Welcome to MilkyWay@home

Very high temps out of two of the VRM's

Message boards : Number crunching : Very high temps out of two of the VRM's
Message board moderation

To post messages, you must log in.

AuthorMessage
Aedazan

Send message
Joined: 30 Jan 11
Posts: 23
Credit: 33,424,579
RAC: 0
Message 62263 - Posted: 6 Sep 2014, 9:33:09 UTC

Hey guys, ive got x2 5970's in my rig crunching away. Problem is that one of the 4 cores has 2 VRM's with very high temperatures, much higher than any of the other VRM's. It was thermally throttling at 140c before so I took the whole thing apart, spent a ton of time getting the old thermal-goop off and repasted it with arctic silver 5 which has cut an awesome 40 degrees c off of the temps but they are still very high. I am noticing that this core specifically seems have a VDDC Current around 50%+ higher than any of the other cores which may mean that its pulling far too much current for the VRM's to manage. Does anyone have any idea why this may be happening?


Here is a screenshot of GPU-Z for that core. I have it undervolted to 1.00v to keep the temps to a point where it doesnt thermal throttle.
http://imgur.com/9OxXIFT

This is the machine:
http://milkyway.cs.rpi.edu/milkyway/show_host_detail.php?hostid=585984
ID: 62263 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Aedazan

Send message
Joined: 30 Jan 11
Posts: 23
Credit: 33,424,579
RAC: 0
Message 62264 - Posted: 6 Sep 2014, 10:18:47 UTC - in response to Message 62263.  

I have done a little bit more experimentation and when I underclock and undervolt it to the point where its VDDC Current is the same as the other cards the VRM temperatures are fairly similar, being only about 12 degrees c higher. This may suggest that the issue is both with these two VRM's as the other VRM maintains a normal temperature. This does not explain why this core is using so much current though so I am still quite unsure where the problem may be originating.
ID: 62264 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Aedazan

Send message
Joined: 30 Jan 11
Posts: 23
Credit: 33,424,579
RAC: 0
Message 62265 - Posted: 6 Sep 2014, 10:53:18 UTC - in response to Message 62264.  

I read around about it possibly being a PSU issue so Ive rebalanced the 12v rails as per the PSU datasheet and each card now has up to 50A from the 12v rail and supposedly an additional 6.25A from the PCIE Bus (Total 56.25A x 12 = 675 per card which is more than double the TDP of 294w). Issue is still persisting.
ID: 62265 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Aedazan

Send message
Joined: 30 Jan 11
Posts: 23
Credit: 33,424,579
RAC: 0
Message 62267 - Posted: 6 Sep 2014, 11:09:11 UTC - in response to Message 62265.  

Just to clarify the temperatures on all of the other cores VRM's are nearly identical to each other in their respective cores (Being located so close to the cores they are around that same temperature). None of the other VRM's make it to 80 c at stock clocks. Most of them are around 60 c. It is just these two in the first core of my second card that are acting up.
ID: 62267 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Aedazan

Send message
Joined: 30 Jan 11
Posts: 23
Credit: 33,424,579
RAC: 0
Message 62277 - Posted: 8 Sep 2014, 1:24:52 UTC - in response to Message 62267.  
Last modified: 8 Sep 2014, 1:25:38 UTC

Here is an image of the board before I removed the old thermal paste and put the new paste onto it: http://i.imgur.com/GenZa4I.png
ID: 62277 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile FalconFly
Avatar

Send message
Joined: 28 Mar 10
Posts: 13
Credit: 13,311,528
RAC: 0
Message 62281 - Posted: 8 Sep 2014, 14:07:51 UTC - in response to Message 62277.  

How is the case cooling for that card ?

The Video card PCB components (especially VRMs) need a good airflow to cool them, for crunching purposes an excellent airflow is mandatory to ensure longlivety of the hardware.
ID: 62281 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Aedazan

Send message
Joined: 30 Jan 11
Posts: 23
Credit: 33,424,579
RAC: 0
Message 62286 - Posted: 9 Sep 2014, 0:10:06 UTC - in response to Message 62281.  
Last modified: 9 Sep 2014, 0:17:30 UTC

Ive got 12 fans in the case, 6 of which are 70+ CFM high performance fans (One of which is the case side fan which blows a small tornado directly onto the cards, its my 79CFM fan).

As you can see the GPU Core itself is very cool (<60c == <140f) (Recommended operating temps are allowed to go upto 90c from what ive read).

According to the sensors the temperature inside the case is around 35 c.

From this toms hardware post it seems that my cooling is above average:
http://www.tomshardware.com/forum/285547-33-hd5970-running
ID: 62286 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile FalconFly
Avatar

Send message
Joined: 28 Mar 10
Posts: 13
Credit: 13,311,528
RAC: 0
Message 62288 - Posted: 9 Sep 2014, 0:42:58 UTC - in response to Message 62286.  
Last modified: 9 Sep 2014, 0:47:49 UTC

Yes, that's very good cooling. Given that, seems weird those VRMs get so hot.

Any chance some of the small resistor(s) got damaged or missing on that card's PCB?

That would be the only reason I could imagine, as video cards often still run fine with damaged or missing resistors - with some components getting out-of-specs Voltage or Ampere as a result. That can create uncommon/random hotspots on other components.
Dry/dead capacitors (typically not visible on modern hardcap variants) could also be a factor - after all those have a limited lifespan depending on operating temperature and time of exposure. That could cause similar effects IMHO but as opposed to damaged resistors, it's very tough to detect (let alone fix).
ID: 62288 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Aedazan

Send message
Joined: 30 Jan 11
Posts: 23
Credit: 33,424,579
RAC: 0
Message 62289 - Posted: 9 Sep 2014, 2:12:09 UTC - in response to Message 62288.  

No resistors are missing to my knowledge, I have only opened the card to repaste it and I have been quite careful. In terms of the capacitors yes, the card is fairly old so it could be a capacitor on its way out :-\
ID: 62289 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote

Message boards : Number crunching : Very high temps out of two of the VRM's

©2024 Astroinformatics Group