3900x / 3950x. Want to improve cooling. Help? - Page 3 - Overclock.net - An Overclocking Community
Forum Jump: 

3900x / 3950x. Want to improve cooling. Help?

Reply
 
Thread Tools
post #21 of 49 (permalink) Old 03-28-2020, 01:20 PM
New to Overclock.net
 
doyll's Avatar
 
Join Date: Aug 2011
Location: UK
Posts: 26,448
Rep: 1689 (Unique: 935)
Quote: Originally Posted by GoldCartGamer View Post
Thanks for the link and chats. I ended up spending a bunch of time on his site and reading his reviews on techpowerup as well. It appears to me the Blacknoise and EK Vadar are the best performing 140mm fans form his tests on both sites. Or are the be quiet! SW3 High fans plenty?

Oh I did. I skipped right over it and kept moving. I've seen the data on AIO/CLC vs Air coolers. Not worth it to water cool unless doing a full custom loop.

How about the EK Vadar and Blacknoise fans? They seem to perform better.

After I read your link about airflow and optimum case cooling, I realized removing that top panel was not a good idea. So much info on here and other places as I dig down this rabbit hole of finding the best for my build. I became cross-eyed again last night. I reminded myself again about the servers I build and will stick to their method of cooling / design. Intake fans, sealed case, and no exhaust fans. It seems to line up with your findings and recommendations. I only got a case with mesh / filter due having pets. I read more about why removing the PCIe slot covers works and it is something I never honestly thought about until I stumbled upon your thread before making this post.

We are on overclock.net and I can promise you my setup is not running on stock settings.

Plus, that is not a guaranteed ambient. The thermostat is on the other side of the house, my office faces the sun in the afternoon, I have a few electronics in my office that generate heat, etc. The CPU cooler I have is fine. I just wanted a gut check on it. The case fans are absolute garbage at full speed, let alone any other speed. Barely blow any air. I wanted to seek out the best case fans and optimal case airflow. Then doyll's posts made it click. My case is no different than a server case. At the end of the day it is still computer parts.
Stock case fans move about as much air at full speed as SW3 PWM move at 500-600rpm.

Honestly Silent Wing 3 PWM high speed are the best, not just because of their airflow but also because of their pleasant sound. Often how a fan sounds is just as important as how much air they move. Because unless your system never needs fans to be running faster than 800-900rpm you will definitely be hearing them .. and SW3 really sound nice as well as moving plenty of air.

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  
Sponsored Links
Advertisement
 
post #22 of 49 (permalink) Old 03-28-2020, 02:18 PM
New to Overclock.net
 
Join Date: Jan 2019
Posts: 422
Rep: 13 (Unique: 12)
Quote: Originally Posted by GoldCartGamer View Post
Long time lurker, recently signed up on the site. I have spent A LOT of time reading through doyll's posts and threads among other places on the internet. Very insightful, but I still had some questions for my particular config / environment.

I currently have the following configuration: https://pcpartpicker.com/user/goldca...r/saved/HnsxGX with a soon to be 3950x that is staring at me as I type this, lol.

I did have to remove a panel to make room for the USB 3.0 cable on the mobo. I have attached a stock photo with a red circle around that panel.

I am looking at case fans and proper air flow on this case. I know the stock case fans are junk. Based on what I have read, it seems it is a good idea to remove the PCIe slots, no exhaust fan, and having 2 to 3 strong intake fans will be plenty.

I was thinking I would have 3 to 4 intake fans and 2 exhaust fans, but my research seems to show that is not necessarily the best route.

I have been down so many rabbit holes tonight on case fans I am a bit cross eyed. Budget and sound are not a concern. I am concerned about performance only. I started out looking at Noctua NF-A14 PWM then read about the Noctua NF-A12x25 PWM. I believe the front of this case blocks part of the air flow of a 140mm fan as well.

I have been happy with my CPU cooler choice, but now worried I picked wrong based on threads on here.

I am just now getting to worry about temps because outside temps will soon go up in my area. It gets pretty hot and humid in my summers. We keep the thermostat set to 68F (20C) during the summer.

I hope I didn't miss anything to get this conversation started. I hope to learn more and perhaps provide a place for others to learn as well.
No exhausts fans, it's best for noise/performance. Pure performance, just go all the noise and fans you want. But you will start to see very aprupt diminishing returns at some point.

Just keep in mind that the more exhausts you have in the top area, the more hot air from the GPU area you will drag into the CPU/VRM area, which is why no exhaust works so great. Don't get me wrong, having an exhaust it's still better than no exhaust at all, but it's just barely, and only when you pushing it hard.

Here's a middle ground. 3 front intakes, +1-2 bottom slow intakes, and 1 rear exhaust that only spins after your CPU hits for example 120w (it's just an example) and goes over. Ofc you can't set fans based on power consumption, but you can see where the temps tend to stay over that power consumption target.

Last edited by Shenhua; 03-28-2020 at 02:22 PM.
Shenhua is offline  
post #23 of 49 (permalink) Old 03-28-2020, 02:25 PM
New to Overclock.net
 
poah's Avatar
 
Join Date: Sep 2016
Location: in a house
Posts: 908
Rep: 7 (Unique: 6)
Quote: Originally Posted by GoldCartGamer View Post


Oh I did. I skipped right over it and kept moving. I've seen the data on AIO/CLC vs Air coolers. Not worth it to water cool unless doing a full custom loop.


We are on overclock.net and I can promise you my setup is not running on stock settings.
you'll want an AIO at least then if you are not running stock settings. arctic freezer II cools better and is quieter than the top end air coolers on the 7nm ryzen. Full custom loop isn't that much better because you can't get the heat out the chiplets as efficiently as you could with the monolithic dies. If you were running stock, any top end air cooler with decent air flow into the case would be fine.

several members on here have swapped out NH-D15 for a 280mm arctic freezerII and have quieter noise and better cooling.
poah is offline  
Sponsored Links
Advertisement
 
post #24 of 49 (permalink) Old 03-28-2020, 02:52 PM
Shooting down fallacies
 
ciarlatano's Avatar
 
Join Date: Aug 2013
Location: NYC Metro
Posts: 7,634
Rep: 438 (Unique: 291)
Quote: Originally Posted by GoldCartGamer View Post
Thanks for the link and charts. I ended up spending a bunch of time on his site and reading his reviews on techpowerup as well. It appears to me the Blacknoise and EK Vadar are the best performing 140mm fans form his tests on both sites. Or are the be quiet! SW3 High fans plenty?






Oh I did. I skipped right over it and kept moving. I've seen the data on AIO/CLC vs Air coolers. Not worth it to water cool unless doing a full custom loop.



How about the EK Vadar and Blacknoise fans? They seem to perform better.
The EK and eLoops both have noise issues. The Vardars, while not loud, have a really irritating sound signature. The eLoops are fantastic fans in the right application, but they howl if they have even the slightest bit of restriction near their intake. That includes dust filters or case mesh. The SW3 give you the upside without the issues.

And glad to see you already realized what you should skip right over.

Yeah....the little girl hentai avatar is really creeping me out....
ciarlatano is offline  
post #25 of 49 (permalink) Old 03-29-2020, 05:09 AM
New to Overclock.net
 
doyll's Avatar
 
Join Date: Aug 2011
Location: UK
Posts: 26,448
Rep: 1689 (Unique: 935)
I would ask poah for data supporting his false claims but I know there are lots of reviews with test results that are not even close to being accurate claiming CLCs are better than air. But the truth is if both are cooling with same temp air top tier air coolers are as good or better than CLCs.

The only top tier air coolers I know of with really high performance fans are Silverstone HE-01 with 181cfm fan and Thermalright's Silver Arrow TR4 Extreme, Silver Arrow IB-E Extreme, andSilver Arrow IB-E Extreme Rev.B with 130cfm fan/s.

We tested other top tier air coolers like NH-D14 and TRUE Spirit 140 Power with hi-performance TY-143 fan/s finding they run 8-9c lower temps at 2500rpm instead of 1200rpm.

That testing was all done on i7 [email protected] (lots of heat) test system.
I7 920 CPU is rated 130w TDP
test system stock setting wattage at wall was 126w idle and 225w test load stock and 400w @ 4.3GHz.
published spec for I7 920 is 130w TDP
Based on above I think it's safe to assume 225w at the wall with 130w CPU spec means system was using 95w. That means CPU was using about 305w of power and producing 300w of heat.

This all means TRUE Spirit 140 Power on 305w of heat was 43.25c delta (air entering coolers was 19-21.1c, recorded same time as CPU temp).
NH-D14 was 8.25c cooler at 2500rpm than at 1200rpm,

TRUE Spirit 140 Power was 9.3c cooler at 2500rpm than at 1200rpm.
The down side is much more noise at 2500rpm. TY-143 are about 48dB(A) @ 2500rpm compared to 25dB(A) at 1200rpm.


A year or so ago Kyle Bennet of [H]ardOCP did extensive testing of NH-U14S TR4-SP3, Silver Arrow TR4 Extreme, his custom loop, I think Liqtech TR4 and a few others on Threadripper 2990WX & 2950X. Silver Arrow TR4 Extreme at full speed cooled almost as well as his custom loop did. Liqtech was not as good. Sadly that entire series of testing and reviewing has been scrubbed from internet. The only things remaining are a few forum threads talking about the results.

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 #26 of 49 (permalink) Old 03-29-2020, 07:33 AM
New to Overclock.net
 
Join Date: Jan 2019
Posts: 422
Rep: 13 (Unique: 12)
Quote: Originally Posted by doyll View Post
I would ask poah for data supporting his false claims but I know there are lots of reviews with test results that are not even close to being accurate claiming CLCs are better than air. But the truth is if both are cooling with same temp air top tier air coolers are as good or better than CLCs.

The only top tier air coolers I know of with really high performance fans are Silverstone HE-01 with 181cfm fan and Thermalright's Silver Arrow TR4 Extreme, Silver Arrow IB-E Extreme, andSilver Arrow IB-E Extreme Rev.B with 130cfm fan/s.

We tested other top tier air coolers like NH-D14 and TRUE Spirit 140 Power with hi-performance TY-143 fan/s finding they run 8-9c lower temps at 2500rpm instead of 1200rpm.

That testing was all done on i7 [email protected] (lots of heat) test system.
I7 920 CPU is rated 130w TDP
test system stock setting wattage at wall was 126w idle and 225w test load stock and 400w @ 4.3GHz.
published spec for I7 920 is 130w TDP
Based on above I think it's safe to assume 225w at the wall with 130w CPU spec means system was using 95w. That means CPU was using about 305w of power and producing 300w of heat.

This all means TRUE Spirit 140 Power on 305w of heat was 43.25c delta (air entering coolers was 19-21.1c, recorded same time as CPU temp).
NH-D14 was 8.25c cooler at 2500rpm than at 1200rpm,

TRUE Spirit 140 Power was 9.3c cooler at 2500rpm than at 1200rpm.
The down side is much more noise at 2500rpm. TY-143 are about 48dB(A) @ 2500rpm compared to 25dB(A) at 1200rpm.


A year or so ago Kyle Bennet of [H]ardOCP did extensive testing of NH-U14S TR4-SP3, Silver Arrow TR4 Extreme, his custom loop, I think Liqtech TR4 and a few others on Threadripper 2990WX & 2950X. Silver Arrow TR4 Extreme at full speed cooled almost as well as his custom loop did. Liqtech was not as good. Sadly that entire series of testing and reviewing has been scrubbed from internet. The only things remaining are a few forum threads talking about the results.
You are generalizing in the first paragraph and you're using an example of the best case scenario for air cooling to prove it in the last paragraph.

I don't disagree with you but you are somewhat manipulating the image of air cooling.
While a u14s it's gonna trash most CLCs "pound for pound" on a threadripper, it won't happen the same on a 3700x for example.

Neither cooling is better than other, each has its own properties and one is more suitable than the other depending on the CPU and the build. Heatpipe cooling disipate heat faster, so where there is no limitation in heat transference between CPU and cooler they do great. Liquid cooling holds heat but due to their mechanism they will always have a colder coldplate, so if the power consumption is not that high but the heat transference between CPU and cooler is poor, they will do great and/or better than any air cooler.

The CLC that poah is talking about will definitely gets its ass handed by any top air cooler on a 200w+ 7-8years OLD CPU, but it will will eat them for breakfast on a 3600-3700x at 1.35v or higher.

Given the setup of the OP here a CLC won't be better than a air cooler, and if it is, it's gonna be by margin of error at the same noise floor.

Last edited by Shenhua; 03-29-2020 at 07:40 AM.
Shenhua is offline  
post #27 of 49 (permalink) Old 03-29-2020, 07:49 AM
New to Overclock.net
 
hazium233's Avatar
 
Join Date: Jul 2017
Posts: 159
Rep: 4 (Unique: 3)
Quote: Originally Posted by poah View Post
you'll want an AIO at least then if you are not running stock settings. arctic freezer II cools better and is quieter than the top end air coolers on the 7nm ryzen. Full custom loop isn't that much better because you can't get the heat out the chiplets as efficiently as you could with the monolithic dies. If you were running stock, any top end air cooler with decent air flow into the case would be fine.

several members on here have swapped out NH-D15 for a 280mm arctic freezerII and have quieter noise and better cooling.
Was this with or without lapping? What was the difference in temps typically, and what was the max Tdie they saw in each case?

COL1
(3 items)
CPU
Ryzen 7 2700X
Motherboard
Asus Strix X370-F (4012)
RAM
Ballistix Sport AT 2666 (D9TZV)
▲ hide details ▲
hazium233 is offline  
post #28 of 49 (permalink) Old 03-29-2020, 08:28 AM
New to Overclock.net
 
poah's Avatar
 
Join Date: Sep 2016
Location: in a house
Posts: 908
Rep: 7 (Unique: 6)
Quote: Originally Posted by hazium233 View Post
Was this with or without lapping? What was the difference in temps typically, and what was the max Tdie they saw in each case?
this is on ryzen 3900x. temps were cooler on both dies with the arctic and quieter particularly at lower fans speeds where a computer is 99% of the time.
poah is offline  
post #29 of 49 (permalink) Old 03-29-2020, 09:35 AM
New to Overclock.net
 
doyll's Avatar
 
Join Date: Aug 2011
Location: UK
Posts: 26,448
Rep: 1689 (Unique: 935)
Quote: Originally Posted by Shenhua View Post
You are generalizing in the first paragraph and you're using an example of the best case scenario for air cooling to prove it in the last paragraph.

I don't disagree with you but you are somewhat manipulating the image of air cooling.
While a u14s it's gonna trash most CLCs "pound for pound" on a threadripper, it won't happen the same on a 3700x for example.

Neither cooling is better than other, each has its own properties and one is more suitable than the other depending on the CPU and the build. Heatpipe cooling disipate heat faster, so where there is no limitation in heat transference between CPU and cooler they do great. Liquid cooling holds heat but due to their mechanism they will always have a colder coldplate, so if the power consumption is not that high but the heat transference between CPU and cooler is poor, they will do great and/or better than any air cooler.

The CLC that poah is talking about will definitely gets its ass handed by any top air cooler on a 200w+ 7-8years OLD CPU, but it will will eat them for breakfast on a 3600-3700x at 1.35v or higher.

Given the setup of the OP here a CLC won't be better than a air cooler, and if it is, it's gonna be by margin of error at the same noise floor.
First paragraph is not best case at all, it's about giving air coolers fair play. All it is about is supplying air cooler with same or at least similar temp of air as is going through radiators on CLCs. To me that is just being reasonable .. not 'best case' at all, just a fair and 'logical case'.
Personally I have never been overly impressed by NH-U14S. It is just another of many big coolers, definitely not as good as many others in it's class .. even TRUE Spirit 140 Direct is within a couple degrees even on 300w TDP and it has 1300rpm fan to U14S's 1500rpm fan/s. TRUE Spirit 140 Power out-performs it by a good margin with similar airflow. Point here is Thermalright Silver Arrow TR4 with fan at about half speed was able to 'trash' all other air coolers he tested .. and when he used TY-143 at full speed it 'trashed' them by a large margin.In his first set of results were with single TY-143 at 1300-1500rpm (TY-143 is 2500rpm) and results were similar to NH-U14S. He later retested with single and double TY-143 at full speed and it not only beat all other air coolers but was able to cool almost as good as his custom loop. Keep in mind in that testing his Threadripper was maxing out and throttling under all cooling tested. Results were not temps but maximum operating speed of CPU.

I don't remember anything about lapping to improve heat transfer from CPU to cooler bases.

I really regret not copying/saving his testing and results, because now I can't find it anywhere on web. He wiped it all off. Like I said, there are a few discussions about it on forums, but sadly not the details and data.

Here are a few links with a little about what Kyle did and results, wish there was more:
https://www.reddit.com/r/Amd/comment...ooler_roundup/
https://hardforum.com/threads/thread...per-h.1968718/



Quote: Originally Posted by poah View Post
this is on ryzen 3900x. temps were cooler on both dies with the arctic and quieter particularly at lower fans speeds where a computer is 99% of the time.
What was air temp into coolers at time CPU temps were recorded?


If anyone has or finds any of the test data from HardOCP TR4 testing of air and water coolers Please post it or links to it.


Edit:
Did find below about that testing:
Quote:
Quote:
So Silver Arrow TR4 is 11c cooler at full speed and 53dB, but still 9.5c warmer than your XSPC Raystorm. What is the noise level of XSPC Raystorm system in these tests? Have your ran any test on 2990WX with Liqtech TR4 and if so what is it's temps and noise level?
The four fans at full RPM on that rad are 45.5db at 4 feet.
Given the quality problems shown on the Liqtech coolers, we have rescinded our awards and have not done any further testing. Those reviews are still up however.
That was Oct 11, 2018. Can't find Liqtec CLC review anywhere now.

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

Last edited by doyll; 03-29-2020 at 10:42 AM.
doyll is offline  
post #30 of 49 (permalink) Old 03-29-2020, 10:27 AM - Thread Starter
New to Overclock.net
 
GoldCartGamer's Avatar
 
Join Date: Sep 2019
Location: US
Posts: 18
Rep: 0
Ordered the SW3 High Speed fans. Excited to install them and make changes to my case as previously discussed (i.e., PCIe slot covers removal).

Does the internet wayback machine not have an archive of the post(s) and results? I'm sure we could scrounge some of the results from there.

3950x, Gigabyte X570 Xtreme, Team Dark Pro 16GB (2x8GB), MSI RTX 2080 Ti Gaming X Trio, Arctic Liquid Freezer II 360.
Build list
GoldCartGamer is offline  
Reply

Tags
3900x , 3950x , Air Cooling , meshify s2

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)
 
Thread Tools
Show Printable Version Show Printable Version
Email this Page Email this Page


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