High VRM temps?

[quadrax]

Hello. I was wondering, what is considered as "high VRM temps" on MOSFETs and VRM Controller. My board is Gigabyte AORUS Gaming K3 and when i set LLC "High", next level Turbo didnt tried, yet - so my temps on VRMs RAISE in LinX temps (i know its kinda extremely loadout and almost unreallistic, but still) to 88°C. Then i stopped. My bios settings are :

vCore : 1.260v
LLC : High
VCCIO : 0.950
VCCSA : 1.050
XMP Profile (3000MHz, 1.35v) on Corsair pair of 2x8GB RAMs.

For monitoring temps of VRM i am using latest stable HWINFO64.

CPU is cooled by Scythe Ninja 4 (used paste MX4) in Fractal Design Define R5 Case with 2 inttake and one outtake fan. All fans running at maximum speed (no temperature setting, just max peformance). And still i am getting this temperatures.

I still have an option to return my board back to shop and change it for something else in that price range.

Could be just bad sitting VRM cooler on board? (Its not based on screws), just some kind of plastic lock / secure, like chipset and so on.

What u guys saying, about 24/7 using temps for VRMs as SAFE?

 

[TJ Hooker]

IIRC a typical max temperature for power MOSFETs used in VRMs is 125 C. I'd say anything under 100 C is leaving plenty of margin. I'm assuming you looked at the VRM temp while running some sort of CPU stress test/benchmark? Your CPU will rarely be running that hard anyway, so that is a worst case VRM temp not a typical one.

 

[quadrax]

Yeah. These measured temps were under extremely stress test for testing stability of my CPU overclock. It was in running LinX with allocated 14GB of RAM. I really dont need extremely overclocks on edge of silicon lottery, but i want something like mid overclock at least to 4.7-4.8 GHz. Maybe in other way i can set AVX negative offset for being sure, there will nothing dangerous happens. Some games using AVX (AFAIK - CS:GO, GRID and maybe more?) Or some programs. From time to time i am rendering some videos and on daily basis i am working with music.

But back VRMs. 125°C is maximum for them? What's about lifespan or damaging something? I really want keep them at least under 100°C if it's possible and if it's not dangerous. Ambient temperature i forget to say is momentally 22°C, but it's winter there, so in summer it could be worse. So need little space for temperature reservation for these days.

Is this safe then? Let's say 90°C in 24/7 without any damaging something or shorten lifespan of MOSFETs / VRMs in long term use? I want to stay at least with this board 2 years or maybe more (not sure about upcoming intel cpus).

These are VRM and MOSFETs : ISL95866 (4+3) 2x 4C10N 2x 4C06N

 

[TJ Hooker]

Looks like those FETs actually have a maximum junction temp of 150 C. So yes, having a max VRM temp of 90 C under max load is totally fine for 24/7 use.

 

[quadrax]

Thanks for the reply and information. I found datasheet as well for MOSFETs, but nothing for used VRMs, any kind of information - what they can survive. But probably i will try to stay under 90°C in extremely stress condition. In normal using it should be okay then, even high using of CPU intensive programs probably will not use AVX instructions and if so, it wouldn't be in so much extreme "Horse" pack. But it's for me little confusing at all, someone says on net under 80°C. But 90°C then should be okay if it's not all time then. Thanks one more time for clarification.

 

[TJ Hooker]

A VRM consists of MOSFETs, chokes (inductors), capacitors, and a PWM controller (the ISL95866 you listed). I'm fairly confident the VRM temperature reading would be measuring the MOSFET temp. I doubt the controller would really heat up much, and the chokes shouldn't be particularly temperature sensitive given they're just passive coils. The caps may heat up a bit, but the MOSFETs are where the action is happening, and thus likely what is being monitored.