Three Thermalright True Spirit Heat Sinks, Reviewed

Round-ups
By Kai Tubbesing
published

Thermalright sent over all three models in its True Spirit heat sink family, complete with 92, 120, and 140 mm fans. We dropped each one onto an FX-8350 processor to gauge its cooling performance and acoustic output. Which one takes the value win?

Benchmark System

Despite their size differences, all three of the models we tested are able to keep AMD's FX-8350 cool at stock clock rates. The two larger coolers will participate in a second round of benchmarking with an overclocked processor for a chance to qualify for a place in our separate Overclocking Charts.

Swipe to scroll horizontally
Benchmark System CPU Cooling, Normal And Overclocked
ProcessorAMD FX-8350 (Vishera) 4M/8T, 8 MB Shared L3 Cache, 4 GHz Base Clock Rate, 4.2 GHz Maximum Turbo Core
AMD FX-8350 (Vishera) All Cores at 4.4 GHz using 1.4325 V
MotherboardGigabyte 990FXA-UD7, Socket AM3, 990FX North Bridge, SB950 South Bridge, BIOS F10
System Memory1 x 4 GB G.Skill DDR3-1333 CL9
Graphics CardXFX Radeon HD 5450 (Passive)
System SSD60 GB Kingston SSDNow V+ 200
Power Supply1000 W Xilence XQ Series Platinum R4 (Semi-Passive)
CaseCooler Master CM Storm Stryker
Case Fans Front and Back: 1000 RPM
Case Fan Top: 900 RPM
OtherAqua Computer Aquaero 5 LT (Firmware 1027)
Arctic Cooling MX-4 (Thermal Paste)
Operating SystemWindows 8 Pro 64-Bit  (Version: May 2013)

The processor’s temperature is measured after loading (and heating) it up for 60 minutes using Prime95 with user-defined settings. Theose thermals are logged for 10 minutes, and the results are averaged. This is more accurate than just measuring at one specific point in time. It should be noted that the processor’s temperature sensors only provide whole numbers, which makes the results a bit less precise than if we had a tenths or hundredths place.

Noise levels are measured with an open case from a distance of 30 cm at a 90-degree angle to the CPU cooler and at the level of the middle of the CPU cooler’s fan. Don’t be alarmed if the numbers seem high; they would be lower in a closed case. Measuring the noise level this way gives us more precise numbers due to reduced interference from ambient noise.

Kai Tubbesing
28 Comments Comment from the forums
  • expl0itfinder
    I would love to see this compared against a wider selection of air coolers. For instance, the Cooler Master Hyper 212, or other fan favorites. Comparing it to 1 or 2 other brands does not give us a lot to look at.
    Reply
  • Novuake
    A single other well known cooler for reference would have been great for easy context.
    Not all sites and reviews used Delta temps.
    Nice review but kind of renders it moot for comparison out of the Silverstone and THermalright Scope.
    Reply
  • Someone Somewhere
    I think you got a bit lost on your graphs. At a guess, you meant to have one as distance from ambient, and one as absolute, or possibly there's idle readings.

    Also, please stick to the same units.

    More comparisons would have been nice.
    Reply
  • SpadeM
    There are a few issues /omissions in the article, first being that the motherboard cap is only used on 939 sockets (i did an install for 970A-D3 AM3+ and it didn't require it) also the anchoring plate is flawed in my opinion because the screw holes that hold it in place aren't actually holes, but more live grooves on the outside of the plate and so the screws actually hold the plate in place with only half of their screw cap. And finally, for users that aren't experienced with multiple cpu heatsinks, tightening the final plate that locks the cooler in place ca very well damage your processor since there's no limiter in place.

    Other then that, those are some fantastic coolers, even though the thermal probe on FX-8350 under windows gives me flawed measurements in idle, under load it's quite good.
    Reply
  • iam2thecrowe
    2000 rpm is noisy? i used to have a 6000rpm fan on my old athlon xp....... I think people these days are too picky about noise.
    Reply
  • Someone Somewhere
    11299459 said:
    2000 rpm is noisy? i used to have a 6000rpm fan on my old athlon xp....... I think people these days are too picky about noise.

    Diameter has a lot to do with it - it's more to do with tip velocity than actual RPM. Your fan was likely about 40-50mm.

    I've got 40mm fans that are near silent at 3K RPM. You can get 40mm fans that do 13K.
    Reply
  • ubercake
    Whether it's the motherboard or the heat sink manufacturers... We're still seeing 3rd-party CPU heat sinks and their fans blocking RAM slots for a decade. There's not much innovation going on here. I would tend to think the heat sink manufacturers need to accommodate the current motherboard designs. If they could solve that part of it in a full-size effective heat sink solution, innovation will have taken place. Until that time we'll keep seeing copper this; nickel-plated that; aluminum fins here; heat pipes there; one fan on this one; two fans on that.

    Same old stuff; different day.
    Reply
  • Someone Somewhere
    There's only so many ways you can build a MB, or a cooler...
    Reply
  • Myrkvidr
    @Someone Somwhere: Sorry, there's a copy&paste issue in the 2nd and 4th chart on page5: It should say "CPU temperature at 20°C ambient" while the 1st and 3rd chart are Delta temps.

    @SpadeM: Seems like I just did not RTFM close enough ;) But the measurement results were not affected by using or not using the plastic cap (I ran two separate measurement series). The anchoring plate is sitting very tight - I installed a lot of Thermalright heatsinks during the past couple of weeks. The contact pressure between the True Spirit heatsinks and the CPU is not excessively high, so it shouldn't cause any damage to the CPU.

    @all: We just startet off with the new system for CPU cooler testing and had to start somewhere - there will be more results and coolers coming soon :)
    Reply
  • mironso
    I hope they will fix this issue with memory in next gen on 140 series.
    Reply