Air Cooling Threadripper (Case Study) - Overclock.net - An Overclocking Community

Forum Jump: 

Air Cooling Threadripper (Case Study)

Reply
 
Thread Tools
post #1 of 29 (permalink) Old 02-04-2019, 08:28 PM - Thread Starter
New to Overclock.net
 
jhsu's Avatar
 
Join Date: Oct 2013
Posts: 12
Rep: 0
Air Cooling Threadripper (Case Study)

So I spent a day going through different air cooling options for my Threadripper 1950x.

Originally I was using an Enermax liquid cooler because it's the only AIO I could find with full cpu coverage for Threadripper, but now I'm on the second RMA and find them totally unreliable at this point. You can read more on this thread here -> https://hardforum.com/threads/enerma...eview.1945204/

At first I installed a Noctua NH-U14S just to have something, but soon started researching other coolers and my shortlist looked like this:
Thermalright Silver Arrow (coolest/second least expensive)
Cool Master Wraithripper (2nd coolest/very expensive)
Noctua U14S (least cooling/least expensive)

It made sense to try the Silver Arrow (it was only $10 more than the Noctua), but I noticed a couple things - the Silver Arrow's fan is a lot faster than the Noctua's, and it also has more heat pipes. I wanted to see if the Silver Arrow performed better in tests because of the extra pipes, or because the of fan. While I was at it, I also wanted to try Noctua's 140mm industrial fans since they could go even faster. On top of all that I wanted to see how loud everything was in person.

Basically my questions were:
Which will perform better with the same fans?
How much effect does fan speed/fan size/number of fans/number of pipes, have on performance?
How loud does everything sound in person?
What will fit with my motherboard+case?

Currently this is my system:
Case: Corsair 300R
Motherboard: ASRock x399 Taichi TR4
CPU: AMD Threadripper 1950x
RAM: G.Skill Ripjaws 4 series 32 GB (4x8GB) DDR4-3000

The parts list that I tested:
Noctua NH-U14S TR4 (6 heat pipes)
Noctua NF-A15 (stock fan, 140mm/120mm holes, 1500RPM)
Noctua NF-A14 iPPC-3000 (140mm, 3000RPM)
Thermalright Silver Arrow TR4 (8 heat pipes)
Thermalright TY-143 (140mm/120mm holes, 2500RPM)
* I list the hole dimensions because they ended-up being a major factor

*** SPOILER ***
The quick answer for those who don't want to read through, my final choice was to keep the Noctua NS-U14S cooler and replace the fan with Thermalright's TY-143 140MM fan. The main factor for my case being the max CPU cooler height of 170mm. If your case has a higher max, then you could consider the NF-A14 industrial or similar higher performing 140mm fan with a normal frame. (see below for height clearance photos)

Last edited by jhsu; 02-17-2019 at 05:21 PM.
jhsu is offline  
Sponsored Links
Advertisement
 
post #2 of 29 (permalink) Old 02-04-2019, 09:21 PM
Tetrapyloctomist
 
Aenra's Avatar
 
Join Date: Feb 2017
Posts: 1,236
Rep: 15
I never understood these kind of posts..
If you have the answer to your question (you even mention it), why post it as a question in the first place?

I mean, free time to waste is nice, but.. that much free time? Not so much.

Pride, honour and purity.
Aenra is offline  
post #3 of 29 (permalink) Old 02-04-2019, 09:25 PM - Thread Starter
New to Overclock.net
 
jhsu's Avatar
 
Join Date: Oct 2013
Posts: 12
Rep: 0
Air Cooling Threadripper (Cont'd)








Methodology:
I ran everything with the side panel open and the case fans off. I used a Prime95 custom torture test with Min/Max FFT size set to 96 and in-place. I started with a 3 min warm-up period, reset the values, then let it run for +10 min recording the average values for clock speed, voltage, Tdie temp, and fan speed.

I used the mic on my phone to measure dB levels placed 12 inches away from the fans, so I can't vouch for it's absolute acccuracy, but useful enough for relative comparisons.

I then subtracted the ambient temperature from the Tdie temps to find the temperature above ambient (ΔT).

Disclaimer: I only have a digital thermometer that reads temps in whole degrees Fahrenheit, so margin of error probably spans a few degrees.


FANS.......................mm.....RPM
-----------------------------------------
NF-A14 iPPC-3000.....140.....3000.....Noctua Industrial
TY-143.....................140.....2500.....Thermalri ght (stock)
NF-A15....................140.....1200.....Noctua (stock)
UCTBP12P-C.............120.....2300.....Enermax


COOLERS
-------------
NH-U14S (12 heat pipes)
Silver Arrow (16 heat pipes)



NH-U14S Stock.............3.3GHz......3.9GHz
Voltage (V)...................1.053........1.153
Temp. Diff. (°C).............37.7..........50.2
Fan Speed (RPM)...........1236.........1402
Loudness (dB)..............38.............38



NH-U14S + TY-143........3.3GHz......3.9GHz
Voltage (V)...................1.054........1.153
Temp. Diff. (°C).............34.8..........45.1
Fan Speed (RPM)...........1982.........2336
Loudness (dB)..............40.............46



NH-U14S + NF-A14.......3.3GHz......3.9GHz
Voltage (V)...................1.053........1.153
Temp. Diff. (°C).............33.0..........45.8
Fan Speed (RPM)...........2306.........2595
Loudness (dB)..............51.............54



NH-U14S + Enermax.....3.3GHz......3.9GHz
Voltage (V)...................1.049........1.152
Temp. Diff. (°C).............37.7..........50.2
Fan Speed (RPM)...........2268.........2269
Loudness (dB)..............42.............42
=============================================



Silver Arrow Stock............3.3GHz......3.9GHz
Voltage (V)......................1.050........1.153
Temp. Diff. (°C)................36.0..........49.4
Fan Speed (RPM)..............1999.........2387
Loudness (dB).................41.............50
*Tdie max (68C) reached @ 3.9GHz



Silver Arrow + NF-A14......3.3GHz......3.9GHz
Voltage (V)......................1.051........1.153
Temp. Diff. (°C)................34.4..........49.4
Fan Speed (RPM)..............2318.........2608
Loudness (dB).................48.............49
*Tdie max (68C) reached @ 3.9GHz



Silver Arrow + NF-A14 + UCTBP12P-C (x2)...3.3GHz.........3.9GHz
Voltage (V)............................................... 1.052............1.153
Temp. Diff. (°C)..........................................32. 9.............48.1
Fan Speed (RPM)........................................2384/2131.....2671/2139
Loudness (dB)...........................................54. ...............53

Last edited by jhsu; 02-04-2019 at 09:52 PM.
jhsu is offline  
Sponsored Links
Advertisement
 
post #4 of 29 (permalink) Old 02-04-2019, 09:36 PM - Thread Starter
New to Overclock.net
 
jhsu's Avatar
 
Join Date: Oct 2013
Posts: 12
Rep: 0
Air Cooling Threadripper (Cont'd)




In the end, it appears the Noctua cooler performed better than the Silver Arrow by a fair margin when using the same fans, which suggests extra heatpipes don't necessarily improve anything. Even after stacking 3 fans onto the Silver Arrow, it still didn't do better than the Noctua with a faster fan.









Having a faster fan on the Noctua cooler seemed to be the key. This narrowed things down to either the industrial fan or the Thermalright fan. Since they essentially performed the same, I decided to go with the Thermright TY-143 because it was a better fit due to it's circular frame.

That about wraps it up, just wanted to share for others with Threadripper systems looking at air cooling options as well provide a case study for air cooling in general. Any constructive input is appreciated.

Thanks for reading

Last edited by jhsu; 02-04-2019 at 11:06 PM.
jhsu is offline  
post #5 of 29 (permalink) Old 02-04-2019, 09:57 PM - Thread Starter
New to Overclock.net
 
jhsu's Avatar
 
Join Date: Oct 2013
Posts: 12
Rep: 0
Quote: Originally Posted by Aenra View Post
I never understood these kind of posts..
If you have the answer to your question (you even mention it), why post it as a question in the first place?

I mean, free time to waste is nice, but.. that much free time? Not so much.
You seem like a nice guy.
jhsu is offline  
post #6 of 29 (permalink) Old 02-04-2019, 11:00 PM
Tetrapyloctomist
 
Aenra's Avatar
 
Join Date: Feb 2017
Posts: 1,236
Rep: 15
Quote: Originally Posted by jhsu View Post
You seem like a nice guy
Smartly phrased
I never try to be, i like practical, honest is even better, not everyone has to like me; not even most really.

As to this case in specific, i'd blame your original wording, not me. There are countless of people doing just that, posting both question and answer, only so that others waste yet more time responding; in vain mind.
I do not divine, i judge based on what's being made available to me.

* given the effort you've put into this, a few comments:
- never mix different fans in the same push-pull setup; ever.
- the above suffices for an informed opinion, but this only validates it; never use squared 'typical' fan frames (120x120, 140x140) when you're interested in max performance; even if said max performance is only of academic interest to you.
- have the TR4 SA run with two TY 143s (i run three personally), then try comparing it again to the Noctua, equally equipped.

Pride, honour and purity.

Last edited by Aenra; 02-04-2019 at 11:07 PM.
Aenra is offline  
post #7 of 29 (permalink) Old 02-04-2019, 11:28 PM - Thread Starter
New to Overclock.net
 
jhsu's Avatar
 
Join Date: Oct 2013
Posts: 12
Rep: 0
Quote: Originally Posted by Aenra View Post
- never mix different fans in the same push-pull setup; ever.
- the above suffices for an informed opinion, but this only validates it; never use squared 'typical' fan frames (120x120, 140x140) when you're interested in max performance; even if said max performance is only of academic interest to you.
- have the TR4 SA run with two TY 143s (i run three personally), then try comparing it again to the Noctua, equally equipped.
I have to admit, this is a much more constructive reply. I'm interested to know what you mean by never use typical fans if you want max performance? Are you saying the shape of the frame? I tried doing a search for other circular 140mm fans and the only 2 I could find were the Noctua and TY-143, what other type of fans are you referring to?

Yea, I wish I had time to have done more tests, but I kind of forced myself to only spend a day because it already felt ridiculous spending as much effort as I did. Even still, using 2 circular fans on the Silver Arrow wasn't an option for me because the ram chips would cause the fans to sit too high, that's why I only used 1 fan in the end to compare.

How is your motherboard laid out? You must have a lot of space between the CPU and RAM.
jhsu is offline  
post #8 of 29 (permalink) Old 02-05-2019, 05:41 AM
New to Overclock.net
 
doyll's Avatar
 
Join Date: Aug 2011
Location: UK
Posts: 24,810
Rep: 1631
@jhsu
Well done bit of testing!

I ignore troll like posts. Seems every forum as some users that do it. I used to get it all the time when I pointed how CLCs were not as good as air cooling while costing much more, making more noise and not lasting as long. Still do on some forums but not near as much as a few years ago .. many pump failures and more users testing good air coolers against CLCs finding out the same things I was preaching.

Were you monitoring the air temp into front of cooler and recording it at same time as CPU temp? What exactly is the thermometer you used?

Reasons I ask is because I've tested normal Silver Arrow IB-E Extreme against TRUE Spirit 140 Power which is similar to NH-U14S but with 6x 8mm heatpipes instead of 6x 6mm heatpipes. SA IB-E Extreme has 2x TY-143 fans which is what was also used on TRUE Spirit 140 Power. My results were both coolers cooling showed similar cooling improvements .. in other words Silver Arrow IB-E and TRUE Spirit 140 Power have similar performance with same fans. Granted, these tests were done on i7 920 @ 4.2GHz, not 1950X and Silver Arrow IB-E has smaller base than Silver Arrow TR4, but it's the same cooler with only difference being size of bottom half of base.

[H]ard|OCP tested NH-U14S, Silver Arrow TR4, Wraith Ripper and a few others including XSPC Raystrom loop on 2990WX PBO and found Silver Arrow TR4 while not giving lower temps did in fact allow 2990WX to process much faster wiht only XSPC being better. NH-U14S did 3.30GHz to Silver Arrow TR4's 3.603GHz with Wraith Ripper trailing at 3.094GHz. Links to testing attached.
https://www.hardocp.com/article/2018...raith_ripper/7
https://www.hardocp.com/article/2018...oler_roundup/8

Long story short, I'm wondering If the larger size of Silver Arrow TR4 compared to NH-U14S fitted in your case effected case airflow resulting is warmer air entering Silver Arrow TR4 than NH-U14S which raised CPU temp with Silver Arrow TR4 because every degree warmer air into cooler is translates almost exactly to same number of degrees hotter CPU temp will be.

Phanteks Owners Club Ways to Better Cooling
i7 980 @ 3.55GHz =PH-TC14PE w/2x TY-143 fans =Crucial Ballistix 3x4GB =GA-X58A-UD5 =ASUS GTX580 DirectCU II =Enermax Modu84+ =Define R2 w/3x TY-140 case intake fans; all PWM controlled by CPU fan socketPhanteks Enthoo Primo MoBo Rampage III Extreme CPUi7 980X
@ 4.0GHz =R1 Ultimate w/2x TY-143 fans =Dominator GT 6x2GB =Rampage III Extreme =ASUS GTX580 DirectCU II =TX850 =Enthoo Primo w/ custom castor base
doyll is offline  
post #9 of 29 (permalink) Old 02-05-2019, 05:55 AM
Shooting down fallacies
 
ciarlatano's Avatar
 
Join Date: Aug 2013
Location: NYC Metro
Posts: 6,486
Rep: 368
Quote: Originally Posted by doyll View Post
Long story short, I'm wondering If the larger size of Silver Arrow TR4 compared to NH-U14S fitted in your case effected case airflow resulting is warmer air entering Silver Arrow TR4 than NH-U14S which raised CPU temp with Silver Arrow TR4 because every degree warmer air into cooler is translates almost exactly to same number of degrees hotter CPU temp will be.
I was wondering the same, as the results of the SA compared to the NH-U14S don't line up with a few tests I have seen, all of which were pretty consistent. There is definitely something else at play here.

Yeah....the little girl hentai avatar is really creeping me out....
ciarlatano is offline  
post #10 of 29 (permalink) Old 02-05-2019, 06:42 AM
New to Overclock.net
 
doyll's Avatar
 
Join Date: Aug 2011
Location: UK
Posts: 24,810
Rep: 1631
Quote: Originally Posted by ciarlatano View Post
I was wondering the same, as the results of the SA compared to the NH-U14S don't line up with a few tests I have seen, all of which were pretty consistent. There is definitely something else at play here.
Still,@ jhsu; is definitely trying to do a good job of testing.
Definitely not the usual 'I tested my XYZ cooler against BUT and AS cooler and XYZ is 47c better' with no details at all as to how testing was done .. then when asked get something link 'XYZ is simply better, intake air temp doesn't matter' kind of garbage

Phanteks Owners Club Ways to Better Cooling
i7 980 @ 3.55GHz =PH-TC14PE w/2x TY-143 fans =Crucial Ballistix 3x4GB =GA-X58A-UD5 =ASUS GTX580 DirectCU II =Enermax Modu84+ =Define R2 w/3x TY-140 case intake fans; all PWM controlled by CPU fan socketPhanteks Enthoo Primo MoBo Rampage III Extreme CPUi7 980X
@ 4.0GHz =R1 Ultimate w/2x TY-143 fans =Dominator GT 6x2GB =Rampage III Extreme =ASUS GTX580 DirectCU II =TX850 =Enthoo Primo w/ custom castor base
doyll is offline  
Reply

Quick Reply
Message:
Options

Register Now

In order to be able to post messages on the Overclock.net - An Overclocking Community forums, you must first register.
Please enter your desired user name, your email address and other required details in the form below.
User Name:
If you do not want to register, fill this field only and the name will be used as user name for your post.
Password
Please enter a password for your user account. Note that passwords are case-sensitive.
Password:
Confirm Password:
Email Address
Please enter a valid email address for yourself.
Email Address:

Log-in



Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)
 


Forum Jump: 

Posting Rules  
You may post new threads
You may post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off