Arctic Cooling NV Silencer 5 GPU Cooler Review!

Discussion in 'Reviews & Articles' started by News Writer, Nov 5, 2004.

  1. News Writer

    News Writer News Writer

    With GPUs producing so much heat, it is inevitable that third-party GPU coolers are becoming more and more popular. However, some GPU coolers sound like jet engines at full blast! Can we still get good cooling without inducing hearing loss?

    Today, we take a close look at the new Arctic Cooling NV Silencer 5 GPU cooler for NVIDIA GeForce cards! Let's see if this GPU cooler actually lives up to its name!


    Link : Arctic Cooling NV Silencer 5 GPU Cooler Review!
    Last edited: Nov 11, 2004
  2. Maximus_Detritus

    Maximus_Detritus Moderator

    PsY, you cant remove your graphics card?? :shock:
  3. PsYkHoTiK

    PsYkHoTiK Admin nerd

    Now I managed to remove it... :thumb: :thumb: :thumb:
  4. Great review.

    Glad you went to all the trouble of looking at the copper plate properly. So it turns out that the plate is made from separate parts and not just one piece of cooper....mmmmm......very interesting.
  5. JayN1

    JayN1 Newbie

    I cant help to notice that its a faulty method youve been using to compare the stock cooler and the NV Silencer 5.

    Besides the wrong method of calculating delta T, youve also made a mistake as 57 - 28 = 29, not 31.

    The correct method of finding the difference between the two coolers would be

    31 + 53 = 84 and

    28 + 57 = 85

    Which means the difference is 1 degree, not 9.

    The difference in idle state is obviously 3 degrees as 57 - 54 = 3

    I would use a similar method to calculate load difference.

    34 + 66 = 100 and

    28 + 78 = 106

    the difference is 6 degrees and not 18

    and the difference at full load is 8 degrees.

    The other method is simply wrong and doesnt give a good comparison between the two coolers.
  6. JayN1

    JayN1 Newbie

    As an example of why your method is wrong, lets say temperatures idle was

    40 and 60 = 20 for the stock, and

    30 and 50 = 20 for the silencer

    Difference is 0. Which means they cool equally well? - Not!

    It is logical that the silencer in my example is 10 degrees cooler in both results, yet the difference with your method would be zero (0).

  7. Chai

    Chai Administrator Staff Member

    The calcuation was a mistake. We will append it ASAP. Thanks for notifying.

    But as for your method, I can't see how that is right.

    In the most ideal room temperature, it should be constant, and never change because room temperature should be controlled. But since we don't have that sort of facility, we need to measure the ambient temperature changes.

    Assuming that NV Silencer 5 is 10C than stock:
    That is impossible. Air cooling depends 100% on the ambient temperature. If the ambient temperature is 10C cooler, the GPU temperature should be 10C cooler too, assuming there is no other factors. Don't forget, the best air cooler can do is to cool down to room temperature, never lower than room temperature because it's not possible, which is why we calculate the delta between the room temperature and the GPU temperature.
  8. JayN1

    JayN1 Newbie

    I see. You learn something new every day. I didnt know that a component (gfx, cpu) inside the case never can go below ambient. :think: (not sarcastic)
  9. Olle P

    Olle P Newbie

    I agree with this one mistakte, but then...
    Huh? Adding T(ambient) to T(GPU), what's the logic behind that?
    It isn't wrong, it just doesn't provide the entire truth.

    My objections/question to the article:
    1) Room temperature (air temperature into the computer) wasn't recorded.
    2) How was "ambient" temperature recorded?
    "All of the temperatures were obtained using the GeForce 6800 GT's on-die thermal sensor."
    This one shouldn't be able to measure ambient temperature, so my guess is that the motherboard sensor was used. That sensor isn't in the path of airflow to the GPU cooler.

    Now, assuming room temperature was about constant, within +-1 degree, for all measurements the one conclusion one can draw is that the VGA Silencer is better at cooling the GPU with that (and similar) hardware setup(s).

    If we also assume that the motherboard was used for measuring ambient temperature we can see that the VGA Silencer "steal" cool air that would otherwise have been used to cool the CPU, thus increasing the temperature in the upper part of the cabinet (as can be seen in the "ambient" readings). That ambient reading is totally useless to compare dT since it doesn't measure what goes into the fan, but rather what comes out (of the stock cooler)!
    I bet my hat that the CPU temperature was higher with the VGA Silencer than with the stock graphics cooler!

    Therefore the correct conclusions, given that my assumptions above are true, are:
    - The VGA Silencer is very good at cooling the GPU.
    - It does so by reducing cooling capacity for the CPU.
    - Make sure you have a sufficient airflow around the CPU cooler if you install a VGA Silencer!

  10. Chai

    Chai Administrator Staff Member

    I think an external thermometer was used to measure the ambient temperature. I might be wrong.
  11. PsYkHoTiK

    PsYkHoTiK Admin nerd

    A external thermometer was used.... :mrgreen:
    I didn't use the onboard ambient temp sensor. I only had one thermometer that time. Now I have two. :mrgreen:

    Because I had only one thermometer, I had to decide which of the onboard sensors were more reliable. I felt that the ambient sensor was a tad bit high. I attached the thermometer to the bottom of the NV Silencer 5 to check the temps, and it showed results that was consistent with the on die sensor. So I went with the thermometer taking temps for ambient. :D

    Well, you would need sufficient airflow to cool anything right? :)
    The NV Silencer 5 also reduces the overall case temps coz the heat from the GPU is exhausted out of the case and not recycled back into the case. ;)
    Last edited: Jan 21, 2005
  12. Olle P

    Olle P Newbie

    Where was that sensor positioned?
    If room temperature was about constant it seems strange (counterintuitive) that the air sucked into the VGA coolers would be warmer for the VGA Silencer.

  13. PsYkHoTiK

    PsYkHoTiK Admin nerd

    The probe was placed below the 6800GT. I used it to get readings from where air would be sucked into the cooler.

    The reason why we use the (t2-t1) is because of fluctuating room temperatures. I did that in summer and it got reasonable warm (especially coz someone turned off the ac! :mad: ) By using the t2-t1, we can still get an objective score as it would eliminate any impact the ambient temp would introduce into the scores. :mrgreen:
  14. Chai

    Chai Administrator Staff Member

    The simplest explaination for this is probably because the ambient temperature changes a lot.
  15. siddiq

    siddiq Newbie

    hmm. interesting. using seperate copper plates. :think:
  16. Olle P

    Olle P Newbie

    That's the way to do it! (Should have been mentioned in the review though.)
    Fluctuating room temperatures is horrible when making this kind of comparisons. You still obviously did as good as can be expected under those circumstances.

    Comparing t2-t1 does provide a useful result if there's a considerable difference between the two tested setups (as was the case here).
    The absolute temperature does also have an impact beyond the simple t(GPU) vs t(ambient) calculation, since increased heat means exponentially higher black body radiation, and thereby more efficient cooling. That difference is normally too small to take into comparison, given the magnitude of measurement errors. (I don't trust your measured temperatures as true, for example, but consistent enough to provide a sufficiently accurate difference in temperature between the two tests.)

    Neither the accuracy nor the precision of the two thermometres is given/known.
    My somewhat educated guess is that precision is way better than accuracy, so that the given temperatures differ within +-3 degrees of actual (true) temperature, but repeated measurements on same temperature differ no more than 1 degree.

    For the "Full load" measurements this means:
    t1(stock) = 28+-3, t2(stock) = 78+-3, Delta(stock) = 50+-6
    t1(NVS5) = 34+-3, t2(NVS5) = 66+-3, Delta(NVS5) = 32+-6
    But (because of the good precision):
    t1(NVS5) - t1(stock) = 6+-2
    t2(NVS5) - t2(stock) = -12+-2
    => Difference Delta(NVS5) - Delta(stock) = -18+-4

  17. Chai

    Chai Administrator Staff Member

    Yup, I totally agree with you on temperature measurements. Which is why we have been investing on better thermometer instead of using cheapo thermometers which merely display numbers!

    I hope Zalman VF700-Cu review which will be published soon is an improvement. ;)
  18. Rodney

    Rodney Official BOG Supporter!

    My down side to the Arctic 5

    By porting the air out the rear of the case will it not disrupt the air flow that the case was designed to for. Have you noticed an increase in CPU temperatures. I have a eVGA 6800 and the rear I/O is covering two slots the AGP and the PCI slot that makes it impossible to use the Artic Cooling Silencer 5 I wish there was a work around but I can't see it. :cry:

    PS: Loved the Book Breaking Thought the Bios Barrier. I have posted many high praises. :thumb:
  19. PsYkHoTiK

    PsYkHoTiK Admin nerd

    Hi Rodney! Welcome to the ARP forums! :thumb: :beer: :arp:

    There is a new revision of the NV Silencer 5 that comes with the PCI cover. It's designed for cards that has a solid two slot cover. You have to remove the stock one and use theirs. :mrgreen:

    I dont see why it would disrupt the case's airflow... :think:
  20. Rodney

    Rodney Official BOG Supporter!

    NV Silencer 5 Version 2?

    Is it the second Version of 5, if it is I have looked high and low for it in the US market. Do you have a clue as to where it is hiding. Thanks for the welcome and after reading the book and personal bios I feel right at home. I'll get to my Bio's this week sometime. Many kind thanks, :wave:

Share This Page