be quiet! Dark Rock TF Review

Test Results & Conclusion

Comparison Coolers

Loading...

Test Results

The Dark Rock TF may be unique among our tested downdraft coolers in scale, but that doesn’t mean there weren’t any other coolers of this height to compare to. The server-oriented NH-U9DX i4 stands 0.1” shorter than the Dark Rock TF, and though it also has tiny fans, it performed excellently in its review. It even outpaced the Dark Rock TF's thermal performance, though we expect server-oriented coolers to have some noise issues.

The Dark Rock TF’s fans run slightly below the rated RPM, due to resistance of the sink fins and backpressure from the motherboard they face. Turning the motherboard’s PWM-based fan controller to 50% duty cycle, we see that the lower air pressure allows them to run faster than 50% of the rated speed.

The Dark Rock TF is quieter than any other cooler in this comparison, and that could give it a competitive cooling-to-noise ratio as well as making it a reasonable choice for quiet environments. It spun down to produce only 19.1 decibels at 1m (calculated) when the motherboard’s fan controller was set to 50% duty cycle. The similarly tall NH-U9DX i4 produced only slightly more noise, and the included resistor wire didn’t spin it down as much.

Looking at the big-fan coolers, the Dark Rock TF has the same price as the Deepcool Assassin II and a 5% worse cooling-to-noise ratio at full speed. It’s even beaten by the tiny NH-U9DX i4, which in turn gets its improved cooling-to-noise ratio from better cooling rather than lower noise.

The Dart Rock TF can’t match the value of the DeepCool Assassin II, but that’s not really a fair comparison since the shorter downdraft cooler is designed to fit more cases. Unfortunately, it also loses to the even smaller NH-U9DX i4. And that might not be a fair comparison at “low” fan speed, since I used Noctua’s resistor wires rather than 50% PWM setting. I did this because I always want to include the manufacturer-supplied fan control in the results, whenever provided. The cost of controllers is part of the price of the product, after all.

Without resistor wires and the motherboard’s fan controller set to 50% duty-cycle, the NH-U9DX i4 allowed CPU temperature to rise to 75° above ambient (Dark Rock TF was hotter at 76°). Sound pressure level dropped to 18.8 decibels (Dark Rock TF was louder at 19.1 dBA), the cooling-to-noise ratio rose to 22% above average (Dark Rock TF was worse at 18%), and NH-U9DX i4 value rose to an incredible 58% above average. Half of that extra value is due to a 25% lower price.

Of course the NH-U9DX i4 does have a few shortcomings, such as its inability to support anything other than the square and narrow ILM version of Intel’s LGA 2011 (including v3). Noctua sells a completely different model called the NH-U9B SE2 for other form factors, yet our review of the NH-U9DX i4 showed that these two products differ only in their mounting kits. That is to say, Noctua’s customers need two different model numbers to figure out what to buy instead of the Dark Rock TF from be quiet! What does this all mean for the Dark Rock TF?

Conclusion

Entering the review process with the assumption that a product will win our stamp of approval, I lose that enthusiasm only when that product comes up significantly short of competing samples. The lack of any similar capacity downdraft coolers to compare should have made it easy for the Dark Rock TF to reach that mark, yet today’s review showed that a 92mm cross-draft cooler can beat a 135mm downdraft cooler in temperature, noise, price and size.

MORE: Best Cooling
MORE: Best CPUs
MORE: Best Cases
MORE: All Cooling Content

Thomas Soderstrom is a Senior Staff Editor at Tom's Hardware, covering Cases, Cooling, Memory and Motherboards. Follow him on Twitter.

Subscribe to Tom's Hardware on Facebook and Google+, RSS, Twitter and YouTube.

This thread is closed for comments
21 comments
    Your comment
  • unixwizard
    I am not certain from the information in this article how the testing was conducted, but I have some serious concerns that the information given here doesn't paint a whole picture of real world effects.

    Lets take one example: Many articles on air and water coolers note that airflow over VRMs and other MB components could be significantly different, yet I never see readings of MB temps, VRM temps, or anything other than CPU. How much of an effect is it? Show the data.

    Likewise, as you mention in text, a cooler like this could perform best when it can vent air directly out the side of a case. You lack any data showing this however. As far as I can tell, the usual rigor for testing isn't applied to coolers.

    In my opinion you should test coolers in a handful of configurations. Maybe pick some of the most popular cases or best performing cases in a handful of configurations and use these as a standard. An open test bench is the one configuration that you should never use, since it will never be similar to an actual usage scenario. Feel free to focus on only a few scenarios, like 'big high airflow case for overclocking to the max on air", and "balanced temp/noise performance in a mid-sized size vented case", and "space restricted case with little venting" or whatever, but once you pick the scenarios, please report more than just CPU temp and noise. Report VRM temps, chipset temp, RAM temp, and use a couple temp probes to measure exhaust temp out the side/back, etc. Better yet, show some thermal images of the cooler/MB in operation if you can. Also, make sure you let the case/system 'heat soak' under load so we know how it works under sustained load rather than just a quick heat pulse.

    Do this for a few coolers of differing type and see if the picture changes or more is revealed with this method. If it isn't then at least you'll have data to say that more detailed reporting doesn't create a different picture, but I have a suspicion that differing coolers will change positions considerably in the various scenarios, revealing strengths and weaknesses that matter to people looking to buy them for the specific scenario tested.

    As an example, I own a Lian-Li V2000B and a Rosewill Throne, both large cases but very different in scenario. The V2000B is a sealed through-flow case with no side vent and the Throne is side/top vent with little airflow out the back. The cooler in this article might work wonders in the Throne, perhaps even better than the other coolers, and be terrible in the V2000B.

    Just my 2 cents, but I don't feel I get enough out of these cooler articles to make a good purchase decision.
  • Crashman
    2005870 said:
    I am not certain from the information in this article how the testing was conducted,
    In a complete machine. It appears the site removed this information, so here's an earlier article that contains it:
    http://www.tomshardware.com/reviews/deepcool-gamer-storm-captain-240-liquid-cpu-cooler,4158.html
    2005870 said:
    but I have some serious concerns that the information given here doesn't paint a whole picture of real world effects.
    Do tell!
    2005870 said:
    Lets take one example: Many articles on air and water coolers note that airflow over VRMs and other MB components could be significantly different, yet I never see readings of MB temps, VRM temps, or anything other than CPU. How much of an effect is it?
    This review discusses the same issues. Perhaps you didn't read the text? It also charts the voltage regulator temperature. Perhaps you didn't read the chart legend?

    PWM means Pulse Width Modulator, which is a far-more-accurate name than "Voltage Resistor Module". This measurement is shown in the chart and discussed below the chart.
    2005870 said:
    Show the data.
    We did. Sorry you missed it.
    2005870 said:
    Likewise, as you mention in text, a cooler like this could perform best when it can vent air directly out the side of a case. You lack any data showing this however.
    Thats because we haven't found a specific case to fit every specific cooler. The generic statement about a downflow fan working best when it can draw air through the side panel comes from prior experience with cases designed specifically for that purpose, but cases aren't designed that way any longer. It's mention as a suggestion to people able to find such a case, or modify it appropriately
    2005870 said:
    As far as I can tell, the usual rigor for testing isn't applied to coolers.
    Cooler tests take less time than motherboard tests, but is that really a problem? We've attempted to create the "most generic" airflow configuration for the case and even tested the cooler at max/min CPU fan setting to show its operating range.
  • photonboy
    My beef:

    I always find reviews for CPU coolers difficult to decipher. Different fan speeds, noise, cooling arrrggh.

    *What I personally want is a simple chart showing cooling performance at say a total 35dB for the system or whatever.

    Go ahead and include whatever charts you want, but I'm about cooling-to-noise ratio and that needs to be easy to find.
  • Crashman
    67821 said:
    My beef: I always find reviews for CPU coolers difficult to decipher. Different fan speeds, noise, cooling arrrggh. *What I personally want is a simple chart showing cooling performance at say a total 35dB for the system or whatever. Go ahead and include whatever charts you want, but I'm about cooling-to-noise ratio and that needs to be easy to find.
    I know it's not the easiest thing to decipher, but the cooling-to-noise chart is in there.

    It shows how much better or worse each cooler's result was compared to the average of all tests.
  • AndrewJacksonZA
    Thanks for the review Crash.
  • Crashman
    545051 said:
    Thanks for the review Crash.
    You're welcome!
  • milkod2001
    Nice review. Looks like Noctua is still the way to go.

    On the side note: Could you guys do tests of some PCspeakers. Im looking for replacement of my Logitech Z 2300. Im looking for something what could match it price/performance wise. Would anyone had some recommendations?

    Thanks
  • firefoxx04
    Overpriced. I dont care if it is smaller than the D15 and Phanteks dual 140mm cooler.., its not worth that much more.
  • firedust
    You mentioned that the review includes voltage regulator temperature. I must be blind because I can't find it anywhere. Is it the first graph?
  • Crashman
    382747 said:
    You mentioned that the review includes voltage regulator temperature. I must be blind because I can't find it anywhere. Is it the first graph?
    Yep, PWM is Pulse Width Modulator as described where you saw me say it above.
    https://www.youtube.com/watch?v=DWZYS4-09wo
  • firedust
    Ah OK. I was confused because I thought Pulse Width Modulation was a term used for only for case fans. I didn't know it could be used to describe motherboard temps. Thanks.
  • photonboy
    382747 said:
    Ah OK. I was confused because I thought Pulse Width Modulation was a term used for only for case fans. I didn't know it could be used to describe motherboard temps. Thanks.


    I'm not aware of any way to describe temperature with PWM.

    Pulse Width Modulation - basically you are pulsing a magnetic signal against part of the fan motor. This is almost exactly like spinning one of those kiddy carousels or whatever they are called by hand. Push. Wait. Push. Wait.

    The length of time you push (width) is proportional to the fan speed.

    That's opposed to Voltage (3-pin) which sends a constant voltage (varying by desired temperature). PWM uses less power.
  • Crashman
    382747 said:
    Ah OK. I was confused because I thought Pulse Width Modulation was a term used for only for case fans. I didn't know it could be used to describe motherboard temps. Thanks.
    It's not being used to describe a temperature, its the type of voltage regulator used in modern motherboards.

    The old term "Voltage Resistor Module" doesn't work because it's not modular and doesn't use resistors. But back in the 1980s and early 90's...
    67821 said:
    I'm not aware of any way to describe temperature with PWM.

    That's because there isn't.
    67821 said:
    Pulse Width Modulation - basically you are pulsing a magnetic signal against part of the fan motor. This is almost exactly like spinning one of those kiddy carousels or whatever they are called by hand. Push. Wait. Push. Wait. The length of time you push (width) is proportional to the fan speed. That's opposed to Voltage (3-pin) which sends a constant voltage (varying by desired temperature). PWM uses less power.
    Ready to have your mind blown?
    Think about the fan analogy you just made, the speed change is buffered by the inertia of the fan.
    PWM voltage regulators work in a similar way, using coils and capacitors to even out the pulses. A shorter pulse and longer pause gets leveled out to a lower voltage, a longer pulse and shorter pause gets leveled out to a higher voltage. Multiple phases make the current even smoother.
  • Moktah
    Does a heat pipe cooler with the motherboard in a horizontal position work more effectively than the same cooler in a vertical position? My theory is that horizontal (a test bed for example) is more effective since it will have a higher chance of liquid migrating to the contact point of the processor. Most effective heat transfer is via phase change between liquid to vapor at the contact point of the cooler resulting in heat being absorbed from the processor. Phase change from vapor back to liquid releases the captured heat in the radiator portion. Liquid migrates back to down via differences in density (maybe easier to picture it due to gravity). It looks like there could be instances of less liquid in the actual contact point of the cooler (in a typical vertical motherboard configuration) therefore possibly causing less heat transfer. This is going to depend upon design from each cooler manufacturer and the amount of liquid contained within the heat pipes of course. To better visualize, picture a glass of water half full. There is a 100% chance of liquid at the bottom (horizontal motherboard configuration will have a cooler with a 100% chance of liquid state at the bottom where the contact point is) What if the liquid in the cooler in a vertical motherboard is at the 'half way' point during operation. You lose a bunch of effectiveness. If you get a chance, please check it out and see if there is a difference between vertical and horizontal orientation. Thanks for the article by the way.
  • Crashman
    2121809 said:
    Does a heat pipe cooler with the motherboard in a horizontal position work more effectively than the same cooler in a vertical position?
    Most of the people following the thread understand the theory. Our tests haven't shown a big difference in recent coolers, and that's probably due to a wicking material in the base
  • firedust
    Thanks for the clarifications Crashman. Am I right to assume that blow down coolers don't cool the motherboard any better than tower coolers?
  • Crashman
    382747 said:
    Thanks for the clarifications Crashman. Am I right to assume that blow down coolers don't cool the motherboard any better than tower coolers?

    They should, but in this case it didn't. I can only guess that the crossdraft cooler was moving sufficient air over the voltage regulator sink. It is a short cooler and a tall sink. And then there's the downdraft problem of air not having anywhere to go when it hits the board: voltage regulator design and DRAM placement probably increases the problem for this board (by sending the warm air back up into the fans).
  • cgigoux
    I apologize if this sounds like a dumb question, but here goes. Given that hot air rises relative to cold air, why not just reverse the fan orientation so that the "blow down" coolers blows upwards (no way I was going to say blows up!) instead? It would seem in theory that having the heat pipes aligned horizontally as Moktah suggested combined with the fans pushing the hot air upwards would make it more efficient. Have this type of cooler ever been tested this way?

    Thx
  • Crashman
    2098962 said:
    I apologize if this sounds like a dumb question, but here goes. Given that hot air rises relative to cold air, why not just reverse the fan orientation so that the "blow down" coolers blows upwards (no way I was going to say blows up!) instead? It would seem in theory that having the heat pipes aligned horizontally as Moktah suggested combined with the fans pushing the hot air upwards would make it more efficient. Have this type of cooler ever been tested this way? Thx
    This seems like it would work, but the air is warmer by the board. In this build I think the DIMMs would block air to the base, but it might work with a board designed to support it.
  • Compuser10165
    For 80$ which is the current price for this cooler on newegg better to go with th 89$ Dark Rock Pro 3 (if you can fit it). P.S. I have the Dark Rock Pro 3 (along with an i7 6700k) cooler and it performs very well (cpu actual temp is around 40 degrees celsius on idle and around 50-60 at load) this in the bequiet silentbase 800 case. P.S. CPU utilization around 30% given the I7.
  • Crashman
    2080651 said:
    For 80$ which is the current price for this cooler on newegg better to go with th 89$ Dark Rock Pro 3 (if you can fit it). P.S. I have the Dark Rock Pro 3 (along with an i7 6700k) cooler and it performs very well (cpu actual temp is around 40 degrees celsius on idle and around 50 at load) this in the bequiet silentbase 800 case.
    Well then, here you go!
    http://www.tomshardware.com/reviews/be-quiet-dark-rock-pro-3-cpu-cooler,4350.html