• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

[Help] High VRM1 Temperature.

r9 290 on heaven after 30mins, ambient temp is 31*C, fan profile set as (Current Temp+20%) to achieve preferred temps, noise not being an issue since i have 2 ceilings fans and a busy street right in front.

umv.png
 
Apples & Apples? I think not.

Look at your fan speed! That bench was on full out 80% fan speed, no wonder temps are lower.

In Furmark screenie your fan was at 44%...

Agree with Vayra, the problem is probably OP's fan speed.

Both of the newer Heaven and Furmark screenshot read around 40% (1800-2080 RPM) on the fan speed, and the old Heaven screenshot is 80% fan speed (3613 RPM).

Maybe Crimson fan speed bug?
 
Agree with Vayra, the problem is probably OP's fan speed.

Both of the newer Heaven and Furmark screenshot read around 40% (1800-2080 RPM) on the fan speed, and the old Heaven screenshot is 80% fan speed (3613 RPM).

Maybe Crimson fan speed bug?

Looks fine too me, my system is crashed or something, i just don't like the high temps and GPU isn't stable in overclocking anymore.
The main problem is VRM1 temperature - 101*c - with stock clocks.
 
Last edited:
Hi.

This is happening too me in last few months.

Card and PC - Ultra clean - i'm cleaning filters and PC almost every week.



Stock clocks Sapphire 290X TRI-X - After ~4 mints in FurMark.
KyjnEky.jpg
Good example for how bad Furmark is for testing the GPU, that's why AMD/NV don't like it. Newer GPUs are throttling down due to the high stress that Furmark sets on the GPU. You can easily see it on the GPU clock, 934 MHz is clearly throtteled, from 1040 (stock). Also VRM temperatures are extremely high, and I know for a fact that this only happens in Furmark (but you've proven it yourself too, nevertheless). Some old GPUs just used a heck ton of power in Furmark, like HD 5000 series, because they didn't have power limit. Even older cards could be destroyed, because they were simply not designed for it. Newer GPUs just stay in their power limit, like 250W on the 290X and throttle if they reach it. But you really see on the VRM temps how bad Furmark still is for the card.

Basically everything is fine in your system, but as Vayra said, you should install the PSU properly so that it sucks air strictly from the outside and doesn't take air away from the graphics card. The mystery how your VRM and OC was better some time ago was solved by the fact that your fan settings were way off at 80% - I don't think its enjoyable when your PC is as loud as a plane lol. ;)
 
Good example for how bad Furmark is for testing the GPU, that's why AMD/NV don't like it. Newer GPUs are throttling down due to the high stress that Furmark sets on the GPU. You can easily see it on the GPU clock, 934 MHz is clearly throtteled, from 1040 (stock). Also VRM temperatures are extremely high, and I know for a fact that this only happens in Furmark (but you've proven it yourself too, nevertheless). Some old GPUs just used a heck ton of power in Furmark, like HD 5000 series, because they didn't have power limit. Even older cards could be destroyed, because they were simply not designed for it. Newer GPUs just stay in their power limit, like 250W on the 290X and throttle if they reach it. But you really see on the VRM temps how bad Furmark still is for the card.

Basically everything is fine in your system, but as Vayra said, you should install the PSU properly so that it sucks air strictly from the outside and doesn't take air away from the graphics card. The mystery how your VRM and OC was better some time ago was solved by the fact that your fan settings were way off at 80% - I don't think its enjoyable when your PC is as loud as a plane lol. ;)

My PSU loud as plane, cuz PSU fan is half broke and touching parts inside the psu.... lol

and temps still pretty high =P maybe is normal for you but 74*c and 82*c after 2 mints in app like heaven is freaking high for me =P.

Thanks any way guys =].
 
Its honestly prob a mix of both @Kanan point and that you overclocked.. chips can def degrade with high voltage and temps. most people are very comfortable with all temps at around 80c but some chips and components can withstand 95c to like 120c and be perfectly fine.
Eh try a rma on the gpu and see if they feel like being nice.
 
Looks fine too me
How does it look fine if GPU fans don't go higher than 44% in long running benchmark? In your own screenshots it's visible it went to 80% a year ago ... did you flash the BIOS on your card and by that changed Saphire's default fan curve? Maybe you used custom fan curve in afterburner a year ago?
 
How does it look fine if GPU fans don't go higher than 44% in long running benchmark? In your own screenshots it's visible it went to 80% a year ago ... did you flash the BIOS on your card and by that changed Saphire's default fan curve? Maybe you used custom fan curve in afterburner a year ago?

44%+ pretty high speed for stock TRI-X.


and 80% was custom fan mode, +207 mhz on gpu and +266 mhz on memory + 50% power and addition voltage on GPU pretty high overclock for air cooled AMD card.
 
Last edited:
The PSU Problem.


 
44%+ pretty high speed for stock TRI-X.


and 80% was custom fan mode, +207 mhz on gpu and +266 mhz on memory + 50% power and addition voltage on GPU pretty high overclock for air cooled AMD card.
Ah, so it's a somewhat worn out, maybe GPU leaks more current now and needs more power at stock clocks and voltages, and on top of that furmark ... that would make VRM toasty ... 80-90 degrees in Heaven sounds acceptable, at least it's stable at stock clocks without voltage increase, so you didn't wear that GPU out completely. It should survive until arctic islands ;)
 
OP, don't overestimate your hardware.

GPU always gets hot under full load, and this is no different than any other GPU I have ever seen in my entire life. Pushing down temps is very hard on air cooling alone. If you want to push full load temps below 65 C, you need water. Tri-X is just open air cooling, a good one at that, but it won't work magic, it just gives the GPU a little bit more 'temperature headroom' than most other cooling solutions. Think in the range of a couple degrees C, 5-8 C tops.

With very frequent use and high voltages chips do degrade ever so slightly over time, and may require slightly more voltage in the future to reach the same clock. Lower temperature = lower required voltage, higher temperature = more leakage. You can see how it may easily look like your GPU is suddenly less good when temperatures rise, because the effect is two-fold: degradation causes more leakage -> GPU still wants to hit high clockspeed and takes more power -> more power = higher temperature -> higher temperature = more leakage.

The behaviour you are seeing, is normal and expected, and the Heaven results show that the GPU is in working order and doing well. VRM temps are normal, GPU temp is as expected.
 
OP, don't overestimate your hardware.

GPU always gets hot under full load, and this is no different than any other GPU I have ever seen in my entire life. Pushing down temps is very hard on air cooling alone. If you want to push full load temps below 65 C, you need water. Tri-X is just open air cooling, a good one at that, but it won't work magic, it just gives the GPU a little bit more 'temperature headroom' than most other cooling solutions. Think in the range of a couple degrees C, 5-8 C tops.

With very frequent use and high voltages chips do degrade ever so slightly over time, and may require slightly more voltage in the future to reach the same clock. Lower temperature = lower required voltage, higher temperature = more leakage. You can see how it may easily look like your GPU is suddenly less good when temperatures rise, because the effect is two-fold: degradation causes more leakage -> GPU still wants to hit high clockspeed and takes more power -> more power = higher temperature -> higher temperature = more leakage.

The behaviour you are seeing, is normal and expected, and the Heaven results show that the GPU is in working order and doing well. VRM temps are normal, GPU temp is as expected.

I Agree about the time thing =]... but like i said the problem was VRM1 not GPU temperature =P. 102*c in cold winter, let see what would happen in hot as hell summer, any way the warranty up to 03/1/2017.
 
Last edited:
I Agree about the time thing =]... but like i said the problem was VRM1 not GPU temperature =P. 102*c in cold winter, let see what would happen in hot as hell summer, any way the warranty up to 03/1/2017.

In the summer, don't you dare to run that Furmark again :D Beyond that, you're fine :)
 
I have similar issue like this with my previous Tri-X R9-290.
VRM1 when gaming is 105C (no OC).
When OC +100mv it hit 120-135C !!!
 
I have similar issue like this with my previous Tri-X R9-290.
VRM1 when gaming is 105C (no OC).
When OC +100mv it hit 120-135C !!!

Provide more information and screenshots of GPU-Z in load condition and we can probably point you towards solution or explanation :)

In a nutshell, if you add voltage and see these temps, STOP adding voltage to the core. It is possible you have a weak GPU that does not take well to additional voltage. Every GPU is different for overclocking.
 
I have similar issue like this with my previous Tri-X R9-290.
VRM1 when gaming is 105C (no OC).
When OC +100mv it hit 120-135C !!!

Really ? my pc crashed after 110*c

Provide more information and screenshots of GPU-Z in load condition and we can probably point you towards solution or explanation :)

In a nutshell, if you add voltage and see these temps, STOP adding voltage to the core. It is possible you have a weak GPU that does not take well to additional voltage. Every GPU is different for overclocking.

TRI-X is best 290x/290 on the market =P
 
Really ? my pc crashed after 110*c



TRI-X is best 290x/290 on the market =P

Apparently no one has seen the image i posted in this thread about my 290, i understand that in your winter those temps are supposed to be impossible compared to my tropical weather. I'd say if you can RMA it go for it otherwise things are only gonna get worse and do something about that PSU :roll:.
 
Really ? my pc crashed after 110*c



TRI-X is best 290x/290 on the market =P

Tri-X is just the cooling solution. The chip is the same 290(x) that is on every other board, and the chip is subject to silicon lottery.
 
Tri-X is just the cooling solution. The chip is the same 290(x) that is on every other board, and the chip is subject to silicon lottery.
*Best Air Cooling Solution on the market. XD.

Like i said many times before, 74 after 2 mints in app is to high.

In last winter i didn't had over 70*c after hours and hours in BF4.

My GPU has over 77*C on GPU and on VRM1 102*C - 4 mint of test.

so Stock 290X can easily reach 95*C on GPU and 120*c on VRM1 ?


Apparently no one has seen the image i posted in this thread about my 290, i understand that in your winter those temps are supposed to be impossible compared to my tropical weather. I'd say if you can RMA it go for it otherwise things are only gonna get worse and do something about that PSU :roll:.
I wish !!! but they can't see the FU* problem in PSU... =[
 
Last edited:
*Best Air Cooling Solution on the market. XD.

Like i said many times before, 74 after 2 mints in app is to high.

In last winter i didn't had over 70*c after hours and hours in BF4.

My GPU has over 77*C on GPU and on VRM1 102*C - 4 mint of test.

so Stock 290X can easily reach 95*C on GPU and 120*c on VRM1 ?



I wish !!! but they can't see the FU* problem in PSU... =[
BF4 is not Unigine Heaven, don't compare different things. ;) And as Vayra already pointed at, temps are good enough, it's okay, don't make this silly. Graphics cards are not about "feelings" ("I feel it's too hot", basically what you're saying) it's about facts and as already stated 70°C is np. Not the first time people complain about "too high temperature", but every time it's the same: it's just a feeling and nothing more. Hardware don't cares about feelings and you shouldn't too. ;)
 
BF4 is not Unigine Heaven, don't compare different things. ;) And as Vayra already pointed at, temps are good enough, it's okay, don't make this silly. Graphics cards are not about "feelings" ("I feel it's too hot", basically what you're saying) it's about facts and as already stated 70°C is np. Not the first time people complain about "too high temperature", but every time it's the same: it's just a feeling and nothing more. Hardware don't cares about feelings and you shouldn't too. ;)
After few mints , in empty server ....

w9n.png



I care about my hardware, so my hardware would care about me later xD HAHA =p.
 
Last edited:
you only love me while im touching your keys and gaming all night :cry:
 
IIRC VRM1 is the auxiliary voltage rail not Vcore. Both VRMs use the exact same components and the MOSFETs are all rated for 125C. The AUX rail is super low power so running it cool is not all that important.

If you want to check that VRM1 really is the aux rail raise core voltage by 100mv and check which VRM temp goes up. If VRM1 temps don't change then it's AUX and you can basically forget about it because it doesn't need to provide much power and uses 70A @ 125 IR MOSFETs.
 
IIRC VRM1 is the auxiliary voltage rail not Vcore. Both VRMs use the exact same components and the MOSFETs are all rated for 125C. The AUX rail is super low power so running it cool is not all that important.

If you want to check that VRM1 really is the aux rail raise core voltage by 100mv and check which VRM temp goes up. If VRM1 temps don't change then it's AUX and you can basically forget about it because it doesn't need to provide much power and uses 70A @ 125 IR MOSFETs.

Thanks for info, but VRM1 always heat up with any small overcloclock, and VRM1 can't handle 110*c, my PC is crashing =].
 
Last edited:
Back
Top