Thermalright is well known for delivering some of the best coolers on the market, usually available at extremely competitive prices. So today’s review of a Thermalright AIO featuring a price tag of $260 USD might come as a shock to some of our readers.

What makes this AIO different from the lower-priced, value options we’re accustomed to seeing? Well,it has a large 2400x1800 6.7-inch anamorphic screen with a slightly oversaturated look that makes colors “pop,” a wide variety of preset customization options – and the screen can be detached and mounted elsewhere in (or outside) the case.

Will this unusual option make our list of best CPU coolers? Well, let's just say we might have to make a new category covering curved LCD AIOs in our recommendations list.

Cooler specifications

Packing and included contents

Thermalright’s Wonder Vision 360 is packaged in a reflective, jet-black box.

The inner contents are protected by molded foam, carboard, and plastic coverings.

Included with the product are the following:

• 6.7-inch anamorphic display
• 360mm radiator with pre-installed fans
• Mounting for AMD and Intel platforms
• Thermal Paste

Features of the Thermalright Wonder Vision 360 UBB ARGB Black

▶️ Impressive 6.7-inch curved display

The standout feature of this AIO is the beautiful 6.7-inch display, featuring an ultrawide 2400x1080 resolution with a curve for an anamorphic 3D effect with moving media. The display’s colors seem slightly oversaturated, making them “pop,” or seem more vivid than reality.

▶️ TRCC Software

To control the display, you’ll need to install Thermalright’s TRCC software. There are an insane number of preset visuals to choose from here — over 200 by my count!

And of course, there are multiple ways to display performance metrics, in addition to the visual themes.

▶️ The screen is detachable

The included screen is detachable and doesn’t have to be connected to the cooler in order to function. This means that you can place the screen in any location inside of your PC’s case, or even externally. If the included USB cable that attaches to the motherboard isn’t long enough, you can simply replace the cable with a standard USB cable and connect it to an external USB port.

Still pictures don’t quite do this screen justice. This is a seriously eye-catching display, which makes the ability to mount it wherever you want all the more appealing.

▶️ Fully rotatable tubing, management clips

The AIO’s tubing is fully rotatable in any direction.

Three management clips are included to keep the liquid tubing organized and clean-looking.

▶️ Pre-installed fans

There’s more to a cooler than just the heatsink or radiator. The bundled fans have a significant impact on cooling and noise levels, as well as how the cooler looks in your case.

The fans included with this cooler have a unibody design, with a single cable for ARGB and PWM connections – simplifying cable management.

The blades of the fans are solid black, but there are ARGB strips along the sides of the fans and on the sides of the radiator.

AM5 installation

The installation of this cooler is simple, and assumes the you’ll have secured the radiator to their case before these steps. This guide will go over the installation on AMD systems.

1. Remove the default AM4/AM5 mounting mechanism.

2. Place the rubber standoffs around the screw holes surrounding the CPU.

3. Take the included mounting bars and place them on top of the standoffs, and then secure them with a screwdriver and the included screws.

4. Apply the included Arctic MX-6 thermal paste to your CPU. If you aren’t sure how to do this properly, please refer to our handy guide on how to apply thermal paste.

5. Press the pump block on top of the CPU, and secure the built-in screws using a screwdriver.

6. Once complete, place the 6.7-inch display on top of the CPU block and connect the included cable to a USB header, as well as the AIO’s PWM and ARGB cords.

Real world testing configuration – AMD AM5 platform

The 9950X3D setup we started testing coolers with last year has suffered an unfortunate death. As such, we’ve begun to test coolers on a system with AMD’s 9950X.

There are some differences in how the 9950X and 9950X3D CPUs are impacted by thermal events. While the heat output of the CCDs of AMD’s 9950X3D are relatively balanced, the 9950X I’m using has one CCD that runs much hotter than the other – with a difference of over 10C in some scenarios, shown below.

The benchmark results shared in these reviews may differ from others because I place an emphasis on having results that are comparable to real-world use.

This means that I generally test CPU coolers inside of a closed desktop case, which increases cooling difficulty compared to other testing methods. Some reviewers will test CPU coolers outside of a case, on an open test bench. Open benches have lowered ambient temperatures, which in turn makes weak coolers appear stronger than they really are. Some publications have also used generic thermal plates to test cooling solutions. I reject both of these methods because they don’t accurately reflect the real-world conditions where a CPU cooler is.

Noise-normalized testing

Most testing is performed with the cooler tied to the default fan curve of our MSI x870E Carbon motherboard. But some of y’all prefer to see tests when the noise levels of coolers are equalized. This is especially important to those of you who prefer silent computers.

We’ll start with the “easiest” of these tests, which runs Cinebench R23 with a “stock” power limit. On my motherboard, this means the CPU will consume ~200W. The results here aren’t quite as impressive as we hoped, falling behind other 360mm AIOs.

The next test is a bit harder, with PBO enabled to allow the CPU to use as much power as can be handled – where it handled 237W, on average.

For our last test, we’ve added the heat of a GPU to the mix, which contributes 295W of added thermals into the PC case. This result is similar to our other noise-normalized tests, and isn’t very impressive. However, when tied to a typical fan curve, this AIO performs much better – as you’ll see in the last section.

Noise levels

We’ve measured noise levels at two points to give you an idea of how the cooler operates when tied to a motherboard’s default fan curve. Our first measurement is when the fans are allowed to run at full speeds, the loudest noise levels possible.

In terms of maximum volume, Thermalright’s AIO reaches 47.2 dBA. This is louder than I would prefer, but seems to be about “middle of the road” in comparison to other high-performance liquid cooling systems these days.

Our next acoustic measurement is the cooler’s volume when the CPU consumes 150W. This might be similar to your “worst case” gaming scenarios, especially if PBO is enabled. Noise levels are very good here, reaching only 38.9 dBA. This means the cooler runs quietly in most common use cases, and will only become noisy in high-intensity workloads.

PBO Performance – full fan speeds

While some of you prefer to run coolers at low noise levels, I know many just want to see maximum performance. This test lets the fans run at their full speeds, show the full potential of the AIO’s cooling potential. Turning on PBO allows AMD’s Ryzen 9 9950X to stretch its legs and guzzle power, pushing the limits of any CPU cooler on the market.

The Wonder Vision 360 UB cooled 255W in this test, only 3W behind the best results we’ve had from a 360mm AIO thus far.

200W thermal benchmarks and noise levels

For the next thermal test, I leave motherboard settings at their defaults, which results in a power limit of 200W when running Cinebench R23. CPU temperatures here averaged 74.7C, which is rather impressive when you consider that our best result – with Arctic’s Liquid Freezer III Pro 420 – was only 1.1C cooler!

150W thermal benchmarks and noise levels

For the next thermal test, I’ve set the power limit to 150W.

In this lower-intensity scenario, the Wonder Vision 360 UB showed off, delivering the best temperatures we’ve seen on this AMD Ryzen 9 9950X system, reaching an average of only 59.1 degrees Celsius.

100W thermal results

Our next test is the least-difficult featured in our cooling reviews, with a workload consuming only 100W. In this lower-intensity scenario, Thermalright’s Wonder Vision 360 UB again had the best results of the 360mm AIOs we’ve tested on this system, with a temperature averaging a ‘chilly’ 47.6 degrees C.

Conclusion

With the Wonder Vision 360 UB ARGB Black, Thermalright has shown that the company isn’t just an effective competitor for budget cooling – it’s also fully capable of producing high-end, premium products.

This AIO’s price of $205.90 might seem steep at first, until you consider that curved-screen competitors from Tryx and others often have prices closer to $300 or more. The included 6.7-inch ultrawide display is gorgeous, and features hundreds of visual theme options, courteously of Thermalright’s TRCC software. And the ability to use the screen independently of the cooler might appeal to those wanting to customize their PC cases in unique ways.

Lack of Linux support for internal PC build components, such as RGB lighting and internal LCD screens, is a big headache that Linux users normally have to get around by installing Windows just to install the control software for these devices to make them function. However, a Redditor has decided to fix this problem for his personal Thermalright LCD cooler and has built a full-fledged Linux port of the Thermalright Linux software all by himself. The Redditor's new Linux build of Thermalright's control center software can be found on GitHub.

The software is feature-complete and supports a wide variety of Thermalright coolers, including the Frozen Warframe/SE, Frozen Vision V2/Core Vision/Elite Vision, Frozen Horizon Pro/Magic Pro, LC1/LC2/LC3/LC5 AIO pump heads, AK120, AX120, PA120 Digital, and Wonder Vision CZTV.

This Linux version of the Thermalright app includes all the features necessary to fully utilize Thermalright's supported devices. It comes with a GUI copied from the Windows version "pixel-for-pixel", featuring local themes, and can download extra themes from the cloud. Video and GIF playback is also supported, with options for overlaying hardware information, such as CPU and GPU temperatures, on the screens of supported devices.

The Redditor who created the app has only tested his Linux version on Fedora so far, but states the app technically supports all Linux distros.

Linux's very low consumer population has made it practically invisible to most component manufacturers that build internal PC components that require dedicated software to customize or control. Most of these apps only support Windows and MacOS.

This problem has forced the Linux community to build its own third-party dedicated apps to rectify support issues on Linux. For instance, a well-known RGB application on Linux is OpenRGB, which can interface with lots of RGB devices, including motherboards, RAM, GPUs, fans, mice, and keyboards. With Linux support now present through a third party, Thermalright owners of the aforementioned supported coolers no longer need to have a secondary Windows installation on hand to control their coolers' screens. That said, this Linux version is not an official version supported by Thermalright, so there could be bugs.

In ancient Rome, the term Pretor (more commonly spelled Praetor) was given to magistrates and commanders of armies – and boy, does the title fit Thermalrights' latest flagship air cooler, the Royal Pretor 130. The performance of this latest product is ferocious, blurring the lines between what you should expect from air and liquid coolers.

Will the Royal Pretor 130 make our list of the best coolers? Yes. As we’ll see in testing, it’s the best air cooler on the market! Let’s take a look at the specifications and features of the cooler, then we’ll go over thermal performance with both Intel and AMD CPUs, as well as noise levels.

Cooler specifications

Packing and included contents

The packaging for the cooler features a slick white and black design. Opening it reveals the parts, protected by molded foam, cardboard, and plastic coverings.

Included in the box are the following:

• One 28x120mm fan
• One 28x130mm fan
• Dual-tower heatsink
• TF7 Thermal paste
• Mounting accessories for modern AMD & Intel platforms
• Installation manual

Features of Thermalright’s Royal Pretor 130

▶ Dual-tower heatsink

The heatsink has two silver towers, with a black etched metal top showcasing Thermalright’s brand name and logo.

▶ High-quality TF7 thermal paste

Thermalright includes its TF7 thermal compound with the cooler, which offers good performance, only a couple of degrees away from the best pastes you can buy.

▶ Six copper heatpipes

The Royal Pretor has six 6mm copper heatpipes to move heat away from the CPU and into the fins of the towers.

▶ Two fans of different sizes – 120mm, 130mm

There’s more to a cooler than just the heatsink or radiator. The bundled fans have a significant impact on cooling and noise levels, as well as how the cooler looks in your case. This cooler arrives with two different types of fans. Both are 28mm thick, but one is 120 mm and the other is 130 mm.

Real world testing configuration: Intel LGA1700 and AMD AM5 platform

My results may differ from others because I emphasize results that are comparable to real-world use. This means that I test CPU coolers inside of a closed desktop case, which increases cooling difficulty compared to other testing methods.

Many will test CPU coolers outside of a case, on an open test bench. Open benches have lowered ambient temperatures, which makes weak coolers appear stronger than they are. Some publications have also used generic thermal plates to test cooling solutions. I reject both of these methods because they don’t accurately reflect the real-world conditions a CPU cooler is used in.

My previous reviews have tested Intel’s latest platform, using the Core Ultra 9 285K Arrow Lake CPU. But we’re retiring this from our testing suite. Between BIOS changes and Windows updates, Arrow Lake’s thermal characteristics have changed in some scenarios, rendering much of our previous testing data useless.

With today’s review, we’re also testing AMD’s Ryzen 9 9950X3D. This is a beast of a CPU, providing the best gaming and multithreaded performance on the market. It can prove quite challenging thermally when PBO is enabled for overclocking.

AM5 and 1851 installation

The installation of this cooler is simple for both Intel and AMD CPUs.

1. You’ll first need to apply the included backplate if you’re using an Intel CPU. AMD users will remove the default mounting mechanism.

2. Next, you’ll set the rubber standoffs on both Intel and AMD systems.

3. Then you’ll want to take the mounting bars and place them on top of the standoffs, securing them with the included screws.

4. Apply the included thermal paste to your CPU. If you have any questions on how to do this properly, please refer to our handy guide on how to apply thermal paste.

5. Mount the heatsink on top of the CPU, using a screwdriver to secure the screws in the middle of the unit.

6. Attach the fans to the heatsink using the included clips, and then use the included PWM cable to connect the fans to the motherboard.

Noise Normalized CPU-only thermal results

Finding the right balance between fan noise levels and cooling performance is important. While running fans at full speed can improve cooling capacity to some extent, the benefits are limited and many users prefer a quieter system. Many consider these tests to be the most important of all, so we’ll start today’s review by looking at noise-normalized performance.

We’ll be looking at two different, distinct tests. We’ll first look at our traditional noise-equalized results with Intel’s i7-14700K, which places a full CPU load with the cooler’s noise levels set to 38.9 dBA.

The performance of Thermalright’s Royal Pretor 130 was astounding on htis test, blurring the line between air and liquid coolers, with 243W cooled during this test – on par with ID-Cooling’s DX360 Max, a budget 360mm AIO. This is a full 7 W better than the best results we’d previously seen with an air cooler, from Chinese manufacturer Jiushark’s JF800 which cooled 236W in the same scenario.

Now let’s take a look at our noise-equalized tests utilizing AMD’s Ryzen 9 9950X3D CPU. Two changes (in comparison to our first test) have been made to increase the difficulty of this challenge. The noise level of the CPU coolers has been reduced to 37.3 dBA, and I’ve also run a full load on MSI’s RTX 4070 Ti Super GPU at the same time to increase the difficulty of this benchmark.

At this moment, I only have a limited amount of data for this noise-equalized test with AMD’s Ryzen 9 9950X3D – so I’m hesitant to make definitive judgments based on these four results. With that disclaimer out of the way, this test has the heat of a GPU added to it, makes it more difficult than typical CPU cooler tests. And yet, Thermalright’s Royal Pretor 130 still managed to cool over 199W! What this means is that this air cooler will also be strong enough to handle the full heat of AMD’s CPU (at stock settings, this CPU does not consume more than 200W), even with low noise levels – if you’re running typical CPU workloads.

CPU-only thermal results with PBO enabled: AMD Ryzen 9 9950X3D

Without power limits enforced, the hottest CPUs on the market will hit their peak temperature (TJ Max) and thermally throttle with even the strongest of air coolers and even most liquid coolers on the market in intensive stress testing. For the best liquid coolers on the market, the results of this test will be shown using the CPU’s temperature. However, when the CPU reaches its peak temperature, I’ve measured the CPU package power to determine the maximum wattage cooled to best compare their performance. It’s important to note that thermal performance can scale differently depending on the CPU it’s being tested with.

We’ll start by looking at the performance of this cooler with AMD’s newly released Ryzen 9 9950X3D CPU, with the disclaimer that I’ve only had a limited amount of time to test coolers with this system.

The results on AMD’s Ryzen 9950X3D were very good. Thus far, I’ve only tested one cooler that can keep AMD’s flagship CPU under its peak temperature in a full-strength workload – Arctic’s Liquid Freezer III Pro!

Of these other coolers tested, Thermalright’s Royal Pretor 130 shows the best performance we’ve seen thus far, on par with the infamous Hyte THICC Q60 AIO!

CPU-only thermal results with power limits removed: Intel’s i7-14700K

Now let’s switch it up and take a look at how this cooler performs with Intel’s i7-14700K. The performance here is insane, matching many 360mm AIOs and almost able to handle the full heat of the CPU. Keep in mind that I test coolers at 23 degrees Celsius.

Maximum noise levels

A cooler’s maximum performance can’t be considered the only important metric, because some coolers run louder or louder than others. The maximum noise level this cooler operates at is 45.4 dBA. Most users will find this an acceptable volume, but some users might want to limit its speed if they are particularly sensitive.

200W Power Limit

My recent reviews have focused more on tests with both the CPU and GPU being stressed, but many of y’all have indicated that you would like to see more CPU-only tests. Listening to that feedback, I’ve tested thermal and noise performance for AMD’s Ryzen 9 9950X3D at stock CPU settings – which limits power consumption to 200W.

Thermalright’s Royal Pretor does very well at the CPU’s default power settings, outperforming a 240mm AIO from LevelPlay by two degrees. Noise levels are slightly reduced here compared to the previous tests, averaging 44.9 dBA.

150W CPU + 290W GPU results

Testing a CPU Cooler in isolation is great for synthetic benchmarks, but doesn’t tell the whole story of how it will perform. If your GPU is running a full load, that heat doesn’t just disappear – it makes it harder for your CPU cooler to do its job due to the increased heat within your computer case.

The CPU power limit of 150W was chosen based on the worst CPU power consumption I’ve seen reported in gaming with AMD’s Ryzen 9 9950X3D. My personal experience tells me most games to use much lower amounts of power, but this might be biased by the games I play.

The results here are slightly flipped compared to our previous chart, with LevelPlay’s 240mm AIO taking a 3-degree lead ahead of Thermalright’s flagship air cooler. This is simply a reflection of the fact that when GPU heat is added to a workload, AIO coolers have a thermal design advantage that air coolers can’t quite overcome.

That said, the noise levels of the Royal Pretor 130 are very reasonable in this scenario, at only 40.9 dBA!

140W CPU results

The last round of test results we’ll look at in this review is a test with a 140W power limit imposed. This level of power is much easier for most CPU coolers to handle. As a result, this is the easiest test we’ll run with AMD’s Ryzen 9 9950X3D CPU for most reviews. But I’ve recorded lower power results for comparisons with SFF and air coolers for future reviews.

Thermalright’s Royal Pretor 130 kept the CPU at an average of 60.9 C in this scenario, which only stresses the CPU. Based on the results collected thus far, I expect most high-end cooling products to maintain a temperature similar to this.

The noise level of the cooler is slightly reduced in this scenario, at 40.3 dBA. As a reminder, if you prefer the lowest noise levels, our noise-normalized results above show that this cooler performs excellently even when limited to 38.9 or 37.3 dBA.

Conclusion

Thermalright’s Royal Pretor 130 shattered my expectations, raising the bar for what I consider possible with air cooling both in terms of maximum theoretical performance and when its fans are set to run quietly! The price of this cooler is extremely reasonable at only $52.90 USD, so if you’re looking for the best air cooler on the market, the Royal Pretor 130 is what you should buy.

Before AIO liquid coolers, the air cooler ruled the roost, and Thermalright's Peerless Assassin is one of the best out there. With fantastic performance on modern CPUs, and at a price that AIO coolers just cannot compete with. This has been a top pick at Tom's Hardware for a long time, thanks to its great benchmarking performance and affordable price.

The cooler received our Editor's Choice award, and scored full marks when we reviewed the Thermalright Peerless Assassin 120 SE. It's more affordable than the competition without holding back on quality and build standards, the Peerless Assassin 120 SE outperformed peers while still offering astounding value for PC builders on a budget.

You can find this excellent deal over at CCL Computers, where the Thermalright Peerless Assassin 120 SE ARGB CPU cooler is just £17. The usual list price for this cooler is £30, which is still a good price, but being able to grab this cooler for just £17 makes it a no-brainer.

Using a solid copper base with six copper heat pipes running to dual aluminum fin stacks, the Peerless Assassin 120 SE dissipates heat away from your CPU and quietly draws cool air over the heatsinks with two 120mm TL-C12C PWM fans. Included brackets and mounting options give compatibility for Intel 115X, 1200, 1700, 17XX sockets, and AMD AM4 and AM5 platforms, if you kept hold of the original AMD motherboard backplate.

Everyone in the PC enthusiast community is likely familiar with cooling manufacturer Thermalright. The company is well known for providing quality competitive coolers at rock-bottom prices and raising the bar for value in the cooling realm. The company can do this because it typically directly manufactures its own products, having strong vertical integration as part of its business model.

Some of my past reviews of their products include the “Thermalright Phantom Spirit 120 Review: Simply the Best”, “Thermalright Phantom Spirit 120 EVO Review: This isn’t a competition. This is a massacre.”, and “Thermalright Grand Vision 360 Review: It’s not a competition, it is a massacre (again).” I only mention this to establish that I genuinely love Thermalright’s products – most of the time.

When I ask the typical “Will Thermalright’s Hypervision make our list of best coolers on the market question,” the answer is a clear “no” in this case, because this SKU is a pointless addition to Thermalright’s lineup which shouldn’t exist and confuses customers.

Let’s take a look at the specifications and features of the cooler, as well as the things I don’t like about this liquid cooling model, then we’ll go over thermal performance and noise levels.

Cooler specifications

Packing and included contents

The Hyper Vision 360 is the second product from Thermalright with packaging that I would describe as “fancy.” As you open the top, the white section with “low temperature, high performance” pops up – just like the previously reviewed Grand Vision 360. The product is secured with plastic wrappings and molded cardboard.

Included with the box are the following:

• Three 120mm fans
• 360mm radiator and CPU block
• Thermal paste
• LCD screen
• Mounting accessories for modern AMD & Intel platforms

Features of Thermalright’s Hyper Vision 360

▶ Thermalright TF7 Thermal Paste

Included with the AIO is a small tube of Thermalright’s TF7 thermal paste. This is one of the better pastes on the market, as you can see in our best thermal paste tests.

▶ 480x480 3.95-inch display

One thing that immediately sets the Hyper Vision 360 apart from its competitors, at least in this price range, is the inclusion of a 3.95-inch LCD screen. While this screen is larger than the one included with the Grand Vision 360, it is the same resolution -- so the quality of the image is effectively lower due to the decreased PPI of the screen being used.

▶ Consumer unfriendly: Accessible refill port - but using it breaks the warranty

I go out of my way to recognize when companies don’t include a consumer-unfriendly “warranty void if removed” stickers on top of the refill port – an action which is technically illegal under the Magnuson-Moss Warranty Act in the USA.

Thermalright includes an accessible refill port hidden behind a sticker with its logo on it, and while they don’t include a scary sticker on top of the refill port, their warranty policy is no different than those that do. If you attempt to service your AIO by refilling it, Thermalright will deny you warranty service.

What a shame!This is a consumer unfriendly move, especially when you consider that competitors such as BeQuiet not only allow this - but encourage it by including additional coolant with AIOs like the Light Loop 360.

▶ Standard 27mm radiator

The radiator included has a standard 27mm thickness, which means it should fit in the vast majority of ATX PC cases, without space constraints.

▶ Discernable pump whine

During testing, I noticed that the AIO's pump emitted a subtle whine. While not overly loud, it almost interfered with noise-normalized assessments which I run at 38.9 dBA – and many users will prefer their pumps to run at 36.4 dBA or quieter.

This type of issue with a liquid pump is more of a hassle to deal with compared to potential problems with an air cooler – it's easier to replace a fan than it is to go through the process of RMAing an AIO for having a noisy pump and being potentially told it is operating within expected performance.

This is a factor worth considering when investing in liquid cooling if you're not fully confident in the ease of a company's warranty process.

▶ Mixed Bag: Software with tons of customization options, but stability may vary

You don’t need to install Thermalright’s software to operate the AIO normally, but you’ll need to download it if you wish to customize the display – and if you paid extra for a screen, why wouldn’t you want to customize it? Thermalright doesn’t make it obvious where to download the software, but this link will take you to the company’s TRCC software.

At first glance, you’ll see the screen above and might think that Thermalright’s software is rather basic. The software is deceptively simple looking, but exploring the options further reveals a wide variety of preset customization options – more than I’ve seen from any other AIO software I’ve used before.

That said, your mileage may vary when dealing with this software. While my own personal experience with Thermalright’s software has been flawless, I’ve received reports from users having problems with the software on social media and the Tom’s Hardware Discord server. Some of these users I’ve been able to help resolve their problems, but others I’ve been unsuccessful in helping to troubleshoot.

▶ Cable management features

Thermalright’s Hyper Vision AIO features pre-installed fans with a quick-connect system; the cables are routed through the tubing of the AIO with clips. This was a good effort. But it ends up a tad bit messier than needed. This is a minor complaint overall though. The bigger problem with this SKU is its inferior performance compared to similarly named Thermalright products.

▶ 120mm TL-K12W fans

There’s more to a cooler than just the heatsink or radiator. The bundled fans have a significant impact on cooling and noise levels, as well as how the cooler looks in your case.

These fans are 25mm thick and have a quick connect system designed to simplify installation and (in theory) reduce cable clutter, and have subtle ARGB accents.

Testing configuration – Intel LGA1700 and LGA1851 platform

There are many factors other than the CPU cooler that can influence your cooling performance, including the case you use and the fans installed in it. A system's motherboard can also influence this, especially if it suffers from bending, which results in poor cooler contact with the CPU.

In order to prevent bending from impacting our cooling results, we’ve installed Thermalright’s LGA 1700 contact frame into our testing rig. If your motherboard is affected by bending, your thermal results will be worse than those shown below. Not all motherboards are affected equally by this issue. I tested Raptor Lake CPUs in two motherboards. And while one showed significant thermal improvements after installing Thermalright’s LGA1700 contact frame, the other motherboard showed no difference in temperatures whatsoever. Check out our review of the contact frame for more information.

I’ve also tested this cooler with Intel’s latest platform, Arrow Lake and LGA 1851.

LGA 1700 and 1851 Installation

The installation of this AIO is simple. The following steps assume that you will mount the radiator to your case first, which is generally a good idea unless you are building in a very cramped case.

1. You’ll first need to place the backplate against the rear of the motherboard.

2. Next, you’ll secure the backplate by attaching standoffs. You’ll then place the mounting bars on top of the standoffs, and secure them with the included screws.

3. Apply the included thermal paste to your CPU. If you have any questions on how to do this properly, please refer to our handy guide on how to apply thermal paste.

4. Place the CPU block on top of the CPU, and secure it with a screwdriver. Attach the LCD screen after securing the CPU block.

5. Connect the PWM and ARGB cables to your motherboard. If you wish to use the display, you’ll need to connect the USB cable to the CPU block on one end, and to a USB and SATA power header on the other.

6. Next you can power on your computer, as installation is complete.

CPU-only thermal results without power limits

Without power limits enforced on Intel’s Core Ultra 9 285K and i7-14700K CPUs, the CPU will hit its peak temperature (TJ Max) and thermally throttle with even the strongest of air coolers and even most liquid coolers on the market. For the best liquid coolers on the market, the results of this test will be shown using the CPU’s temperature. However, when the CPU reaches its peak temperature, I’ve measured the CPU package power to determine the maximum wattage cooled to best compare their performance. It’s important to note that thermal performance can scale differently depending on the CPU it’s being tested with.

Looking at the results here with Intel’s “Raptor Lake” Core i7-14700K, you’ll notice the chart is measuring the CPU package power instead of the average temperature. This is because while Thermalright’s Hyper Vision 360 is a strong AIO, it isn’t strong enough to keep the CPU from reaching its peak temperature in stress testing.

It is important to remember that CPU coolers can scale (and perform) differently depending on the CPU it is paired with due to differences in manufacturing processes and the location of hotspots on the CPU – and in this case, the Hyper Vision 360 performs even worse with Arrow Lake. In fact, in my stress testing, FSP’s MP7 air cooler actually performed better than Thermalright’s AIO on Intel’s Core Ultra 9 285K CPU.

Worse yet, this cooler runs LOUD by default, reaching a whopping 53 dBA! I don’t understand why Thermalright let these fans run so loudly, especially since – as the next section will show – they perform well even when normalized to low noise levels!

CPU-only thermal results with noise normalized to 38.9 dBA

Finding the right balance between fan noise levels and cooling performance is important. While running fans at full speed can improve cooling capacity to some extent, the benefits are limited and many users prefer a quieter system.

While this cooler’s default behavior is loud with underwhelming maximum performance, when noise-normalized, it is very strong, losing only 1W of performance – a figure so small as to only be relevant for benchmarking.

253W results

My recent reviews have focused more on tests with both the CPU and GPU being stressed, but many of y’all have indicated that you would like to see more CPU-only tests. In response, I’ve started testing Intel’s “Arrow Lake” Core Ultra 9 285K with a 253W limit.

The results of Thermalright’s Hyper Vision when a more reasonable 253W limit is imposed on Intel’s Core Ultra 9 285k aren’t as good as I had expected. At 80 degrees C, it’s 3C ahead of the worst result I have from an AIO in this benchmark and 4C behind the best result I’ve recorded. My results are somewhat limited here, but I expect this to be more or less “average” for AIOs I test in the future.

135W CPU + 290W GPU results

Testing a CPU Cooler in isolation is great for synthetic benchmarks, but doesn’t tell the whole story of how it will perform. I’ve incorporated two tests with a power limit imposed on the CPU, while also running a full load on MSI’s GeForce RTX 4070 Ti SUPER 16G VENTUS 3X.

The CPU power limit of 135W was chosen based on the worst CPU power consumption I observed in gaming with Intel’s Core Ultra 9 285K, which was in Rise of the Tomb Raider.

In this test, the Hyper Vision 360 actually performs very well – at least in terms of overall temperature. It averaged 61C, only 2C behind the best result we’ve recorded. In terms of noise levels (when tied to the default fan curve of my motherboard) it wasn’t too bad either, measuring at 41.4 dBA – though this is technically the third-loudest result in this chart.

85W CPU + 290W GPU results

Our second round of CPU + GPU testing is also performed with Arrow Lake. The power limit of 85W was chosen based on typical power consumption in gaming scenarios using the Core Ultra 9 285K CPU. This should be fairly easy for most coolers, the main point of this test is to see how quietly (or loudly!) a cooler runs in low-intensity scenarios.

With a CPU temperature of 53 degrees C, the thermal performance was alright. What’s frustrating about the results in this section is the noise levels. Looking at the chart, you might think: But Albert, that reading of 38.9 is even better than the Grand Vision 360!

And that technically would be true. But what this chart is incapable of expressing is that in the case of the Hyper Vision 360, the noise level is caused by the pump whine of the unit I tested. If the pump wasn’t noisy, the rating here would be even quieter – and that annoys me because this cooler would be a better value if the pump was properly tuned! Pump whine has a more irritating pitch and frequency to its noise than any PC fan I’ve heard.

Conclusion

Unless you really need a slightly larger LCD screen, I advise passing on Thermalright’s Hyper Vision 360. Instead, most users should consider purchasing Thermalright’s superior Grand Vision 360 instead. Both AIOs cost the same, but the Hyper Vision 360 has worse overall performance and minor but noticeable pump whine. Most of the time, thanks to test results, I gladly recommend Thermalright products. But in this case, the Hyper Vision 360 is an example of unnecessary SKU spam that only serves to confuse customers. It’s not worth your time or hard-earned money.

The last time I used the title of it’s not a competition, it is a massacre was when Thermalright released its Phantom Spirit 120 EVO air cooler. To this day I consider it one of the best options in air cooling. And after testing it, I’m equally impressed with Thermalright’s latest AIO, the Grand Vision 360.

This cooler not only has some of the best thermal performance I’ve seen from any AIO I’ve tested, but it also has a fancy LCD display with tons of preset customizations available, while retaining a reasonable price tag of $129. You can bet this one will soon be added to our list of the best coolers we’ve tested. But before it lands there, let’s take a look at the cooler’s features and specs, followed by its impressive results in our testing.

Cooler specifications

Packing and included contents

The Grand Vision 360 is the first product from Thermalright with packaging that I would describe as “fancy.” As you open the top, the white section with “low temperature, high performance” pops up. The product is secured with plastic wrappings and molded cardboard. The fans are preinstalled for convenience.

Included with the box are the following:

• Three 120mm fans
• 360mm radiator and CPU block
• Thermal Paste
• LCD Screen
• Mounting accessories for modern AMD & Intel platforms

Features of Thermalright’s Grand Vision 360

*️⃣Thermalright TF7 Thermal Paste

Included with the AIO is a small tube of Thermalright’s TF7 thermal paste. This is one of the better pastes on the market, as you can see in our best thermal paste tests. 

*️⃣ Accessible refill port

I go out of my way to recognize companies when that don’t include a consumer-unfriendly “warranty void if removed” sticker on top of the refill port – an action which is technically illegal under the Magnuson-Moss Warranty Act in the USA. Thermalright includes an accessible refill port hidden behind a sticker with its logo on it, so that users who wish to service their equipment can do so. Most users won’t have to do this during the three-year warranty rating, but it’s nice to have the option should the need arise. 

*️⃣ 27mm radiator 

The Grand Vision 360 includes a radiator 27mm thick, which is standard for most liquid coolers.

 *️⃣ Fancy 3.4-inch display 

One thing that immediately sets the Grand Vision 360 apart from its competitors, at least in this price range, is the inclusion of a 3.4-inch, 480x480-resolution screen. 

*️⃣Software with tons of customization options 

You don’t need to install Thermalright’s software to operate the AIO normally, but you’ll need to download it if you wish to customize the display. Thermalright doesn’t make it obvious where to download the software, but this link will take you to the company’s TRCC software.

At first glance, you’ll see the screen shown above and might think that Thermalright’s software is rather basic. The software is deceptively simple looking, but exploring the options further reveals a wide variety of preset customization options – more than I’ve seen from any other AIO software I’ve used before!

*️⃣ Cable management features 

Thermalright’s Grand Vision AIO features pre-installed fans with a quick-connect system; the cables are routed through the tubing of the AIO with clips. This was a good effort. But it ends up a tad bit messier than needed. This is a minor complaint in an otherwise stellar product. 

 *️⃣ TL—H12-X28-S 120mm PWM fans 

There’s more to a cooler than just the heatsink or radiator. The bundled fans have a significant impact on cooling and noise levels, as well as how the cooler looks in your case. The fans included are TL—H12-X28-S (that’s a mouthful!) 120mm PWM fans, which, as the benchmarks will show, enable strong cooling performance both at full speed and when restricted to low noise levels.

These fans are pre-installed and feature a quick-connect system, designed to save the user time and offer tidy cable management. They’re also slightly thicker than normal: 28mm instead of the normal 25mm fans.

*️⃣ Full RAM Compatibility 

Like almost every other AIO on the market, the Grand Vision 360 doesn’t interfere with or overhang RAM DIMMs in any manner – allowing for all sizes of RAM, no matter how tall.

Testing configuration – Intel LGA1700 and LGA1851 platform

There are many factors other than the CPU cooler that can influence your cooling performance, including the case you use and the fans installed in it. A system's motherboard can also influence this, especially if it suffers from bending, which results in poor cooler contact with the CPU. 

To prevent bending from impacting our cooling results, we’ve installed Thermalright’s LGA 1700 contact frame into our testing rig. If your motherboard is affected by bending, your thermal results will be worse than those shown below. Not all motherboards are affected equally by this issue. I tested Raptor Lake CPUs in two motherboards. And while one of them showed significant thermal improvements after installing Thermalright’s LGA1700 contact frame, the other motherboard showed no difference in temperatures whatsoever! Check out our review of the contact frame for more information.

I’ve also tested this cooler with Intel’s latest platform, Arrow Lake and LGA 1851.

LGA 1700 and 1851 Installation

The installation of the AIO is simple. The following steps assume that you will mount the radiator to your case first, which is generally a good idea unless you are building in a very cramped case.

1. You’ll first need to place the backplate against the rear of the motherboard.

2. Next, you’ll secure the backplate by attaching standoffs. You’ll then place the mounting bars on top of the standoffs, and secure them with the included screws.

3. Apply the included thermal paste to your CPU. If you have any questions on how to do this properly, please refer to our handy guide on how to apply thermal paste.

4. Place the CPU block on top of the CPU, and secure it with a screwdriver.

5. Connect the PWM and ARGB cables to your motherboard. If you wish to use the display, you’ll need to connect the USB cable to the CPU block on one end, and to a USB and SATA power header on the other ends.

6. Afterwards, power on your computer as installation is complete.

CPU-only thermal results without power limits

Without power limits enforced on Intel’s Core Ultra 9 285K and i7-14700K CPUs, the CPU will hit its peak temperature (TJ Max) and thermally throttle with even the strongest of air coolers and even with most liquid coolers on the market. When the CPU reaches its peak temperature, I’ve measured the CPU package power to determine the maximum wattage cooled to best compare performance between coolers. It’s important to note that thermal performance can scale differently depending on the CPU it’s being tested with.

The results we observed with Thermalright’s Grand Vision 360 were extremely impressive, outperforming our expectations.

With an average temperature of 89 degrees Celsius, Thermalright’s Grand Vision 360 outperformed its closest competitor by just a hair. The strongest result I’ve recorded yet, this alone would earn this new AIO a spot on our best AIO coolers list.

But remember: how well a cooler performs can vary depending on what CPU it is paired with. So how’s performance with a CPU that is harder to cool like Intel’s i7-14700K? The same as above, holding the crown for top performance. Note that the Be Quiet cooler in second place costs about $20 more and doesn’t have an LCD screen.

To achieve this level of thermal dominance, the fans of the Grand Vision can get a bit noisy, at 48.9 dBA in our testing. This isn’t the loudest AIO I’ve tested, but it is one of the louder ones. I’d advise limiting the maximum speed of the fans if that might bother you. As we’ll see in the next test, this cooler doesn’t need to get noisy to perform well.

CPU-only thermal results with noise normalized to 38.9 dBA

Finding the right balance between fan noise levels and cooling performance is important. While running fans at full speed can improve cooling capacity to some extent, the benefits are limited and many users prefer a quieter system.

With this noise-normalized test, I’ve set noise levels to 38.9 dba using the i7-14700K system. This level of noise is a low volume level, but slightly audible to most people. The Grand Vision 360 cooled an average of 258W in this test, taking the crown away from NZXT’s Kraken Elite 360 for the strongest AIO liquid cooler on the market!

253W results

My recent reviews have focused more on tests with both the CPU and GPU being stressed, but many of y’all have indicated that you would like to see more CPU-only tests – so I’ve started testing Intel’s Arrow Lake Core Ultra 9 285K with a 253W limit. My current results for this test are limited, but this will grow as I have time to test more coolers.

Testing with the default power limits of 253W shows the Grand Vision 360 at its best, outperforming its closest competitor by 3 degrees C!

165W CPU + 275W GPU results

Last year, I updated my testing methodology to better represent the scenarios a user might actually encounter in real-life use. The first of these new tests is designed to emulate, though not perfectly, a CPU in an intensive game with a full GPU load using ASRock’s Steel Legend 7900 GRE GPU and Intel’s Core i7-14700K CPU.

Most of the time, Intel’s i7-14700K will use less than 165W in games. This value was chosen because it was the peak power consumption I observed in Far Cry V. I would consider the results of this test closer to what you might see with an overclocked setup.

In this scenario, the Grand Vision 360 takes the performance crown again – performing 2 degrees C cooler than its closest competitor, the Tryx Panorama 360. Noise levels are moderate here, at 42.9 dBA.

135W CPU + 290W GPU results

Our second round of CPU + GPU testing is performed with Arrow Lake. The power limit of 135W was chosen based on the worst CPU power consumption I observed in gaming, which was in Rise of the Tomb Raider.

This scenario isn’t quite as thermally demanding as our previous results, and the AIO retains a similar lead over its competitors, this time outperforming the ID-Cooling FX360 INF by 2 degrees C! Noise levels registered at 40.3 dBA, a moderately quiet noise level. Noise levels could be better here, but I’m not complaining.

85W CPU + 290W GPU results

Our third round of CPU + GPU testing is also performed with Arrow Lake. The power limit of 85W was chosen based on typical power consumption in gaming scenarios using the Core Ultra 9 285K CPU. This should be fairly easy for most coolers. The main point of this test is to see how quietly (or loudly!) a cooler runs in low-intensity scenarios.

In terms of thermals, the Grand Vision 360 again performed excellently at only 50 degrees C. As with previous results, this is the strongest performance we’ve seen yet. But honestly, thermals in this type of scenario really don’t matter – especially when the most basic of air coolers is managing to keep the CPU under 60C. Noise matters much more here.

When it comes to noise levels, Thermalright’s Grand Vision technically runs the loudest of those shown here, at 39.3 dBA. But this is still a relatively low noise level, and I’m not complaining given the superb thermal performance and value this cooler brings.

Conclusion

Thermalright proves once again that it is a market leader, raising the bar for performance and value with the Grand Vision 360. Its thermal results are better than any previous AIOs I’ve tested. It includes a fancy LCD screen with tons of preset customization options, and its price tag is reasonable at $129 USD. If you’re looking for a strong LCD AIO that doesn’t break the bank, Thermalright’s Grand Vision 360 is the liquid cooler you’re looking for.

Thermalright is well-known by PC enthusiasts in recent years, for pushing the bar of both performance and value with its low-cost air coolers. I consider their Phantom Spirit 120 to have the best overall value on the market when you consider performance, cost, and noise levels. 

With this review we’re looking at one of Thermalright’s cheapest desktop coolers, the Assassin X 120 R SE. This is a basic, entry-level cooler with a single-tower radiator, four copper heatpipes, and a single 120mm fan. 

This cooler won’t make our list of best CPU coolers, but it is definitely a great value in entry-level cooling.  Let’s take a quick look at the unit’s specifications, then we’ll dive into its features and the testing.

Cooler specifications

Packing and included contents

The packaging for this cooler is fairly basic, as to be expected for a low-cost product.

Opening the box reveals the cooler and fan, protected by a combination of molded foam and plastic wrap.

Included with the cooler are the following:

• 1x 120mm fan
• Tube of Thermalright TF4 thermal paste
• Single-tower heatsink with four direct-touch heatpipes
• Mounting for modern Intel and AMD platforms
• Installation manual

Features of the Assassin X 120 R SE

▶ Four 6mm direct-touch copper heatpipes

Direct-touch heatpipes are incorporated into this design to most-effectively transfer heat away from the CPU.

▶ Low noise levels

This cooler runs quietly in all workloads, with a maximum noise level of 41.9 dBA. 

▶ Thermalright TF4 thermal grease

Thermalright includes a tube of their TF4 thermal grease for use with the Assassin X 120 R SE. This is an entry-level paste from Thermalright, but it still performs rather well – only 3 degrees Celsius behind the best performing pastes we’ve benchmarked in our thermal paste tests.

▶ Low price

You won’t find many coolers with a cheaper price than the Assassin X 120 R SE. It has an official MSRP of $19.90 USD, but can often be found for as cheap as $15 at Amazon and other retailers.

▶ TL-C12C 120mm fan

There’s more to a cooler than just the heatsink or radiator. The bundled fans have a significant impact on cooling and noise levels, as well as how the cooler looks in your case. Thermalright includes one of their TL-C12C 120mm fans with the Assassin X 120 R SE. It is a very basic fan, but it provides essential cooling with low noise levels.

Testing Methodology, and how my testing differs vs the competition

I do a few things to ensure my results are useful. To begin, I strictly regulate the ambient temperature to 23 degrees C during testing. I do not accept results at 22 or 24C as valid, even though in theory there won’t be much of a difference. 

Some reviewers test CPU Coolers on an open test bench, the idea being to remove any variables other than the cooler. The problem with this type of testing is that with typical usage – you know, with an enclosed case – the ambient temperature within will be higher than that of outside air, which adds difficulty to the cooler’s job.

Some test CPU coolers using a thermal plate instead of using a CPU. This suffers from all the drawbacks of open bench testing, but also doesn’t effectively emulate the heat from a CPU. A thermal plate generally evenly distributes heat, with a low overall thermal density. CPUs, on the other hand, have most of their heat generated in concentrated hotspots, making them much more difficult to cool.

Older CPUs are also poor candidates for cooler testing for similar reasons – they were produced with larger manufacturing processes and may have lower thermal density in their hotspots as a result.

The problem with CPU-only testing

Testing with CPU only loads provides useful information, but it doesn’t paint a complete picture of how a cooler will perform. Many workloads, especially gaming, will stress both the CPU and GPU. 

To test how well a CPU’s cooler will perform in these situations, I’ve added two tests: one with a light CPU and one with a moderate CPU load – both while running a full load on the GPU.

Testing configuration – Intel LGA1700 platform

There are many factors other than the CPU cooler that can influence your cooling performance, including the case you use and the fans installed in it. A system's motherboard can also influence this, especially if it suffers from bending, which results in poor cooler contact with the CPU. 

In order to prevent bending from impacting our cooling results, we’ve installed Thermalright’s LGA 1700 contact frame into our testing rig. If your motherboard is affected by bending, your thermal results will be worse than those shown below. Not all motherboards are affected equally by this issue. I tested Raptor Lake CPUs in two motherboards. And while one of them showed significant thermal improvements after installing Thermalright’s LGA1700 contact frame, the other motherboard showed no difference in temperatures whatsoever! Check out our review of the contact frame for more information.

LGA 1700 Installation

1. Place the backplate against the rear of the motherboard, underneath the CPU.

2. Place the blue rubber standoffs on top of the CPU cooler mounting holes.

3. Place the mounting bars on top of the blue standoffs and secure them with the included screws.

4. Apply the included Thermalright TF4 thermal paste to the top of the CPU. If you’re unsure of how to do this, check out our guide on how to apply thermal paste.

5. Place the heatsink on top of the CPU and secure it with a screwdriver.

6. Attach the TL-C12C fan to the heatsink using the included fan clips. Then connect its PWM cord to your motherboard and installation is complete.

CPU Thermal results without power limits

Without power limits enforced on Intel’s i7-14700K, the CPU will hit its peak temperature (TJ Max) and thermally throttle with even the strongest of air coolers. When the CPU reaches its peak temperature, I’ve measured the CPU package power to determine the maximum wattage cooled to best compare their performance.

I wasn’t expecting a lot out of this cooler, so it is no surprise that it is the worst performing result on this list when you consider that the comparison coolers I’ve also tested are higher-end, more-expensive products.

What is surprising is how closely it performed to Noctua’s original NH-D15, trailing by only 5W! This is very impressive for a single-tower, single-fan cooler.

While the Assassin X didn’t top the charts in terms of thermal performance, it runs nice and quietly. With a maximum volume of 41.9 dBA, the Assassin X 120 R SE runs quieter than all of the nine comparison coolers I’ve tested for this review.

Yes, the chart above starts at 35dBA rather than zero. That’s for a couple key reasons: First, noise level measurements are logarithmic. A person with typical hearing will perceive a doubling of noise for every 10 dBA increase, but the impact can vary from person to person. Subjectively, I feel that a 5dBA increase feels like twice as loud – but I suffer from conditions which can make my ears more sensitive than others.

Secondly, 35dBA was chosen as our “zero” point because it is the lowest level my noise meter can accurately measure.

Thermal results with noise normalized to 38.9 dBA

Finding the right balance between fan noise levels and cooling performance is important. While running fans at full speed can improve cooling capacity to some extent, the benefits are limited and many users prefer a quieter system. With this noise-normalized test, I’ve set noise levels to 38.9 dba. This level of noise is a low volume level, but slightly audible to most people. 

Thermalright’s Assassin X 120 R SE cooled an average of 213W in this scenario, only 5W behind Noctua’s first-generation NH-D15!

165W CPU + 275W GPU results

I’ve recently updated my testing to better represent the scenarios a user might actually encounter in real life usage. The first of these new tests is designed to emulate, though not perfectly, a CPU intensive game with a full GPU load.

Thermalright’s Assassin X 120 R SE gets a bit toasty in this workload, averaging 83C during the course of testing. This is fine for CPUs at their stock speeds, but you wouldn’t want to try overclocking. 

Because of the higher CPU temperature, the Assassin X 120 R SE reaches its maximum noise level of 41.9 dBA – but this is actually quieter than many of the competitors I’ve tested it against, with only Arctic’s Liquid Freezer III AIO running quieter.

110W CPU + 275W GPU results

Our second round of CPU + GPU testing is performed with a lower CPU power limit, and in theory should be similar to games that are moderately intensive for a CPU. The Thermalright Assassin X 120 R SE reached 70C in this scenario, 4C behind Noctua’s NH-D15.

120W Cinebench results

With my previous testing methodology, I used to record the temperatures and noise measurements using a 125W CPU-only test and I would emphasize how temperatures do not matter in this scenario. 

As such, for this section I’m only going to report noise levels – because that’s the only information that really matters for this test. The temperature of the CPU in this scenario is largely irrelevant, but the results for liquid coolers is similar to those reported above in the 110W + GPU tests if you are particularly concerned about this information.

The cooler reached only 38.9 dBA here, a low noise level that shouldn’t bother most folks. Only Arctic’s Liquid Freezer III ran quieter than Thermalright’s Assassin X 120 R SE here.

Conclusion

For users on a strict budget, Thermalright’s Assassin X 120 R SE is probably the best option on the market at the moment. This isn’t a product you’d want to use while overclocking, but for less than $20 USD, it delivers essential cooling performance sufficient for most users, while also keeping noise levels low. 

Thermalright is well known by PC enthusiasts for pushing the bar in terms of both performance and value with its low-cost air coolers. I consider their Phantom Spirit 120 to be one of the best air coolers on the market when you consider performance, cost, and noise levels. 

With this review we’re looking at Thermalright’s latest entry in the fan-favorite Peerless Assassin series of air coolers. There are a few different models of the Peerless Assassin currently on the market, paired with either 120mm or 90mm fans depending on the version in question. As the name of this model implies, what sets the Peerless Assassin 140 apart from the other variants on the market is the use of a larger heatsink and the use of a 140mm middle fan, instead of the 120mm fans used on other units.

Does the PA140 have what it takes to make our list of best coolers? Let’s take a quick look at the unit’s specifications, then we’ll dive into its features and the benchmarks.  

Cooler specifications

Packing and included contents 

The Peerless Assassin 140 arrives in a black and white box, with molded foam and plastic coverings to protect the heatsink and fans. The accessories are contained in a small cardboard box, placed in between the two heatsink towers.

Included with the cooler are the following:

• 120mm fan and 140mm fan
• Tube of Thermalright TF7 thermal paste
• Dual tower heatsink with six heat pipes
• Mounting for modern Intel and AMD platforms
• Installation Manual
• PWM splitter cable

Features of the Assassin X 120 R SE

▶ Six 6mm copper heatpipes

To move heat away from the CPU into the heatsink, the Peerless Assassin utilizes six copper heatpipes.

▶ Available in black or white

The model we’re reviewing today is black, but there is also a white version available for those who desire a different aesthetic. 

▶ Thermalright TF7 thermal grease 

Thermalright includes a tube of its TF7 thermal paste. This isn’t the best thermal paste available, but it works rather well – in our thermal paste testing, it’s less than 2 degrees Celsius behind the best-performing pastes we’ve benchmarked.

▶ Low price

Thermalright is known for having some of the best prices on the market for cooling equipment. The Peerless Assassin continues this tradition, with an MSRP of $39.90. If you’re really on a tight budget, you might even be able to find it for a lower price in the future – sales of Thermalright products on Amazon are quite frequent.

▶ RAM Clearance

The Peerless Assassin’s 120mm side fan rests above most RAM DIMMs. But if you have particularly tall DIMMs, you may need to raise the front fan a few millimeters.

▶ Heatsink photo

Some of y’all like to see the design of the heatsink and its fins, so I’ve included a photo below.

▶ Hybrid 120mm and 140mm fans 

There’s more to a cooler than just the heatsink or radiator. The bundled fans have a significant impact on cooling and noise levels, as well as how the cooler looks in your case. The Peerless Assassin 140 uses a hybrid fan setup, with a 120mm TL-C12B V2 fan on the side of the unit and a 140mm TL-D14CB in the middle.

Testing Methodology, and how my testing differs from the competition 

If you’ve been reading my reviews for a while, you’ll know that I do a few things to ensure my results are useful. To begin, I strictly regulate the ambient temperature to 23 degrees C during testing. I do not accept results at 22 or 24C as valid, even though in theory there won’t be much of a difference. 

Some reviewers test CPU Coolers on an open test bench, the idea being to remove any variables other than the cooler. The problem with this type of testing is that with typical usage – you know, with an enclosed case – the ambient temperature within will be higher than that of outside air, which adds difficulty to the cooler’s job. 

Some test CPU coolers using a thermal plate instead of using a CPU. This suffers from all the drawbacks of open bench testing, but also doesn’t effectively emulate the heat from a CPU. A thermal plate generally evenly distributes heat, with a low overall thermal density. 

CPUs, on the other hand, have most of their heat generated in concentrated hotspots. This is much more difficult to cool. The best way to explain the difference is this: A thermal plate is like a cigarette lighter, you can take that flame and put it against your hand for a quick moment and while uncomfortable – you won’t be hurt seriously. On the other hand, a CPU is like a concentrated butane torch lighter that produces a blue flame – if that touches your skin at all, you’re in for a painful time. Older CPUs are also poor candidates for cooler testing for the same reason – they were produced with larger manufacturing processes and may have lower thermal density in their hotspots as a result.

The problem with CPU-only testing

Testing with CPU-only loads provides good and useful information, but it doesn’t paint a complete picture of how a cooler will perform. Many workloads, especially gaming, will stress both the CPU and GPU. The heat added from the GPU increases the difficulty of cooling the CPU in these scenarios. 

To test how well a CPU’s cooler will perform in these situations, I’ve added two tests – one with a light CPU and one with a moderate CPU load, both while running a full load on the GPU. 

Testing configuration – Intel LGA1700 platform

There are many factors other than the CPU cooler that can influence your cooling performance, including the case you use and the fans installed in it. A system's motherboard can also influence this, especially if it suffers from bending, which results in poor cooler contact with the CPU. 

To prevent bending from impacting our cooling results, we’ve installed Thermalright’s LGA 1700 contact frame into our testing rig. If your motherboard is affected by bending, your thermal results will be worse than those shown below. Not all motherboards are affected equally by this issue. I tested Raptor Lake CPUs in two motherboards. And while one of them showed significant thermal improvements after installing Thermalright’s LGA1700 contact frame, the other motherboard showed no difference in temperatures whatsoever! Check out our review of the contact frame for more information. 

LGA 1700 Installation

1. Place the backplate against the rear of the motherboard, underneath the socket.

2. Place the blue rubber standoffs on top of the CPU cooler mounting holes.

3. Place the mounting bars on top of the blue standoffs, and secure them with the included screws.

4. Apply the included Thermalright TF7 thermal paste to the top of the CPU. If you’re unsure of how to do this, check out our guide on the best practices for applying thermal paste. 

5. Place the heatsink on top of the CPU, and secure it with a screwdriver.

6. Attach the fans to the heatsink using the included fan clips, and then connect the fans to the PWM splitter cable.

7. Connect the PWM splitter cable to your motherboard, and installation is complete.

CPU Thermal results without power limits

Without power limits enforced on Intel’s i7-14700K, the CPU will hit its peak temperature (TJ Max) and thermally throttle with even the strongest of air coolers. When the CPU reaches its peak temperature, I’ve measured the CPU package power to determine the maximum wattage cooled to best compare their performance.

The Peerless Assassin has very good performance for an air cooler, trailing behind a top-end 240mm AIO by only 4%. Compared to air coolers, the PA 140 offered slightly better performance than BeQuiet’s Dark Rock Elite.

Maximum Noise Levels

Of course, maximum thermal performance benchmarks aren’t complete without taking into effect noise levels. The Peerless Assassin also ran slightly quieter than BeQuiet’s Dark Rock Elite, with a maximum noise volume of 43.8 dBA compared to the Dark Rock Elite’s 44.6 dBA.

Observant readers may notice that the chart above starts at 35dBA, and might be worried that this starting point may cause graphs to be deceptive. There are a few reasons that this is not something to worry about. First, noise level measurements are logarithmic. A person with typical hearing will perceive a doubling of noise for every 10 dBA increase, but the impact can vary from person to person.

Subjectively, I feel that a 5dBA increase feels like twice as loud – but I should add the disclaimer that I suffer from conditions which can make my ears more sensitive than others. 35dBA was chosen as the “zero” because it is the lowest level my noise meter can accurately measure.

Thermal results with noise normalized to 38.9 dBA

Finding the right balance between fan noise levels and cooling performance is important. While running fans at full speed can improve cooling capacity to some extent, the benefits are limited and many users prefer a quieter system. With this noise-normalized test, I’ve set noise levels to 38.9 dba. This level of noise is low, but slightly audible to most people.

When noise normalized, the Peerless Assassin 140 punches above its weight – on par with high-quality 240mm AIOs!

165W CPU + 275W GPU Thermal and Noise results

I’ve recently updated my testing methodology to better represent the scenarios a user might actually encounter in real-life use. The first of these new tests is designed to emulate, though not perfectly, a CPU-intensive game with a full GPU load.

Because of the higher CPU temperature, the Peerless Assassin 140 runs almost at its maximum noise level, dropping 0.4 dBA. However, keep in mind this is for a shorter testing duration. You will likely reach the full noise levels of 43.8 in extended gaming sessions if they are CPU intensive.

110W CPU + 275W GPU Thermal results

Our second round of CPU + GPU testing is performed with a lower CPU power limit, and in theory, should be similar to games that are moderately intensive for a CPU.

The Peerless Assassin 140 reached 64 C in this scenario, 2 degrees behind BeQuiet’s Dark Rock Elite, but also 2 degrees ahead of Noctua’s NH-D15. In this scenario, I feel like noise levels are more important, which brings us to the next section.

120W Noise results

With my previous testing methodology, I used to record the temperatures and noise measurements using a 125W CPU-only test and I would emphasize how temperatures do not matter in this scenario.

As such, for this section, I’m only going to report noise levels – because that’s the only information that really matters for this test. The temperature of the CPU in this scenario is largely irrelevant, but the results for liquid coolers are similar to those reported above in the 110W + GPU tests if you are particularly concerned about this information.

The cooler reached 40.3 dBA, a moderately low noise level that shouldn’t bother most folks, but it is an audible noise level. This is on par with Noctua’s NH-D15 G1, but a bit louder than the results from BeQuiet and Scythe. However, if you’re particular about low noise levels our previous noise-normalized results show that this cooler performs well even when set to run quietly!

Conclusion

Thermalright’s Peerless Assassin 140 is a high-performance air cooler with good thermal performance and moderate noise levels. With a price of only $39.90 in the U.S., it is one of the best values on the market and will certainly earn a spot on our list of best CPU coolers.

Thermalright is arguably the most competitive cooling brand on the market right now, their products having better value and lower price tags than any of their competition. When it comes to air coolers, it is hard to recommend any other competitor – their Phantom Spirit 120 Evo  offers the best combination of performance, noise levels, and price of any air cooler currently on the market.   

With Thermalright’s air coolers performing so competitively, we just had to test out some of their liquid coolers. Today’s review will cover two of their latest 360mm AIOs, the Frozen Notte 360 and Aqua Elite 360 White V3. With price tags of less than $65 USD, these AIOs are available for half the price of many competitors. Thermalright’s air coolers have earned spots on our best coolers list, but do their liquid coolers also have what it takes? We’ll have to put it through testing to find out. But first, here are the specifications from Thermalright.

Cooler specifications

Packing and included contents

The packaging of each AIO is similar to most other units on the market, with molded cardboard, foam, and plastic coverings for the protection of the contents.

 Included with the cooler are the following:

• 360mm radiator and CPU block
• Three 120mm fans, preinstalled
• Mounting for modern AMD and Intel Platforms
• Thermal paste

LGA 1700 installation

The installation of the coolers is simple. The fans of the unit arrive pre-installed, which saves time during installation.

1. Press the backplate against the rear of the motherboard, and then secure it using the included standoffs. There is one difference between the installation of these units at this stage. The Frozen Notte has a pre-assembled backplate, the Aqua Elite’s must be assembled by the user.

2. Place the mounting bars on top of the standoffs, and secure them with the included thumbscrews. You won’t need to follow this step on the Frozen Notte 360, as its design doesn’t require mounting bars.

3. Next you’ll need to apply thermal paste to the CPU – and if you're unsure how to do that, see our How to Apply Thermal Paste primer.

4. Place the CPU block on top of your CPU, and secure it with a screwdriver.

5. Finally, connect the ARGB and PWM headers to your motherboard and power on your computer.

Features of Thermalright’s 360mm AIOs

▶ Large copper contact plate

Both units include large copper CPU contact plates to help move heat away from the CPU.

▶ 27mm thick radiator

The radiators included with the liquid coolers are 27mm thick, which is typical of most units currently on the market.

▶ In-line AIO pump (360 Notte only)

Most AIOs on the market feature liquid pumps integrated on the CPU block. The Frozen Notte 360 incorporates an in-line liquid pump, which is different from many other units on the market. In the past, many AIOs with this style of pump have underperformed – but as you’ll see in the benchmarks below, that is not the case with the Frozen Notte 360 as it performs very well!

▶ Fully rotatable, braided tubing and full RAM compatibility

As an AIO does not interfere or overhang DIMM slots in any manner, all sizes of RAM, no matter how tall, are compatible Thermalright’s AIOs. The tubes of the of the AIO are braided, secured with metal fittings, and are fully rotatable for ease of installation and setup.

▶ Cable management clips

Included with the AIOs are clips to help keep the tubing of the AIO straight and tidy.

▶ ARGB accents on CPU Block (Aqua Elite White V3 only)

The CPU block of the Aqua Elite White V3 has thin ARGB strips incorporated into its design, giving it a unique aesthetic accent.

▶ Infinity Mirror CPU Block (Frozen Notte 360 only)

The Frozen Notte 360 features an infinity mirror CPU block to give it a catchy flair.

▶ Accessible coolant refill port

Some AIO manufacturers don’t include refill ports on their AIOs. The manufacturers that do often try to scare you away from servicing your own equipment with a “warranty void when removed” sticker that is technically illegal in the U.S. under the Magnuson-Moss Warranty Act.

While you shouldn’t need to refill Thermalright’s AIOs during the lifespan of their 3-5 year warranties, but should you feel the need to refill the unit there is an accessible port available without any stickers to scare you away from servicing the equipment.

▶ 3x 120mm fans

I say this on almost every cooler review, but there’s more to a cooler than just the heatsink or radiator. The bundled fans have a significant impact on cooling and noise levels, as well as how the cooler looks in your case.

The fans on each unit are different. The Frozen Notte includes 7-bladed fans that are powerful and can run loudly. The Aqua Elite’s fans have 9 thinner fan blades and are tuned for lower noise levels.

LGA1700 Socket Bending

There are many factors other than the CPU cooler that can influence your cooling performance, including the case you use and the fans installed in it. A system's motherboard can also influence this, especially if it suffers from bending, which results in poor cooler contact with the CPU.

In order to prevent bending from impacting our cooling results, we’ve installed Thermalright’s LGA 1700 contact frame into our testing rig. If your motherboard is affected by bending, your thermal results will be worse than those shown below. Not all motherboards are affected equally by this issue. I tested Raptor Lake CPUs in two motherboards. And while one of them showed significant thermal improvements after installing Thermalright’s LGA1700 contact frame, the other motherboard showed no difference in temperatures whatsoever! Check out our review of the contact frame for more information.

Testing methodology

Today's highest-end CPUs, whether Intel or AMD, are difficult to cool in intensive workloads. In the past. reaching 95 degrees Celsius or more on a desktop CPU might have been a cause for concern. But with today’s top-end CPUs, this is considered normal operation. Similar behavior has been present in laptops for years due to cooling limitations in tight spaces.

All testing is performed with a 23 degrees Celsius ambient room temperature. Multiple thermal tests are run on each CPU to test the cooler in a variety of conditions, and acoustic measurements are taken with each result. These tests include:

1. Noise-normalized testing at low noise levels

2. “Out-of-the-box”/default configuration thermal & acoustics testing

     a. No power limits enforced

     b. Because CPUs hit TJMax in this scenario, the best way to compare cooling strength is by recording the total CPU package power consumption.

3. Thermal & acoustic testing in power-limited scenarios

     a. Power limited to 175W to emulate a medium-intensity workload

     b. Power limited to 125W to emulate a low-intensity workload

The thermal results included are for 10-minute testing runs. To be sure that was sufficiently long to tax the cooler, we tested both Thermalright’s Assassin X 120 R SE and DeepCool’s LT720 with a 30-minute Cinebench test with Intel’s i9-13900K for both 10 minutes and 30 minutes. The results didn’t change much at all with the longer test: The average clock speeds maintained dropped by 29 MHz on DeepCool’s LT720 and 31 MHz on Thermalright’s Assassin X 120 R SE. That’s an incredibly small 0.6% difference in clock speeds maintained, a margin of error difference that tells us that the 10-minute tests are indeed long enough to properly test the coolers. 

Testing configuration – Intel LGA1700 platform

 MORE: How to Buy the Right CPU Cooler

MORE: How to Check CPU Temperature

MORE: All CPU Cooling Content

Thermal results without power limits

Without power limits enforced on Intel’s i7-13700K, the CPU will hit its peak temperature (TJMax) and thermally throttle with even the strongest of air coolers. When the CPU reaches its peak temperature, I’ve measured the CPU package power to determine the maximum wattage cooled to best compare their performance.

The general exception to this comes with the strongest AIOs on the market, which can keep Intel’s i7-13700K under TJMax. This is no small task, as most 360mm AIOs still fail this test.

The Frozen Notte 360 does very well here, matching the performance of the 360mm AIOs which are tied for my 2nd best thermal results – but it runs rather loudly to do so, reaching up to 52.5 dBA. The Aqua Elite 360 V3 doesn’t do as well thermally, but it runs much quieter with a maximum noise level of 46 dBA. 

Thermal results with noise normalized to 38.2 dBA

Finding the right balance between fan noise levels and cooling performance is important. While running fans at full speed can improve cooling capacity to some extent, the benefits are limited and many users prefer a quieter system. With this noise-normalized test, I’ve set noise levels to 38.2 dba. This level of noise is a low volume level, but slightly audible to most people.

Both of Thermalright’s AIOs perform well when noise normalized to a low 38.2 dBA. Thermalright’s Aqua Elite 360 cooled 232W in this test, the Frozen Notte performed slightly better cooling 235W.

175W Cinebench results

Most coolers on the market can keep Intel’s i7-13700K under its peak temperature if the power consumption is limited, so for this test, we’ll be looking at the CPU’s actual temperature.

In a more moderate 175W thermal load, both of Thermalright’s offerings had middle of the road results for an AIO. The Frozen Notte’s noise levels were on the higher end compared to other AIOs, reaching 46.3 dBA. Users wanting a quieter experience should look to the Aqua Elite 360 V3 instead, which only reached 41.9 dBA – one of the quieter results in this scenario.

125W Cinebench results

The lowest power limit I test with Raptor Lake CPUs is 125W. This is a high enough limit to allow the CPU to maintain its base clock speeds even in the most intensive tests, and most coolers should be capable of keeping the CPU below TJmax (the max temperature before throttling) – even low-end coolers.

Really, thermals do not matter in this scenario. Even Intel’s stock cooler can handle a load like this with ease. Noise levels, rather than CPU temperature, are the most important factor here. The noise levels of both coolers were acceptable, with the noisier Frozen Notte 360 only reaching 38.9 dBA. The Aqua Elite 360 ran just a little quieter, reaching 38.2 dBA.  

Conclusion

Just as with its air coolers like the Phantom Spirit 120 EVO, Thermalright’s liquid coolers provide a great value with low prices and strong thermal performance. The Frozen Notte 360 runs a bit louder than it needs to out of the box, so I’d advise manually tuning the fan curve if you’re considering that model. The Aqua Elite 360 V3 has much better overall noise levels, but it isn’t quite as strong.

 MORE: How to Buy the Right CPU Cooler

MORE: How to Check CPU Temperature

MORE: All CPU Cooling Content

Thermalright's AXP90-X53 air cooler is getting the company's signature all-copper design for the existing the 53mm top-flow cooler. Other than the retention clips and fan, the cooler is all-in on copper. It's pretty exciting as copper is known to be one of the best heat transfer materials, but it's usually limited to use with copper heatpipes — mixed with aluminum or nickel plated to prevent oxidation. Aluminum normally gets used for the rest of the parts, including the fin stack, but Thermaltake makes special all-copper editions for some of its CPU coolers.

The AXP90-X53 all-copper air cooler uses four 6mm heatpipes with one end connecting the C1100-grade copper base and another passing through its copper fin stacks. While the whole CPU cooler uses bare copper, the base does have a colour-matching nickel plating. The low-profile CPU cooler is made for mini-ITX systems, with a 92mm x 94.5mm length and width and a height of 38mm. The weight of the all-copper variant is 540g without a fan, while the standard nickel-plated AXP90-53 is 320g including the fan. There are two more variants as well, the AXP90-X53 black and the full black.

To blend with an all-copper heatsink tastefully, Thermalright includes a TL-90152 92mm fan with orange fan blades inside the red frame. The fan uses a fluid dynamic bearing with a max airflow rating of 42.58 CFM and a max noise level of 22.4 dbA.

We're a bit puzzled as to why Thermalright excludes the weight of the included fan in its all-copper variant while including the fan's weight in other models, but the TL-9015R fan weighs 55g, bringing the all-copper edition to 595g. Copper is heavier than aluminum so it makes sense for an all-copper version to be heavier than the standard AXP90-53. On the bright side, it's not an all-copper tower CPU cooler, so the total weight isn't obscene.

All variants of the AXP90-53 CPU cooler are compatible with LGA115x/1200/1700 sockets for Intel, and AMD AM4/AM5 mounts. The LGA1200 and LGA115x mounting holes are the same, and the installation instructions cover the details. Thermalright doesn't seem to limit its users to pairing its low-profile CPU cooler with CPUs under a certain TDP, something most low-profile coolers typically do. That's likely the benefit of the superior heat conductivity of copper.

Does it make a difference?

The pricing and performance of the AXP90-53 is not known, but Thermaltake has another all-copper low-profile CPU cooler, the AXP90-47. According to tests done by a user, the Thermalright AXP90-47 full-copper hit 60C with a 105W TDP CPU using an AMD Ryzen 7 3800x, while its nickel-plated variant reached 67.5 degrees Celsius, both with 100% fan speed and excluding ambient temperature. For a low-profile CPU cooler, the difference between the variants looks significant. We'd be interested in seeing tests with a higher power Intel CPU like the Core i9-13900K, or even the previous gen 12900K.

The difference between the AXP90-53 and AXP90-47 is that the new model is 38mm tall, while the AXP90-47 is shorter at 32mm. It will also be interesting to see how these all-copper versions compete with other low-profile air coolers and liquid coolers made for mini-ITX systems. The Noctua L9x65 certainly comes to mind.

Love and obsession with all-copper CPU coolers

The love for the all-copper CPU coolers has a cult-like following because of what Thermalright has done in the past. It started perhaps with the legendary Thermalright Ultra 120 Extreme, colloquially called TRUE120-E. Apart from its vast cooling advantage in comparison to the standard variant, the cooler was equally well-known for its 1.9kg weight — without fans. That tower of power required a horizontal mount so as to avoid breaking the motherboard supports.

The Noctua NH-D15s had a dual fin stack and a 140mm fan and 'only' weighed 1150g, much lighter than TRUE-E all-copper. Those were the good old days, when Noctua was relatively new in the market with the U12P and the well-known U12P-SE2 CPU cooler taking the crown away from most 120mm fan tower coolers. Nobody sold closed loop liquid coolers at the time. Other all-copper variants included the Thermalright AXP-100 with fancy designer ends on the copper heatpipes.

The main caveat of copper is that it oxidizes relatively easily, hence the use of nickel plating on copper surfaces on CPU coolers. Some CPU cooler manufacturers use copper-composite heatpipes to ensure the copper 'shine' sticks around. We typically saw these coolers paired with an Intel Q6600 quad-core CPU, or later the Core i7-920. CPU enthusiasts wanted the extra overclocking performance, and it was worth the effort of cleaning off the copper oxidation on occasion.

Where might Thermalright go next? Would it be too crazy to see its HR-10 2280 M.2 cooler get an all-copper treatment? Probably not, considering Phison E26 PCIe 5.0 controllers are known to get pretty hot. Thermalright already has multiple variants of that cooler, including one with a pre-installed fan.

Thermalright hasn't listed the price for its AXP90-53 all-copper CPU cooler. But as a benchmark, the AXP90-47 Black costs $18.32 while the all-copper version costs $37.99. We'll have to see where the AXP90-53 copper variant lands before we can answer the question of whether it's worth purchasing or not.

When we mention Thermalright, we think of high-performance air or liquid cooling systems as well as accessories for DIY PCs and, therefore, products aimed at enthusiasts. But it looks like even Thermalright wants to address a more mainstream market, which is why this week it quietly unveiled its first thermal pads, as FanlessTech noticed.

While thermal pads often offer lower thermal conductivity and performance than the best thermal paste, Thermalright rates the thermal conductivity of its Heilos pads at 8.5W/mK, and thermal resistance at 0.04°C cm2/W. To put the numbers into context, the thermal conductivity of Arctic's MX-4/MX-5 is also spec'd at 8.5W/mK, and this is considered to be a quite decent level of performance for inexpensive pastes, as far as our list of the best thermal pastes goes. 

Thermalright offers two types of Heilos thermal pads: one measuring 40x40 mm for AMD's AM5 CPUs and another measuring 40x30 for Intel's LGA1700 processors. Both thermal pads are 0.2 mm thick.

Applying thermal paste optimally could be challenging for novices in DIY PC building, so specialists from Thermalright thought that advanced thermal pads could be a viable option for them. While thermal pads tend to offer lower performance than the best thermal pastes, which are usually designed for overclocking and tailored for particular types of cooling systems (air, liquid, liquid nitrogen, etc.), they are easy to use. Furthermore, Thermalright's Helios promises performance that is akin to good mainstream thermal pastes, so for DIY PC rookies, they might be just what the doctor ordered.

Thermalright established itself as a cooling contender in decades past with options like its all-copper SP94 cooler. The company, also known for unique products like the HR-07 memory cooler, which provided a full heatsink with two heatpipes for DDR2 RAM, and the LGA 1700 Contact Frame which we recently reviewed.

Today we’ll be looking at Thermalright’s compact AXP120-X67 SFF air cooler, giving it quite the challenge by pairing it with Intel’s i9-13900K. Will the AXP120-X67 rise to the challenge and earn a spot on our Best CPU Coolers list? We’ll have to put it through testing to find out, but first here are the specifications from Thermalright.

Cooler Specifications

Packing and Included Contents

Thermalright’s AXP120-X67 comes in a small box, with cardboard and molded foam for the protection of the inner contents.

Included with the package are the following:

• SFF Heatsink
• 1x TL-C12015 fan
• Fan clips for the included fan
• Fan clips for standard 120mm fans
• Mounts for all modern CPU sockets, including AM5 & LGA1700
• A tube of TF7 thermal paste
• Manual

Features of the AXP120-X67

SFF Heatsink

The slim 2.05-inch height of the AXP120-X67 allows it to fit inside of the smallest cases, making it an appealing option for SFF builds.

TL-C12015 fan

There’s more to an air cooler than just a heatsink. The fan included has a significant impact on cooling, noise levels and performance. Included with Thermalright’s AXP120-X67 is a slim TL-C12015 fan.

Additional fan clips

Thermalright also includes fan clips for standard 120mm fans for those who want to use a stronger, full sized fan, for additional cooling performance.

Testing Methodology

While it was fairly easy with previous generations of CPUs for coolers to keep the processor well under TJ max (the maximum temperature a CPU can sustain before throttling) in demanding workloads, this is no longer realistically possible on current generation CPUs without extreme cooling (or enabling power limits).

When I began testing coolers on Intel’s i9-12900K, I found that some products that cooled the i9-10900K well struggled in some scenarios when paired with the Alder Lake CPU – and Raptor Lake is even more difficult to cool in these situations. While in the past a CPU hitting its peak temperature was cause for concern, enthusiasts are going to have to learn to accept high temperatures as “normal” while running demanding workloads with Raptor Lake and Ryzen 7000 CPUs. 

These higher temperatures might cause concern for some, but there isn’t any need for worry. Modern AMD & Intel CPUs are designed to run fairly hot without any problems – up to 95 degrees Celsius for AMD Ryzen 7000 CPUs, and up to 100 degrees C for Intel’s 13900K. Similar behavior has been standard in laptops for years due to the cooling limitations there. 

Furthermore, Intel’s i9-13900K supports Adaptive Boost Technology (ABT), which allows Core i9 processors to dynamically boost to higher all-core frequencies based upon available thermal headroom and electrical conditions. This allows multi core loads to operate at up to 5.5 GHz if the necessary amount of thermal dissipation is there. This feature works in a way that actively seeks high temperatures: If the chip sees that it is running underneath the 100-degree C threshold, it will increase its performance and power consumption until it reaches the safe 100 degree C limit, thus sustaining higher clocks (and providing better performance) for longer periods.

The increased cooling challenges posed by Raptor Lake means that we’ve had to change some of the ways we test coolers. Some coolers were able to pass Cinebench R23 multicore testing with Intel’s 12th Gen i9-12900K when power limits were removed, although only the strongest models were able to pass that test. Most liquid coolers and all air coolers I’ve tested “failed” that test because the CPU reached TJ max in this scenario. 

With Raptor Lake’s 13900K, not a single cooler tested (at least so far) has been able to keep the CPU under TJ max in this test. We’ll compare performance instead by comparing total benchmark scores and clock speeds maintained. 

I’ll be testing Intel’s i9-13900K CPU using Asus’ TUF Gaming Z690 Gaming Plus WIFI motherboard and Cooler Master’s HAF 700 Berserker computer case, with case fans limited to 35% speeds. The motherboard’s default fan curve is used for the CPU Cooler’s fans.

In addition to testing Cinebench without power limits enforced, we’ll also be showing results when the CPU’s power consumption is limited to a more reasonable 200W. We’ll also show results at 125W for those who prefer whisper-quiet cooling, at the cost of some performance. For both of these results, we’ll show traditional delta over ambient temperature results.

We’ll provide noise level measurements recorded using a PSPL25 Sound Meter for all three power levels tested to compare how much noise each cooler makes in different scenarios. We expect most coolers to run silently at 125W.

LGA1700 Socket Bending

Please note there are many factors other than the CPU cooler that can influence your cooling performance, including the case you use and the fans installed in it. A system's motherboard can also influence this, especially if it suffers from bending, which results in poor cooler contact with the CPU.

In order to prevent bending from impacting our cooling results, we’ve installed Thermalright’s LGA 1700 contact frame into our testing rig. This means that if your motherboard is affected by bending, your thermal results will be worse than those shown below. Not all motherboards are affected equally by this issue. I tested Raptor Lake CPUs in two motherboards. And while one of them showed significant thermal improvements after installing Thermalright’s LGA1700 contact frame, the other motherboard showed no difference in temperatures whatsoever. Check out our review of this contact frame for more information.

Testing Configuration

• MORE: Best CPU Coolers
• MORE: Raptor Lake All We Know
• MORE: How to check CPU Temperature

Cinebench With No Power Limits

With Raptor Lake’s i9-13900K pushing all-core turbo speeds of 5.5ghz+, even the strongest of coolers hit TJ Max while running Cinebench R23 and other demanding scenarios. As no cooler that we’ve tested thus far can pass R23 testing using a traditional delta over ambient measurement, we’ll be comparing the overall benchmark score and the CPUs clock speeds instead. 

The results below are for a 10-minute testing run, but to be sure that was sufficiently long we also retested both Thermalright’s Assassin X 120 R SE and DeepCool’s LT720 with a 30-minute Cinebench test. The results didn’t change much at all with the longer test: The average clock speeds maintained dropped by 29mhz on DeepCool’s LT720 and 31mhz on Thermalright’s Assassin X 120 R SE. Looking at this another way, that’s an incredibly small 0.6% difference in clock speeds maintained, a margin of error difference that tells us that the 10 minute tests are indeed long enough to properly test the coolers.

Because of the difficulty of cooling the i9-13900K in this workload, peaking at roughly 335W if the cooler is capable, we feel the best way to compare coolers here is to record the average power consumption of the CPU.

In this test the AXP120-X67 is the worst performing example, but this is to be expected of a SFF cooler. Given it’s small profile, I didn’t expect it to be able to handle over 150W. I was pleasantly surprised to see the cooler was actually capable of cooling almost 200W on average.

Looking at the wattage results posted earlier, you might think that there would be a significant gap in the benchmark results using the different coolers. DeepCool’s LT720 AIO is cooling nearly 60% higher watts than Thermalright’s AXP120-X67 - so surely performance has a similar gap, right? Well, you’d be wrong. Despite DeepCool’s LT720 cooling 117W more than Thermaltake’s AXP120-X67, it’s benchmark score is only 13.7% faster.

Normally we would also show acoustic results here, but because this cooler can’t sustain loads above 200W it would not be fair to compare the results against other products which are cooling heavier workloads. Instead, we’ll show the acoustic results in the 200W and 125W sections.

200W Cinebench Results

Thermalright’s AXP120-X67 is an SFF cooler, and as such it’s not designed to handle high TDP loads. But it surprised us in being able to sustain 198W in a long term load, though it did run at or near TJMax for the duration of this test.

Looking at acoustic levels here, Thermalright’s SFF cooler does well. It’s barely louder than the company’s Assassin 120 R SE, and quieter than DeepCool’s offerings. Considering it’s smaller heatsink and slim fan, being able to run almost as quietly as the Assassin 120 R SE while still cooling nearly 200W is indeed impressive.

125W Cinebench results

The lowest power limit I test at is 125W, mainly because this is also the lowest level where I can measure reliable noise measurements. Lower power consumption causes the noise of the CPU cooler to fall below the noise created by the system fans (even while restricted to 35% speed). 

When limited power to 125W the AXP120-X67 runs silently and is just barely distinguishable from the rest of the system fans, if you listen closely. It does run a bit warmer than the other coolers tested, but this isn’t a cause for concern given that the increased temperature here doesn’t impact performance and the cooler runs quite silently here.

Conclusion

Intel officially recommends at least a 280mm AIO cooler (or air-cooling equivalent) to get the best possible performance from the Core i9-13900K, but Thermalright’s AXP120-X67 shows that even a small form-factor cooler can handle the Raptor Lake flagship in most scenarios. In the worst case scenarios, the AXP120 provided 88% of the benchmark performance of the same chip using a high-end AIO. Of course, that’s not taking overclocking into account, but that’s pretty darned good for a slim cooler that will fit in even the smallest of SFF cases.

• MORE: Best CPU Coolers
• MORE: Raptor Lake All We Know
• MORE: How to check CPU Temperature

Thermalright established itself as a cooling contender in decades past with options like its all-copper SP94 cooler, and unique products like the HR-07 memory cooler, which provided a full heatsink with two heatpipes for DDR2 RAM. And the company is back with another intriguing cooling-focused product. 

After posting our previous review of Thermalright’s Peerless Assassin 120 SE, some of our readers asked us to test the company’s LGA 1700 contact frames, which are designed to prevent and impove issues surrounding CPU and socket board bending with Intel’s current-gen socket. We reached out to Thermalright and the company sent us a sample for testing. Is the frame as good as it's rumored to be? The short answer is yes, but with a few asterisks. 

What's the purpose of this product? 

Due to the elongated shape of LGA 1700 CPUs, and how they are secured in the socket, the CPUs have been known to bend or warp while installed in the motherboard’s socket. This can cause higher temperatures due to uneven contact with CPU coolers. 

There are a few solutions available for enthusiasts who wish to fix this issue in pursuit of lower CPU temperatures, but they all involve potentially warranty breaking modifications. Intel previously told Tom’s Hardwre:

"We have not received reports of 12th Gen Intel Core processors running outside of specifications due to changes to the integrated heat spreader (IHS). Our internal data show that the IHS on 12th Gen desktop processors may have slight deflection after installation in the socket. Such minor deflection is expected and does not cause the processor to run outside of specifications. We strongly recommend against any modifications to the socket or independent loading mechanism. Such modifications would result in the processor being run outside of specifications and may void any product warranties." —Intel Spokesperson to Tom's Hardware.

Buildzoid and Igor’s Lab tested a washer modification to alleviate the bending issue, and Thermal Grizzly, Thermalright, and others have released complete replacements for the ILM (Independent Loading Mechanism, the CPU locking hardware) of the LGA1700 socket. There are two primary manufacturers of the socket: Lotes and Foxconn. In theory, motherboards may be affected differently depending on the socket manufacturing. 

We’ll be testing Thermalright’s LGA 1700 contact frame on two systems. The first will use MSI’s Z690 A Pro motherboard, which I’ve been using for most of my thermal tests since Alder Lake’s release. The second is a recently purchased Asus TUF Gaming Z690 Plus WIFI D5. I’ve upgraded both with Intel’s brand-new Raptor Lake CPUs for this test. 

Specifications for the Thermalright LGA1700 Contact Frame

Packing and Included Contents

Thermalright’s LGA1700-BCF contact frame ships in a tiny cardboard box that’s smaller than a keyboard number pad.

The bottom of the box includes basic instructions for installation.

Included with the package are the following:

• Contact Frame
• Thermalright TF7 paste
• L-Shaped screwdriver

Product Installation

I was a little nervous about the setup, because I’ve heard reports of folks having problems with stability after installation of similar products. However, in my experience this was an extremely easy product to install, as long as you take care during setup.

First you’ll use the included screwdriver to remove the screws from the standard CPU socket. Once the screws have been removed, you can ditch the old frame - but don’t get rid of the screws! Then it’s a simple matter of mounting the frame around the CPU and securing them with the screws. Before securing the screws, rotate them backward in the thread until you hear a click, and then secure the screw. This will ensure the socket and screw are properly secured. Then, just attach your cooler of choice (hopefully one of the best CPU coolers), as normal.

Testing Configurations

Asus + i9-13900K Thermal Results

I theorized that since Intel’s i9-13900K can run hotter than other LGA1700 chips, it should see more of a temperature improvement with Thermalright’s frame. So I installed it on Asus’ TUF Gaming Z690 Plus WIFI DDR5 motherboard, pairing it with DeepCool’s AG620 air cooler.

The results were utterly underwhelming; I didn’t see a difference in cooling effectiveness whatsoever. There wasn’t any problem with cooler contact to begin with in this configuration, and as such I didn’t see any improvement in temperature. 

The temperature results of every single thermal scenario I tested here were, within margin of error, exactly the same. I was a bit disappointed with this result, because Intel’s i9-13900K can be difficult to keep cool in some scenarios.

MSI + i5-13600K Thermal Results

After finishing the less-than-exciting testing mentioned above, I thought I might have wasted my time. However, I decided to give it another chance and test the frame with MSI’s Z690 A PRO DDR4 motherboard. I’ve been using that board since Alder Lake’s launch for testing thermal results with air and AIO coolers, and I have definitely noticed bending issues on this setup.

I’m glad I tried testing again, because on this system I noticed significant differences in thermal performance. I tested Intel’s i5-13600k, paired with an Arctic Liquid Freezer II 240, running Cinbench R23’s ten minute multi-threaded benchmark. 

I tested thermal performance at three levels: With power limits removed, which consumed  roughly 200W during this test, as well as at both 125W and 95W.

Running Cinebench without power limits can be difficult to cool, and even with Arctic’s Liquid Freezer II 240, Intel’s i5-13600K peaks at TJ max (the highest temperature before throttling occurs, 100 degrees C) in this test. Even with Thermalright’s frame installed, the i5 CPU still peaked at TJ max in this test, but it did take longer to reach that limit.

On average, temperatures dropped by 6 degrees C in this ten-minute test. A stronger cooler might show a bigger difference in this scenario, if it were able to stay under TJ max. Things were much more interesting when limited to a more reasonable 125W. There, temperatures dropped by a massive 12 degrees Celsius when paired with Thermalright’s contact frame.

When limited to a low(er) power 95W TDP, the differences were less pronounced. There, the delta between the configurations was only 3 degrees C.  

There’s another benefit not shown in the chart above: Lower temperatures mean that fans run slower, which in turn means they make less noise. I didn’t take detailed noise readings for this test (which would obviously vary quite a bit from cooler to cooler), but in each of these tests the average fan RPM speed was lower when paired with Thermalright’s contact frame.

Conclusion

Thermalright’s LGA 1700 contact frame can greatly improve thermal performance if your CPU suffers from bending. But it won’t make a difference if your cooler contact is already good. While testing it with Intel’s i5-13600K, I observed reductions of up to 12 degrees C in CPU temperature and saw lowered fan speeds as a result. The frame was very easy to install, and is easy on the wallet, costing less than $20. 

If you’re more concerned about CPU temperatures and system noise levels than you are about (probably) voiding your warranty, Thermalright’s LGA1700 contact frame is worth considering. Even if you don’t have bending issues yet, the company says its Contact Frame should also help prevent the problem from happening in the first place. Time will tell how much that holds true, but $20 isn’t much to pay for a little peace of mind for your expensive CPU and motherboard investment.

Cooling specialist Thermalright has launched a new accessory for early adopters of the AMD Ryzen 7000 platform. It's pitching its new AMD AM5 Secure Frame at PC DIYers and enthusiasts looking to optimize their CPU socket. The design certainly has some unique AMD Raphael appeal, as it is precision-made with a notched design to prevent any thermal grease from oozing into your CPU’s crevices.

CPU frame accessories have gained some favor with Intel PC DIYers in 2022 due to how the LGA1700 socket and processors (Alder Lake, Raptor Lake) are designed. Products such as the Thermal Grizzly Contact Frame, or, indeed, some Thermalright designs, are said to lower Alder Lake CPU temperatures by as much as 10 degrees Celsius by preventing CPU integrated heat spreader (IHS) flexing and bending. 

As far as we know there is no need for this type of support frame on the AMD AM5 platform. AMD’s new CPU has an extremely thick IHS, but it comes with other drawbacks.

Thermalright doesn’t offer any explanation about why Socket AM5 users will be attracted to its new AM5 Secure Frame, which is available in black or red. It released the product without fanfare (or a press release) and the product page is extremely spartan, with just a smattering of tech specs. However, the design plainly features an internal void that echoes the shape of the new Ryzen 7000 series CPUs. The notches, as long as the design is precise enough, will act very much like the Noctua Paste Guard for Ryzen 7000, but with a much more rigid and premium design.

The AM5 Secure Frame is made from a block of aluminum alloy, with black or red finish. It measures 75 x 56 x 7.5mm and weighs 45g. Thermalright bundles it with 2g of its TF7 thermal grease and an L-shaped screwdriver for fitting. A warranty of six years is provided, but we don’t currently have pricing and availability. For reference, Thermalright's latest Intel LGA1700-BCF CPU 'Bending Corrector Aluminum Frame' is listed for $14.99 on Amazon, so it will likely be priced similarly.

Thermalright doesn’t offer up any purported benefits for using the new AM5 Secure Frame, which is an interesting new marketing ploy. Hopefully, we'll get to see some independent third party reviews or investigations regarding its qualities soon.

Thermalright was founded in Taiwan in 2001, and established itself as a cooling contender in decades past with options like its all-copper SP94 cooler, and unique products like the HR-07 memory cooler, which provided a full heatsink with two heatpipes for DDR2 RAM. The company’s current lineup includes air & liquid coolers, fans, thermal pastes, and a variety of accessories. 

After posting the previous review of DeepCool’s AK500, some of our readers asked us to test Thermalright’s Peerless Assassin 120 SE. We reached out to Thermalright, and the company sent us a sample for testing. It features a dual-tower radiator and two 120mm fans, but is that enough to tame Intel’s 12900K and earn a spot on our Best CPU Coolers list? The short answer is yes, but we’ll have to put it through testing to find out just how good it is. First, let’s check out the cooler’s specifications from Thermalright. 

Specifications for the Thermalright Peerless Assassin 120 SE

Packing and Included Contents

The Peerless Assassin 120 SE arrives in a medium-sized brown cardboard box, with cardboard and molded foam for protection.

Included with the package are the following:

• Dual-tower heatsink
• 2x TL-C12C 120mm fans
• Mounts for modern Intel & AMD platforms
• A medium-sized tube of thermal paste
• Fan splitter
• Installation guide

Cooler Installation

Installing the Peerless Assassin was not hard. To begin, you’ll first want to secure the backplate to your motherboard. Next, mount the stand-offs and then secure the mounting brackets with the provided screws. After applying thermal paste, place the cooler on the brackets and secure them with the captive screws. Finally, attach the fans to the towers using the provided clips.

New Testing Configuration

What's different than other coolers?

• Budget-friendly price of $36 USD 

While it’s not unheard of to find a single tower air cooler for $40, most dual-tower air coolers are priced higher at $60 and more.

• Generous Packaging

I don’t normally comment on a cooler’s packaging in this section of the review, the Peerless Assassin is packaged exceptionally well, with cardboard cutouts and molded foam which make it highly unlikely it will get damaged during shipping. Its brown exterior and black ink are far from flashy, but what’s here is all about utility, and as we’ll see soon in testing, that goes deepter than just the packaging. 

• Dual TL-C12C 120mm Fans

There’s more to an air cooler than just the heatsink. The fan included has a significant impact on cooling, noise levels, and performance. Included with Thermalright’s Peerless Assassin 120 SE are a pair of  TL-C12C 120mm fans. There’s no RGB or lighting at all, but again, this cooler is all about performance at a low price. And as our testing will show, it’s truly unmatchable on that front.

I'll be testing Thermalright’s Peerless Assassin 120 SE with Intel's Core i9-12900K. Due to the increased thermal density of the Intel 7 manufacturing process, as well as changes to core and component layouts, Alder Lake CPUs are more difficult to cool than previous generation CPUs in the most heat-intensive of workloads. 

This means that coolers that kept previous-gen products like the i9-10900K nice and cool sometimes struggle to keep Intel's i9-12900K under Tj max–the maximum temperature before the CPU starts to throttle. Many coolers, air coolers in particular, fail to keep the i9-12900K under TJ max when power limits are removed in workloads like Cinebench and OCCT when I’ve tested them, including many coolers I previously considered top of the line.

Please note there are many factors other than the CPU cooler that can influence your cooling performance. A system's motherboard can especially influence this, as there are boards on the market with CPU sockets that aren't up to Intel's spec, which can cause warping or poor contact with the CPU. The case you use will also influence cooling results. 

With Alder Lake's cooling demands in mind, I'll be rating CPU Coolers in 3 different tiers.

Tier 1: These coolers are able to keep the i9-12900K below TJ max in most loads, with no power limits enforced. I expect only the best liquid coolers to meet this standard. 

Tier 2: These coolers are able to keep the i9-12900K under the TJ max threshold with CPU power limits of 200W enforced. I expect most liquid coolers and the best air coolers to meet this standard.

Tier 3: These coolers are able to keep the i9-12900K under TJ max with CPU power limits of 140W enforced.

Testing Methodology

To test the limits of a cooler's thermal dissipation capabilities, I run two primary stress tests: Cinebench and OCCT, each for 10 minutes. While this may be a short amount of time, it is sufficient to push most coolers–air and liquid–to their limits. 

While stress testing in Cinebench, I run both with power limits removed and with an enforced 200W CPU power limit,  using MSI’s Z690 A Pro DDR4 Motherboard and Be Quiet’s Silent Base 802 Computer Case. Only the most capable coolers are able to pass Cinebench testing when power limits are removed. 

I don’t test OCCT without power limits, because attempting to do so results in CPU package power consumption jumping to over 270W and instantly throttling with even the best AIO coolers. Instead, I test at 200W to give coolers a chance at passing. I also include 140W to give data comparable to cooling a CPU like Intel’s i5-12600K. 95w results are also included to give data closer to low TDP CPUs such as AMD’s Ryzen 5600X or Intel’s i5-12400.

Cinebench Test Results

I first attempt to run Cinebench without power limits enforced. If the cooler can handle it, this load will use a little more than 230W. Unfortunately, I haven’t found an air cooler capable of passing this test, and Thermalright’s Peerless Assassin 120 SE is no exception.

When testing a more reasonable 200W CPU power limit, Thermalright’s Peerless Assassin 120 SE performed exceptionally well. In this workload, the CPU temperature averaged 61 degrees Celcius over ambient – the best resultI’ve seen from any air cooler I’ve tested, beating DeepCool’s AK500 by seven degrees C!

Thermalright’s Peerless Assassin 120 SE is also the only air cooler that has passed my 200W tests when limited to 50% fan speeds,  though it did get quite toasty when doing so.

200W and 140W OCCT Test Results

I usually like to run OCCT's small set stress testing for stability when overclocking, but on Alder Lake, I haven't found a cooler that's capable of handling OCCT without throttling unless power limits are enforced.

I test OCCT at 200W to demonstrate a thermally demanding load, but also with a 140W power limit enforced to show how these coolers might perform with a CPU that's easier to cool, like Intel's i5-12600K or AMD's Ryzen 5800X.

With OCCT, Thermalright’s Peerless Assassin SE continued to deliver outstanding results. While it wasn’t able to pass 200W with reduced fan speeds in this scenario, at the default fan curve it averaged 64 degrees C over ambient – once again degrees C cooler than DeepCool’s AK500!

Where things get really interesting with the Peerless Assassin 129 SE is in lower wattage loads. In fact, I had to retest these results multiple times because I couldn’t believe my own results. When limited to 140W and with fans running at the default fan curve, the Peerless Assassin SE outperformed every cooler I’ve tested – including many high-end AIOs. At 140W with fans reduced to 50% speeds, Thermalright’s cooler tied with BeQuiet’s Pure Loop 2 FX and Cougar’s Poseidon GT360 AIOs for the best cooling performance.

95W OCCT Test Results

Due to requests from our readers, I’ve begun to include 95W results, which should be broadly applicable to CPUs like AMD’s Ryzen 5 5600V or Intel’s Core i5-12400. In these lower wattage tests, the Peerless Assassin once again lives up to its name, outperforming it’s nearest rival by six degrees C when fans ran at the default fan curve, and a whopping 12 degrees C when set to 50% fan speeds.

These results are even more interesting when you consider that most coolers do worse when limited to 50% fan speeds in this test. But the Peerless Assassin does better. This is because it runs at less than 50% fan speeds in this test using the motherboard’s default fan curve.

Noise Levels and Acoustics

To test noise levels, I used the SLM25TK Sound Level Meter positioned 18 inches behind the rear of the Be Quiet Silent Base 802 PC case, and recorded early in the morning to achieve the lowest noise floor possible. The chart below shows averaged results, measured over the course of five minutes, to account for sudden sound spikes.

In all workloads tested, the Peerless Assassin 120 SE ran very quietly. When running at maximum fan speeds, it is the second-quietest cooler we’ve tested thus far. At an enforced 50% fan speed, it runs quieter than all coolers we’ve tested except for Cooler Master & Corsair’s AIOs.

Conclusion

I was truly amazed by the performance of Thermalright’s Peerless Assassin 120 SE, especially when you consider its low $36 price. I had to retest this cooler multiple times before I would believe my own testing results! It is by far the best-performing airr I’ve tested, capable of handling 200W+ with Intel’s Core i9-12900K, while running whisper silent, and it absolutely murders its competition in lower TDP workloads. Whether you care about value or not, you should consider this cooler for your next system build.

Thermalright's new pair of M.2 heatsink models, which resemble a scaled down kitchen gadget, promise to funnel heat from even the Best SSDs through their vertical design. The HR-09 2280 is available in a plain and Pro version, with the Pro version essentially doubling the available aluminium surface area that could make a difference between your SSD throttling - or leisurely facing whatever workloads you can throw at it in what should be the hardware equivalent of a poolside summer day.

The Thermalright HR-09 Pro is a blast from the past for hardware enthusiasts, and keep the company's distinct visual style. A pair of 6mm heatpipes conduct heat away from the nickel-plated copper base that attaches to your M.2 SSD, dispersing the heat through thematically-coherent, fire-shaped aluminium fins (33 per heatpipe) that are all jagged edges and protrusions. With dimensions of 74 x 24 x 86mm and weighing in at 90 grams, the HR-09 Pro is taller than some low-profile CPU heatsinks the company has already delivered.

The non-Pro HR-09, on the other hand, is likely to appeal to users that like a cleaner top view of their SSD heatsink, sporting a single 6 mm heatpipe and 32-piece aluminium fin array. The reduced footprint results in smaller dimensions overall, standing at 48 x 24 x 79mm and weighing in at 80 grams. It still keeps the flame aesthetic of the Pro version, albeit only in a side-view; the unit's top features the Thermalright branding and a much cleaner aesthetic.

While current top-end SSDs don't tend to struggle with overheating - which results in throttling and lowered performance in order to keep temperatures in check - PCIe 5.0 SSDs are expected to land in a cooling requirement territory of their own. Phison, one of the industry-leading controller manufacturers for SSDs, has indicated that PCIe 5.0 SSDs would likely require active cooling in order to keep their tremendous performance characteristics in check. While the controllers themselves can reach 120 Celsius, NAND chips work best - and last the most - while operating under 70 Celsius - but would ideally operate between 35 Celsius and 50 Celsius.

None of Thermalright's current solutions feature active cooling of any kind, but ingenious users could fashion fan clips for a 60 mm cooling solution on the HR-09 Pro. We've seen some outlandish SSD cooler designs already - and the number of those is sure to go up as PCIe 5.0 SSDs hit the market with their increased cooling requirements.

Thermalright might not be the most predominant name in PC cooling, but the company has been making cooling solutions since 2001. Today, the company revealed the Peerless Assassin 120 Black, which is a new dual-tower CPU cooler that's just 157 mm tall. 

Of course, that's not super-compact, but using only 120mm fans, it's a lot smaller than something like Noctua's NH-D15, a cooler that pretty much sets the benchmark in the category but is rather sizeable in pulling off that achievement.

So while you won't get the same cooling performance, you'll still get pretty close with two cooling towers and 120mm fans. The assembly measures 125 mm wide, 135 mm long and is 157 mm tall, featuring six, 6 mm heatpipes to bring heat to the 53-fins on each fin stack. The base is made from nickel-plated copper, and the cooler weighs a total of 750 grams.

Thermalright installed two TL-C12B fans on the unit, which can spin at a speed of up to 1550 RPM, when they produce 25.6 dBA of noise, 66 CFM of airflow, and generate 1.53mm of static pressure. They feature an S-FDB bearing and can make use of PWM signals so you'll be able to run them quietly when needed.

The Peerless Assassin 120 will work on all modern CPU sockets including LGA 1200 and AMD's AM4 sockets, though there is no word on pricing or availability just yet.

Not satisfied with the temperatures of the memory on his GeForce RTX 3090 Founders Edition, a determined YouTuber (CryptoAtHome) replaced the factory thermal pads for aftermarket ones. The results are impressive as he managed to improve the temperatures by up to 25 degress Celsius — even while doing Ethereum mining.

Even though the GeForce RTX 3080 and GeForce RTX 3090 are two of the best graphics cards, their memory chips are notorious for running a bit hot if you stress the GPU long enough. Evidently, heat has been a problem from the beginning. An early investigation into the GeForce RTX 3080 already showed the memory hitting dangerous temperatures that surpassed 100C. In our own tests, the memory inside the GeForce RTX 3080 and RTX 3090 peaked at temperatures of 94 degrees Celsius and 104 degrees Celsius, respectively.

Micron rates its GDDR6X chips for operational temperatures up to 95C. Running the memory out of spec during prolonged durations is a recipe for disaster. Cryptocurrency mining takes an even bigger toll on the graphics card and was probably the primary motivation for the YouTuber to swap the thermal pads to improve its thermals.

Before surgery, the YouTuber's GeForce RTX 3090 Founders Edition was pushing a hash rate up to 82 MH/s mining Ethereum. The performance is a bit underwhelming since the GeForce RTX 3090 can easily reach 100 MH/s, and aftermarket models with better GDDR6X cooling can put hash rates up to 125 MH/s. Even though the YouTuber dropped the memory speed to 18Gbps and cranked the fan speed up to 88%, his GeForce RTX 3090 Founders Edition's memory was still hitting 110 degrees Celsius.

The YouTuber replaced the factory thermal pads with Thermalright's Odyssey Thermal Pad 85x45x1.5mm. Admittedly, the thermal pads aren't the best aftermarket offering that money can buy, as their thermal conductivity rating is only 12.8 W/mk. However, they appear to have done wonders for the memory chips inside the GeForce RTX 3090 Founders Edition.

After replacing the thermal pads, the YouTuber was able to restore the memory speed to 10,577 MHz (21.15 Gbps) and lowered the fan speed to 70% to pump out 100 MH/s. The graphics card's memory was dancing around the 84C–86C range during an entire day of cryptocurrency mining.

The Thermalright Odyssey Thermal Pad 85x45x1.5mm retails for $14.99 a piece on Amazon. Although the YouTuber bought four of them, he only needed three to completely substitute all the thermal pads on the GeForce RTX 3090 Founders Edition. It's a pretty good investment no matter which way you look at it. For $30, one could shave off as much as 25C off the memory's operating temperatures.

The memory's thermals shouldn't be as big of a concern if you're not into cryptocurrency mining. We can't generalize, but we expect the majority of custom GeForce RTX 3080 and RTX 3090 graphics cards on the market to come with better memory cooling solutions than Nvidia's wacky Founders Edition design. If you suspect that GeForce RTX 3080 or RTX 3090 is suffering from thermal throttling, the latest version of HWiNFO64 now shows the temperature for the GDDR6X memory chips.

Spotted by FanlessTech, and posted on Chiphellforums, it looks like Thermalright is working on a new CPU cooler called the Frost Spirit 40. The unreleased offering seemingly follows the same classic design of many dual-tower coolers today.

The fin stack of the unit is symmetrical, and it comes with clearance for memory. At the front of the unit, a 120mm fan pushes air into it, and the two fin towers sandwich a 140mm spinner to move air through the cooler. The heat is pushed into the fin stacks through four very thick 8mm heatpipes, which certainly helps the unit look hefty –- which it is, weighing 2.2 pounds (1kg). 

The base of the unit is made from copper for better thermal transfer, but nickel-plated to prevent corrosion. Naturally, it also has a mirror finish. We also have reason to believe that there will be RGB support. 

From the looks of it, the unit will support modern Intel LGA115X and AMD AM4 CPU sockets.

For now, that’s all we know. We’ll probably learn more in due time when Thermalright makes a formal announcement. Knowing Thermalright, this product should come at a very competitive price point.

With the high price of copper, seeing full-copper CPU coolers isn’t a common sight anymore. According to TechPowerUp, Thermalright has decided it’s time again for one and introduced the AXP-90. It’s not very big , but it’s full-copper, except for  some screws and brackets.

The AXP-90 is a simple low-profile cooler that comes with a copper cold plate that leads to four copper heatpipes, which lead into a copper fin stack with 57 fins, and a neat orange fan with a burgundy/maroon fan casing. The unit weighs in at 490 grams in total and is no taller than 47mm, so unless you’re using it in a very cramped system, clearance shouldn’t be an issue.

The 15 mm thick fan can push up to 42.58 CFM of air when spinning at its maximum of 2700 RPM, but with PWM control can spin at much lower speeds.

If you’re interested, you’ll have to watch out for which version you purchase as it comes in a variant for the LGA115X sockets and another one for AM4. Pricing in China is set at 299 Chinese Yuan, or about $43 USD. We’re curious if this little unit will make it over to the states.

Thermalright, a long time air cooling specialist, has finally bitten the bullet and ventured into the AIO (all-in-one) liquid cooling realm. The company recently announced its Turbo Right 240 C and Turbo Right 360 C liquid coolers.

The new Turbo Right liquid cooling systems employ Thermalright's own Turbo Right CPU copper waterblock that employs 0.1mm microchannels with a large contact area for optimum liquid flow. The Turbo Right pump has been placed on the outside of the waterblock which, according to Thermalright, helps reduce vibrations and stress on the processor and motherboard. The waterblock has a transparent window where you can gaze at the rotating blade, liquid flow, coolant level and incorporated RGB lighting. 

The Turbo Right 240 C and Turbo Right 360 C utilize 240mm and 360mm radiators, respectively. The radiators exhibit a high-density fin design that should help with better heat dissipation. Thermalright's TY-121BP PWM fans provide the radiators with active cooling. The 120mm fan spins between 600 and 1,800 RPM (revolutions per minute) with an airflow between 25.76 and 77.28 CFM (cubic feet per minute). It's rated with an air pressure of 0.30 to 2.27 mmH₂O (millimeter of water) and a noise level ranging from 19 to 25 dBA depending on the speed that the fan is spinning at.

Similar to other AIO liquid coolers on the market, the Thermalright Turbo Right 240 C and Turbo Right 360 C come pre-filled with coolant and require little maintenance. However, they do have a fill port in case you want to flush out the coolant or add more to compensate for coolant loss. As a matter of fact, Thermalright even includes a 100ml bottle of Turbo Right coolant.

Support on AMD platforms is limited to the AM4 socket. On the other hand, the Turbo Right liquid coolers are compatible with a wide range of Intel sockets including LGA 775, LGA 115X, LGA 1366, LGA 2011, LGA 2011-V3 and LGA 2066.

Thermalright hasn't released the pricing or availability for the Turbo Right 240 C and Turbo Right 360 C.

This week, Thermalright announced a new revision of the Thermalright Silver Arrow IB-E Extreme cooler, tacking on a "Rev. B" at the end and creating the Thermalright Silver Arrow IB-E Extreme Rev. B. The claim to fame for this heatsink, outside of its massive size and 320W capacity, is the offset heat pipe and fin array also around the topmost slots, giving users more freedom to choose the memory they want without sacrificing cooling performance. 

First, its size. The IB-E Extreme Rev. B measures 155x130x163mm (LxWxH). The CPU cooler itself features a mirror finish, nickel-plated copper base along with eight 6mm heatpipes, which meander through both of the heatsink towers. Not only is the heatsink large, but it is also has a lot of mass, weighing in at 850g without fans. The offset design should allow for taller memory to fit underneath, as well as giving a bit more room for the topmost expansion slot on the motherboard. 

The heatsink includes two TY-143 fans, which run from 600~2500 RPM ±10 percent. The PWM-powered dual ballbearing fans can push 31.4~130 CFM with a static pressure of 0.32~3.0mm H20 with noise levels said to be in the 21~45 dBA range. Using higher static pressure fans like these is a good match considering the fin gap is pretty tight at 2mm (with 0.4mm thick fins). 

The Thermalright Silver Arrow IB-E Rev. B heatsink will support Intel sockets 775, 115X, 1366, 2011, 2011-v3 and 2066. For AMD, only socket AM4 is listed, but it's possible AM3+ on down is supported as well, since its predecessor supported them.

Neither pricing nor availability was listed at the time of writing, but it is expected to be in the ballpark of the outgoing Silver Arrow IB-E, which is currently $99.95 on Amazon. 

Introducing The True Spirit 140 Direct

Specifications

A premiere European manufacturer of air-based CPU coolers, Thermalright initially launched its value-based solutions under the Cogage brand. The True Spirit model line is what remains of those early attempts to crack the mainstream market, as Thermalright eventually put its own name on the original Cogage model. Years of price-conscious performance optimizations have led to it its latest iteration, the True Spirit Direct 140.

The package includes a heat sink solution with 1.5”-thick (38mm), 140mm radiator, an oversized 152x140mm fan that’s 27mm thick, two sets of wire clips that enable installation of a second 140mm (or 152x140mm) fan, and an installation kit that goes from current CPUs all the way back to AMD socket AM2 and Intel LGA-775.

Compared to early models, the half-based True Spirit 140 Direct uses direct contact, rather than encapsulated heat pipes. While some of Thermalright’s competitors have claimed performance superiority of this design for many years, our comparisons of similar coolers have proven those claims baseless. I’ve postulated that direct-contact pipes were a way to cut manufacturing cost without hurting performance, and Thermalright is pushing the second half of that hypothesis by not even flattening the pipes enough to fill most of the void between them.

Rather than attempt to make its True Spirit 140 Direct base completely flat, Thermalright uses extreme contact pressure to get maximum thermal transfer from a reduced surface area. The installation kit therefore includes a heavy duty back brace designed to push against both the holes and the socket plate of LGA 115x/1366, a center plug to make it work on boards that have no support plate such as LGA-775 and AMD (using the four holes of the original clip bracket), and a separate set of standoffs for LGA-2011x (aka, v1 and v3). The small plastic washers provide additional clearance for components on the back of a motherboard, which is a fairly common concern on mini-ITX motherboards, and a separate (larger) set of plastic washers are used on AMD motherboards.

Thermalright addresses the “extreme contact pressure” concern of Kaby Lake CPU owners with a shim that allows the CPU clip to press against the entire land grid, preventing it from bending back away from contacts. The large volume of damaged processors we’ve seen could make this shim a viable standalone product! And if the rumors of “resurrected” processors prove true, a shim like this could be used to reinforce the weakened LGA package.

Special standoffs screw directly onto the integrated support bracket of our LGA 2011-v3 motherboard. Users of other sockets will find a different set of standoffs, with plastic insulating washers stuck on one side, to secure the support plate shown two images above. The standoffs are topped with the chrome ring-shaped bracket, which is secured to the top of the standoffs with another four screws. After applying thermal paste to the point that it fills the gaps between heat pipes, the CPU cooler’s integrated bracket screws to the ring-shaped bracket, as shown.

Small grooves in the heat sink’s fins are designed to hold included rubber tubes, which reduces the transmission of fan vibration.

Clipping the fan to the heat sink and plugging it into the board finalizes the True Spirit 140 installation. It’s important to note that the clip wires should be inserted prior to heat sink installation, as outlined in the manual, to assure motherboard component clearance.

MORE: Best CPU Cooling

MORE: How To Choose A CPU Cooler

Test Settings, Results, And Final Analysis

We continue using our 2015 Reference PC minus its open test bed (and obviously the reference cooler) to test the H220-X in a closed system. The Core i7-5930K is overclocked to a fixed 4.2GHz core frequency at 1.20V core voltage.

Test System Components

Comparison Coolers

Setting the performance baseline are the two largest slim coolers of previous reviews, along with a budget spoiler from DeepCool.

Results

The True Spirit 140 Direct’s cooling at full fans appears mediocre, but spinning the fan down has minimal impact. That’s certain to make things interesting in our cooling-to-noise comparison.

The True Spirit 140 Direct has the lowest fan speed, which again should make the cooling-to-noise ratio interesting.

Paydirt comes in the noise test, where the True Spirit 140 Direct is whisper quiet. Those who noticed that I didn’t mention the ultra-cheap Gammaxx 400 in the cooling test will wish I hadn’t in the noise test, since decibels use a logarithmic scale. Note that while our results were 3db higher than Thermalright’s rated values, having the cooler mounted can increase the amount of noise reflected back toward the meter.

Our acoustic efficiency (cooling-to-noise ratio) chart uses a direct scale, and adding a logarithmic function for the noise portion would have made the worst coolers look even worse. The True Spirit Direct is the clear leader.

A superb cooling-to-noise ratio gives Thermalright’s True Spirit 140 Direct the edge it needs to combat DeepCool’s low price, at least at full fan speed.

Thermalright’s True Spirit 140 Direct produced excellent noise and efficiency marks, but mediocre cooling performance (by big-air standards) might be its biggest accomplishment. After all, the heat pipes of this particular design were only flattened by roughly 25% of their diameter, making contact with only around half of the CPU heat spreader’s surface area. While I can’t recommend it to everyone, its extremely favorable cooling-to-noise ratio combines with an excellent price to put this on our approved list.

MORE: Best CPU Cooling

MORE: How To Choose A CPU Cooler

MORE: All Cooling Content

MORE: In Pictures: 20 Clever Liquid-Cooled PC Setups

Even A Macho Changes Over Time

For the past three years, Thermalright's HR-02 Macho has been a stalwart in the competitive CPU cooler segment. We've seen various Macho models offer good price-to-performance ratios, and as fan speeds are reduced, they show their real strength. In a recent review on Tom's Hardware Germany, we documented this by testing the black nickel-plated HR-02 Macho Zero, which doesn’t even ship with a fan.

Over the years, most of the changes to the HR-02 Macho family were minor adjustments. Sometimes the bundled fan was replaced by something newer. Occasionally, the heat sink's looks were updated. Moreover, Thermalright created smaller and larger derivatives of its box-shaped cooler.

The Macho 120 (we called it the Mini Macho) offers reduced height, helping it fit into slim, budget-priced enclosures. The huge HR-22 was intended to break into the semi-passive segment. Surprisingly, though, the standard-sized Macho Zero wound up beating it.

More recently, Thermalright introduced the HR-02 Macho Rev. B, the latest iteration of the standard model with a 14cm fan. Compared to its predecessor (the HR-02 Macho Rev. A), the updated model not only comes with a few design improvements like nickel-plated heat pipes and a black paint job for the topmost cooling fin, but it also sports two technical innovations.

Until recently, Thermalright used its own TY-147 fan, which is replaced by the TV-147A. The new cooler complements the good performance of the HR-02 Macho at low fan speeds by offering a much wider RPM range (from 300 to 1300RPM) than its predecessor, which had a 900RPM floor.

The second technical innovation is the CPU contact plate. Thermalright's HR-02 Mach Rev. B uses a significantly larger base. In fact, it's the same one that adorns the Macho Zero. This is intended to improve cooling performance for the latest LGA 2011-v3 processor generation.

Inspired by the introduction of Thermalright's refresh, we arranged a family meeting of the various Macho models. Mounted on top of an Intel Core i7-5820K, they can demonstrate their individual strengths and weaknesses.

Test Platform And Overview

Our (Temporary) Test System

Since one of the test’s objectives is the assessment of Thermalright's new base plate, expected to improve cooling performance on Haswell-E-class processors, we're using an X99- and LGA 2011-v3-based system.

Our subject is a six-core Intel Core i7-5820K equipped with Hyper-Threading technology. Selling for around $390, it constitutes an affordable entry point into the company's highest-end platform. We expect many readers to choose this processor for price and performance reasons.

Sporting a base clock of 3.3GHz and a maximum Turbo Boost frequency of 3.6GHz, this CPU is actually slightly slower than Intel’s four-core LGA 1150 flagship, the Core i7-4790K. But for many applications outside gaming, the two extra cores improve performance markedly.

We picked Gigabyte's X99-Gaming 5 for our supporting board; it's able to handle most of the overclocking scenarios we'll throw at the processor, and it offers one distinct advantage over larger motherboards: the cooler’s heat pipe doesn’t extend over the rear RAM slot. That's a good thing, since a large heat sink's pipes dissipating thermal energy over your memory can be counterproductive.

The operating system and the test programs reside on a fast, inexpensive Transcend SSD370. Its 128GB capacity is almost overkill for this test.

Because this machine is partly the author’s personal PC, picking RAM brand was easy. We went with 32GB of Crucial DDR4-2133. While these four 8GB DIMMs are budget-priced, they're a solid choice since they can typically be overclocked to DDR4-2400, even at 1.2V. At $100 per DIMM, the Crucial RAM is an excellent choice from a price/performance perspective.

The real surprise may be our PSU choice. We went with a 400W be quiet! Straight Power E8. Since this is not a graphics-heavy story, a 400W PSU is more than sufficient, and this specific model's fan is nice and quiet.

The graphics card we used, an XFX Radeon HD 5450, won’t inspire envy, either. But at least its passive cooler is completely silent, allowing us to more accurately judge the CPU fan.

We conducted all measurements with the system mounted on a Cooler Master Test Bench.

Last but not least, we used an Aqua Computer Aquaero 6 controller to set and measure the fan speed. While most motherboards offer some kind of fan control built-in, they're not as fine-grained as the Aquaero 6. Nor are their read-outs as precise.

Overview of the Test Candidates

In order to make this Macho family comparison more exciting, we invited lots of derivatives to the table, in addition to Thermalright's brand new HR-02 Macho Rev. B. One of them is the HR-02 Macho Rev. A (BW). It offers less eye candy, a different fan and an older version of the base plate.

We could have omitted the Macho Zero, since it's basically the same as the HR-02 Macho Rev. B with black nickel plating and without the latter’s fan. However, we're including it for the sake of completeness.

The biggest and baddest participant is the huge HR-22. A few days before we started testing, we received a TY-121 fan, which now ships with the small Macho 120 Rev. A. Hence, we put the latest fan on our older Macho 120 and included it in our test.

Like the Macho Zero, the HR-22 doesn’t ship with a fan. It's intended for (semi-)passive operation, after all. But in order to level the playing field and because fan mounting components are included with all of these coolers, we put a TY-147A fan on both of them.

As mentioned, we also included the Macho 120 (our Mini Macho). We had an older version in our lab and upgraded it with the fan of the current Macho 120, for fairness' sake.

Test Procedure

It takes 70 minutes to measure the CPU temperature. After a soak time of 60 minutes, the remaining 10 minutes are logged. We plot the average of the six core temperatures. Note that we report the difference to the ambient temperature, subtracting the room temperature from that average.

If you want to know what CPU temperature to expect in summer conditions, add your room temperature to our reported temperature value. Of course, this is only an approximation, since cooling performance degrades as ambient temperature increases.

We put the ambient temperature sensor 8“ in front of the CPU fan’s intake, within the top third of the fan. We took care to position the PSU in such a way that its exhaust air is not fed back into the processor fan. In fact, we reversed the PSU and put it at the bottom of the test bench. Thus, the ambient temperature sensor measures actual room temperature.

Comparison Test: Cooler And Fan

Cooler Comparison

While the changes made to the latest Macho cooler aren't earth-shattering, revision B of the HR-02 Macho does sport a few modifications.

The overall construction of the HR-02 Macho Rev. A (BW) and Rev. B is identical. A huge cooling block and a relatively large fin-to-fin distance of 3mm (0.012“) help these coolers perform well in a semi-passive application or at drastically reduced fan speeds.

Thermaltake punches U-shaped holes into the cooling fins and bends up the tiny tabs, which is supposed to break up the laminar air flow and thus enhance performance.

The most recent version of the HR-02 Macho comes with an updated design; its black top and nickel-plated heat pipes reflect the most recent design trends.

All told, Thermalright offers three different cooler sizes based on the same basic design. These range from the 6"-high Macho 120 with five 6mm heat pipes to the 6½"-high Macho HR-02 to the wider, deeper, but at 6.6"-high, only slightly taller HR-22, which sports a whopping eight 6mm heat pipes.

In spite of their large dimensions, the models are all compatible with front DIMM slots; the Macho family members stay clear of your system memory. This is achieved through the coolers' asymmetrical design. Half of the heat pipes (three on the two smaller models and four on the HR-22) are bent backwards, allowing the cooling tower to be moved back. This configuration is mostly geared towards the mainstream interfaces from Intel and AMD, where the only DIMM slots are in front.

While the coolers overlap some of the DIMM slots at the rear, they're also high enough to accommodate the installation of RAM with heat spreaders. For the Macho HR-02, DIMMs can be up to 2" tall. Incompatibilities should arise rarely, if ever.

However, the huge HR-22 overlaps all rear DIMM slots. And the first slot is right underneath the heat pipes, which doesn’t give you room for tall heat spreaders. If you own modules with big spreaders, make sure they don't touch the heat pipes. Otherwise, the RAM will be heated up beyond its spec.

The smaller Macho 120 Rev. A offers less cooling fin real estate, mostly due to the fact that it is lower, and sports only five heat pipes. This makes it an ideal choice for cases that are too slim for the regular-size HR-02. We used the first-generation Macho 120 in this test, the one without the Rev. A suffix. The only difference between them are the nickel-plated heat pipes and the black paint job on top. For a fair comparison, we mounted the Rev. A’s fan on the old Macho 120 cooler.

The real innovation of the HR-02 Macho Rev. B compared to its predecessor is the larger base plate. Since LGA 2011 processors have a large heat spreader, you want the base plate to approach or even match the heat spreader’s size. The new base was taken from the Macho Zero. In fact, the Macho HR-02 and the Macho Zero are all but identical if you disregard the black nickel plating of the former and the enclosed white gloves.

While the Thermalright HR-02 Macho Rev. A is available for $53, expect the Rev. B to cost a few dollars more. But the optimized fan, better looks and larger base plate are certainly worth a small premium.

Fan Comparison

Because the HR-22 is shipped without a fan, we also equipped it with the new TY-147A, the TY-147’s successor. Armed with identical fans, comparing the HR-22 and the HR-02 should be fair and accurate.

The new fan and its predecessor look the same, and even the TY-147 was famed for its quiet operation. While the TY-147A has the same frame, the same rotor blades and even sports the same maximum fan speed (1300 RPM), its minimum is now a mere 300 RPM, down from 900. If you want to do some testing of your own, keep in mind that these numbers vary in real life. Plus/minus 10% is not uncommon.

Our TY-147 sample operated down to a tad above 500 RPM, which is well below the minimum speed in its spec sheet. We expected the TY-147A to dip down to 330 RPM, allowing for 10% variation. We'll cover its actual floor in more detail shortly.

We also looked at the Macho 120 Rev A’s fan, the TY-121. This 12cm fan sports the same design of its brethren, but it is smaller and employs seven rotor blades. Its speed range is quoted as 600 to 1300 RPM.

Results: Cooling Performance, RPM Curve And Sound Level

Cooling Performance

Not surprisingly, the HR-02 Rev. A and B are similar. By precisely controlling the test environment, we were able to confirm that the newer revision does enjoy a minor performance advantage, as a slightly more expensive version should.

In spite of utilizing eight instead of six heat pipes, the HR-22 doesn’t perform much better. In semi-passive mode, the HR-22 "merely" performs on a par with the HR-02 in active mode, but it limits your RAM choices and even fares slightly worse than the HR-02 at reduced fan speeds.

However, we were positively surprised by the little Macho 120 Rev. A. While it has only five heat pipes, is shorter and less wide than the HR-02, its cooling performance is only slightly lower. That's an admirable feat for a 6“ cooler.

RPM Curves

The latest Macho revision sports the TY-147A fan, which can operate at a low 300 RPM. The difference between our TY-147 and TC-147A samples wasn’t that significant at all. Perhaps that's because our TY-147 worked down to 497 RPM (a whopping 400 RPM less than its spec sheet promises). Don’t count on this, though. Deviations of 10% are expected; any more than that and you’re just lucky (or unlucky).

Thus, we expected the TY-147A to spin down to 330 RPM. And it did. In fact, it continued on down to 317 RPM.

Note that some motherboard fan controls cannot go below a certain minimum speed. Thus, if you cannot get the TY-147A to run at 300 or 330 RPM, your platform may be to blame.

When we determined the RPM curves, we started at 10 percent and cranked up the power to 100 percent. If you go the other way, the curves will look slightly different. Thus, even measurements determined by the Aquaero 6 fan controller may contain some inaccuracies. But they should be good enough for reference.

When you look at the fan speed curves, you'll notice that both fan generations were optimized for quiet operation. The PWM signal needs to go beyond 40% to 50% for the fan to gradually gain speed.

When your primary goal is low noise, the brand-new TY-147A behaves even better than its predecessor. While the TY-147 starts at a higher speed and goes beyond 1000 RPM at a duty cycle of approximately 80%, the TY-147A stays quiet much longer and only exceeds 1000 RPM slightly above 90% duty cycle, and then rapidly gains another 300 RPM.

The TY-147A’s top speed is higher than the TY-147’s, which means that the TY-147A stays quiet in most situations and only revs up when high temperatures really call for more fan airflow.

Sound Level Measurements

As we try to measure minor differences between quiet fans, we modified the sound level test procedure slightly. This time, we measure at a mere 8“ distance straight in front of the fan, at the height of the fan bearing. If you think the reported sound level is alarmingly high, remember that these values cannot be compared to results gathered from our regular test method, which sports a closed case, not an open test platform. A close-range, open-air measurement like this one is greatly affected to the exact position and angle of the sound level meter. Change the position slightly and massive deviations can result. Thus, take these test results with a grain of salt. Our primary motivation for this unusual test setup was the comparison of both TY-147 generations.

Both fans are pleasant to listen to. We had to hold our ears to the fan to detect a slight bearing noise. This is as good as it gets, and the TY-147A sounds slightly better than its predecessor. But a difference as small as this one may even arise from manufacturing variations.

We record the sound level at each fan's maximum speed, at 1000 RPM and at 800 RPM. Any lower than that and a fan inside a case is not audible except, perhaps, in a sound-proof room. Whatever we tried, the TY-147 fans proved to be equally loud. The faster-turning model is ever-so-slightly louder, but this can be attributed to typical manufacturing variations.

As we crank up the RPM duty cycle towards 90%, the TY-147A turns slower and is thus also quieter. Once you go beyond that, both fans get louder. However, at 800 and 1000 RPM both fans exhibit identical sound levels.

Thermalright's Macho Evolves

The brand-new Thermalright HR-02 Macho Rev. B improves on its predecessor by offering a slightly better fan, a few design enhancements and an enlarged base plate that's aimed at LGA 2011-based systems. Since the price tag is only up a few dollars, the question of whether the new revision is worth a premium seems moot. But let’s discuss this anyway. Is revision B worth a surcharge?

Consider the most important fact: the HR-02 Macho Rev. B does offer slightly better cooling performance than its predecessor. Normally, we’d dismiss a temperature advantage of approximately one Kelvin as insignificant, caused, perhaps, by measurement inaccuracies. But for this comparison of two almost identical revisions of the same cooler model, we pulled out all the stops and ensured a perfectly stabilized test environment, accounting for ambient temperature and even humidity. We also removed and re-attached each cooler on the Intel Core i7-5820K repeatedly, thus minimizing the influences of thermal paste thickness variations. While one Kelvin is not an earth-shattering difference, we are certain that the new cooler revision performs better than the old one.

The larger base plate of the HR-02 Macho Rev. B confers another advantage: it is easier to mount the cooler on the CPU, since its tendency to slide forward or backward is reduced. This makes it easier to attach to the mounting frame.

Speaking of the mounting frames, the Rev. B model adopted the Macho Zero's slightly slimmer frame, improving compatibility with some high-end AMD motherboards. While the smaller Macho 120 holds its own compared to the larger models, we were irritated by the fact that Thermalright's significantly larger HR-22 does not achieve the thermal performance of its mainstream brother, the HR-02, even as it costs a whopping $88!

We like the HR-02 Rev. B's design. Nickel-plated heat pipes and a black top plate are pure eye candy. Since the competition is also working on aesthetics, the HR-02 Rev. A is looking dated.

Last but not least, the new fan offers a small improvement, too. While the visual and mechanical design of the previous TY-147 and new TY-147A appear identical, the more modern model spins at a lower minimum speed (300 RPM) and serves up a more attractive RPM curve.

The massive HR-02 Macho is ideally suited for slow-turning fans, and while the TY-147A is PWM-controlled, it intentionally postpones revving up. This minimizes noise, especially when the CPU fan is controlled by the motherboard, since integrated fan controllers are known for aggressively revving up the fan speed.

Verdict

All in all, we're dealing with several slight improvements. Taken together, they are not insignificant, making the Thermalright HR-02 Macho Rev. B a better choice than what came before. Indeed, the HR-02 Macho Rev. B stands out from the crowd. Its combination of an ultra-quiet fan, excellent cooling performance with low air flow and the option of semi-passive operation is hard to find.

Once again, Thermalright raises the bar with its HR-02 Macho Rev. B and the previously-introduced Macho Zero. It is up to you to decide which of the two models to buy. The Macho Zero doesn’t ship with a fan, but sports an all-black look and comes with a fan duct, whereas the HR-02 Rev. B comes with the ultra-quiet fan we just finished characterizing.

Be that as it may, both variants deserve our recognition. These two Thermalright heat sinks remain the benchmark for competing brands in the CPU cooler marketplace.

As CPUs are becoming more efficient, you may wonder why cooling manufacturers still insist on making large air-cooling towers. Thermalright's latest number, the updated HR-02 "Macho Zero," seems to raise that question at first sight, but there's more to it than meets the eye. Well, when we say more, we actually mean less, as it doesn't ship with any fans. A big tower is useful, then, as it allows you to cool today's high-performance CPU without the immediate need for a fan.

The Macho Zero is a standard tower cooler with six copper U-shaped 6 mm-thick heatpipes feeding heat to the aluminum fin array. Most of the unit is plated with black nickel, which mostly makes the unit look better than if left untreated, but also offers a little bit of protection from corrosion. The top fin of the stack is anodized black, and the entire construction is 162 mm tall.

Included with the Macho Zero are both Intel and AMD mounting mechanisms, a pair of white cotton gloves for clean installation, two grams of Chill Factor 3 TIM (Thermal Interface Material), two black anodized fan brackets, and a Thermalright screwdriver with a magnetic tip.

You can approach mounting the Macho Zero passively, semi-passively, and actively. Passive mode simply means that you don't mount a fan to it and let your case airflow take care of heat dissipation, while active mode means that you mount fans to it using the included fan brackets and a fan you pick yourself.

"Semi-passive" is the interesting one, as it uses a fan duct to use the rear-exhaust fan of your case to draw air through the Macho Zero. Unfortunately, the Macho Zero with fan ducts will only be available with the unit from MindFactory.de, in Germany, where it costs €54.99 ($60 USD).

There's no word on U.S. pricing or availability yet.

Follow Niels Broekhuijsen @NBroekhuijsen. Follow us @tomshardware, on Facebook and on Google+.

Look Mom, No Optical Drive Bays

We've watched the usefulness of optical drives taper off over the years due to two important trends. First, hard drives are larger, offering massive capacities at lower cost. That means you can cram a lot more of your digital information onto disk. Second, we're consuming content differently. Movies are streamed, games are downloaded, and files are backed up to those bigger, cheaper hard drives, either locally or in the cloud.

Much of the data that was transferred from CD or DVD to your hard drive now goes from the Internet to local storage. For those few cases when an optical drive is needed, one external device can serve all of your systems sufficiently.

PC case manufacturers are starting to react, bidding 5.25-inch optical drive bays farewell (even though those aren't the only peripherals this decision affects). The first representative example in our labs is NZXT's mid-tower H440, which closes the gap between the company's H230 and H630 models. You can buy the H440 in either a white finish with black accents or a black finish with red accents. It also features noise dampening, a large window, and ample fans, straight from the factory.

In the place of those scrapped 5.25-inch bays, NZXT includes more interesting features in this chassis' design. For instance, hard drives are installed differently, there are power connectors for up to 10 fans, and the interior is compartmentalized. If you like your systems clean and tidy inside, the H440 is probably going to be right up your alley. Fittingly, the case’s slogan, straight from the manufacturer’s website, is "Remarkably Clean."

Although it sold out fairly quickly after launch, the H440 is now available again online for $120 in black and $140 in white.

NZXT’s H440 features an elegant design with lighting that looks good and serves a practical purpose. Enthusiasts who can live without 5.25-inch drive bays and are in the market for a quiet (albeit not silent) enclosure with lots of features will find NZXT's H440 to be both unique and elegant.

Packaging and In The Box

When we received the H440 from NZXT, it hadn't yet been made public. Consequently, our sample didn't arrive in retail packaging, rendering us unable to unbox the case and introduce its bundle piece by piece.

NZXT's manual hadn’t been printed yet, so we only got a pile of screws, an above-average number of cable ties, a bit to help with motherboard spacer installation, and an NZXT sticker made from stiff metal foil.

Technical Specifications and Front Panel

Up top, the NZXT H440 features two USB 3.0 and two USB 2.0 ports, as well as the requisite HD Audio connectors, all of which are clearly marked with fittingly modern labels. The power and reset buttons are located on the far-left side of the case's top.

Controls for the optional case lighting are moved to the top of the enclosure's back in order to maintain the smooth, uninterrupted appearance of the most visible panels.

First Impressions And 360° Picture Gallery

The H440 employs a simple, yet modern, design with only a few edges breaking up the plain surfaces. These accents are finished in black for the white case and in red for the black case.

There's a side window cut in such a way as to hide the drive bays, but also to facilitate a good view of the interior's most interesting components. Consequently, a large CPU cooler can deliver its full visual impact. And even though that front might look like a door, it's actually just a cover. Oh, the things you can do when you tear the 5.25-inch peripherals out of an enclosure.

A first look through the side window makes it obvious that the power supply is housed in its own compartment on the floor of the case, and not viewable from outside. Not only does this guarantee a certain uniformity between H440-based systems, but it also helps you hide the unused leads from partly-modular or non-modular PSUs. You can even use that space to hide other cables.

As far as quality goes, we're left with a good first impression. NZXT's finish is clean and without identifiable flaws. The front panel and top cover don't have any give to them. Moreover, the case doesn’t rock or slide.

And here's a handy little design detail: the thumb screws responsible for holding the side panels and hard drive bays in place stay attached to their mounts once they're unscrewed, preventing them from getting lost.

Discreet Lighting That Also Increases Usability

Ah ha; you thought you had us on misusing discreet, didn't you? In actuality, all of the H440's lighting elements are in the back of the case. They consist of the manufacturer’s name on the PSU compartment, as well as a number of LEDs in the area around the rear case fan and expansion slots. Subtle, yet classy, the lighting reinforces the H440's high-quality design.

In the picture below, it looks like the LEDs are illuminating the inside-rear of the case. Really though, they're well-concealed and attached to the outside of the case. This makes for a nice visual effect, plus has bonus utility when you're looking for a port or plug back there and need a little extra light.

Air Cooling, Dust Protection, And Noise Dampening

Air Cooling and Dust Protection

Those black accent panels on the front and top covers aren’t just there to look good; they also provide a way for air to enter (front cover) and exit (back cover). This means that there are only two places vulnerable to dust entering the case: the front cover with its three 12-cm fans and the PSU area. Both are protected by dust filters, though.

The filter covering the power supply air intake opening sits securely in place and can be pulled out toward the back.

Getting to the front filter is a bit more involved. If you need to clean it, you'll first have to remove the front cover. The filter is held in place by magnets and a retention rail, which keeps it from protruding and ruining the H440’s smooth looks. We can’t help but think that this could have been achieved with a simpler solution, though.

NZXT's H440 has space for three 12-cm case fans in the front, and comes with all three already installed from the factory. The fans are NZXT’s own FN V2s, spinning at up to 1200 RPM. They can optionally be replaced by two 14-cm models.

The lack of 5.25-inch drive bays give NZXT a big advantage: hard disks can be positioned farther apart than they usually are. In addition, the top fan is at the level of the CPU cooler, providing a much more direct route for air to reach its destination than you typically find in other PC cases.

All of this results in much more direct airflow from the front of the case to installed components. In addition, bays that aren’t needed can be removed, which effectively eliminates all obstruction.

The top of NZXT's H440 features the same fan options as the front, though none of the fans that fit are included.

The H440 comes with another NZXT FN V2 fan installed in the rear. This case fan is of the 14-cm variety, though, and runs at 1000 RPM.

The bundled FN V2s have a frame with a unique shape that reflects the basic design of the NZXT H440. They’re not decoupled with rubber pads though, which is different from the retail version.

A small board on the back of the motherboard tray provides a total of 10 three-pin fan power connectors, enabling plenty of connectivity for coolers you add in yourself. Of course, when you add or replace case fans, bear in mind that a maximum of 30 W is supported.

Sound Dampening

The sound dampening material is approximately 6 mm thick and covers all surfaces except for the case's side window, which is to say that it's on the front cover, top, right side panel, and those areas on the left not cut out for the window. An argument can be made that combining sound dampening and a side window doesn't take acoustic isolation seriously enough. However, for many enthusiasts, the H440 combines elements commonly requested together, and that includes sound dampening material and a window.

Water Cooling

The H440 can accommodate a small 12- or 14-cm radiator in place of the rear case fan, as well as two larger 28- and 36-cm radiators at the top and front of the case, respectively.

At the top, wider 28-cm radiators can be a tight fit, as they might collide with taller motherboard components. Fans can only be installed on the bottom of the radiator.

Since the H440 doesn’t have 5.25-inch drive bays, the entire front can be used for radiators up to 36 cm-long. The hard drive bays can be taken out completely by unscrewing them. Or, you can pull them out separately, giving you a choice in how much space to clear in the enclosure's front area.

After removing the hard drive bays, there’s enough room for radiators with fans on both sides. Naturally, though, you won't have room for 5.25-inch reservoirs, since there aren't any bays to drop them into.

Installation

Power Supply Installation and Cable Routing

Power supply placement is one of the H440’s unique features. The PSU is installed on the case's floor, in its own compartment, hidden from view. This design decision gives the case two practical advantages. First, the entire bottom area can be used to hide unused cables, which makes for a tidy-looking system. Second, you can use some of the empty space above the PSU to hold two additional 2.5-inch drive bays.

The PSU isn’t decoupled from the case, which isn’t a big deal. Most modern power supplies don't require this anymore.

NZXT does make PSU installation easier, however, by implementing a frame around the power supply cavity. It's held in place by four thumb screws and, once you pull it off, the power supply slides right into the chassis.

Screw the frame back into place, then secure the PSU onto the frame. Everything's accessible externally, giving you plenty of room to work.

Tactful-looking cable routing is easy, too. Not only does NZXT include the usual rubber-framed cable openings, but also a number of extra smaller cut-outs in convenient places. Openings behind the 2.5-inch drive bays make it easier to hide the USB and HD Audio cables.

Two additional openings on the bottom of the motherboard tray are available if you use a microATX motherboard. Several openings toward the top allow the power cable and fan leads to be routed out of view, too.

Space for the PSU cables can be a bit tight due to the sound dampening material. NZXT does include a number of built-in cable clips to keep them neatly in place, though. In the area next to the rubber-framed cable openings, the motherboard tray is bent inwards to create some extra room. All of the cables that aren’t needed can be tucked away in the void next to the PSU.

Rare is it that we're able to create a clean and tidy interior so quickly. NZXT's H440 makes a good-looking build possible for any level of enthusiast, though.

Hard Drive Installation

Most of the hard drives are installed on their own sledges. They’re accessed from the right side of the case, which is somewhat unique. The idea is that you want an area on the side of the case, next to the window, hidden by the side panel. Each sledge is screwed to the case separately. We had to tighten all their screws a little, but once that was done, they sat securely in place.

All five hard drive bays have small rubber grommets that are held in place by glue, but can be moved to accommodate 2.5-inch drives, for example. The advantage of this approach (at least compared to the drive cages we're accustomed to) is that each bay can be removed separately to create space for longer peripherals or cooling solutions.

Even if you yank out every single hard drive bay to make room for a radiator or completely unobstructed airflow, there’s still the option of installing a 3.5- or 2.5-inch drive on the bottom of the case.

Additionally, two 2.5-inch drives can be installed in part of the empty space hidden from outside view, next to the PSU.

Motherboard, Graphics Card, and CPU Cooler

The motherboard spacers came installed from the factory, saving us the trouble of screwing them in ourselves. The H440 accommodates CPU coolers up to 18 cm-tall, which is enough for any model in our test lab.

There's a big opening in the motherboard tray that makes switching out CPU coolers with backplates a breeze. You won't need to pull your platform out for a cooler swap.

As it ships, NZXT's H440 accommodates graphics cards of up to 30 cm-long. This number increases to 40 cm once you pull out the appropriate hard drive bay.

Benchmark System And Conditions

Our benchmark system features an Intel Core i5-4670K and Radeon HD 7970 graphics card. The CPU is cooled by a Thermalright True Spirit 140, which offers a great price/performance ratio. We're expecting impressive thermals, even under load, thanks to the H440's design and the reference-class Radeon HD 7970 that exhausts waste heat externally.

We benchmarked NZXT's fans at 12 V, since the H440 doesn't include a controller. Even still, noise levels should remain fairly conservative as a result of low rotational speeds and the case's sound dampening material.

We’re using our usual combination of Prime95 (Small FTTs) and MSI Kombustor (Furry and Tessy tests) to apply a load. Once we crest the 60-minute mark, we measure temperatures.

Temperatures And Noise

Temperatures

It would have been nice if the small board with its 10 three-pin fan power connectors offered more than just a 12 V setting. A 9 or 7 V option might have allowed us to run the case's fans a little slower. Still, the H440 stays pretty quiet thanks to its moderately low 1000 and 1200 RPM fan speeds. It doesn't hurt to have all of that sound dampening material in there as well.

Sound dampening materials don’t just keep noise from getting out; they also serve to insulate the chassis. This can result in higher interior and component temperatures. NZXT knows this and compensates, though, including plenty of space for additional fans in the event that you feel the four bundled ones aren't enough.

Even in its stock configuration, the H440 posts some nice performance numbers. Taking ambient temperature into account, it manages to keep our system at about the same temperatures as the more mesh-heavy Cooler Master Cosmos SE. At a comfortable ambient temperature of 22 °C, our little Haswell-based space heater stays under 65 °C, and our 230 W graphics card peaks at 75 °C with its fan speed at 40 percent duty cycle. The 3.5-inch hard drive positioned right in front of the fan predominantly comes in at 24 °C.

Noise

As usual, our graphics card is the largest contributor to system noise. That's hardly surprising, given that we're using a reference-class Radeon HD 7970. The Tahiti-based board pushes overall system noise to peaks ranging from 42.5 to 43.1 dB(A). Even though the H440’s sound dampening material results in slightly higher graphics card fan speeds compared to non-dampened cases, it still manages to help muffle sound levels outside of the enclosure.

It's important to note that the sound dampening material is especially effective for diminishing high-frequency noise, which is the most annoying kind. The sound pressure measurement numbers don’t tell the whole story. Subjectively, fan noise is neither loud nor annoying.

The NZXT FN V2 fans, running at up to 1200 RPM, generate a noise level between 38.1 and 38.7 dB(A). Those numbers are certainly too high for an ultra-quiet case, but they are completely acceptable for a gaming chassis. Even if you use passively-cooled components, that's what you're going to hear. Again, the option to slow down the case fans for more casual use would have been nice.

Fortunately, even though the fans generate a bit of motor noise, vibrations aren't propagated through the H440's frame. The fact that the fans aren't decoupled from the case with grommets ends up not mattering.

Great Looks, Useful Features, And No 5.25-Inch Bays

We can already hear it now. Some enthusiasts are going to love NZXT’s decision to get rid of 5.25-inch drive bays on the H440; others will hate it for depriving them of the option to mount additional peripherals like fan controllers and reservoirs. NZXT certainly makes good use of the reclaimed space though, providing a unique, well thought-out, and feature-rich case that’s built solidly and attractive to boot. Those top and front covers are particularly nice-looking thanks to their smooth lines and large metal surfaces.

In addition to a side window, NZXT's H440 features full sound dampening. Thoughtful accents like the bent motherboard tray allow data and power cables to be routed more easily. Cable management is one of this enclosure's strong suits. Many small, well-placed openings make it possible to optimize routing without much effort. Along the same lines, the PSU has its own compartment at the bottom of the case that’s hidden from view and illuminated by a stylish NZXT logo. Not only can you hide unused power supply cables in there, but also other loose leads as well.

The covered areas next to the side window conceal the separately-removable hard drive bays too, which can barely be seen from the outside. Since there aren’t any external 5.25-inch drive bays, the hard drive bays are spaced farther apart than usual, allowing for largely unobstructed airflow generated by three bundled case fans up front. And if you want to use a graphics card longer than 30 cm, simply remove one of the hard drive bays. Or, if you're using a 36-cm radiator, pull them all out. Even then, there's still space for one 3.5-inch or two 2.5-inch drives at the bottom of the case.

NZXT's H440 has several other handy features. For instance, there’s a small board with power connectors for up to 10 three-pin fans and additional spaces for SSDs on the PSU cover. Lighting inside of the case is stylish, while outside it's functional, helping you see the expansion slots and I/O shield.

We like that we can use CPU coolers up to 18 cm-tall and, with the previously-mentioned caveats, radiators up to 28 or 36 cm-long.

Really, we're not left with much to complain about. You don't get a ton of room for cable routing behind the motherboard tray due to the sound dampening material. But bends in the tray and hidden cable space in the PSU compartment make up for this. We really did miss the option to control case fan speed. They aren't overly noisy, since the fans don’t transfer vibrations to the frame, but a controller would have been an appropriate inclusion. Finally, the front dust filter is a bit difficult to access. You have to remove the entire front cover before it can be freed from its magnetic supports.

Bottom Line

NZXT’s H440 is perhaps the most uncompromising built-for-looks PC case currently available. It features an elegant design with lighting that looks good and serves a practical purpose. The large window lets your go-fast components take center stage, as the power supply and storage devices are strategically hidden behind solid side panels. Finally, cables can be hidden from view almost entirely.

Enthusiasts who can live without 5.25-inch drive bays and are in the market for a quiet (albeit not silent) enclosure with lots of features will find NZXT's H440 to be both unique and elegant. The white-black model will run you $140 on Newegg, as the black-red version sells for $120. Even though we wish it had a fan controller, we're perfectly comfortable recognizing the H440 with our Tom's Hardware Smart Buy award.

Introduction

We recently tested Raijintek's Ereboss CPU cooler, which was also the company's first CPU cooler, period. That one was a massive tower cooler that left us with a good impression, particularly for low-fan RPM applications. We’re now taking a look at how Raijintek fares in the ITX CPU cooler market.

Raijintek is sticking with Greek mythology for nomenclature. Its first CPU cooler was the Ereboss, named after the ancient Greek god of darkness. The second one is named after a titan.

That choice of naming might raise eyebrows, since this heat sink is quite the opposite of a titan. Its advantage comes from a low, flat profile, which stands just 6.8cm tall with the fan already installed. So long as your motherboard can accommodate it, the Pallas becomes ideal for high-performance mini-ITX platforms.

Raijintek's sink offers six 6mm heat pipes, which should be sufficient to keep even enthusiast-oriented CPUs cool in confined spaces. And while it might look a lot like Thermalright’s AXP-200, there are some significant differences. The Pallas is 5mm shorter than the current segment leader. It also has more fins that are closer to each other and shaped differently. You should be able to find it for sale online for just under $40.

Technical Specifications

MORE: Best Gaming CPUs For The Money

MORE: How To Build A PCMORE: All CPU Content
MORE: All Cooling Content

Packaging And Accessories

The Raijintek Pallas comes in a relatively small package. Its contents are kept in place by folded paperboard so that they don't bounce around. It's a truly efficient minimalist design.

The included accessories are similar to what you get with the Raijintek Ereboss. This isn’t surprising, since the two CPU coolers use the same backplate and retention frame that’s screwed in place. For the ITX model, two retention clips take the place of the large tower cooler’s four rubber fasteners.

This is also not the first time we’re encountering the low-profile fan with its 1.3cm-thick and 14x15cm tall and wide frame.

Design

The Pallas' sink is very different from the Ereboss that we recently tested.

Back to the Raijintek Pallas’ body. A direct comparison to Thermalright's AXP-200 shows Raijintek’s offering with a lot more fins. It has 69, whereas Thermalright employs 49. Naturally, then, the fins aren’t spaced as far apart. Surprisingly enough, the Pallas weighs less without its fan than the AXP-200.

The fins have a somewhat wave-like pattern and are interlaced in the center.

The cooler’s sides have notches to hold the fan’s retention clips.

The aluminum fins are bent in some places to prevent damage during installation. This is the only place where the otherwise-sturdy sink needs to be handled with care.

The Pallas is 6mm shorter than Thermalright's AXP-200. This is mostly achieved through the heat pipe design. Both coolers use 6mm heat pipes of identical length, but Raijintek’s stick out of the base a bit further, which shortens the part of them that bends up. There’s less space between the base plate and the cooling fins, which is to say that there’s less space under the cooler. We’ll take a look at how this affects practical considerations later.

In the height comparison above, Raijintek's Pallas is on the left, while Thermalright's AXP-200 is on the right.

The Pallas’ heat pipes are integrated cleanly with the cooler’s base plate. The contact surface with the processor is polished and plated with nickel.

The ends of the heat pipes are covered where they leave the cooler at the top. There’s another interesting detail here: the aluminum fins right above each of the pipes have little holes in them, which means that there’s a small air tunnel above each pipe. There are small knobs on the heat pipes themselves as well, increasing their surface area a bit.

We already know the bundled Aeolus fan from the Raijintek Ereboss. It’s slim at 1.3cm wide, and has a frame that measures 14x15cm. The red-and-white fan is fastened to the cooler with the help of retention clips.

Both the color scheme and the functional design are different from Thermalright’s competing offering. The fan as a whole is a bit more robust, but also a bit heavier. A cable length of 30cm is plenty for mini-ITX applications and should be ample in ATX-based systems as well. The Raijintek Aeolus has an RPM range between 650 and 1400, and it’s controlled via a PWM signal.

Installation

The Pallas’ installation mechanism is a carbon copy of the Ereboss’. It’s based on a backplate made of hard and sturdy plastic universally applicable to all processor interfaces. The advantage of this is that there are no parts isolating the backplate from the motherboard. However, this can also create problems on some boards due to the unconventional shape (see our compatibility notes below).

Screws with hexagonal heads are pushed through holes in the backplate and around the motherboard's CPU interface. The screw heads sit in fitting indentations in the backplate. On the front, plastic spacers are screwed on.

Next, the included retention bars for the cooler, which are different for Intel and AMD systems, are fastened in place with thumb screws.

Finally, a crossbar is fastened to the cooler and screwed to the retention bars. The screws are easy to reach through holes in the cooler.

The crossbar has a pair of protrusions that snap into place in two holes in the cooler’s baseplate. This keeps the cooler from sliding.

Notes On Compatibility

Similar to the Ereboss, installing Raijintek's Pallas in 90 degree steps is only possible on Intel systems. Of course, this typically isn't an important distinction for coolers that blow down onto your motherboard, though.

The backplate can cause problems with ITX motherboards. If your board of choice has surface-mount components behind the CPU interface, in a worst-case scenario, the Pallas could crush them. You'll need to check for compatibility before trying to seat the backplate. This usually isn’t a problem for the larger form-factor motherboards, but, unfortunately, it’s a common issue for smaller ones.

Raijintek's Pallas fit loosely on an Asus H87I-Plus motherboard. There were no problems with it on an ASRock B75M-ITX, though. It should be noted that the Asus board does diverge quite a bit from Intel’s reference design.

Raijintek could avoid these problems in many cases if it had the backplate touch the motherboard in just a few places, as opposed to it lying flat. This could be achieved by using additional spacers between the backplate and motherboard. Thermalright uses that sort of configuration for many of its CPU coolers. In talking to Raijintek, the company promises to examine this possibility and upgrade the Pallas if it works out.

Memory modules without heat spreaders and those with smaller ones that don’t extend above the upper edge of the PCB by more than a millimeter can be placed under the Raijintek Pallas.

How We Tested

In order to measure the processor’s temperature, it first endures 60 minutes of Prime95 with custom settings, heating it up. Then, its core temperatures are measured for 10 minutes. Finally, all of the temperature measurements across the entire interval are averaged.

This procedure yields more accurate results than recording the temperature at just one point in time. Unfortunately, the processor’s internal sensors provide relatively imprecise values, since they exclusively use whole numbers. This means that the results might have a somewhat higher measurement error.

Noise measurements are conducted with an open PC case from a distance of 30cm perpendicular to the center of the CPU cooler. Don’t panic after looking at the results; a closed case and a more reasonable distance from it will significantly reduce the coolers’ noise level. A greater distance would have meant that ambient noise would play a larger role, increasing measurement error.

We’ll adjust our benchmark system for ITX-oriented CPU coolers soon so that we can test them in a fitting PC case. Even without a purpose-built platform, Raijintek's Pallas turned out to be a strong contender in our tests. It didn’t just manage to complete our usual 125W CPU suite without breaking a sweat, but it also left us amazed after seeing the results of our overclocked 180W workload.

It’s not every day that a diminutive 6.8cm CPU cooler, which was designed for cramped spaces in tiny PC cases, can take on established tower coolers priced at up to $50. We didn’t even trust our results at first, but several more test runs didn’t change them.

The Pallas’ main competitor proved easy to find: it’s the Thermalright AXP-200, which is our go-to cooler for extremely high-performance systems that are supposed to be built with a very small form factor. Alternatively, we use the Deepcool Lucifer, which operates on a similar level as the HR-02 Macho, and the be quiet! Shadow Rock 2, which might not be able to provide quite the same cooling performance but has a leg up on the competition when it comes to noise level.

With the exception of the Thermalright AXP-200, we’re putting Raijintek's Pallas up against supposedly stronger tower coolers in its price range. Again, the results turned out to be a big surprise!

Stock Performance

Looking at the results achieved with a common 125W stock CPU, it quickly becomes clear that the Pallas is, in fact, a tiny titan. At maximum RPM, it beats the Thermalright AXP-200 by three degrees Celsius. The two CPU coolers perform about the same at 1000 RPM.

These results can be explained by a difference in design. Thermalright's AXP-200 has fewer cooling fins that are spaced further apart from each other, which causes it to lose less performance at lower fan speeds. The Pallas, on the other hand, capitalizes on higher air pressure with its many tightly-packed aluminum fins. Its 14cm fan also spins approximately 100 RPM faster, resulting in a lead over the Thermalright AXP-200.

Raijintek's Pallas is able to compete with the tower coolers surprisingly well too. It beats the huge be quiet! Shadow Rock 2, which weighs more than a kilogram. It doesn’t manage to keep up with the Deepcool Lucifer, but comes in two to five degrees Celsius higher. It’s a fraction of the size, though.

Overclocked Performance

Raijintek made a point of telling us that the Pallas would be able to cool a 180W processor before the company shipped us one. This usually earns a skeptical reaction from us as we make a mental note to run the benchmarks in spite of them probably being a waste of time.

In these cases, the CPU cooler is usually unable to make good on the manufacturer's promises. Raijintek's Pallas proves to be the exception. It doesn’t just keep our overclocked AMD FX-8350 cool at maximum fan RPM, but it even does so after slowing its fan down. The heat sink's direct competitor, Thermalright's AXP-200, fails both of these tasks. The same goes for be quiet!'s Shadow Rock 2.

It does get tight at 1000 RPM with a measurement 45.8 degrees Celsius above ambient temperature, which might not be quite enough if your air conditioning unit breaks in the middle of a hot summer. At the maximum fan speed of 1423, this number drops to 39 degrees Celsius. That's only five degrees higher than Deepcool's giant Lucifer.

Noise

High cooling performance achieved over a small surface is always accompanied by more noise. The Pallas’ maximum fan speed is 1423 RPM, which is faster than the AXP-200’s 1295 RPM ceiling. Those speeds result in noise levels of 45.5 and 44 dB(A), respectively. At 1000 RPM, the Thermalright cooler is still quieter than Raijintek’s offering by 0.5 dB(A). The newcomer comes in at 39.7 dB(A) after slowing down its fan. This doesn’t have to be the last word, though. Its fan can be further slowed all the way to 650 RPM via the PWM signal.

Raijintek's motor noise is noticeable in open or mesh-heavy PC cases; there’s some room for improvement in this area to quiet things down. Unfortunately, there’s (literally) less room for these improvements than there would be for a 2.5cm-thick fan. Then again, the AXP-200’s fan isn’t completely quiet either.

In spite of using the same fan as Raijintek's Ereboss, the Pallas’ results are somewhat worse than its sibling’s. One of the reasons for this is the different position of the downward-facing fan in relation to the measurement microphone. Using a different fan isn’t possible without major alterations either, since the included retention clips are specifically designed for low-profile fans. Raijintek doesn’t include an installation frame for larger fans like Thermalright does with its AXP-200 either.

Conclusion

Raijintek achieves the seemingly impossible with its Pallas low-profile CPU cooler. The company produces a brand-new entry in the mini-ITX-specific category that beats our previous recommendation, Thermalright's AXP-200, in several key areas and costs less to boot. Numerous small optimizations give it better performance and less height at the same time.

The Pallas’ higher number of aluminum cooling fins, which are also closer to each other, lets it convert its fan’s higher maximum RPM straight into more performance. The more you slow the fan down though, the more that advantage shrinks. This is due to the different underlying design concepts of the two CPU coolers.

In any case, the bottom line here is that Raijintek's Pallas can handle an overclocked 180W CPU. Performance reserves like these might not always be necessary, but they’re certainly impressive. Enthusiasts who’d like to overclock their processors in a confined space and want to stick with air cooling will find the right tool for the job with the Pallas. Its reserves also make this heat sink a true all-arounder. Even some tower coolers with massive surface area can’t handle our overclocking scenario.

The Pallas’ build quality can also be recommended without any reservations. Thermalright's AXP-200 sports a smaller number of thicker fins, making the sink seem more solid and robust. It also includes the option to use a full-size fan, fares better in the looks department and is quieter. Finally, Thermalright’s backplate has better compatibility with ITX motherboards. Raijintek might take care of this shortcoming in the future, though.

Those factors aside, the Pallas is a clear winner when it comes to height. Standing 6.8cm tall (including the fan), it beats the competition by 0.5cm.

Bottom Line

This year’s first real surprise in the CPU cooler category is here: the Raijintek Pallas. It takes the crown, becoming the most powerful mini-ITX-specific CPU cooler. That’s not all, though. It has the unique ability to cool overclocked 180W CPUs that even some tower coolers can’t handle, making it an option for use in larger systems as well.

At $40, the Pallas is also less expensive than its direct competition. The only caveats are a backplate that’s not compatible with all mini-ITX motherboards and the missing option to install a full-size fan. Neither of these change the fact that Raijintek's Pallas is the new reference for cooling performance in the low-profile space.

MORE: Best Gaming CPUs For The Money

MORE: How To Build A PCMORE: All CPU Content
MORE: All Cooling Content

Kai Tubbesing is an Associate Contributing Editor for Tom's Hardware Germany, covering Cooling.

Follow Tom's Hardware on Twitter, Facebook and Google+.

Making Z87 Express Three-Way SLI-Capable

We've seen incremental updates to Intel's processor architectures yield small performance benefits in games for the past several generations. However, a great many enthusiasts dislike some of the platform restrictions Intel put in place. For instance, CPUs built around the LGA 1156, 1155, and 1150 interfaces have integrated PCI Express controllers limited to 16 lanes of connectivity. Of course, that's a great way to help minimize latency. But Nvidia's official stance is that those available lanes support a maximum of two GeForce graphics cards. A third would force you to split into x8, x4, and x4 links. So what's a fan of three-way SLI supposed to do?

Stepping up to LGA 2011 gives you access to 40 lanes of third-gen PCIe, but you also have to take an architectural step backwards. Up until Intel launched its Ivy Bridge-E-based chips, you actually had to take two steps back. And beyond higher-priced components, you're talking about using more power and generating more heat, too.

But there's a solution, which isn't particularly new. Used to enable AMD's and Nvidia's fastest dual-GPU graphics cards, PLX Technology's ExpressLane PEX8747 48-lane PCIe 3.0 switch acts a lot like a repeater hub, broadcasting the same data to multiple devices, turning 16 lanes on one end of the switch into 32 lanes on the other, without dividing bandwidth (since CrossFire and SLI require all GPUs to receive identical data).

The usual marketing caveats apply, beginning with PLX calling this a 48-lane switch when, again, it takes 16 lanes on one end and facilitates 32 on the other. Moreover, three- and four-card configurations do cause bandwidth division, cutting dual 16-lane links into eight-lane interfaces. And finally, there’s the slight latency penalty of adding another component, though that seems minor considering both graphics vendors drop the switch onto their GeForce GTX 690 and Radeon HD 7990 without negative consequence.

Three of the above boards have a fifth PCIe x16 slot, but they're all limited to four x16 cards. Located between the first and third x16 slots, the fifth x16 interface is wired directly to the CPU, bypassing the PEX8747 bridge. Anyone not interested in three- or four-way SLI would be better off buying a less expensive motherboard without the PLX switch, so the folks in the market for these platforms can just go ahead pretending the extra slot isn't even there.

Two of the boards in today’s test offer Intel’s Thunderbolt technology, which pairs DisplayPort graphics and PCIe in a single serial signal. Compatible devices eat into both cable bandwidth and the Z87 Express chipset’s eight PCIe 2.0 lanes however, requiring ASRock and Gigabyte to add yet another PLX bridge, the four-lane PEX8605. Adding storage devices to the mini-DisplayPort connectors forces displays down from 4K-capable to a maximum of 2560x1600, while on-board devices and PCIe x1 slots are similarly made to share bandwidth over the PEX8605.

ASRock Z87 Extreme9/ac

ASRock hides its two Thunderbolt interfaces under four USB 3.0 ports, maintaining enough room to stick a CLR_CMOS button next to one of them as well. The full-sized DisplayPort connector passes an external signal through to a Thunderbolt port; the HDMI connector can be used for another display.

Both GbE ports interface with Intel controllers to support teaming mode, and ASRock also adds Broadcom’s 802.11ac-capable mini-PCIe card for wireless networking. Those are features that could make the Z87 Extreme9/ac perfect in an enthusiast's principal workstation. We only seem to be missing antenna connectors on the rear I/O panel...

Instead, we get a plastic-faced 5.25” adapter tray, which replaces the firm’s familiar 3.5” USB 3.0 bay adapter, to host internal wireless antennas in addition to a pair of USB 3.0 ports. The bay adapter we knew so well could take a 2.5” SSD internally. In this version, it's able to accept a pair of solid-state drives. But if you want to swap out the internal antennas for high-gain or directional alternatives, you'll have to find yourself a slot adapter for the back of the board. This shouldn't be an issue unless you're fighting for more range.

The second x16 slot is wired directly to the CPU; using it bypasses the PEX8747 PCIe 3.0 bridge and all four corresponding x16-length slots. Though we see four dual-lane pathway switches beneath that slot, ASRock tells us that there's no way to go CPU-direct for x8 and x8 two-way configurations. If you want to use two cards, you instead need to go x16 and x16 using PLX Technology's switch, and slots one and four. The company's bundled SLI bridge supports that spacing arrangement.

Other layout features include a second eight-pin EPS12V connector to feed extra current to the 12-phase voltage regulator (if desired), a second front-panel USB 3.0 header behind the first header at the front edge for easy four-port front-panel configuration, an extra four forward-facing SATA 6Gb/s ports (totaling 10), a second replaceable BIOS I/C, and an extra USB 2.0 port that fits external drives.

Other than the lack of external Wi-Fi antenna connectors, the only insurmountable layout problem is a front-panel audio header that’s a little too far back in the bottom-rear corner to fit the short cables of some cases.

My assumption when I started writing this story was that once the pieces were in place to support three-way SLI, accommodating four cards should be easy. ASRock's Z87 Extreme9/ac doesn’t include a four-way SLI bridge however, instead requiring you to connect the first card to the second, the third to the fourth, and the fourth to the second using separate bridges. This arrangement is illustrated in the company's user manual.

What you do get is a three-way SLI bridge, ten internal SATA cables, and the previously-mentioned bay adapter, which includes two front-panel USB 3.0 ports, an internal Wi-Fi antenna, and support for up to two 2.5” SSDs.

Z87 Extreme9/ac Software

ASRock includes a variety of freeware and trial software on its installation CD, in addition to ODM-supplied and its own branded applications. These include XFast USB by FNet, XFast LAN packet prioritization by cFos, Restart to UEFI, and various freeware and shareware discussed in a previous review. Realtek’s DTS-enabled audio control interface replaces the non-DTS-enabled Creative software found in some of the firm’s other high-end boards.

We were able to confirm much of ASRock’s OC Tweaker functions through CPU-Z and DIMM slot measurements. The program’s limits are tied to motherboard hardware limits, which are fairly consistent for ASRock’s full-sized enthusiast-grade motherboards.

ASRock A-Tuning also has an automatic overclocking algorithm with a built-in stability test. It pushed our Core i7-4770K to 4.30 GHz at 1.20 V, which is a perfectly acceptable target for this heat-constricted CPU. Our tests at 1.25 V generally yield a limit of around 4.6 GHz.

The Thunderbolt device manager is new to the Z87 Extreme9/ac, and accurately shows that I have no Thunderbolt devices to connect. Fortunately, our storage editors handle those details, and the hardware on this board is consistent with previously-tested configurations.

Z87 Extreme9/ac Firmware

The Z87 Extreme9/ac OC Tweaker menu leads with a list of factory-configured overclocking presets from 4.0 to 4.8 GHz, but most of those require too much voltage to run for more than a few seconds on our CPU before thermal throttling kicks in and knocks it back down. Next on the menu are CPU ratio and BCLK settings for manual O/C configuration.

The motherboard properly detects this memory’s DDR3-3000 XMP profile and Haswell’s DDR3-2933 memory controller limit, choosing that ratio and increasing the base clock to 102.3 MHz to compensate.

Everyone major motherboard company has figured out ways to override the CPU’s integrated voltage control. The 1.235 V setting produces a nominal voltage of 1.25 V on the Z87 Exteme9/ac, and load compensation pushes it a little past 1.26 V. Even if a company were to figure out a way to fool the volt meter, full-load temperatures would be a dead giveaway.

All major manufacturers also cheat on DIMM voltage, pushing 10% or more past the set limits to allow a higher overclock at a setting users believe is safe. Our volt meter showed an actual 1.65 to 1.655 V when using the motherboard’s 1.635 V setting.

Primary, secondary, and tertiary memory timings are adjustable over a wide margin to allow super-high memory overclocks.

The Tools menu shows an image of detected components (System Browser), allows flashing firmware from the UEFI GUI (Instant Flash), and stores up to three O/C configurations as user profiles.

Asus Z87-WS

Asus understands that a user paying for the PEX8747 bridge will probably use it, and smartly leaves out the fifth slot that would have stolen the bridge’s lanes. The Z87-WS also lacks Thunderbolt, allowing Asus to remove two controllers (the Thunderbolt controller and PLX Technology's PCIe 2.0 bridge) compared to several of its competitors. Buyers who don't need those features are rewarded with a $45 price reduction compared to ASRock's board on the previous pages.

Also missing from this reduced-price model are a couple of USB 3.0 hubs, which means that more of the remaining ports get full bandwidth. On the plus side are two eSATA ports with their own controller, two additional USB 2.0 ports on the I/O panel, and no reason to complain about any missing antenna connectors.

What appears to be a CLR_CMOS button is actually used for an Asus-exclusive feature called USB BIOS Flashback. A custom IC allows the board to update its firmware from a thumb drive with nothing more than a power supply attached. This is particularly useful when pairing motherboards that've been sitting in inventory a while with brand new CPUs, since the old BIOS often doesn't recognize the new CPU. I recently watched a technician use the feature to recover from a bad flash, as I listened to his call with Asus tech support.

The extra four internal SATA 6Gb/s ports (for a total of ten) are fed by a two-lane PCIe-based controller for added bandwidth compared to ASRock's board, which helps fulfill the Z87-WS’s “workstation” theme. Asus expects this board to be fully loaded with devices. The only sharing necessary is for mSATA, which disables one of the port connectors when it's populated.

Layout highlights include a secondary EPS12V input to support additional CPU current when overclocking, a second status indicator panel to further aid diagnostics, an internally-mounted external USB 2.0 port for ReadyBoost fanatics, a front-panel IEEE-1394 header for users of old scanners and digital cameras, and a front-panel audio header that’s been moved approximately one inch forward from its traditional rear-corner location. Asus appears to understand that the cables of many cases are too short to reach the corner, and compensates accordingly. Bravo!

The company is also fairly proud of the way it supports end users, and likes to remind us about its MemOK memory underclocking button (to assist re-configuration), DirectKey button (to boot directly to UEFI), EPU mode switch to allow auto-underclocking for energy savings, and TPU mode switch that enables a slight (ratio-based) to moderate (ratio- and BCLK-based) overclock for enhanced performance.

In addition to its ten SATA 6Gb/s cables and four-, three-, and two-way SLI bridges, the Z87-WS includes breakout plates for nine-pin serial port, USB 2.0, and IEEE1394 FireWire.

Z87-WS Software

Asus’ standard applications suite includes Digi+ voltage regulator modes, AI Charger+ USB charging booster, EZ Update software updater, EPU energy-saving underclocking profiles, Fan Xpert2 enhanced fan control, USB 3.0 Boost for UASP and USB Turbo modes, Boot Setting reboot to firmware GUI, Network iControl packet prioritization, “Wi-Fi Go!” launching point for wireless networking and AP-mode controls, and TurboV Evo software-based overclocking. The easiest way to get a list of applications included in Asus AI Suite is from its Uninstall menu. Most of these applications must be installed jointly but can be uninstalled separately if desired.

Google and Norton applications are installed with drivers and must be manually de-selected from the Asus InstAll to avoid later annoyances.

The TPU submenu of AI Suite’s Dual Intelligent Processors 4 menu provides Windows-level access to most of the overclocking controls found in the UEFI. These include CPU ratio and base clock controls, as well as voltage levels for most DRAM, chipset, and CPU interfaces.

There are also four ways to set automatic overclocking: as “Ratio Only” or “BCLK-First”, and as “Fast Tuning” or “Extreme Tuning”. These automatic overclocking techniques are also carried into software from the firmware side, where on-board switches provide similar functionality.

In “Fast Tuning” mode, ratio-first overclocking gets us to 4.3 GHz at 1.185-1.240 V. That completely realistic overclock is destroyed when ratio-first is set to “Extreme Tuning” mode, as the increase to 1.36 V causes instantaneous thermal throttling. Forget the shown 4.90 GHz; you’ll only see it at idle.

“Fast Tuning” in “BCLK-First” mode gives us 4.25 GHz at 125 MHz BCLK and a fixed 1.175 V. That’s perfectly stable, even on air cooling.

“Extreme Tuning” in “BCLK-First” mode pushes 4.8 GHz at 1.275 V, but our full-load test crashes so hard the motherboard can't recover on its own. We had to discharge the board and use its CLR_CMOS button to restore functionality.

Z87-WS Firmware

Asus’ UEFI includes a custom-configurable start page with XMP and fan settings, called Easy Mode. But if you really want to make changes easily, you'll instead set the boot options to kick you into the “Advanced Mode”.

Enabling XMP in Advanced mode means choosing XMP as an overclocking baseline, and making additional adjustments from there. The Ai Tweaker menu starts off with that setting, strap ratios for BCLK overclocking, and eventually BCLK frequency.

Further down the Ai Tweaker menu are CPU multiplier, DRAM frequency, and DRAM timing controls.

We’re three pages down the list before we reach CPU voltage controls, but are instantly made happy by one of our findings. While most enthusiast-class motherboards purposely add somewhere around 10% to both set and reported voltage, the Z87-WS gives us 1.25 V at its 1.25-volt setting.

DIMM voltage, on the other hand, turns up the same old tricks we've been observing for a while now from many board vendors. In order to achieve the best memory stability, Asus' Z87-WS adds around 30 to 35 millivolts to our chosen setting. Test consistency and accuracy demands that we choose 1.62 V to achieve the proper 1.65 volts.

Three pages of DRAM settings cover the expected range of primary, secondary, and tertiary options, along with a few additional wave form controls.

Default settings for the Digi+ Power menu provided our hardware with excellent voltage stability.

Gigabyte Z87X-UD7 TH

Gigabyte attempts a features coup with its Z87X-UD7 TH, beginning with the Thunderbolt headers found on ASRock’s sample and continuing into the two-lane RAID controller supported by Asus. We also find the pair of USB 3.0 hubs used by ASRock, along with the two I/O panel and two front-panel ports those hubs supply.

Gigabyte ups its Wi-Fi ante with a $55 dual-band 802.11ac / Bluetooth module manufactured by Intel. Supporting 867 Mb/s transfer rates, the module rests upon a PCIe 2.0 x1 expansion card.

It turned out that the U.S. had a brief exclusive with the board at a lower-than-expected price. After its recent worldwide launch, pricing info looks a lot more realistic. Today, Newegg lists the Z87X-UD7 TH for sale at $430.

Further inspection proves that the Z87X-UD7 TH looks the part of a $430 motherboard. In addition to the expensive PCIe 3.0 switch needed to make three- and four-way SLI work, the motherboard-down RAID controller, the Thunderbolt controller needed to pipe data across the mini-DisplayPort connector, and the PCIe 2.0 switch needed to give Thunderbolt devices their connectivity without disabling other on-board devices, we also find a 16-phase voltage regulator. Oh, and about Thunderbolt: this is the first board from Gigabyte to be Thunderbolt 2-certified, combining both 10 Gb/s channels into a single 20 Gb/s channel.

Anyone who thinks they might be able to push the limits of that 16-phase voltage regulator also knows that they’re going to make a lot of heat. Gigabyte equips the Z87X-UD7 TH voltage regulator with both a fan and a liquid-cooling channel.

Buttons at the front of the board control both base clock and multiplier ratios through a background application, which is perfect for tuners who don't want to reboot or open up another GUI while making adjustements. Other buttons control base clock increment size (0.1 or 1 MHz), enable a pre-configured overclock setting, and save current settings with a tag after resetting BIOS. Yet another button sets the system to retain component power after shut down…which would be perfect if the firm would introduce a DDR3-based version of its i-RAM.

A few things are imperfect though, including a front-panel audio header that’s around 0.5” too far into the rear corner for the cables of some cases to reach, a second front-panel USB 3.0 header that’s sure to be blocked by a fourth graphics card, and the problem that the third graphics card covers up the last PCIe x1 slot. If you do use three or more double-slot graphics cards, you'll be forced to give up the valuable Intel 802.11ac/Bluetooth combo card.

If we were to give up the Wi-Fi card for the sake of three- and four-way SLI, the Z87X-UD7 TH would still look like a fairly solid sub-$400 board. It still includes two-, three-, and four-way SLI bridges, a CrossFire bridge, a 3.5” bay adapter for front-panel USB 3.0, and six SATA cables.

Z87X-UD7 TH Software

Gigabyte’s software remains consistent from previous reviews, though the Z87X-UD7 TH also includes Creative’s X-Fi MB3 application. Tuning software also remains the same, though limits are motherboard-specific.

We still get the pre-configured overclocks of 4.1 GHz at stock voltage to 4.5 GHz at 1.30 volts.Custom settings are easier to read after we crop out the items that aren't configurable.

BCLK can be adjusted up to 266.66 MHz, DRAM data rates to 2933 MHz, CPU multiplier to 64x, and CPU Vcore to 2.0 V, just like in firmware. We were able to confirm several of these settings functional through CPU-Z and a volt meter.

Automatic tuning pushed our CPU to 4.60 GHz at 1.545 V, causing instantaneous thermal throttling under load. Worse still, it also set our DDR3-3000 memory’s XMP profile without regard to the motherboard’s incapability to run it past DDR3-2800 at 1.65 volts. We got this screen shot only after replacing the DDR3-3000 with some DDR3-2200.

Z87X-UD7 TH Firmware

Gigabyte’s UEFI opens to a high-resolution home page with system status bars on both sides and firmware information at the bottom. Frequently-adjusted settings are found in its customizable home menu.

A press of the F6 switch toggles to low-resolution mode, where we can more easily show actual settings. The frequency tab of the Performance menu, for example, begins with automatic overclock presets in 20% intervals, for which anything over 20% is unrealistic given our Haswell-based CPU's heat problems.

Also adjustable above are base clock and clock strap (for high BCLK frequencies), integrated GPU and CPU ratios, and DRAM ratios.

The “Advanced CPU Core Settings” menu offers Turbo Boost, integrated power savings, and power limit controls.

The “Memory” tab has redundant XMP and manual multiplier controls, where we are forced to reduce our DDR3-3000 from its correct 29.33x ratio to 28x in order to reach the board’s maximum stable frequency. Changing “Memory Timing” mode to “Manual” enables combined timing manipulation, while “Advanced Manual” configuration has per-channel timings.

Primary, secondary, and tertiary memory timings are adjustable over a wide range to provide the ultimate stability for custom DRAM overclocking.

The “Voltage” tab is nothing more than a launching point for various submenus that could have easily been combined into a single menu.

Our CPU reached 1.25 V at the motherboard’s 1.235 V settting, and our DRAM reached 1.65 V at its 1.625 V setting.

MSI Z87 XPower

MSI has no trouble with crowded components, spreading its Z87 XPower to an XL-ATX size that could have potentially supported up to nine expansion slots. It omits the top two slot connectors, though, to make more room around the CPU socket. With the same slot layout as competing samples from ASRock and Gigabyte, we have to ask: what does all that space provide?

The I/O panel has eight USB 3.0 and two USB 2.0 ports, yet it’s no larger than standard, and at least one competitor matches that array of connectivity with an ATX-sized board. CPU-based graphics connectors include two HDMI ports and one DisplayPort connector, but without the additional Thunderbolt controller offered by two competitors. Gigabit Ethernet also gets kicked down to a single port and controller, and there is a noticeable gap above the CLR_CMOS button.

That gap gets filled with the same 802.11n single-band Wi-Fi/Bluetooth module seen in our Z87 MPower review, which is less valuable than the dual-band 802.11ac card installed on ASRock’s sample. On the other hand, it resolves the slot issue faced by Gigabyte in three- and four-way SLI configurations. Also important is that it doesn’t require the use of a 5.25” bay for its antennas.

MSI includes overclocking control buttons for CPU ratio and BCLK, along with a power discharge button for CLR_CMOS, an auto-OC button, a power button, and a collection of voltage detection pins at the front edge. But those additions cause the board to be wider from front to back, not from top to bottom, and are matched by Gigabyte’s ATX-sized sample.

A look at the top shows what this motherboard's layout includes. Part of the PCB is consumed by an mSATA slot, but that’s not where the Z87 XPower was stretched. Heat sinks fill most of the area that MSI expanded, but the 32-phase voltage regulator covered by two of those sinks is the hidden prize. Fed by two eight-pin EPS12V connectors, that voltage regulator should be able to put out enough current for serious overclocking attempts. It's only a shame that Intel's Haswell-based processors don't appear very scalable on air or water cooling. While we might question the usefulness of such a large voltage regulator, the extra space it consumes also enables owners of 10-slot cases to fill them up with an appropriately-sized motherboard without any extra space at the bottom.

Can we assume that any manufacturer specialized enough to offer nine- and 10-slot enclosures would also be wise enough to make their front-panel audio cables long enough to reach the bottom-rear corner of a motherboard? MSI doesn’t leave this chance, moving its front-panel audio connector just to be safe. The remaining layout is good as well, including ten forward-facing SATA 6Gb/s connectors that are fed by a combination of the chipset’s controller and two PCIe x1 dual-port controllers.

Vacant slot positions at the top of the board should be filled with something, so MSI includes a two-port USB 3.0 breakout plate for one of the Z87 XPower’s two front-panel headers. Buyers also get two wireless antennas, six SATA cables, and three flexible SLI bridges that can be used for four-way SLI by connecting the first card to the second, the third card to the fourth, and the fourth back to the second. Because they’re flexible, three-way SLI is also supported by connecting cards one to two, cards two to three, and cards three to one (cross-ways).

Z87 XPower Software

Though MSI relies primarily on on-board buttons (Direct OC control) and its co-branded Intel Extreme Tuning Utility (XTU) for on-the-fly overclocking, it still has a more basic utility that’s able to address MSI-specific functions. It even gets a CPU Core Voltage knob this time, though that control was already available in XTU.

OC Genie can be set by either firmware or software, and the software version responds to the on-board button. We were happy to find a super-efficient 4.0 GHz fixed overclock at a lower-than-stock 1.10 V fixed voltage, though we actually had to change memory to get there. OC Genie automatically set our memory’s XMP-3000 profile, but the board couldn’t push this RAM past DDR3-2965.

The Killer E2205 GbE controller also gets a new management interface, with a handy pie chart for traffic analysis. Packet prioritization is a top feature here, as expected from a firm famed for its low response times.

Z87 XPower Firmware

MSI still relies on its tried-and-true firmware GUI with oversized clock and buttons to offset smaller-font settings, but at least that font can be read after shrinking screenshots to fit this Web page. Its OC menu starts off with CPU base clock and quickly progresses to CPU and DRAM ratios.

Voltage settings are further down the menu. We were able to run our Core i7-4770K at 4.60 GHz without thermal throttling by using the board’s 1.230 V setting to reach an actual 1.25 V.

DRAM similarly reached 1.65 V at the board’s 1.63 V setting, though it could only push our DDR3-3000 kit to a 2965 MT/s data rate at full stability. DDR3-3000 mode wasn’t even bootable, though the Z87 XPower did recognize the CPU’s maximum multiplier and choose the appropriate 102.3 MHz base clock to reach its full rating.

Primary, secondary, and tertiary memory timings are all adjustable, as is on-die termination.

MSI’s DRAM Training menu lets you change the way the memory controller interacts with memory modules. MSI hasn’t written a tutorial, but there is a possibility that a few tweaks here might have gotten our modules over the hump to their rated 3000 MT/s data rate.

The DigitALL Power submenu provides many voltage offset and droop compensation controls, none of which needed to be altered for our air-cooled tests. Extreme overclockers will probably find this menu most useful.

Other related menus include a visual map of detected component locations, several pages of SPD and XMP timing detection reports, a page to store up to eight overclocking configurations as user profiles, and even a manually-adjustable fan ramp map.

Hardware And Benchmark Configuration

Thermalright’s classic MUX-120 remains competitive with our recent review of high-end heat sinks, even when using its original clip-on mounting system. It’s that easy-to-mount mechanism that wins me over in motherboard round-ups, and the good performance points to a good design.

Alternatively, we can point to problems with the Core i7's heat spreader as a reason why larger coolers couldn’t give us significantly better thermal performance in that review.

G.Skill’s F3-17600CL9D-8GBXLD is the only memory kit in our lab that defaults to our DDR3-1600 CAS 9 test standard. Faster RAM always uses slower defaults, and slower RAM requires XMP to get there. The problem is that some boards automatically enable other overclocking features when XMP is enabled. Consistency rules these tests.

We replaced the slower memory with G.Skill’s DDR3-3000 kit for our overclocking stability tests.

Corsair sent its 80 PLUS Plantinum-rated AX860i for our benchmark needs, citing enhanced support of Haswell's C7 state.

Results: 3DMark And PCMark

Intel has integrated so much hardware into its CPU and PCH that we’re now left with essentially the same performance between all boards, unless something is configured incorrectly. Some manufacturers even sneak in an overclock just to win our performance benchmarks (though we're pretty quick to catch those attempts). Essentially, we’re only looking for broken stuff and/or cheating.

Minor differences between consecutive runs account for the miniscule variance between Z87-based motherboards in this round-up. So far, so good!

Results: SiSoftware Sandra

Sandra’s CPU benchmarks also show all four motherboards essentially on-par, though its memory test shows the Z87X-UD7 TH slightly behind. With so few other optimizations to choose from, memory optimization has become the “secret sauce” of motherboard makers concerning LGA 1150 performance and memory overclocking.

That slight difference at least gives us something to look for in our real-world benchmark set, which begins on this comparison’s next page.

Results: Audio And Video Encoding

MSI’s Z87 XPower comes up a second behind the pack in both iTunes and LAME, even after we double-check for rounding errors in iTunes. The actual performance difference is miniscule, and these are the only two applications in our benchmark suite where this occurs.

We were looking for applications where the Z87X-UD7 TH’s slightly slower memory performance might hamper real-world performance, and might have found one in HandBrake. The difference is still too tiny for most users to perceive in real-world use, however.

Results: Adobe Creative Suite

The Z87X-UD7 TH gains a second in After Effects, but then loses it in the test of OpenCL-based Photoshop filters. All four boards perform almost identically on average.

Results: Productivity

Still looking for any real-world indicators of the Z87X-UD7 TH’s slightly lower memory bandwidth result, we find Gigabyte's board slightly behind in both 3ds Max and Visual Studio 2010. The same motherboard’s good performance in Blender, along with a competitor’s stumble in ABBY FineReader, still lead us towards essentially comparable overall performance.

Results: File Compression

No difference in file compression performance indicates that none of these motherboards has a configuration issue or is secretly overclocked. Boring charts make for an easier feature-based analysis.

Power, Heat, And Efficiency

Packed with add-on SATA controllers, USB hubs, and a heafty PEX 8747 PCIe bridge, we expected all of these boards to consume more power than those reviewed in our mini-ITX round-up. From a competitive standpoint, the Z87X-UD7 TH’s high idle energy use is a little disconcerting.

MSI’s huge voltage regulator runs coolly under the low load of a non-overclocked CPU. Lower CPU temperatures may also be the result of smoother current that can come from all those extra phases.

A couple pages into the benchmarks, I anticipated that Gigabyte’s lower synthetic memory score would have little to no effect on its overall performance. Indeed, it’s dead center of a very tight pack. However, poor idle power consumption triggers a loss in this efficiency comparison.

Overclocking Results

Some manufacturers happen upon their overclocking victories by tampering with set voltage levels, sneaking in 20 to 30 millivolts more than you specify. But we usually catch that, either by voltmeter or by CPU temperature (often both). We weren’t surprised that three of the boards hit exactly the same overclock, and only one of these allowed us to bump up base clock by a measly 1 MHz at 46x.

Speaking of base clock, the most significant setting for users with multiplier-locked processors is at the 100 MHz strap. That’s because both ratio adjustment and base clock strap are locked out of those same CPUs. Certain manufacturers have ways to boost base clock overclocking capability, included reductions in System Agent/Cache and integrated GPU ratios. We left those settings to the manufacture to choose, and Asus surprised us with a 114 MHz BCLK.

ASRock and Gigabyte experienced the same phenomena seen by EVGA in our mini-ITX motherboard round-up, where increased base clock caused the PCIe graphics card to malfunction. I’m starting to wonder if it’s a problem with Intel’s latest base firmware, which companies can modify to suit their board’s configuration differences.

We used our voltmeter to set 1.650 to 1.655 volts to the memory

Which Premium Z87 Motherboard Takes Top Honors?

What can be said of price-to-performance that hasn’t already been stated? It certainly doesn’t account for added features, and the board that has the fewest features also has the lowest price. I might like that Asus’ Z87-WS isn’t packed with stuff I don’t need, personally, but you might need those things.

What does ARock’s Z87 Extreme9/ac give you for an extra $45? How about Thunderbolt, complete with the added four-lane PCIe switch needed to make the on-board devices and slots work after four of the PCH's lanes are devoted to the technology. The 802.11ac module is worth nearly as much as the price difference, though Asus would probably point out workstation-oriented features like the mid-range 88SE9230 add-on SATA 6Gb/s controller, the lack of sharing between those ports and eSATA, and the USB BIOS Flashback feature that I finally saw required and implemented in a real-world scenario.

Rather than try choosing between two compelling platforms, both receive our Approved recognition. It's up to you to decide which feature set best fits your needs.

Gigabyte’s Z87X-UD7 TH sets the top of this comparison scale for price at $430, but also has the most features. Typically, coming to the table with the most features without any noteworthy flaws, price aside, qualifies a product for our most prestigious Elite award. As it turns out, though, this platform is competing in a comparison of motherboards with three- and four-way SLI support. When you utilize its full feature set, dropping in a trio or quartet of graphics cards, its last PCI Express x1 slot is covered, preventing the use of Gigabyte's bundled 802.11ac/Bluetooth combo card. Great overclocking, a pair of quiet fans for voltage regulation and chipset components, a liquid cooling channel on the voltage regulator, and even the lauded Thunderbolt 2 controller can't distract me from that flaw. It's not minor, either. The high-end Wi-Fi card that three-way SLI builds lose is worth at least $50.

The Z87 XPower is runner-up for Tom's Hardware Elite. At $400, it’s the most elaborate board in this round-up that supports all of its features and three- or four-way SLI simultaneously. Its 300 Mb/s 802.11n Wi-Fi solution is decidedly low-cost, but at least it doesn’t get in the way. It also has a more premium Killer E2205 GbE controller, though most of its competitors have two gigabit-capable interfaces. And it’s hard for us to prove the value of a 32-phase voltage regulator on a CPU that really needs to be de-lidded and cooled with liquid nitrogen before the super-beefy power circuitry really becomes a factor.

Cool stuff abounds in MSI's Z87 XPower, yet the lack of two Ethernet controllers is at best offset by the higher-priced single gigabit chip, and its integrated overclocking features are perhaps offset by the use of a lower-cost wireless solution. That means its larger voltage regulator bears the entire burden of its $70 price premium over the Z87 Extreme9/ac, without consideration for the cheaper board’s Thunderbolt capability. Given the board’s focus, we think it'd fare best in a competition of overclocking platforms designed with extreme cooling in mind.

The Mini-ITX Market Is Small, But Growing...

Every time we hear about the death of the desktop PC, that conclusion seems to be based on slowing sales from big tier-ones like Dell and HP. But we know from talking to boutique builders that the demand for high-end gaming PCs and workstations continues to increase. It's a good time to be an enthusiast, and games like Battlefield 4 help illustrate why. 

At least some of the excitement comes from high-performance, high efficiency hardware, which is allowing the community to construct faster machines in smaller enclosures. It's no longer necessary to drop a big case next to your desk with two or three 250 W graphics cards for playable frame rates in the latest titles. Now we can get plenty of speed in small machines that look more like living room appliances, but are every bit PCs. 

Up until recently, the mini-ITX form factor was all about compromise. You'd give up most of your motherboard's slots and room for integrated extras to fit into tighter spaces. But as this segment grows, more and more companies are recognizing a desire for premium-class platforms with all of the on-board features expected from a larger ATX board. Drop in a high-end, overclockable processor, one dual-slot graphics card with multiple display outputs, and some solid-state storage. Sounds about as sexy as any monolithic gaming box we've ever seen.

So, which motherboard would we choose to serve as the foundation for a build like that? There are certainly many options on the market, and five of the industry’s top performance-oriented brands sent us a sample to compare.

ASRock Z87E-ITX

ASRock fills the mini-PCIe slot of its high-end Z87E-ITX with Broadcom’s dual-band BCM4352 controller card, adding 802.11ac Wi-Fi with a peak data rate of 867 Mb/s and Bluetooth, in addition to Intel’s reputable gigabit Ethernet PHY for networking. The full set of six SATA ports is also available internally, though two of those ports are shared with other connectors.

One shared connection shows up on the I/O panel as eSATA, alongside DisplayPort and HDMI display outputs. DVI-I allows ASRock to get rid of the ancient VGA connector, while retaining compatibility through an included adapter block. The space that might have once been dedicated to VGA is now occupied by the wireless controller's antenna bracket.

The rear I/O panel also provides a CLR_CMOS button, four of the chipset’s six USB 3.0 ports, two USB 2.0 ports, five analog audio jacks, and optical audio ouput. ASRock expands multi-channel digital audio functionality to live streams by licensing DTS Connect capability for its Realtek ALC1150 codec.

The Z87E-ITX hides an mSATA interface under the circuit board to save space and confuse spectators. Could that system really be running with no visible hard drive or SSD? Crazy! That connector does steal one of the top-side SATA 6Gb/s ports when it's occupied, lowering the total number of unshared internal ports to four. Several competing boards are limited to four ports, period, with no other interfaces. So, ASRock can count this as a win.

As with SATA connectivity, the Z87 chipset is packed with more USB 2.0 ports than most manufacturers can fit onto such a small form factor. ASRock again does one better than most of its competitors by offering two internal headers, for a total of four front-panel USB 2.0 ports, in addition to the Z87E-ITX’s dual-port USB 3.0 header.

The Z87E-ITX’s CPU interface is positioned to the far right, putting more space between the CPU cooler and graphics card. Placing it there forced ASRock to move its eight-pin EPS12V connector to the left of the CPU’s input voltage regulator, where it could be unreachable under mid-sized coolers. Builders may be forced to attach this cable before installing a wide, low profile cooler.

Larger coolers might also overlap the last two SATA ports. Knowing that some folks won't be able to reach them, ASRock picked those ports to share with the mSATA and eSATA interfaces.

Four SATA cables, a DVI-I-to-VGA adapter, an I/O shield, and a dual-band antenna complete the Z87E-ITX hardware list.

Z87E-ITX Software

The Z87E-ITX contains ASRock’s standard software suite, including XFast USB by FNet, XFast LAN packet prioritization by cFos, Restart to UEFI, and various freeware and shareware discussed in a previous review. The Z87E-ITX reverts to Realtek’s DTS-enabled audio control interface, but the firm's A-Tuning application adapts to the new board’s hardware configuration.

Confirmed working through CPU-Z, base clock, CPU ratio, and voltage controls match firmware settings.

Additional voltage settings are also adjustable through ASRock A-Tuning, though most of these changes aren’t detectable through software. DRAM and CPU Input were verified functioning with a volt meter.

A-Tuning System Info shows installed hardware on a motherboard-specific image. The Z87E-ITX gets two images, since its mSATA connector is on the opposite side.

An automatic overclocking algorithm pushed our CPU to 4.3 GHz at 1.20 V. Though those settings initially appeared ideal for compact gaming systems, Prime95 revealed power throttling down to 3.5 GHz.

Z87E-ITX Firmware

The Z87E-ITX's firmware is loaded with several overclocking profiles, including 20% and 30% over stock and manual configurations up to 4.8 GHz. We already know where we’re going, jumping down to the CPU Ratio “All Core” multiplier of 46x at the stock 100 MHz BCLK frequency.

Because the Haswell CPU’s integrated memory controller lacks a 30x multiplier, the Z87E-ITX correctly chooses its DDR3-2933 setting and adjusts BCLK to 102.3 MHz.

We reached the expected 4.6 GHz overclock at the Z87E-ITX’s 1.24 V setting. Though CPU-Z now reports set voltage (rather than actual voltage), CPU core temperature was consistent with our actual target of 1.25 V.

DRAM voltage is easier to check with a volt meter, reaching an actual 1.65 V at the board’s 1.635 V setting.

Primary, secondary, and tertiary timings are adjustable over a wide range. XMP timings for our DDR3-3000 modules are shown.

Asus Z87I-Deluxe

Asus packs its Z87I-Deluxe I/O panel with six USB 3.0 and four USB 2.0 ports, but loses PS/2 and eSATA compared to ASRock’s competing product. Internal connectors are moved around to make room for a far-larger 12-phase CPU input voltage regulator.

The I/O panel also gains a USB BIOS Flashback button. This Asus-exclusive feature allows builders to flash a new firmware version using nothing more than a power supply and USB thumb drive, and is particularly useful when installing a newer, unsupported CPU on an older board. We shouldn’t need to worry about that for a while, since all of the Haswell-based processors compatible with LGA 1150 are pretty recent.

Analog connectivity on the rear panel drops to three jacks, though DTS Connect technology encodes live audio streams to 5.1-channel surround over a single digital output.

Asus saves space by placing Broadcom’s dual-channel, 867 Mb/s-capable 802.11ac Wi-Fi + Bluetooth controller on a vertical I/O panel riser. And since the vertical voltage regulator impedes access around the CPU socket, the Z87I-Deluxe’s eight-pin CPU power connector is relocated to the board’s front edge. Large CPU coolers may force builders to connect fan header cables before placing the board into a case.

The USB 3.0 front-panel header is a little close to the CPU socket, which may cause clearance issues with some coolers. Though most large heat sinks can be rotated out of the way, the upright voltage regulator may prevent low-profile models (those with little motherboard clearance) from being rotated. Those factors make cooler offset and vertical clearance equally important, which is why they’re listed on page one of Eight Low-Profile CPU Coolers For Your Compact PC, Reviewed.

Asus surprises us by including all six SATA 6Gb/s internal cables with its six-port board. An I/O shield, Wi-Fi antenna, and front-panel lead extender round out the kit.

Z87I-Deluxe Software

The Z87I-Deluxe includes Asus’ standard applications suite with Digi+ voltage regulator modes, AI Charger+ USB charging booster, EZ Update software updater, EPU energy-saving underclocking profiles, Fan Xpert2 enhanced fan control, USB 3.0 Boost for UASP and USB Turbo modes, Boot Setting reboot to firmware GUI, Network iControl packet prioritization, “Wi-Fi Go!” launching point for wireless networking and AP-mode controls, and TurboV Evo software-based overclocking.

TurboV Evo represents the full range of multiplier and voltage controls already found in UEFI, and changes that are applied directly to UEFI require a reboot.

Automatic Overclocking is also an option, with the Auto Tuning algorithm quickly testing stability at increased multiplier, voltage, and clock settings. Choosing “By Multiplier” mode got us a 4.30 GHz max O/C at 1.155 V that drops to 4.1 GHz as additional cores are loaded.

A few more millivolts would have allowed the system to remain stable at its full overclock with all four cores under load. Choosing “By Base Clock” mode, the algorithm pushed a fixed 4.25 GHz O/C at 1.175 V.

Z87I-Deluxe Firmware

Asus' firmware has several overclocking modes, and the Z87I-Deluxe follows that theme. Selecting XMP overclocking mode before making other adjustments enables both XMP-memory and manual CPU overclocks simultaneously.

The Z87I-Deluxe properly selects the highest-functioning DDR3-2933 memory ratio for our Haswell-based processor, increasing BCLK to 102.3 MHz to enable its DDR3-3000 rating. We had to back it down to 100 MHz to find our CPU’s 4.6 GHz maximum frequency.

Heat is the reason our CPU tops out at 4.6 GHz, as core voltage levels greater than 1.25 V result in thermal throttling. We reached that limit at this motherboard's 1.245 V setting.

DRAM is even more aggressively over-volted; our voltmeter showed a 1.65 V output at the board’s 1.635 V setting.

Primary, secondary, and tertiary timings are adjustable. The top three settings switched from “Auto” to our memory’s XMP ratings when XMP mode was enabled.

Because the Haswell architecture's core voltage is internally regulated, we left load-line calibration settings at motherboard defaults.

EVGA Z87 Stinger

While most of its competitors scrambled to find room to retain all of the Z87 Express PCH's natively-supported features using a mini-ITX form factor, EVGA made space for power and reset buttons, plus a dual-digit system status display. Dedicated to gamers, and by extension to overclockers, the company got there by focusing on the unique needs of the mini-ITX-based gaming system market.

The first thing we noticed was that the board has only four internal SATA 6Gb/s ports, which makes sense when most small cases have no more than four drive bays anyway. A tiny heat sink serves a similarly downsized four-phase voltage regulator, which also works for us since Haswell’s heat issues normally prevent it from capitalizing on the additional capacity of larger regulators. EVGA also gets rid of rarely-used connections like legacy communications ports and the TPM.

That’s not to say the board is stripped. We still find, for example, the full range of front-panel headers, including USB 3.0, USB 2.0, and analog audio I/O. In fact, front-panel audio is a significant addition to the Z87 Stinger, since EVGA neglected to include it on the company's Z77-based predecessor. Also, one of the missing SATA 6Gb/s ports gets re-routed to eSATA.

EVGA even includes a decorative cover so you don't have to look at parts of the board that don't add anything aesthetically. In fact, Wi-Fi is the only feature notably missing from the Z87 Stinger compared to competing products.

The Z87 Stinger has a mini-PCIe slot into which you can drop your choice of wireless networking controller or other compatible device. EVGA even includes a place on the board's I/O shield for mounting antenna leads. On the other hand, if the company wanted to provide its customers with added configuration flexibility, it should have bundled the break-out cables that aren’t included with most mini-PCIe-based wireless modules.

The Z87 Stinger's manual is universal, and based on EVGA's other products. You'll need to rely on its additional configuration sheet to determine the location and function of various connectors, and know to dismiss anything listed in the manual but not found on the board.

Z87 Stinger Software

Equipped with Creative’s Core3D audio processor, the Z87 Stinger features the same audio control interface as the previously-reviewed M8 mini PC, with one exception: there are no digital inputs in its mixer panel.

EVGA E-Leet is updated for the company's Z87 line-up, adding tuning menus to the familiar CPU-Z interface. An entire row of five tabs is devoted to hardware monitoring.

Base clock settings far exceed UEFI limits, though both exceed the capabilities of any Haswell-based CPU that we've seen. We verified that smaller changes worked though, in addition to ratio multiplier adjustments.

Another menu provides controls for CPU input, DIMM, and chipset voltages. We were able to confirm DIMM voltage changes with a volt meter.

Two more menus provide core affinity for specific processes and overclocking profile settings.

Z87 Stinger Firmware

The Z87 Stinger is a formidable competitor in multiplier-based overclocking. But if you're doing any tuning with the BCLK, then beware. A “d6” error code had this sample locked up for hours after setting 104 MHz with integrated graphics disabled.

It turns out that whenever the board experiences overclocking issues on its PCIe bus, it reverts to integrated graphics. If integrated graphics are disabled, it stops. And the CLR_CMOS button does not reset the base clock nor re-enable on-board graphics. Pulling the battery with the power disconnected and allowing it to drain down for several hours didn’t solve the issue, nor did changing the CPU or graphics card. Even EVGA’s own GeForce GTX 760 in our lab didn’t solve this show-stopper.

Forcing configuration errors is my last-ditch approach to sneaking into a board's firmware without replacing its CMOS ROM. In theory, a platform should (and usually does) jump to default settings when major hardware changes take place. The problem is that all of my LGA 1150-based processors are the same model!

I tried booting without memory and then booting with memory, but the UEFI remained stuck at 104 MHz with on-board graphics disabled and a PCIe error to the graphics card. I even tried removing the CPU, starting without the CPU, and then re-installing the CPU. Many of these changes would cause the board to temporarily use default settings, but it would then detect the hardware and reboot with the previous problematic settings before POSTing.

Thank goodness for bad memory, though. After pondering the type of error that would force the Z87 Stinger to remain in default mode until I could enter the UEFI, I remembered that some very old DDR3 modules aren’t properly programmed for newer memory controllers. I reached into my scrap pile, pulled out some discarded memory, and installed it. Viola, the error was detected, the board couldn’t figure out how to re-configure it at previous settings, and I was dropped into the firmware for manual configuration.

Remember that the diagnostics process started with many attempts to CLR_CMOS? Upon entering the UEFI, I found that the 104 MHz setting was still locked and loaded for application upon the next boot. I also found that on-board graphics would still be disabled. Resetting to defaults from inside the firmware got me back up and running.

The Z87 Stinger reached our expected 4.6 GHz at 1.25 V using a motherboard setting of 1.245 V, but couldn’t configure our memory to its XMP-3000 settings. It instead chose a 30x memory multiplier that’s not supported by any LGA 1150 processor. The board was, however, able to reach DDR3-3025 by choosing its DDR3-2933 setting and then increasing BCLK to 103 MHz.

Primary, secondary, and tertiary timings are all configurable from the Z87 Stinger’s memory menu. A 1.635 V setting produced a detected 1.65 volts.

I’ve never used memory training algorithms to search for higher DRAM clocks, but welcome the opinions of advanced memory tweakers in our comments below. The Z87 Stinger provides a lengthy selection of these settings.

Gigabyte Z87N-WiFi

While some companies add features and others reduce prices to win over different types of consumers, Gigabyte tries to do both with its Z87N-WiFi. We found this board, with its two GbE and Wi-Fi controllers (with Bluetooth support), for a mere $127.

There’s no such thing as a free lunch though, and 802.11ac capable of 867 Mb/s performance is certainly a big cost-adder. We instead find Intel’s 802.11n-based single-band module on-board, which is rated for up to 300 Mb/s in a 2x2 configuration. Hop online and you can buy that card for as little as $15 or so. And the second GbE controller isn’t free either, as the board also has a reduced-cost four-phase voltage regulator.

Gigabyte drops two of the chipset’s six SATA 6Gb/s ports. Even though the four remaining ports are more than adequate for most builds, the loss of a pair internally also highlights the fact that the board lacks eSATA. And the VGA output capability of its DVI-I connector is lessened by the installation kit’s lack of an adapter block.

None of those exclusions dissuade us from seeing high value in the Z87N-WiFi’s low price though, since available connections and power still surpass the maximum needs of most miniature builds. The Haswell architecture's heat issues even mean that most overclockers will run out of thermal headroom before they reach the voltage regulator’s maximum output.

Mini-ITX chassis- and CPU-imposed limitations don't save the Z87N-WiFi from harsher feedback over only including two fan connectors. Most of our tiny cases have both intake and exhaust fans. And the CPU cooler still needs at least one fan header.

Close spacing between the Z87N-WiFi’s CPU socket and graphics slot raise a second concern: cooling. Eight Low-Profile CPU Coolers For Your Compact PC, Reviewed probably has enough offset data to help an owner of this board figure out what will work, but we had to turn our tower-style heat sink sideways in order to fit a graphics card. Then, memory clearance becomes a concern.

The four-pin CPU power connector (ATX12V) is also uncomfortably close to the CPU socket, and will likely compel owners of low-profile coolers to install the cable, heat sink, and motherboard in that order.

The Z87N-WiFi includes a wireless networking antenna, two SATA 6Gb/s cables, and an I/O shield.

Z87N-WiFi Software

Gigabyte’s APP Center resides under the Windows task bar, popping up when selected to reveal a scrolling menu. The image below shows these menu items pasted side-by-side for a quicker view, and the applet links a variety of handy applications such as USB Blocker for port security and Live Update to poll Gigabyte’s server for new drivers.

Gigabyte @BIOS allows firmware updates within windows, polls from Gigabyte’s server, and includes its “Face Wizard” boot image modification application.

Gigabyte’s EZ Setup provides a simple interface for the functions of Intel’s SATA controller.

Smart Recovery 2 provides file backup and system recovery functions.

Gigabyte Smart TimeLock gives PC owners additional user management options, as described on its info page.

The company's EasyTune hasn’t changed significantly since our previous verbose description, but its tuning options are still governed by each motherboard model's feature set. We find the full range of clock, multiplier, and voltage controls repeated from firmware into this easy Windows-based interface:

Automatic overclocking is also an option, with both pre-defined overclocking profiles and an overclocking algorithm at your fingertips. The top profile (OC Extreme) increases the CPU frequency range to 3.7 through 4.5 GHz, depending on the number of cores in use, at 1.30 volts.

Gigabyte’s Auto Tuning algorithm managed to get our CPU stable at 4.40 GHz with all of its cores loaded. The voltage level for this somewhat-impressive O/C was on the high side at 1.475 V unfortunately, causing near-instant thermal throttling from our Haswell-based Core i7.

Z87N-WiFi Firmware

Gigabyte’s UEFI is optimized for 1080p displays, showing system status in various menus that surround its main settings. Fortunately, the F6 key toggles the screen to low-resolution mode, where only the main menus remain.

The user-configurable home page is set by default to provide the most common overclock settings, including base clock, multipliers, and key voltage levels.

Z87N-WiFi firmware defaults to 99.77 MHz in Auto mode, but selecting a 100.01 MHz base clock results in an actual 100.00 MHz.

The Z87N-WiFi had no difficulty reaching the 4.6 GHz we were expecting, but our XMP-3000 memory profile didn’t work. We had to drop the memory multiplier to 24 x 100 MHz.

The Z87N-WiFi memory menu has redundant multiplier and XMP settings. Switching from Auto memory timings to Manual provides combined (both channel) timing controls, while per-channel settings can be made using Advanced Manual timings.

Primary, secondary, and tertiary memory timings are all adjustable. We were able to run our RAM at DDR3-2400 CAS 9, but loosening it to slower XMP timings did not allow stability at any higher memory ratio.

The Voltage menu is a launching point for three submenus that, combined, would have still fit onto one page. We reached our desired 1.25 V CPU core voltage at the Z87N-WiFi’s 1.23 V setting, and our memory’s 1.65 V rating at the board’s 1.62 V setting.

MSI Z87I

MSI’s Z87I puts Gigabyte in a tight spot by providing a nearly identical feature set and price. Using the same 802.11n wireless card from Intel as its competition, MSI adds teaming support to its dual GbE ports by leaning on a matched pair of less expensive controllers.

MSI adds DisplayPort to facilitate higher resolutions from the board’s second video output than Gigabyte enables, but joins the Z87N-WiFi in leaving out the rarely-needed DVI-to-VGA adapter block. We even find four USB 3.0 and two USB 2.0 ports on the I/O panels of both boards.

You get six analog audio outputs from the Z87I, compared to Gigabyte's five. This is achieved by relocating the Z87I’s optical audio output. MSI also finds room for two easy-access buttons on its I/O panel: CLR_CMOS and the firm’s aptly-named “Go2BIOS”.

Similarities between competing models don’t end at the I/O panel. The Z87I also has the same CPU-to-graphics slot crowding issue, and is even limited to two fan headers. I'd like to offer the same tip about offset measurements from Eight Low-Profile CPU Coolers For Your Compact PC, Reviewed, though the Z87I does have slightly more room between its CPU interface and ATX12V connector.

We're pointing out an awful lot of similarities here, but also like Gigabyte's board, MSI only exposes four internal SATA 6Gb/s ports, highlighting a lack of eSATA on the Z87I. Both platforms go so far as to use the same ALC892 audio codec, so we’re left to check overclocking and efficiency before we can choose a leader.

You only get two SATA cables with the Z87I, an I/O shield, and a pair of Wi-Fi antennas to complete the installation. Aside from antenna style, this is an almost exact match to its closest competitor.

Z87I Software

MSI’s Command Center software is now updated for better Z87 motherboard support, providing several overclocking, monitoring, and system function controls.

CPU multiplier and BCLK settings are adjustable within firmware limits, though our CPU kept our verification process constrained to a few steps over stock. Smart fan controls appear on the same page, letting you set a speed curve that best matches thermal needs to acoustic desires.

DRAM and GPU overclocking menus appear to have available adjustments, though the settings were locked in our specific configuration.

MSI now has its own complementary RAMDisk utility.

The company couldn't find the space to drop an OC Genie button on its Z87I, but the software application still works. It set our CPU to 4 GHz at 1.10 volts, and offers several pop-up menus for additional adjustments.

The Motherboard icon opens a pop-up menu labeled Voltage that also has a ring ratio setting. MSI's CPU icon opens a redundant fan setting pop-up, and the Memory icon opens a DRAM Timings page. Voltages are software-adjustable within firmware limits, but changes to DRAM timings prompt a reboot or cause the software to crash.

MSI’s sensor recorder and status alerts are available from more pop-up menus.

Four more menus are devoted to system information.

Z87I Firmware

MSI surrounds its menu settings with a status bar and several large icons, though the setting fonts are still large enough to read even after shrinking from the native 1024x768 to 600x450 scale.

The Z87I didn’t reach the stable 4.6 GHz at 1.25 V we were expecting from our processor, but it got close at 4.55 GHz using a 45x multiplier and 101 MHz BCLK. Our XMP-3000 memory was also properly configured within memory multiplier limits at 22x 1.33x 102.3 MHz.

Our 1.25 volt CPU core target was reached at the Z87I’s 1.240 V setting, and the memory’s 1.65 V rating was similarly reached at the board’s 1.635 V setting.

Primary, secondary, and tertiary memory timings are all configurable over a wide range, with our XMP-3000 defaults shown in grey text.

The Z87I also features DRAM training configuration.

MSI’s DigitALL Power menu still contains CPU Vdroop offset, though this has primarily become an integrated function of the CPU’s input-to-core voltage controller. Default settings gave us excellent voltage stability.

Test Hardware And Benchmark Settings

Thermalright’s classic MUX-120 remains competitive with our recent review of high-end heat sinks, even when using its original clip-on mounting system. It’s that easy-to-mount mechanism that wins me over in motherboard round-ups, and the good performance points to a good design.

Alternatively, we can point to problems with the Core i7's heat spreader as a reason why larger coolers couldn’t give us significantly better thermal performance in that review.

G.Skill’s F3-17600CL9D-8GBXLD is the only memory kit in our lab that defaults to our DDR3-1600 CAS 9 test standard. Faster RAM always uses slower defaults, and slower RAM requires XMP to get there. The problem is that some boards automatically enable other overclocking features when XMP is enabled. Consistency rules these tests.

We replaced the slower memory with G.Skill’s DDR3-3000 kit for our overclocking stability tests.

Corsair sent its 80 PLUS Plantinum-rated AX860i for our benchmark needs, citing enhanced support of Haswell's C7 state.

Results: 3DMark And PCMark

Tighter integration makes it difficult for any product to stand out in a performance comparison, though companies can still cheat a little on base clock or force a modified Turbo Boost mode, where the highest (single-core) turbo ratio is applied even to quad-core loads. We disable that tomfoolery when settings are available, and enable Intel’s full power-saving feature set to the best of each motherboard’s capability. That means benchmarks do nothing more than show if anyone is cheating or accidentally crippling their product.

Storage scores are the only synthetic performance measurements that carry though to our final analysis. They range from 4993 to 4997 will have little overall impact there.

Results: SiSoftware Sandra

Gigabyte’s small lead in Sandra's Arithmetic module appears strange, especially since we locked the board in at an exact 100 MHz base clock and made sure that its full range of Turbo Boost ratios were represented at various loads. EVGA’s stumble in Sandra's Cryptography test appears equally out of place, and retesting confirmed that this data was consistent.

Suddenly, EVGA's finish in the Cryptography test makes a lot more sense. Poor memory performance curbs the rate at which our CPU can churn through AES-related instructions. Our Z87 Stinger operates at the same DDR3-1600 CAS 9 defaults as the other boards, but perhaps EVGA is playing with other timings to boost memory stability?

Results: Audio And Video Encoding

We try not to worry about one-second differences in these benchmarks, since very small absolute values can cause rounding errors anyway. However, the two-second delta between ASRock's and EVGA's samples in Lame is at least noteworthy (though still tiny).

Results: Adobe Creative Suite

EVGA’s Z87 Stinger takes a few more small victories in Adobe Creative Suite. This is the same board that had trouble in a couple of our synthetics, so the company is obviously doing something different.

Results: Productivity

MSI's Z87I slightly leads in 3ds Max and ABBY FineReader, but again by an amount that you’d probably need a performance chart to recognize. EVGA’s Z87 Stinger sweeps Visual Studio by a similar margin.

Results: File Compression

The Z87 Stinger leads most of our file compression suite, but takes a noticeable stumble in the OpenCL-accelerated WinZip benchmark.

Energy, Heat, And Efficiency

The Z87E-ITX puts ASRock in the lead for lowest idle power consumption, while EVGA registers the highest idle and lowest full-load power numbers. Inconsistent Turbo Boost behavior could explain the small leads in a few benchmarks, though the Z87 Stinger's overall performance is just 1% above average.

Gigabyte’s Z87N-WiFi demonstrated high full-load power consumption, and showed us similarly higher core temperatures. With its processor interface nudged up close towards the graphics card, we credit cooler orientation with part of the heat increase.

High full-load power consumption hurts Gigabyte’s overall efficiency, while low full-load power consumption boosts EVGA’s average efficiency in spite of idle power issues.

Overclocking

Five of the six motherboards we tested pushed our Core i7-4770K sample to 46 x 100 MHz stably under an eight-thread, AVX-optimized Prime95 load. MSI’s Z87I was down by half a tick at 45 x 101 MHz.

We generally limit ourselves to multiplier-based overclocks whenever possible, so the only base clock that matters to most of us is at the processor’s 100 MHz strap. That’s because locked cores also have locked straps. A non K-series processor can still get a 9% overclock from Asus’ Z87I-Deluxe, in addition to the four 100 MHz bins of headroom available to those processors.

ASRock falls slightly behind the top three boards in DRAM overclocking, but Gigbyte’s Z87N-WiFi had a more-serious issue of instability at high memory multipliers. But we’re not sure this matters, since our recent DRAM tests showed that timing-optimized DDR3-2133 offers slightly better performance than timing-optimized DDR3-2400 with Haswell-based CPUs.

Which Mini-ITX Motherboard Would We Buy For Haswell?

Who has the best mini-ITX-based motherboard? And which board offers the best value? Can those two competing concepts be tied together in a single product?

Gigabyte and MSI present significantly similar platforms, which offer features and faults that largely mirror each other. Naturally, they sell for almost identical prices, too. The Z87N-WiFi had a tough time pushing memory data rates beyond 2400 MT/s, while the Z87I was over 50 MHz down on CPU core overclocking. Recent memory tests indicate that CPU core frequency is the more important variable, though anyone who disagrees will want to see these motherboards achieve a value tie. Our Smart Buy recognition doesn't have to be exclusive in a round-up, so they both receive our recognition.

At the other end of the pricing scale, EVGA’s Z87 Stinger includes fewer features and sells at a higher price. This leaves Asus and ASRock competing for best-of-the-best status.

Both boards are also equipped similarly, with Broadcom’s 867 Mb/s 802.11ac controller topping their feature sets. The Z87I-Deluxe adds an overkill voltage regulator suitable to processors twice this size, but the Z87E-ITX has a hidden mSATA slot that some builders will love (in addition to its adequately-oversized voltage regulator).

The Z87I-Deluxe’s beefier power circuitry costs more than ASRock’s alternative sneakily-placed SATA connector, and probably justifies its $30 price difference. Then again, Asus would tell us that the hardware needed to enable its USB BIOS Flashback feature is also worth a chunk of that difference, even if most of us won't have occasion to use it. The Z87E-ITX comes out on top of a practical features-to-price comparison, while Asus’s more-expensive extras finish first in a cost-versus-price analysis.

When Tom's Hardware looks to crown a product the best in a category, however, price gets limited weighting. Asus' features are probably more valuable than ASRock’s, even though we're asked to pay more for them. Nevertheless, the Asus Z87I-Deluxe receives our coveted Tom's Hardware Elite award.

On the other hand, the flexibility of eSATA and mSATA connectors on ASRock's Z87E-ITX truly does make it difficult to crown another board unquestionably superior. So, competing closely for top-value and top-features, many of us would rightly consider the ASRock platform as a stunning offering as well. The only award we have left to bestow on the Z87E-ITX is Tom's Hardware Approved, which is recognition from our editor of a job well done.

Introducing The Lian Li PC-Q30

Lian Li's PC-Q30 chassis is specifically designed to show off the guts of your PC in a place where everyone is going to see them. It boasts a unique shape, a large window, and, surprisingly, more space than you would think a small chassis could offer. At the very least, it's a conversation piece.

Because of its unique form factor and purpose, we thought we'd do more with this case than simply list off its pros and cons. With a little bit of effort, you can turn this thing into a striking little example of how proud you are to be a PC gamer.

The Lian Li PC-Q30 is built around the mini-ITX form factor. Of course, this limits the amount of technology you can shove inside. But don't underestimate the enclosure's roominess, either. We approached this build with AMD's most powerful APU, effective cooling, and a mind for aesthetics. We're going through each component and how to get it installed. If you want to build a system like this, you'll be able to walk through with us.

This case will never be the final resting place for all of your gaming gear. But since it does accommodate discrete graphics, we're also looking at some add-in options. At the end of the build, though, I ended up yanking all of the cards I tried and stuck with just the APU. Even still, it was interesting to see how fast this system could go (hint: it's further than you might think after a cursory glance). Fortunately, AMD's processor is quick enough for light gaming, and I wasn't trying to build something high-end. It might even turn out that Lian Li's PC-Q30 is a good foundation for a do-it-yourself Steam Machine.

Packaging

You (and whoever delivers your package) will immediately marvel at how light the PC-Q30 really is. It comes in a familiar brown cardboard box that's thick and sturdy enough to protect the aluminum contents from even the most mean-spirited couriers.

Inside, the PET foam used to hold the case in place is far superior to the Styrofoam we're accustomed to. You don't end up with a bunch of small white chunks to clean out of your carpet. This really should be standard packing material for all cases.

The case’s acrylic window has a sheet of recycled paper in front of it to prevent scratches.

In The Box

Apart from the case, the box contains everything else you need. There's even a piezo speaker included, which is far from expected these days. In addition to the usual screws, cable binders, and manual, Lian Li also includes a USB 3.0-to-USB 2.0 adapter that allows the case to be used with older motherboards.

Technical Specifications And Features

Lian Li's PC-Q30 employs horizontal airflow. Cool air enters through the two openings on either side of the case and is blown out the back, along with the system’s heat, via a fan. There’s no opening at the top, which means passive cooling isn't in the cards.

The power button on the front of the case features a blue LED to convey the system’s status and a red LED to show storage activity. It’s the same button used on many of Lian Li's other cases. There is no reset button.

The back of the enclosure's front cover exposes the attached cable, which you can detach. This actually makes installation easier, since the front cover can be set aside without detaching cables from the motherboard.

The front panel is more of a side. It sports two USB 3.0 and audio I/O (microphone input and headphone output).

The only potential issue we found was the PC-Q30’s four metal feet. They won't discolor the surface they're set on like the soft plastic ones tend to do, but they will scratch your desk if you're not careful. The lack of insulation against vibration didn't turn out to be a problem, surprisingly, due to the small footprint making contact.

360-Degree Video Of The Empty Case

360 Degree Video of the Empty Case

Picture Gallery

The aluminum surface looks great, of course, but it does encourage you to use thin cloth gloves during the installation process. Our sample had some annoying dust in it, so I decided to thoroughly clean it after taking my first few pictures.

Lian Li's PC-Q30 is truly unique, and I really appreciate yet another case from the company that doesn't follow mainstream convention, especially when it's well-executed like this one. You can love or hate the aesthetic, but you have to admit it's different, just like the chassis I reviewed in Take The Train: Lian Li PC-CK101, Built, Tested, And...Driven?

PSU: Be Quiet! SFX Power 350 W

Naturally, a case this small requires an SFX-based power supply. There are a number of these available, though they aren't particularly common. Our U.S. team has had good luck with SilverStone's offerings in the past. But in Germany, be quiet! is a familiar vendor. The team there went with that company's BN134, a 350 W, 80 PLUS-certified unit with a PCI Express auxiliary connector for discrete graphics.

Opening be quiet!'s SFX Power 350 Watt reveals the power supply itself, as well as a power cable, manual, screws, and cable ties. In short, it doesn't leave you wanting for anything.

The cables are sleeved well and long enough for our installation.

The SFX Power 350’s fan is audible when the system is under full load. However, most of that noise is redirected back into the case, so you won't hear the power supply over the other fans in taxing situations. If you're only using an APU, and no discrete graphics card, the PSU shouldn't be audible at all.

PSU Installation: Nimble Fingers Needed

Installing the power supply is a little annoying because the case's center of gravity shifts from the front to the back once its front cover is removed, particularly if the PSU is the only thing in there. Your best bet is to hold onto the chassis when you slide the power supply into place, or keep it off the rubber insulating strips until it's almost there.

The PSU’s cables are more than long enough, so you won't have any trouble making connections.

From there, the installation should be fairly uneventful. If you aren't using an add-in graphics card, hide the cables you don't need in Lian Li's SSD cage.

Installing The SSD

Choosing the Right SSD

Any SSD with at least 256 GB of capacity should be fine. Even a roomy hard drive is suitable if you're willing to tolerate the less responsive storage performance. We have a bunch of Samsung's 840 Pros in the lab, so that's what we used in this build. It was a nice little bonus that the drive comes with a data migration tool, allowing me to clone the contents of an old HTPC quickly.

As a side note, it doesn't matter how thick your SSD is in Lian Li's PC-Q30. Any 2.5” device will fit.