks	(2) 2.5 GbE
Audio Jacks	(5) Analog + SPDIF
Legacy Ports/Jacks	(1) PS/2
Other Ports/Jack	✗
PCIe x16	(2) v5.0 (x16/x0 or x8/x8)
PCIe x8	✗
PCIe x4	(1) v4.0 x4
PCIe x1	✗
CrossFire/SLI	Supports Nvidia SLI
DIMM Slots	(2) DDR5 6600+(OC), 64GB Capacity
M.2 Sockets	(1) PCIe 4.0 x4 (64 Gbps) / PCIe (up to 110mm)
	(1) PCIe 4.0 x4 (64 Gbps) / PCIe (up to 110mm)
	(1) PCIe 4.0 x4 (64 Gbps) / PCIe (up to 110mm)
U.2 Ports	(1) PCIe 3.0 x4 (32 Gbps)
SATA Ports	(8) SATA3 6 Gbps (Supports RAID 0/1/5/10)
USB Headers	(1) USB v3.2 Gen 2, Type-C (10 Gbps)
	(1) USB v3.2 Gen 1 (5 Gbps)
	(2) USB v2.0 (480 Mbps)
Fan/Pump Headers	(8) 4-Pin (CPU, Pump, Aux, Pwr, System fans)
RGB Headers	(2) aRGB (3-pin)
	(2) RGB (4-pin)
Diagnostics Panel	(2) 2-character Debug LEDs, 22 LED display
Internal Button/Switch	Power/Reset buttons and PCIe/BIOS, Slow Mode switches
SATA Controllers	ASMedia ASM1061
Ethernet Controller(s)	(2) Intel i225V (2.5 GbE)
Wi-Fi / Bluetooth	Intel AX210 Wi-Fi 6E (2x2 ax, MU-MIMO, 2.4/5/6 GHz, 160 MHz, BT 5.2)
USB Controllers	ASMedia ASM3142
HD Audio Codec	Realtek ALC1220/NU Audio + Savitech SV3H615 Amp
DDL/DTS Connect	✗ / X
Warranty	3 Years

Inside the Box of the EVGA Z690 Dark Kingpin

Inside the large EVGA Z690 Dark Kingpin box, the company includes a slew of accessories along with the board. You get common items such as SATA cables and a Wi-Fi antenna, as well as some more unique items, including a full-size PCB that displays and labels the major parts on the board. It even comes with a PCIe Release Assist bar (a long piece of PCB to reach the PCIe latch) and additional thermal pads for M.2 modules. Below is a complete list of the included extras:

Case badge
Driver USB disk
Tall standoffs
Screw sets (M.2)
PCIe Release Assist
Probe IT Cable
Thermal pads
(4) SATA cables
Wi-Fi antenna
PCB Legend

Design of the Z690 Dark Kingpin

EVGA

Taking the board out of the box, we’re greeted by a 10-layer black PCB, with sparse gold highlights dotting the board. There are large, actively cooled VRM heatsinks connected via a heat pipe, covered by a vented shroud, with the Z690 Dark branding in white. The biggest difference between this board and others is the rotated socket area, with the RAM slots up top and the VRM bank on the left (normal) and bottom. Along the right edge of the board, you’ll find a large notch cut out and horizontal connectors to help with cable management, and several buttons and switches for overclocking and other board functions. The bottom portion of the board is covered in shrouds and heatsinks, with the Dark branding cut out out of the large heatsink covering all three M.2 sockets.

While there are literally dozens of LEDs on this board, none are RGB (they're mostly white) or designed specifically for looks. If you want to add RGB lighting to your case, you’ll need to use the headers hidden along the bottom edge. Overall, I like the look of the Dark Kingpin. While it won’t stick out on its own, the stealthy black look makes for a great centerpiece for your build.

(Image credit: EVGA)

On the upper-left side, we get a better look at the vented shroud, socket area and the right edge. The circular vents in the shroud allow fresh air to reach the small fan hidden below. The tiny fan was inaudible during our testing, blowing the fresh air over the VRMs and kept the hardy power bits running cool.

Above the socket are the two unreinforced DRAM slots that lock the sticks down on both sides. The Z690 Dark supports DDR5 up to 64GB with speeds listed up to DDR5-6600(OC). As always, your mileage may vary on reaching those speeds but with two slots and using a maximum of two sticks; the potential to reach that high is greater.

Above the DRAM slots on the top edge, instead of seeing 8-pin EPS connectors we run into the Probe IT header, which is a convenient way to measure the system's voltage. EVGA includes an adapter cable that connects to the header, allowing you to check the voltage with a multimeter easily. This is a great feature for the sub-ambient/extreme overclocking crowd or anyone looking to get every last Mhz out of their processor. Just off to the right are a slew (22, to be specific) of status LEDs for varying functionality. It shows the status of PCIe slots, M.2 sockets, voltage, and more. Additionally, there are another dozen or so LEDs on board that also indicate the status of hardware. Please refer to the manual to see what each does. Just off to the right is a vertical USB 2.0 port used for updating the BIOS (this board has THREE) without a CPU installed.

The board's right edge is where most connectivity is located. First, in the upper-right corner, we run into a multi-function POST indicator with two two-character LEDs. You can configure the display to show data in decimal format (default) or hexadecimal. After POST, the same LEDs can display hardware monitoring details, including CPU temperatures and voltages. The Safe Boot and power/reset buttons are near the display.

Next, we hit the first two (of seven available) fan headers (CPU_FAN1/2). These headers support PWM fans and output up to 2A/24W. To the right of the socket between the VRMs and the right edge are two more fan headers (AUX_Fan and PUMP1), while the last two are tucked in the bottom-right corner. The AUX header supports PWM and DC-controlled fans, with the PUMP1 header able to output 3A/36W.

Continuing down the right edge, we spy the 24-pin EPS connector to power the board and the two 8-pin EPS connectors (one required). Below is a USB 3.2 Gen1 header and a front-panel USB 3.2 Gen 2 Type-C header. Moving toward the middle of the board, we run into three different switches. The first (top) is a BIOS slow mode switch. Directly below that is a three-position switch to access all three BIOSes. Finally, five DIP switches are between the BIOS switch and the USB 3.2 Gen1 connector to disable/enable PCIe slots. The only active switches are one and two, which control the two full-length PCIe slots.

(Image credit: Tom's Hardware)

The Dark Kingpin's overclocking pedigree comes with a 21-phase VRM, with 18 phases dedicated to Vcore. Power goes from the two 8-pin EPS connectors (one required) onto a Renesas RAA229131 20-channel controller. Power flows to 18 Intersil ISLL99390 90A power stages. Some napkin math says there is 1,620A available for just the processor, which is plenty for ambient and extreme overclocking. You’ll be limited by the temperatures long before this power delivery hinders progress.

(Image credit: EVGA)

Looking at the left-hand side of the board, near the bottom, we see a mostly hidden Realtek ALC1200 audio codec, along with three large yellow Bennic brand audio capacitors. Also hidden under the shroud is a Savitech SV3H615 amp. EVGA works with NU Audio to help bring high-quality sound to their boards. Although we’d like to see the current-generation audio codec used on this flagship-class motherboard, most users should be pleased with this offering.

In the middle of the board are three PCIe slots and three M.2 sockets. Two of the PCIe slots are reinforced, full-length slots for graphics. Both slots support PCIe 5.0, with the top slot capable of running x16, and the bottom maxing out at x8. If both slots are populated, both run x8/x8. This configuration supports 2-way NVIDIA SLI. (AMD Crossfire users need not apply.)

Sitting between the two full-length slots are three M.2. sockets for storage. Each slot supports up to 110 mm modules and runs up to PCIe 4.0 x4 (64 Gbps). You’ll only be able to use PCIe-based devices, as SATA support is not listed. I would like to see a PCIe 5.0 M.2 socket, however.

Moving right, over the chipset heatsink and Kingpin branding, we hit the U.2 port (why does this still exist?) and eight SATA ports. If the U.2 port is populated, the PCIe x4 slot gets disabled. Six of the SATA ports run from the chipset and support RAID0/1/5/10. The other two ports (top set) run from an ASMedia chip that splits a PCIe x1 lane into two SATA ports. None of the SATA ports or M.2/U.2 sockets get disabled, so if you need a lot of storage, the bandwidth and connectors are here.

Across the bottom are several headers, all hidden from above by shrouds, yielding a clean appearance hiding those unsightly protrusions. The strip includes USB, RGB and other ports/headers. Below is a complete list, from left to right:

Supplemental 6-pin PCIe power
3-pin ARGB header
4-pin RGB header
System Fan header
(2) USB 2.0 headers
Front panel header(s)

(Image credit: EVGA)

The rear IO area sports a pre-installed and perforated IO plate that matches the Dark Kingpin’s black theme. The black background gives way to white writing for the ports and EVGA branding. The stark contrast makes the port labels easy to read.

There are seven USB ports total here: A single USB 3.2 Gen 2x2 (20 Gbps) Type-C header, four USB 3.2 Gen2 (10 Gbps) ports, and two USB 3.2 Gen 1 (5 Gbps) ports. If you’re looking to use the integrated graphics on most Intel processors, it won’t work with this board as there are no video outputs. There’s a legacy PS/2 keyboard/mouse combo port, Wi-Fi antenna connections and two 2.5 GbE ports. There are convenient Reset CMOS and BIOS update buttons (for one of THREE BIOSes) as well. Last is the audio stack: a 5-analog plus SPDIF configuration. The only concern I have here is the seven USB ports here on the rear IO may not be enough for some users (me).

MORE: Best Motherboards

Firmware

EVGA’s BIOS, like other board makers', hasn’t changed much compared to the Z590 generation. Once the system POSTs, you’re still presented with four options (Setup, Default, Gamer Mode, and EVGA OC Robot) to access different functionality. After entering the Setup portion of the BIOS where you can tweak settings, there’s an informative system summary up top. There are headings for different sections below that, with the rest of the screen taken up by options for each header. For the most part, there isn’t much digging around in the sub-menus to find the most frequently accessed items, but CPU and memory overclocking are in different sections. Gamer Mode provides a slight boost to the CPU, while the EVGA OC Robot finds a faster clock speed by raising clocks and playing with voltage while stress testing. Overall, I like EVGA’s BIOS implementation for Z690.

Tom's Hardware

Software

Tom's Hardware

Test System / Comparison Products

We’ve updated our test system to Windows 11 64-bit OS with all updates applied. We kept the same Asus TUF RTX 3070 video card from our previous testing platforms but updated the driver to version 496.13. Additionally, our game selection has been updated, as noted in the table below. We use the latest non-beta motherboard BIOS available to the public unless otherwise noted. The hardware we used is as follows:

CPU	Intel Core i9-12900K
Memory	Kingston Fury DDR5 5200 CL40 (9KF552C40BBK2-32)
	GSkill Trident Z DDR5 5600 CL36 (F5-5600U3636C16GX2-TZ5RK)
	ADATA XPG DDR5 6000 CL40 (AX5U6000C4016G-FCLARBK)
GPU	Asus TUF RTX 3070
Cooling	MSI MEG Coreliquid S360
PSU	EVGA Supernova 850W P6
Software	Windows 11 64-bit (21H2, Build 22000.282)
Graphics Driver	NVIDIA Driver 496.13
Sound	Integrated HD audio
Network	Integrated Networking (GbE or 2.5 GbE)

(Image credit: Tom's Hardware)

EVGA supplied our Supernova 850W P6 power supply (appropriately sized and more efficient than the outgoing 1.2KW monster we used) for our test systems, and G.Skill sent us a DDR5-5600 (F5-5600U3636C16GX2-TZ5RK) memory kit for testing.

EVGA

GSkill

Benchmark Settings

Synthetic Benchmarks and Settings
Procyon	Version 2.0.249 64
	Office Suite, Video Editing (Premiere Pro), Photo Editing (Photoshop, Lightroom Classic)
3DMark	Version 2.20.7290 64
	Firestrike Extreme and Time Spy Default Presets
Cinebench R23	Version RBBENCHMARK330542
	Open GL Benchmark - Single and Multi-threaded
Blender	Version 3.0.1
	Full benchmark (three sub-tests)
Application Tests and Settings
LAME MP3	Version SSE2_2019
	Mixed 271MB WAV to mp3: Command: -b 160 --nores (160Kb/s)
HandBrake CLI	Version: 1.2.2
	Sintel Open Movie Project: 4.19GB 4K mkv to x264 (light AVX) and x265 (heavy AVX)
Corona 1.4	Version 1.4
	Custom benchmark
7-Zip	Version 21.03-beta
	Integrated benchmark (Command Line)
Game Tests and Settings
Far Cry 6	Ultra Preset - 1920 x 1080, HD Textures ON
F1 2021	Ultra Preset - 1920 x 1080, HBAO+, RT Med, TAA + 16xAF, Bahrain, FPS Counter ON

MORE: Best Motherboards

Our standard benchmarks and power tests are performed using the CPU’s stock frequencies (including any default boost/turbo) and with all power-saving features enabled. We set optimized defaults in the BIOS and the memory by enabling the XMP profile. For this baseline testing, the Windows power scheme is set to balanced (default), so the PC idles appropriately.

To get the most out of the Intel Alder Lake chips, you need to be on Windows 11 with its updated scheduler. In most cases, Windows 10 performs well. However, some tests (Cinebench R20, Corona and POVRay) take a significant hit. In short, if you’re going with Alder Lake, you must upgrade to Windows 11 for the best results across the board. That may change with patching and updates in the future, though.

Synthetic Benchmarks

Tom's Hardware

Results in our synthetic benchmarks were above average. In fact, in many of the multi-threaded tests, this board was the fastest tested so far. Single-threaded results in Cinebench and POV-Ray were around average. Memory bandwidth testing in AIDA64 also showed this to be one of the fastest along with the Procyon suite. There’s absolutely nothing to worry about in the synthetic testing.

Timed Applications

Tom's Hardware

In our timed applications, the Dark Kingpin did well overall, tying for the fastest time in Corona (51 seconds) and the x264 and x265 Handbrake tests (111 and 293 seconds, respectively). LAME testing, another lightly threaded application, was slightly below average but not alarming. Here again, great multi-threaded results and average single-threaded.

3D Games and 3DMark

Tom's Hardware

Starting with the launch of the Z690 chipset, we’ve updated our game tests, updating to Far Cry 6 and shifting from F1 2020 to F1 2021. We run the games at 1920 x 1080 resolution using the Ultra preset (details listed above). As the resolution goes up, the CPU tends to have less impact. The goal with these settings is to determine if there are differences in performance at the most commonly used resolution with settings most people use or strive for. We expect the difference between boards in these tests to be minor, with most falling within the margin of error differences. We’ve also added a minimum FPS value, as that can affect your gameplay and immersion experience.

In F1 2021, the Dark averaged 165 frames per second (fps) with minimums of 142 fps, an average result. In Far Cry 6, it managed to be the fastest average at 139 fps, with the minimum right up there with the rest. In the synthetic benchmarks, the board scored 14,478 in Time Spy and 17,049 in Fire Strike Extreme - slightly above-average results. If you plan to game on this motherboard, it’s just as fast, if not negligibly faster, than the rest.

Power Consumption / VRM Temperatures

(Image credit: Tom's Hardware)

We used AIDA64’s System Stability Test with Stress CPU, FPU and Cache enabled for power testing, using the peak power consumption value. The wattage reading is from the wall via a Kill-A-Watt meter to capture the entire PC minus the monitor. The only variable that changes is the motherboard; all other parts are the same.

At idle, the Kingpin consumed 70W, which is about average among Z690 boards (most B660 boards use a lot less). Load wattage peaked at 326W, also around the average of these overclocking-capable boards. There’s nothing out of line when it comes to power consumption.

Tom's Hardware

During stock stress testing, for the first time in a while, we saw no thermal throttling! This is one of the rare boards not to throttle the processor out of the box due to high temperatures. Additionally, the VRMs were some of the coolest we’ve tested, partly because of the tiny VRM fan and large finned heatsinks with a lot of surface area and direct contact with the heat pipe.

Overclocking

Overclocking with Alder Lake is slightly different from what we’re used to due to the hybrid core configuration. Now, you can overclock the P and E cores separately, though they both use the same voltage domain. You can push one and not the other, or both, so there’s some flexibility. We set an overclock to 5.1 GHz across all P-cores and 4.1Ghz on all E-cores. This yields a 200 MHz bump on the P-cores and a 400 MHz increase on the E-cores. We’re topped out on the P-cores, primarily due to temperatures, but the more efficient E-cores still have some headroom.

(Image credit: Tom's Hardware)

Overclocking our Intel Core i9-12900K with the Dark Kingpin proved quite simple. With most BIOS functions in one area (memory and CPU overclocking are in different sections), we manually set our voltage, and off we went. Setting 1.25V yielded around 1.20V according to the multimeter and Eleet X1 software.

On the memory side, we dropped our ADATA DDR4-6000 sticks in, enabled XMP, and off we went without issue. Since this is an overclocking board, I tried my intermediate hand at overclocking these sticks and was able to set DDR5-6400 and lower the CL rating to 30, just like the Tachyon. There’s a lot more headroom available for those with the skills and proper memory kit.

Bottom Line

We like what EVGA brought to the table with the Z690 Dark Kingpin. There's something for everyone, from the all-black styling to the slew of overclocking/monitoring features. The Z690 Dark has plenty of storage options, including eight SATA ports, three M.2 sockets, fast networking with dual 2.5 GbE and integrated Wi-Fi 6E, and last-gen flagship-class audio. If you’re one of the few who need more than 64GB of RAM, you’ll have to look elsewhere since there are only two slots here. Some may find this limiting, but this is an overclocking-focused board that includes all the basics and much more.

That said, the Kingpin is not without some shortcomings. Positioned as a flagship-class motherboard, I wanted to see one PCIe 5.0 M.2 socket, but EVGA chose to support SLI instead. The company designed the board for benchmarking and overclocking, so I certainly understand the choice, especially considering a lack of PCIe 5.0 M.2 devices on the market for this generation. I also wanted to see the latest Realtek audio codec, but most users won’t complain as the last-gen flagship should be more than adequate for all but the most critical listeners.

If we compare flagship to flagship, EVGA’s Z690 Dark Kingpin is the least-expensive (by far) among the other major board partners (ASrock Z690 Aqua $999, Asus ROG Z690 Maximus Extreme $1,099.99, Gigabyte Z690 Extreme $899.99 and MSI Z690 Godlike $1,199.99). And even if you put aside prices for a moment,, it’s a toss-up in terms of features. Some have fewer SATA ports but more M.2 sockets (Gigabyte and MSI). Others integrate a 10GbE and use the latest audio codec. Like many motherboard decisions, it will come down to the features you need, with the look you want at a price you are willing to pay. It’s basically the sameif we compare the Dark Kingpin to the overclocking-focused boards like the Gigabyte Z690 Tachyon ($528.99) and Asus ROG Maximus XIV Apex ($859.99), in that which one is best for you is going to come down to features you need, price and use case. If you’re in the market for a premium Z690 motherboard and perhaps want to try extreme overclocking, or you just want to know all the details, the EVGA Z690 Dark Kingpin has almost all the bases covered. It's a solid option to build your Alder Lake system with, especially if you plan to push the limits of your hardware.

MORE: Best Motherboards

EVGA's $2,500 RTX 3090 Ti Comes With Free 1600W Power Supply

editors@tomshardware.com (Aaron Klotz) — Sat, 02 Jul 2022 16:37:15 +0000

After months of rumors, EVGA has finally released its top-tier overclocking-focused variant of the GeForce RTX 3090 Ti under the Kingpin Hybrid Gaming branding. This card caters to extreme GPU overclockers but competes with the best graphics cards on the market. The Kingpin GeForce RTX 3090 Ti sells for an eye-watering $2,499.99, but to sweeten the deal EVGA has bundled a free 1,600W SuperNova P2 power supply to the 3090 Ti, which is an exceptional deal considering it costs $399.99 all by itself.

Coincidentally, EVGA is also advertising the same deal with the GeForce RTX 3090 Ti FTW3 Ultra Gaming for a limited time with even more stuff to celebrate EVGA's 23rd anniversary. With the FTW3 deal, you get a free EVGA SuperNova 1600 P+ PSU, a 23rd-anniversary shirt, and three months of Xbox Game Pass for PC. It is even better, considering this deal with the GeForce RTX 3090 Ti is on an attractive discount of $1,899.

The RTX 3090 Ti Kingpin is the most advanced model from EVGA's lineup. The card features a 360mm AIO liquid cooler for the GPU and a secondary heatsink and fan for cooling PCB components on the card, such as the power delivery system. Aesthetically the card is decked out in a matte black finish with a large silver accent, with all three fans on the AIO featuring RGB lighting. For the select few customers who won't be putting an LN2 pot or custom water block on this card, you will be greeted by EVGA's K|NGP|N logos everywhere to make it stand out from the rest of EVGA's hybrids. An OLED display is also mounted on the side of the card to show off unique animations or pictures.

EVGA

The most significant selling point of the Kingpin is its incredibly overbuilt power delivery system and very high-quality componentry. In addition, the Kingpin is one of the only GeForce RTX 3090 Ti models to feature dual 16-pin power connectors, giving the card a whopping 1,200W power for overclocking.

It is highly doubtful that the GeForce RTX 3090 Ti core will ever take advantage of 1,200W of power or anything close to that. However, the twin 16-pin connectors ensure the card will never be power throttled under any scenarios, including liquid nitrogen overclocking.

To nobody's surprise, the GeForce RTX 3090 Ti Kingpin also comes with the highest factory overclock of any EVGA GeForce RTX 3090 Ti, with a 1,950 MHz boost clock by default. However, with its monstrous cooling solution, you should be able to overclock the card well past that limit for regular gaming use.

Inflated GeForce GTX 1630 Pricing Strikes Early US and UK Buyers

Mark Tyson — Fri, 01 Jul 2022 15:41:59 +0000

Even before its launch, the Nvidia GeForce GTX 1630 grabbed headlines for the wrong reasons. Since becoming official, the news has become even worse; with China pricing revealed, by making the AMD Radeon RX 6400 look like a hero, and for making the six-year-old GeForce GTX 1050 Ti look like a contender in 2022. Now we have seen popular resellers in the US and UK price various GTX 1630 models at preposterous levels.

EVGA GeForce GTX 1630 SC Gaming - US pricing

The EVGA direct retail page clearly illustrates the absurdity of current GeForce GTX 1630 graphics card pricing. If you visit the EVGA sales page for its GTX 16 series GPUs, you will see the solitary GTX 1630 priced at $199, sitting alongside much more capable GPUs, some at the same or even lower price.

(Image credit: Future)

Of course, cards like the GTX 1660 Ti and 1660 Super are more expensive than the lowly GTX 1630. However, we see that a far more capable GTX 1660 is only $50 more than a GTX 1630, thanks to a $60 instant rebate. Meanwhile, you can purchase GTX 1650 models direct from EVGA starting at $179. With some patience, you could wait for a re-stock of a $159 EVGA GeForce GTX 1650 SC Ultra Black Gaming model. Lastly, an eyebrow must be raised at the "limit 2 per household" notice on EVGA's GTX 1630.

Zotac GeForce GTX 1630 Gaming – UK pricing

There is no need for UK-dwellers to feel left out concerning Nvidia GeForce GTX 1630 price gouging. Popular UK PC maker and components retailer Scan has listed the Zotac GeForce GTX 1630 Gaming at £179.99 (equivalent to $215, but UK prices include 20% VAT).

(Image credit: Future)

Zotac's card is definitely cheaper looking than the EVGA SC Gaming model. It is a compact single fan model, which is entirely acceptable given the TDP. But such designs are usually marketed at MSRP or slightly lower.

The GeForce GTX 1630 might find a home in many compact PCs that don't have a spare 6- or 8-pin power connector. However, if you look at the models available from MSI, for example, you can get compact ITX and low-profile versions of the GTX 1650 (and GTX 1050 Ti) fitting the same criteria on the new and used market.

The feeling that the GeForce GTX 1630 will likely be a misfire from Nvidia isn't helped by these prices. Before launch, we had it pegged as a "below $149" graphics card. However, after launch, we've seen the likes of the Radeon RX 6400 deliver almost 60% better performance. And you can buy these Radeon cards new for about $169 from several brands.

If you are looking for a graphics card this summer, please check out our regularly updated Best Graphics Cards for Gaming in 2022 feature. The Tom's Hardware GPU Benchmarks and Hierarchy 2022 is another tool you should certainly check out.

EVGA Starts to Kill GPU Queue Orders as Inventory Normalizes

editors@tomshardware.com (Aaron Klotz) — Tue, 21 Jun 2022 21:43:39 +0000

EVGA has begun cutting the cord on its GPU Queue program for good as graphics card inventory returns to normal. But, instead of deleting the program all at once, EVGA is killing off its GPU queue orders per GPU model, starting with the EVGA FTW3 models, which have the highest availability. In our case, we received an email that the GeForce RTX 3080 FTW3 Gaming - one of the best graphics cards for gaming, will no longer be available in the GPU queue starting on the 23rd.

With the recent cryptocurrency mining crash and the reduction of GPU prices, it was inevitable that EVGA would no longer need the GPU queue order for buying and selling goods. As a result, we fully expect EVGA to delete the queue at some point. But that will depend entirely on when EVGA receives enough volume for every single graphics card product it sells.

If you are still in the queue for a product, don't worry, you will receive a notification that your queue order will get deleted and that you can buy your graphics card from EVGA's store directly. In our case, we were notified two days in advance that EVGA removed all pending orders for the GeForce RTX 3080 FTW3 Gaming due to an influx of GPU stock for this particular model.

GeForce RTX 3080 (Image credit: EVGA)

We don't know how many other GeForce RTX 30-series (Ampere) GPUs will drop out of the program; all we know about is our GeForce RTX 3080 order. But there's a good chance many other models are getting cut, especially the higher-tier products like the GeForce RTX 3090 Ti and GeForce RTX 3090, which don't see as much selling traffic as lower-end models.

But, despite the high inventory for the GeForce RTX 3080 FTW3 and other models we might not know about, don't expect any discounts on EVGA's store. Most cards are still selling for high prices, with no cards at MSRP. As a result, we recommend you check out other retailers such as Newegg for better GPU prices on RTX 30-series GPUs.

Either way, this is yet another good sign that GPU prices are hurdling down to MSRP, and now is the best time to buy a modern gaming GPU if you don't want to wait for next-generation GPUs.

EVGA Discounts RTX 3090 Ti Below Retail

editors@tomshardware.com (Aaron Klotz) — Wed, 25 May 2022 18:25:48 +0000

With the GPU shortage in full decline, EVGA has introduced some serious price discounts to its RTX 3090 Ti FTW3 lineup featured in the EVGA store. Discounts for the two highest-end FTW3 variants hit the $200 mark, putting these flagship cards in direct competition with the $2000 reference RTX 3090 Ti GPUs. The "entry-level" RTX 3090 Ti FTW3 Black Gaming model is selling for even cheaper, with a discount that nearly cannibalizes EVGA's own vanilla RTX 3090 offering.

EVGA's 3090 Ti FTW3 Ultra Gaming is currently selling for $1999.99 on the EVGA store, $200 below retail. This is the same price as Nvidia's reference RTX 3090 Ti, making this model, with its 1920 MHz overclock and beefy triple fan heatsink a particularly great deal for anyone considering an RTX 3090 Ti GPU.

The EVGA RTX 3090 FTW3 Gaming (non-Ultra variant) is also packing a $200 discount and is currently selling for $1949.99 on the EVGA store. The only difference between the RTX 3090 FTW3 Ultra Gaming GPU and the RTX 3090 FTW3 Gaming GPU is a 30 MHz deficit in boost frequency, which is barely noticeable in gaming applications (and makes this card a confusing option in EVGA's lineup). You'll have to decide whether a 30 MHz difference in boost frequency and the bragging rights of having "Ultra" in your GPU's name is worth an extra $50.

The most interesting GPU in EVGA's arsenal is the baseline RTX 3090 Ti FTW3 Black Gaming GPU. This card is on sale for $1899.99 — it's a smaller discount of $100, but it's the cheapest available RTX 3090 Ti and it undercuts the reference retail price by a full $100. The only feature you'll be sacrificing with the Black Gaming model is a factory overclock.

At $1899.99, the RTX 3090 Ti FTW3 Black Gaming is a very attractive offering in the $1500 to $2000 price bracket, especially because that prices puts it very close to the RTX 3090 (including the RTX 3090 FTW3 Ultra Gaming). It's currently almost impossible to find the RTX 3090 at its MSRP of $1499.99. The EVGA GeForce RTX 3090 FTW3 Ultra Gaming is currently selling for $1799.99 on EVGA's site (discounted $120), which means that for just $100 customers can upgrade from an RTX 3090 to an RTX 3090 Ti. When we're talking about GPUs that cost between $1500 and $2000, $100 is almost nothing in terms of budget.

In this case, that extra cost will net you more performance than going to a higher-end RTX 3090 SKU: The 3090 Ti offers a much higher power limit, more cores, and up to a 10% performance improvement in 4K gaming. If you find the 3090 Ti's power limit to be too great, you can always undervolt for some impressive efficiency gains.

If you're looking for a different GPU, take a look at our breakdown of the current best GPU deals.

EVGA RTX 3060 XC Now Available Near MSRP, Nvidia RTX FE Restock Incoming

ashilov@gmail.com (Anton Shilov) — Thu, 21 Apr 2022 21:07:03 +0000

It has been impossible to purchase a modern gaming graphics card at its recommended price for the past six or seven quarters, but with recent restock campaigns by Nvidia and EVGA, things are slowly getting back to normal. Right now, EVGA is selling its custom GeForce RTX 3060 XC Black Gaming graphics board with a $30 discount from its list price.

The GeForce RTX 3060 XC Black Gaming (12G-P5-3655-KR) is currently in stock at EVGA's online store for $369.99, which is $30 below EVGA's list price (due to a $30 instant rebate). At present, GeForce RTX 3060 boards are selling for around $500 and higher by retailers. Cards like these cost at least $470 on eBay, but most offerings are still around $500 or higher, so EVGA's offering looks quite competitive. Nvidia's MSRP for the GeForce RTX 3060 is $329, so EVGA's RTX 3060 XC Black Gaming is very close to that.

Nvidia and EVGA's restock campaigns are clearly aimed to bring GeForce RTX 30-series graphics boards to gamers at more or less decent prices. It looks like Nvidia has managed to increase supply of its Ampere GPUs either by improving yields at Samsung Foundry or by placing additional orders as prices of graphics board have been declining for several weeks now, according to our frequently updated GPU Pricing Index.

GeForce RTX 30 Series GPUs are returning to shelves! Keep an eye on our Restocked & Reloaded landing page here as we continue to make more supply available →https://t.co/d0ziJIop64April 21, 2022

Interestingly, Nvidia's chief financial officer Colette Kress implied several times that shortages of its up-to-date products would begin to ease around middle-2022 and as we see from pricing and availability of GeForce RTX 30-series offerings, this is what may be happening. In fact, supply situation may get even better in the second half of 2022, Kress said back in January.

(Image credit: EVGA)

EVGA's GeForce RTX 3060 XC Black Gaming graphics card is powered by Nvidia's GA106 GPU with 3584 CUDA cores clocked at up to 1777 MHz (boost), which is in line with Nvidia's recommendations. The board carries 12GB of GDDR6 memory with a 192-bit interface and a 15 GT/s data transfer rate. The card has four display outputs (three DisplayPort 1.4a, one HDMI 2.1) and comes with a rather minimalistic (by today's standards) dual-slot dual-fan cooling system. The device is about 20 cm long, so it will fit into most modern gaming desktops.

EVGA Officially Announces RTX 30 Series Restock For Most High-End Models

editors@tomshardware.com (Aaron Klotz) — Thu, 14 Apr 2022 18:25:34 +0000

EVGA appears to have received a boatload of GPU shipments, with enough volume for the company to announce full availability for a number of its RTX 30 series products. Pricing is also decent, with most cards being available at close to MSRP levels.

EVGA's availability list at the time of this writing consists of RTX 3090 Tis, a single RTX 3090, RTX 3080 Tis, a single RTX 3080, a single RTX 3070 Ti, and a single RTX 3050 model. Most SKUs are also available with a $50-$270 rebate depending on the model.

EVGA is only selling three of its air-cooled RTX 3090 Tis at the moment, including the $1,999 RTX 3090 Ti FTW3 Black, $2,149 RTX 3090 Ti FTW3, and $2,199 RTX 3090 Ti FTW3 Ultra. No liquid cooled models were mentioned.

Only a single RTX 3090 SKU is available, that being the RTX 3090 FTW3 Ultra Gaming for $1699.

For RTX 3080 Tis, EVGA is selling five total SKUs, including air-cooled, hydro copper, and AIO versions. The 3080 Ti XC3 is the cheapest at $1279, and $1299 for the FTW3 Ultra variant. The 3080 Ti XC3 Ultra Hybrid with AIO cooling is going for $1379, the Hydro Copper XC3 for $1429, and the FTW3 Hydro Copper for $1529.

The most amount of SKUs available from the EVGA store is its RTX 3080 lineup with six different models to choose from including five 12GB models and a single 10GB model.

But unfortunately, most of these models are liquid-cooled variants. The only air-cooled model in stock is the RTX 3080 12GB XC3 Ultra for $979.99. AIO variants include the RTX 3080 12GB XC3 Ultra Hybrid for $1299.99 and the FTW3 Ultra Hybrid for $1199. Hydro Copper versions include the RTX 3080 12GB XC3 Ultra Hydro Copper for $1199, and the FTW3 Ultra Hydro Copper for $1249.

The only vanilla RTX 3080 10GB model in stock is the FTW3 Ultra Hydro Copper variant for a respectable $919. Among mid-range and budget-friendly products, it only has the RTX 3070 Ti FTW3 Ultra for $759 and the RTX 3050 XC Gaming for $299 in stock right now.

It's a real shame EVGA doesn't have more mid-range cards, which arguably is what most gamers are shopping for. However, this has been a common issue not only with EVGA's store but with other retailers.

The most elusive cards by far seem to be Nvidia's RTX 3070 8GB models, which we presume have been put on the back burner by Nvidia in favor of building RTX 3070 Tis, since both GPUs are made from the same GA104 die.

EVGA's RTX 2060 KO Ultra Available at $330 MSRP for Budget Gamers

brandon.hill@futurenet.com (Brandon Hill) — Wed, 13 Apr 2022 14:37:32 +0000

As we’ve noted in our frequently updated GPU Pricing Index, graphics card prices are on a steady decline (by roughly 12.5 percent last month), much to the delight of PC gamers. However, pricing is still not where we’d like it to be, compared to MSRP. For example, the GeForce RTX 3080 has an MSRP of $699 (roughly $799 for custom cards), yet the average eBay transaction is just under $1,300.
If you don’t mind grabbing a previous-generation card, EVGA has its GeForce RTX 2060 KO Ultra Gaming card available at its MSRP of $329.99.

While the RTX 2060 isn’t exactly a cutting-edge GPU currently (it’s three years old as of this writing), it is at least faster than Nvidia’s entry-level, Ampere-based RTX 3050. It’s also interesting to note that the average street price for the RTX 3050 is $416, which is nearly $100 more expensive than EVGA’s pricing for the RTX 2060 KO Ultra Gaming.

The RTX 2060 KO Ultra Gaming comes factory overlocked, with a boost clock of 1755 MHz compared to 1680 MHz for standard cards. You’ll also receive 1920 CUDA cores, 6GB of GDDR6 memory, along with DVI-D, HDMI and DisplayPort connections on the I/O plate. This is a dual-fan card with a metal backplate, and it forgoes RGB lighting flourishes for those that prefer a more “stealth” appearance in their gaming rig.

(Image credit: EVGA)

We should note that this card could go out of stock at any moment, but it has been available for sale on EVGA’s website for at least the past 24 hours (and is still going). EVGA limits availability to two cards per person, which seems fair given the current GPU climate.

The one downside relates to shipping, which can be pricey depending on where you live. Standard ground shipping to North Carolina is $17.89, which might be a bit off-putting to those used to free shipping from retailers like Newegg, Amazon and Best Buy. However, beggars can’t be choosers these days when it comes to GPU. The fact that EVGA is offering this at MSRP is a huge plus for budget gamers, and we can only hope that the vendor’s current-generation “mainstream” Ampere cards will also come back in stock soon.

EVGA’s Flagship RTX 3090 Ti Ampere Cards Still in Stock at MSRP

brandon.hill@futurenet.com (Brandon Hill) — Tue, 05 Apr 2022 14:28:19 +0000

Last week, Nvidia launched the most powerful video card for gamers: the GeForce RTX 3090 Ti. The RTX 3090 uses a fully-enabled GA102 GPU delivering 10,752 CUDA cores and 336 Tensor cores compared to 10,496 and 328 for the vanilla RTX 3090, respectively. As expected, the RTX 3090 Ti is an absolute beast on the benchmark circuit, and it comes with an equally beastly MSRP of $1,999 compared to $1,499 for the RTX 3090.

Given that this is currently the fastest consumer graphics card on the market destined for our list of best graphics cards for gaming, you would expect pricing to be through the stratosphere with scalpers looking to take advantage of early adopters. However, EVGA is holding the line on pricing as its RTX 3090 Ti FTW Black Gaming is in stock at its MSRP of $1,999.99. If you want to step into a slightly higher-performing card (1,890MHz boost clock versus 1,860MHz), the RTX 3090 Ti FTW Gaming is in stock at $2,199.99.

Somewhat comically, EVGA is limiting customers to two each of its RTX 3090 Ti cards, but we'd imagine that gamers won't be buying multiples of these cards given their exorbitant pricing.

(Image credit: EVGA)

If you're looking to purchase a Founders Edition version of the RTX 3090 Ti, you're out of luck. Best Buy is the exclusive retailer for Nvidia's Founders Edition graphics cards and currently shows that it is out of stock for shipping. It's a similar situation over at Newegg, as every single RTX 3090 Ti card (from board partners like Asus, Gigabyte, Zotac, MSI) is out of stock, with prices ranging from $1,999 to $2,199.

Interestingly, EVGA has been the only major graphics card OEM that regularly keeps Ampere graphics cards in stock while its competitors have struggled. In addition to the seemingly ample supply of RTX 3090 Ti cards, it also has RTX 3060, RTX 3070 Ti, RTX 3080, RTX 3080 Ti and RTX 3090 cards in stock.

With that said, GPU prices have been on a steady decline for the past few months, which is welcome news to battle-weary gamers who have endured inflated prices for well over a year. And we heard good news from Asus last week when it said that it might slash GPU prices by up to 25 percent this month.

MSI MPG A1000G Power Supply Review

Aris Mpitziopoulos — Sat, 02 Apr 2022 10:00:33 +0000

MSI continues its good path in the PSU market with the MPG A1000G. Still, there is much road to cover until it starts to trouble the intense competition by Corsair and EVGA. The overall performance of the A1000G is high enough, and its build quality is top-notch, something that is shown by its ten-year warranty, too. With a few fixes here and there, it could earn a place in our best PSUs article. Other good alternative options in this category are the Corsair RM1000x (2021) and the EVGA 1000 G6.

MSI enriches its MPG series with a 1000W power supply, with model number A1000G. The flagship of the MPG line uses a different exterior design, which is appealing to our eyes, at least. With the new Intel ATX spec and the PCIe 5.0 standard raising the Wattage bar for upcoming GPUs, all brands have to focus on stronger PSUs that will cope with the increased energy demands.

The MPG A1000G uses Japanese capacitors only, aiming at the higher-end of the Gold efficiency 1kW category. The OEM remains the same as the other MPG members, Channel Well Technology (CWT), one of the best and largest PSU manufacturers, with its headquarters in Taiwan and manufacturing facilities in China and Vietnam.

Tom's Hardware

The MPG A1000G is 80 PLUs Gold and Cybenetics Gold certified in efficiency, while in noise, it is Cybenetics Standard++, meaning that its average noise output is within the 30-35 dBA range. It has compact dimensions with 150mm depth, and it is equipped with a 135mm FDB fan, which will last long. There is also a semi-passive operation which can be disabled, thankfully, if you don't want any heat build-up at the PSU's internals.

Tom's Hardware

Specifications

Manufacturer (OEM)	CWT
Max. DC Output	1000W
Efficiency	80 PLUS Gold, Cybenetics Gold (87-89%)
Noise	Cybenetics Standard++ (30-35 dB[A])
Modular	✓ (Fully)
Intel C6/C7 Power State Support	✓
Operating Temperature (Continuous Full Load)	0 - 50°C
Over Voltage Protection	✓
Under Voltage Protection	✓
Over Power Protection	✓
Over Current (+12V) Protection	✓
Over Temperature Protection	✓
Short Circuit Protection	✓
Surge Protection	✓
Inrush Current Protection	✓
Fan Failure Protection	✗
No Load Operation	✓
Cooling	135mm Fluid Dynamic Bearing Fan (HA13525H12SF-Z)
Semi-Passive Operation	✓ (selectable)
Dimensions (W x H x D)	150 x 85 x 150mm
Weight	1.73 kg (3.81 lb)
Form Factor	ATX12V v2.52, EPS 2.92
Warranty	10 Years

Power Specifications

Rail		3.3V	5V	12V	5VSB	-12V
Max. Power	Amps	22	22	83.5	3	0.3
Watts			120	999.6	15	3.6
Total Max. Power (W)	1000

Cables and Connectors

Description	Cable Count	Connector Count (Total)	Gauge	In Cable Capacitors
ATX connector 20+4 pin (600mm)	1	1	16AWG	No
4+4 pin EPS12V (700mm)	2	2	18AWG	No
6+2 pin PCIe (600mm)	3	3	16AWG	No
6+2 pin PCIe (600mm+150mm)	2	4	16-18AWG	No
SATA (500mm+150mm+150mm+150mm)	3	12	18AWG	No
4 pin Molex (500mm+150mm+150mm+150mm) / FDD (+150mm)	1	4 / 1	18-20AWG	No

There are plenty of cables and connectors to allow the PSU to deliver its full power without issues. It is nice to see three dedicated PCIe cables with thicker 16AWG gauges. The 24-pin ATX connector also uses 16AWG wires for lower voltage drops.

There are no in-cable caps, and all cables are long enough not to have any problems in large chassis. Moreover, the distance between the peripheral connectors is ideal at 150mm.

Tom's Hardware

Component Analysis

`General Data`
Manufacturer (OEM)	CWT
PCB Type	Double Sided
`Primary Side`
Transient Filter	4x Y caps, 2x X caps, 2x CM chokes, 1x MOV
Inrush Protection	1x NTC Thermistor SCK203R0 (3 Ohm) & Relay
Bridge Rectifier(s)	2x GBU1506 (800V, 15A @ 100°C)
APFC MOSFETs	2x Infineon IPA60R099P6 (600V, 24A @ 100°C, Rds(on): 0.099Ohm)
APFC Boost Diode	1x On Semiconductor FFSP1065B (650V, 10A @ 139°C)
Bulk Cap(s)	1x Nichicon (400V, 820uF, 2,000h @ 105°C, GL)
Main Switchers	2x Infineon IPA60R125P6 (600V, 19A @ 100°C, Rds(on): 0.125Ohm)
APFC Controller	Champion CM6500UNX & CM03X
Resonant Controller	Champion CU6901VA
Topology	Primary side: APFC, Half-Bridge & LLC converter Secondary side: Synchronous Rectification & DC-DC converters
`Secondary Side`
+12V MOSFETs	8x International Rectifier IRFH7004PBF (40V, 164A @ 100°C, Rds(on): 1.4mOhm)
5V & 3.3V	DC-DC Converters: 2x UBIQ QM3054M6 (30V, 61A @ 100°C, Rds(on): 4.8mOhm) & 2x UBIQ QN3107M6N (30V, 70A @ 100°C, Rds(on): 2.6mOhm) PWM Controller(s): uPI-Semi uP3861P
Filtering Capacitors	Electrolytic: 2x Nippon Chemi-Con (105°C, W), 1x Nichicon (2-5,000h @ 105°C, HD), 2x Nippon Chemi-Con (1-5,000h @ 105°C, KZE), 2x Nichicon (4-10,000h @ 105°C, HE), 1x Nippon Chemi-Con (4-10,000h @ 105°C, KYA), 1x Rubycon (4-10,000h @ 105°C, YXJ) Polymer: 10x United Chemi-Con, 18x FPCAP
Supervisor IC	Weltrend WT7502R (OVP, UVP, SCP, PG)
Fan Controller	Microchip PIC16F1503
Fan Model	Hong Hua HA13525H12SF-Z (135mm, 12V, 0.5A, Fluid Dynamic Bearing Fan)
`5VSB Circuit`
Rectifier	1x D10PS45L SBR (45V, 10A)
Standby PWM Controller	On Bright OB2365T

Tom's Hardware

This looks to be CWT's CSZ platform, which goes up to 1200W, offering Gold efficiency in both 80 PLUS and Cybenetics standards. The PCB is small for a 1000W PSU, so it is packed with parts. This will make airflow harder, so the fan will have to work more to cool down the internals. Nonetheless, the high-efficiency levels will help in this. The building quality is high despite the leaning heat sink holding the bridge rectifiers.

Tom's Hardware

The transient/EMI filter has all necessary parts, including an MOV and an NTC thermistor supported by a bypass relay. However, the thermistor's resistance should be higher to restrict inrush currents with 230V input more effectively.

Tom's Hardware

The pair of bridge rectifiers is bolted on a heavily leaning heat sink.

Tom's Hardware

The APFC converter uses two Infineon FETs and a powerful boost diode provided by On Semiconductor. The bulk cap is by Nichicon and has enough capacity to offer more than 17ms hold-up time.

Tom's Hardware

The primary switching Infineon FETs are installed in a half-bridge topology. They are accompanied by an LLC resonant converter for lower power losses on both primary and secondary sides.

Tom's Hardware

Eight FETs regulate the 12V rail, and the minor rails use four UBIQ FETs and have a common PWM controller.

Tom's Hardware

All filtering caps are of good quality. Besides electrolytic caps, many polymer ones are used for ripple filtering purposes.

Tom's Hardware

The standby PWM controller is an On Bright OB2365T IC, and the rectifier on the 5VSB circuit's secondary side is a D10PS45L SBR.

Tom's Hardware

Several bus bars for power transfer on the modular board, along with a number of polymer caps and a large electrolytic.

(Image credit: Tom's Hardware)

The main supervisor IC is a Weltrend WT7502R.

Tom's Hardware

Soldering quality is good.

Tom's Hardware

The cooling fan measures 135mm across and uses a fluid dynamic bearing. Hong Hua delivers good products at reasonable prices, and this is why it dominates the PSU market.

MORE: Best Power Supplies

To learn more about our PSU tests and methodology, please check out How We Test Power Supply Units.

Corsair RM1000x (2021)

Super Flower Leadex Pro Platinum 1000

EVGA SuperNOVA 1000 G6

Primary Rails And 5VSB Load Regulation

Tom's Hardware

Load regulation is tight enough at 12V and pretty tight at 5VSB. On the 5V and 3.3V rails, the PSU doesn't meet the competition, though.

Hold-Up Time

Put simply; hold-up time is the amount of time that the system can continue to run without shutting down or rebooting during a power interruption.

Tom's Hardware

The hold-up time is slightly above 17ms, but the power ok signal's hold-up time doesn't reach 16ms.

Inrush Current

Tom's Hardware

The inrush current is increased with 230V. A higher resistance NTC thermistor would help here.

Leakage Current

(Image credit: Tom's Hardware)

Leakage current is at normal levels.

10-110% Load Tests

These tests reveal the PSU's load regulation and efficiency levels under high ambient temperatures. They also show how the fan speed profile behaves under increased operating temperatures.

Test	12V	5V	3.3V	5VSB	DC/AC (Watts)	Efficiency	Fan Speed (RPM)	PSU Noise (dB[A])	Temps (In/Out)	PF/AC Volts
10%	6.483A	1.993A	2.011A	0.991A	100.005	86.193%	0	<6.0	44.97°C	0.977
	12.094V	5.018V	3.282V	5.045V	116.025				40.59°C	115.16V
20%	14.007A	2.993A	3.021A	1.192A	199.962	90.913%	0	<6.0	45.89°C	0.986
	12.069V	5.013V	3.277V	5.034V	219.949				41.03°C	115.16V
30%	21.881A	3.494A	3.528A	1.371A	300.013	91.562%	0	<6.0	46.83°C	0.987
	12.063V	5.009V	3.274V	5.105V	327.66				41.45°C	115.16V
40%	29.732A	3.997A	4.037A	1.569A	399.679	91.333%	0	<6.0	47.43°C	0.986
	12.057V	5.004V	3.27V	5.099V	437.604				41.62°C	115.15V
50%	37.253A	5A	5.051A	1.769A	499.432	90.775%	416	7.5	42.34°C	0.987
	12.051V	5.001V	3.267V	5.088V	550.188				48.41°C	115.15V
60%	44.842A	6.004A	6.067A	1.97A	599.952	90.06%	643	18.1	42.49°C	0.989
	12.045V	4.998V	3.264V	5.078V	666.168				49.2°C	115.14V
70%	52.377A	7.009A	7.085A	2.171A	699.692	89.206%	854	27.2	43.77°C	0.991
	12.039V	4.995V	3.261V	5.067V	784.353				50.97°C	115.14V
80%	60.001A	8.002A	8.103A	2.274A	799.652	88.422%	1041	33.1	43.96°C	0.992
	12.030V	4.992V	3.258V	5.057V	904.361				52.32°C	115.14V
90%	67.959A	8.522A	8.602A	2.377A	899.506	87.501%	1427	42.4	44.5°C	0.993
	12.023V	4.988V	3.255V	5.048V	1027.993				53.55°C	115.14V
100%	75.722A	9.031A	9.133A	2.982A	999.504	86.457%	1767	48.1	45.63°C	0.994
	12.016V	4.983V	3.251V	5.031V	1156.079				55.62°C	115.14V
110%	83.431A	10.043A	10.254A	2.987A	1100.15	85.091%	2149	52.8	46.69°C	0.995
	12.008V	4.978V	3.247V	5.022V	1292.906				57.56°C	115.14V
CL1	0.116A	14.407A	14.621A	0A	121.298	84.076%	560	13.2	43.23°C	0.986
	12.097V	5.011V	3.262V	5.042V	144.271				48.43°C	115.18V
CL2	0.115A	21.846A	0A	0A	111.39	82.211%	868	27.8	41.61°C	0.981
	12.101V	5.035V	3.276V	5.055V	135.494				48.69°C	115.17V
CL3	0.115A	0A	22.242A	0A	73.988	75.94%	940	30.3	40.18°C	0.977
	12.095V	5.008V	3.264V	5.042V	97.43				49.29°C	115.17V
CL4	83.215A	0A	0A	0A	1000.08	87.244%	1504	43.9	42.62°C	0.994
	12.018V	5.005V	3.272V	5.094V	1146.31				54.03°C	115.13V

No problems in full power delivery under tough conditions, but efficiency takes a hit and the cooling fan has to spin at high speeds, so it makes a loud noise.

20-80W Load Tests

Test	12V	5V	3.3V	5VSB	DC/AC (Watts)	Efficiency	Fan Speed (RPM)	PSU Noise (dB[A])	Temps (In/Out)	PF/AC Volts
20W	1.228A	0.496A	0.5A	0.197A	19.999	49.037%	0	<6.0	40.07°C	0.932
	12.099V	5.044V	3.297V	5.065V	40.783				36.85°C	115.17V
40W	2.702A	0.694A	0.701A	0.296A	39.997	76.284%	0	<6.0	41.29°C	0.959
	12.097V	5.043V	3.297V	5.062V	52.432				37.81°C	115.16V
60W	4.176A	0.895A	0.903A	0.395A	59.996	81.677%	0	<6.0	41.78°C	0.968
	12.097V	5.029V	3.288V	5.059V	73.455				38.02°C	115.16V
80W	5.648A	1.095A	1.105A	0.495A	79.952	85.456%	0	<6.0	43.55°C	0.977
	12.095V	5.022V	3.285V	5.055V	93.559				39.49°C	115.16V

Efficiency is bottom low with 20W, and we would like to see over 80% with 40W load.

2% or 10W Load Test

12V	5V	3.3V	5VSB	DC/AC (Watts)	Efficiency	Fan Speed (RPM)	PSU Noise (dB[A])	Temps (In/Out)	PF/AC Volts
1.472A	0.255A	0.255A	0.053A	20.199	50.084%	0	<6.0	32.45°C	0.932
	12.094V	5.044V	3.297V	5.068V	40.33			31.46°C	115.17V

Efficiency with 2% load is too low. It should be above 60% and ideally above 70%.

Efficiency & Power Factor

Tom's Hardware

Efficiency is high enough with normal loads, but low with light loads and bottom load with super-light loads.

5VSB Efficiency

Test #	5VSB	DC/AC (Watts)	Efficiency	PF/AC Volts
1	0.1A	0.507W	76.499%	0.065
	5.069V	0.663W		115.16V
2	0.25A	1.267W	78.796%	0.147
	5.066V	1.608W		115.16V
3	0.55A	2.785W	79.512%	0.259
	5.061V	3.503W		115.17V
4	1A	5.055W	79.438%	0.347
	5.054V	6.363W		115.17V
5	1.5A	7.57W	79.819%	0.396
	5.045V	9.484W		115.17V
6	3A	15.063W	78.297%	0.46
	5.021V	19.239W		115.16V

Tom's Hardware

The 5VSB rail is not efficient.

Power Consumption In Idle And Standby

Mode	12V	5V	3.3V	5VSB	Watts	PF/AC Volts
Idle	12.079V	5.045V	3.299V	5.07V	4.14	0.35
						115.17V
Standby					0.025	0.002
						115.17V

Tom's Hardware

Vampire power is low and this is good, of course, since it means that the PSU has minimized energy needs in standby.

Fan RPM, Delta Temperature, And Output Noise

All results are obtained between an ambient temperature of 37 to 47 degrees Celsius (98.6 to 116.6 degrees Fahrenheit).

(Image credit: Tom's Hardware)

The fan speed profile is linear, but due to the compact dimensions of the populated PCB, it applies high speeds at increased loads, which impacts noise output.

The following results were obtained at 30 to 32 degrees Celsius (86 to 89.6 degrees Fahrenheit) ambient temperature.

(Image credit: Tom's Hardware)

At normal operating temperatures, close to 30 degrees Celsius, the PSU is silent up to 600W loads. The 30 dBA mark is passed at around 740W and the noisy operating starts with 940W.

MORE: Best Power Supplies

Protection Features

Check out our PSUs 101 article to learn more about PSU protection features.

OCP (Cold @ 24°C)	12V: 104A (124.55%), 11.966V 5V: 30.7A (139.55%), 5V 3.3V: 30.2A (137.27%), 3.248V 5VSB: 4.7A (156.67%), 4.990V
OCP (Hot @ 41°C)	12V: 104.2A (124.79%), 11.983V 5V: 30.8A (140%), 5.004V 3.3V: 30.2A (137.27%), 3.247V 5VSB: 4.7A (156.67%), 4.988V
OPP (Cold @ 26°C)	1243.38W (124.34%)
OPP (Hot @ 45°C)	1243.4W (124.34%)
OTP	✓ (154°C @ secondary side)
SCP	12V to Earth: ✓ 5V to Earth: ✓ 3.3V to Earth: ✓ 5VSB to Earth: ✓ -12V to Earth: ✓
PWR_OK	Accurate but lower than 16ms
NLO	✓
SIP	Surge: MOV Inrush: NTC Thermistor & Bypass relay

OCP is properly set at 12V, but we cannot say the same for the minor rails, especially 3.3V. There is no point in delivering so many Amps at 3.3V. Lastly, OPP is properly configured, too, and there is over temperature protection.

DC Power Sequencing

Tom's Hardware

No problems here since the 3.3V rail is always lower than the other two.

Cross Load Tests

Load Regulation Charts

Tom's Hardware

Efficiency Graph

(Image credit: Tom's Hardware)

Ripple Graphs

The lower the power supply's ripple, the more stable the system will be and less stress will also be applied to its components.

Tom's Hardware

Infrared Images

Tom's Hardware

The hottest part is the board holding the 12V FETs. With the fan in operation, the temperatures on this board will be under control, but it requires good airflow.

MORE: Best Power Supplies

Advanced Transient Response Tests

For details about our transient response testing, please click here.

Advanced Transient Response at 20% – 20ms

Voltage	Before	After	Change	Pass/Fail
12V	12.051V	11.878V	1.43%	Pass
5V	5.011V	4.910V	2.01%	Pass
3.3V	3.277V	3.158V	3.63%	Pass
5VSB	5.032V	4.966V	1.32%	Pass

Advanced Transient Response at 20% – 10ms

Voltage	Before	After	Change	Pass/Fail
12V	12.054V	11.888V	1.38%	Pass
5V	5.011V	4.911V	2.01%	Pass
3.3V	3.277V	3.158V	3.63%	Pass
5VSB	5.032V	4.967V	1.29%	Pass

Advanced Transient Response at 20% – 1ms

Voltage	Before	After	Change	Pass/Fail
12V	12.056V	11.942V	0.94%	Pass
5V	5.012V	4.922V	1.80%	Pass
3.3V	3.277V	3.155V	3.74%	Pass
5VSB	5.032V	4.982V	1.00%	Pass

Advanced Transient Response at 50% – 20ms

Voltage	Before	After	Change	Pass/Fail
12V	12.027V	11.914V	0.94%	Pass
5V	4.998V	4.915V	1.65%	Pass
3.3V	3.266V	3.144V	3.72%	Pass
5VSB	5.084V	5.037V	0.93%	Pass

Advanced Transient Response at 50% – 10ms

Voltage	Before	After	Change	Pass/Fail
12V	12.030V	11.921V	0.91%	Pass
5V	4.999V	4.911V	1.75%	Pass
3.3V	3.266V	3.138V	3.92%	Pass
5VSB	5.085V	5.048V	0.74%	Pass

Advanced Transient Response at 50% – 1ms

Voltage	Before	After	Change	Pass/Fail
12V	12.032V	11.936V	0.80%	Pass
5V	4.999V	4.913V	1.71%	Pass
3.3V	3.266V	3.138V	3.91%	Pass
5VSB	5.085V	5.044V	0.80%	Pass

Tom's Hardware

Transient response is average at 12V and 5V and good at 5VSB, which doesn't matter much anyway. At 3.3V, the deviations are not large, but the rail drops way below 3.2V when the transient load is applied because of the low initial voltage.

Turn-On Transient Tests

Tom's Hardware

There is a small voltage overshoot at 5VSB, which is nothing to worry about. Both 12V slopes are almost perfect.

Power Supply Timing Tests

There are several signals generated by the power supply, which need to be within specified, by the ATX spec, ranges. If they are not, there can be compatibility issues with other system parts, especially mainboards. From year 2020, the PSU's Power-on time (T1) has to be lower than 150ms and the PWR_OK delay (T3) from 100 to 150ms, to be compatible with the Alternative Sleep Mode.

PSU Timings Table
T1 (Power-on time) & T3 (PWR_OK delay)
Load	T1	T3
20%	50ms	140ms
100%	50ms	140ms

Tom's Hardware

The PWR_OK delay is within the 100-150ms region, so the PSU supports the Alternative Low Power Mode (ALPM) recommended by the ATX v3.0 spec.

Ripple Measurements

The ripple limits, according to the ATX specification, are 120mV (+12V) and 50mV (5V, 3.3V, and 5VSB).

Test	12V	5V	3.3V	5VSB	Pass/Fail
10% Load	4.9 mV	6.2 mV	4.7 mV	6.2 mV	Pass
20% Load	11.4 mV	5.9 mV	4.8 mV	6.0 mV	Pass
30% Load	9.1 mV	6.2 mV	6.7 mV	6.8 mV	Pass
40% Load	9.2 mV	6.5 mV	5.4 mV	7.4 mV	Pass
50% Load	9.5 mV	6.8 mV	5.7 mV	7.3 mV	Pass
60% Load	10.3 mV	7.6 mV	5.8 mV	7.3 mV	Pass
70% Load	10.5 mV	8.1 mV	6.2 mV	8.4 mV	Pass
80% Load	11.8 mV	8.6 mV	10.3 mV	8.7 mV	Pass
90% Load	12.2 mV	9.3 mV	10.8 mV	8.9 mV	Pass
100% Load	21.0 mV	10.1 mV	11.5 mV	10.5 mV	Pass
110% Load	21.2 mV	10.6 mV	12.0 mV	11.1 mV	Pass
Crossload 1	10.1 mV	8.7 mV	11.8 mV	10.6 mV	Pass
Crossload 2	7.7 mV	11.2 mV	5.7 mV	10.1 mV	Pass
Crossload 3	6.6 mV	5.8 mV	15.0 mV	9.6 mV	Pass
Crossload 4	19.4 mV	8.8 mV	6.7 mV	13.4 mV	Pass

Tom's Hardware

Ripple suppression is good.

Ripple At Full Load

Tom's Hardware

Ripple At 110% Load

Tom's Hardware

Ripple At Cross-Load 1

Tom's Hardware

Ripple At Cross-Load 4

Tom's Hardware

EMC Pre-Compliance Testing – Average & Quasi-Peak EMI Detector Results

Electromagnetic Compatibility (EMC) is the ability of a device to operate properly in its environment without disrupting the proper operation of other nearby devices.

Electromagnetic Interference (EMI) stands for the electromagnetic energy a device emits, and it can cause problems in other nearby devices if too high. For example, it can be the cause of increased static noise in your headphones or/and speakers.

(Image credit: Tom's Hardware)

EMI emissions are high, in the 522 to 882 KHz range.

MORE: Best Power Supplies

Performance Rating

(Image credit: Tom's Hardware)

The overall performance is not that far from other offerings, but it is not competitive.

Noise Rating

The graph below depicts the cooling fan's average noise over the PSU's operating range, with an ambient temperature between 30 to 32 degrees Celsius (86 to 89.6 degrees Fahrenheit).

(Image credit: Tom's Hardware)

Given its capacity, it is not super-noisy, but there is room for improvement in this section. The king, by far, in noise output is the Asus Rog Thor model, thanks to its beefy heatsinks.

Efficiency Rating

The following graph shows the PSU's average efficiency throughout its operating range with an ambient temperature close to 30 degrees Celsius.

(Image credit: Tom's Hardware)

The average efficiency is high enough for a Gold-rated unit.

Power Factor Rating

The following graphs show the PSU's average power factor reading throughout its operating range with an ambient temperature close to 30 degrees Celsius and 115V/230V voltage input.

Tom's Hardware

The APFC converter does a good job, especially with 230V input.

MORE: Best Power Supplies

MSI strengthens its cooperation with Channel Well Technology, one of the largest PSU OEMs, enriching its MPG line with a 1000W model. With the upcoming generation of GPUs and with the introduction of ATX12V v3.0 spec, it became clear that strong PSUs will be a one-way road to deal with GPU power spikes effectively. This means that we will see more high-capacity PSUs in the following months hitting the market, which will have to meet the newest ATX spec if they want to remain in the game.

For units like the MSI MPG A1000G, which is not ATX v3.0 compatible, we guess there will be revisions with all necessary changes. Please keep in mind that the new ATX spec changes will affect the PSU prices, so if you don't plan on upgrading any time soon, once the new GPUs become available, it would be more affordable to get one of the current PSUs.

(Image credit: Tom's Hardware)

The MSI MPG A1000G offers good performance, but it doesn't pose a threat to the tough competition. That said, these are the first steps of MSI into the PSU business and so far things go well, unlike its major opponent, Gigabyte. The MPG A1000G has enough power to meet the demands of the current generation GPUs and thanks to its compact dimensions, it will fit in any ATX case. If you want something with lower noise output, a solid choice is the Asus Rog Thor 1000 (2021), which is more expensive. Other good alternative options are the Corsair RM1000x (2021) and the EVGA 1000 G6.

MORE: Best Power Supplies