Help required - Extreme GPU Idle power draw and crash

[rving]

Guys,

 

Having too much time on hand and trying to push things with my PC, I thought I would use a Kraken G12 with a Corsair H80i on my Zotac Twin fan 2080 Super. The thing ended up failing on load test, so I went back to the original heatsink, thinking the VRMs might be the issue (Not enough cooling?). However the stock heatsink showed the same issue - desktop is fine, but any time a load is placed - like a game - a small electric hiss comes from the case and the display goes black. The GPU fans start whirring at max. Only a hard switch off helps.

I though I needed more thermal paste, so did that. I also thought new thermal pads should do the trick, and they're in the mail. However, when I was still digging around for an answer, noticed that my idle power consumption was 65W! 

I reinstalled my graphics driver and it went down to 25W, which is still extremely high compared to when new.

My CPU alsp for some reason seems to run much hotter than usual on idle at 40degrees C(9900k, H115i Platinum), with an idle power draw of 33W!!

Tried reinstalling motherboard BIOS, and clearing CMOS, no change

Does anyone have an idea what went wrong here?

[emosun]

what is the actual wall power draw as opposed to gpu-z becuase nobody should be using software based power readings.

[rving]

I don't have a meter/device to measure that

[rving]

Where does the control for the GPU VRM reside at? Can MOBO influence that?

 

[rving]

If I had nicked something on the board during teardown, can something like this happen?

[chckovsky]

A similar thing happened to me when old thermal paste lost contact with the heat sink on an Rx570. Slightly higher than normal idle temps, rushing up to 95c (shutoff temp for Polaris) within 5 seconds of load. Cleaning and repasting fixed the issue completely.

EDIT: I'm only saying this because you haven't expressively stated that you cleaned and repasted the card after disassembly.

[rving]

I cleaned and repasted with Nh-1. But I'm using the older thermal pads as of now. The newer thermal pads will arrive today, will check with them as well

[rving]

But will the power consumption be higher as well with inefficient cooling?

[rving]

VDDC is at 0.64V vs 0.6V at normal idle earlier. Before the driver reinstall, it was 0.8V!!

I was using MSI Afterburner when I did the G12 mod. I disconnected both the fans and the small RGB logo while running it on the G-12. That and the change in voltage post driver reinstall gets me thinking whether something in the GPU BIOS got screwed, messing with the Voltage controller mechanism. Is that even possible? 

Another thing to note is that the GPU temps don't hit 38-40degrees straight at boot. It heats up over ten minutes, and I can literally feel the backplate exactly above the VRM heating up first, and then radiating towards everywhere else slowly, raising the temperature.

 

Will VRM cooling inefficiency do that? Mess with power/voltage delivery?

Or have I nicked something on the board? 

[rving]

Flashed VBIOS. Didn't help

[rving]

Changed the thermal pads - didn't help.

 

Anyone? Help please?

[rving]

Tried running stress tests with reduced power limits. The buzzing starts at 120W power draw and system crashes around 140W-150W. Definitely nicked something I think. Any idea what it could be?