VRM/Mosfet temperature overheating

[wah123]

Hello guys,

I wanted to know what happens when the VRM on the motherboard overheats?

Does it start throttling the CPU, blow up and ruin the motherboard or take any other components with it?

My Motherboard is GA-970a-ud3p with a fx-8320 CPU.

In HWiNFO the "VR T1" and "VR T2" temperature headings under the motherboard are showing as 96c under load on Prime95 test for 30 minutes, and 38c when idle.

I've contacted Gigabyte and their reply was that the third party software temperature readings aren't accurate, but the CPU and GPU temperatures seem normal under load at 48c (CPU) and 64c (GPU).

I can't find any info regarding this, any help appreciated

Thanks

 

[MrGenius]

Smell something burning? No? Then don't worry about it.

Wondering what the VRM temps really are? Buy a laser/IR thermometer.

 

[wah123]

Smell something burning? No? Then don't worry about it.

Wondering what the VRM temps really are? Buy a laser/IR thermometer.

Thanks MrGenius.

I just wanted to know if it's dangerous that's all, like if it's on its way out it wont harm anything else.

 

[xvi]

Failing VRMs will likely cause damage to the card. If the card is under warranty, just let it fail and RMA it without touching anything. It's the manufacturer's responsibility to provide a working card.

It depends on the type of VRMs, but ~90c temps aren't uncommon.

 

[suraswami]

It's probably not the VRM temps that is shown, it might be the NB temp, put a fan on the NB sink and it helps. Putting a fan on VRM helps too.

 

[[Ion]]

It's probably not the VRM temps that is shown, it might be the NB temp, put a fan on the NB sink and it helps. Putting a fan on VRM helps too.

I doubt that--I'd suspect that 'Temperature 2' or 'Temperature 3' is the chipset. The NB temperature won't vary by 60C from idle to load--whereas the VRMs easily could.

 

[theonedub]

If you can put your finger on any of the heatsinks, whether it be VRM or NB, without burning yourself immediately then there is no way any of the components are at 96C.

 

[wah123]

Failing VRMs will likely cause damage to the card. If the card is under warranty, just let it fail and RMA it without touching anything. It's the manufacturer's responsibility to provide a working card.

It depends on the type of VRMs, but ~90c temps aren't uncommon.

Card? The VRM I'm referring to is on the motherboard. Thanks for your reply

It's probably not the VRM temps that is shown, it might be the NB temp, put a fan on the NB sink and it helps. Putting a fan on VRM helps too.

Thanks for the advice. I've read that they're safe up to about 100c, so if I remove the OC and revert to stock clocks I should get away with it. While gaming it only goes up to 70ish, but the higher temps are only with stress testing.

If you can put your finger on any of the heatsinks, whether it be VRM or NB, without burning yourself immediately then there is no way any of the components are at 96C.

That could be risky lol. If by any chance it's really at 96c it will burn me. I think I'll just try with those laser temp sensors. Thanks

 

[theonedub]

If the sink is 96C, you'll be able to feel the heat as you get into close proximity to it.

An IR Thermometer is very useful for the PC and for other tasks around the house, car, etc. I picked one up for ~$10 on Amazon and have used it for this very purpose.

 

[wah123]

If the sink is 96C, you'll be able to feel the heat as you get into close proximity to it.

An IR Thermometer is very useful for the PC and for other tasks around the house, car, etc. I picked one up for ~$10 on Amazon and have used it for this very purpose.

Thanks.

I've found this..

http://www.amazon.co.uk/dp/B007Q87J3U/?tag=tec053-21

It will come in handy for other things too.

 

[natr0n]

mount a fan blowing over them.

 

[wah123]

mount a fan blowing over them.

Will do, if the temp is really that high but hopefully it's not.

 

[Batou1986]

Those temps are way too high my VRM/NB are no where near that hot, Full load IBT stock volts 8 cores @ 4 ghz
I think the sensors are wrong but can't confirm because I have the UD3 non P and it has no VRM sensors in HWiNFO
If they were running at 90*c they would have burnt out by now.
yes its calibrated for reading anodized aluminum before anyone asks and yes the fan got my thumb knuckle

Spoiler

 

[suraswami]

I doubt that--I'd suspect that 'Temperature 2' or 'Temperature 3' is the chipset. The NB temperature won't vary by 60C from idle to load--whereas the VRMs easily could.

Based on 'VR T1', it seems like it might be VRM temps, I would just stick a fan on the VRM sink and make sure the case has good airflow.

 

[Batou1986]

For the record temperatures 1,2,3 are SB,NB,CPU in that order on gigabyte AMD 9xx boards.
Those sensors are definitely for the VRM if they are accurate or not I dunno.
Most of the Gigabyte boards that i've had have incorrect temp sensors.

 

[wah123]

Those temps are way too high my VRM/NB are no where near that hot, Full load IBT stock volts 8 cores @ 4 ghz
I think the sensors are wrong but can't confirm because I have the UD3 non P and it has no VRM sensors in HWiNFO
If they were running at 90*c they would have burnt out by now.
yes its calibrated for reading anodized aluminum before anyone asks and yes the fan got my thumb knuckle

Spoiler

Thanks for that. I'll just have to make sure by testing it with an IR thermometer.

I think it is accurate because the idle reading is at around 38c which everyone else is getting. The only difference is that mine gets way hotter under load.

I've emailed Gigabyte asking them what the safe temps are for the VRM on this board, then I can go from there. If the maximum is around 125c then I think I can live with 90c.

 

[wah123]

Based on 'VR T1', it seems like it might be VRM temps, I would just stick a fan on the VRM sink and make sure the case has good airflow.

That's what I don't understand. The other temperatures are normal, it's just this temperature that's high.

The CPU core temp doesn't exceed 48c under load, and the GPU temp doesn't exceed 66c under load. I've seen people much higher temps than me with the same CPU (fx-8320) and GPU (gtx 970).

 

[suraswami]

That's what I don't understand. The other temperatures are normal, it's just this temperature that's high.

The CPU core temp doesn't exceed 48c under load, and the GPU temp doesn't exceed 66c under load. I've seen people much higher temps than me with the same CPU (fx-8320) and GPU (gtx 970).

I just ran 4 cycles of IBT stress testing and this is what temps looks like.

I have a 70mm fan on the VRM sink and for CPU I run TT AIO cooler.

 

[wah123]

I just ran 4 cycles of IBT stress testing and this is what temps looks like.

I have a 70mm fan on the VRM sink and for CPU I run TT AIO cooler.

View attachment 64329

Thanks for taking the time to check this. It looks like the fan works wonders with cooling the high temps. I'll stick a fan on there and then test it for a while.

 

[suraswami]

and BTW, what is the CPU volts and what speed you are OCing to? May be leave the CPU volt at default and keep raising the CPU clock until it fails to boot and run stable.

 

[wah123]

and BTW, what is the CPU volts and what speed you are OCing to? May be leave the CPU volt at default and keep raising the CPU clock until it fails to boot and run stable.

About a week ago I OC'd my CPU to 4.2GHZ @ 1.296 Vcore with LLC set to auto and APM set to off, all the other power saving features were on as default.

I installed HWiNFO yesterday and noticed the 96c temp, then I reset the CPU back to stock clocks @ 3.5ghz with everything set back to default, with Vcore set to Auto. Then I ran the stress test again and it didn't go above 90c so it looks like I lost 6c by doing that.

When gaming or doing other tasks that temp doesn't exceed 70c.

I think I can live with it and monitor my temps, and if it gets worse I can stick a fan on it. I don't really need to run stress tests anyway as I can just monitor my temps in HWiNFO.

 

[wah123]

Okay so today I looked at undervolting my CPU and the temperatures went down significantly. I undervolted the CPU to 1.28750V @ 3.5GHZ and stress tested Prime95 (small FFTS) for 45 minutes and it was stable with no warnings or errors, and my temps have improved a lot with the CPU temp @ 41C under load and the VR T1 and T2 temps staying at 80C under load. It looks like I have sorted out this issue. See the attached screenshots, does everything look fine?

I also received a reply from what is supposed to be Gigabytes technical support team. Their reply to me...

"Dear Customer,

The VR T1/VR T2 is one for CPU temperature and one for System temperature, when run the Prime95 and the temperature below 100 degree is normal. ".

 

[peche]

temps seems to be pretty weird, I know that most of your problems seems to be solved on your last post, but I would like to know some facts or info:

how is your fan Setup? which case do you have ?

Airflow is pretty important, most times people only take care about setting a intake fan and a exhaust fan, but sometimes you have to take the time and look for the correct setup, correct fans and also improving with couple of extra fans over here and over there…also as somebody said even fans blowing directly hot components, I'm about to make a mod to my Gigabyte Z68X motherboard on the chipset itself,


Regards,

 

[wah123]

temps seems to be pretty weird, I know that most of your problems seems to be solved on your last post, but I would like to know some facts or info:

how is your fan Setup? which case do you have ?

Airflow is pretty important, most times people only take care about setting a intake fan and a exhaust fan, but sometimes you have to take the time and look for the correct setup, correct fans and also improving with couple of extra fans over here and over there…also as somebody said even fans blowing directly hot components, I'm about to make a mod to my Gigabyte Z68X motherboard on the chipset itself,


Regards,

Hello,

My case is a Cooler Master K280, and I have 2 fans in there. The rear top has a exhaust 120mm fan and the front bottom has a 120mm intake fan, both are bitfenix spectre fans. I may be able to put a side panel fan on but it would be very close or touching the GPU.

I think it could be the CPU cooler fan as currently the set up is that the CPU fan is facing towards the bottom of the case and it's pulling air from the bottom to the top. I could have had it so the fan is facing towards the back of the case and is blowing air out towards the exhaust fan, but I had RAM clearance issues so could not do it that way. I did have a lot of trouble installing this cooler the first time so won't have much confidence in taking it off again and seeing if I could change the position.

It seems like you're very experienced in mods, it must be easy for you to do these adjustments..

Thanks for your reply.

 

[peche]

There we have your case right?

there is the recommended fan setup, do you have the same or have a diferent set up?
also I may ask another question, in the airflow chart there is no top fans … does the case have them?

Regards,