Question Trouble with 3080 after new pads and PTM7950 ?

[DasReelGood]

Hello,

I recently repadded my EVGA XC3 10GB LHR 3080. I used Kritical Pads from their website. I followed their instructions from their website on where each pad goes and applied PTM7950 on my GPU's core.

I experienced a number of issues after repasting my GPU.

Code 43:
I managed to fix this by using DDU and reinstalling a graphics driver using NVCleanstall. I also removed and then re-enabled XMP on my RAM, and lowered my cache overclock from 4.8 to 4.4.

Now my current issue is that my PC will blackscreen and crash entirely when there is a heavy load from gaming and some benchmarks on it. I cannot figure why exactly this is.

For some reason MSI Afterburner is allowing my card to use voltage up 1.094v. This is insanely high from my understanding and I've tried to undervolt it to 1950MHz at 0.950v. This undervolt was stable beforehand. It never crashed during countless hours of gaming and withstood OCCT and Unigine Heaven [I know Unigine Heaven is not a good test for stability but I noticed it does generate a LOT of heat].

These crashes happen with or without my CPU overclocked, or XMP enabled. When I was debugging the code 43 I noticed some times the GPU only worked with my XMP and CPU/cache overclocks enabled.

I have a Corsair RM850x [850 Watt Gold] power supply.

Images from OCCT monitor tab; the high values are from trying to play Call of Duty Black Ops 6. There was heavy artifacting and I was certainly going to crash but I managed to close it down in time to also get these readings.

Edit:

Because I'm not allowed to insert photos here some possible answers to questions:

Max Temp / HotSpot Temp: 49.34C / 60.35C
GPU Mem Temp: 56C
GPU Max Power: 187.81Watt
GPU Max Clock: 1980MHz
GPU Max Memory Clock: 1187.5MHz
GPU Video Clock: 1740MHz

 

[beyondlogic]

Hello,

I recently repadded my EVGA XC3 10GB LHR 3080. I used Kritical Pads from their website. I followed their instructions from their website on where each pad goes and applied PTM7950 on my GPU's core.

I experienced a number of issues after repasting my GPU.

Code 43:
I managed to fix this by using DDU and reinstalling a graphics driver using NVCleanstall. I also removed and then re-enabled XMP on my RAM, and lowered my cache overclock from 4.8 to 4.4.

Now my current issue is that my PC will blackscreen and crash entirely when there is a heavy load from gaming and some benchmarks on it. I cannot figure why exactly this is.

For some reason MSI Afterburner is allowing my card to use voltage up 1.094v. This is insanely high from my understanding and I've tried to undervolt it to 1950MHz at 0.950v. This undervolt was stable beforehand. It never crashed during countless hours of gaming and withstood OCCT and Unigine Heaven [I know Unigine Heaven is not a good test for stability but I noticed it does generate a LOT of heat].

These crashes happen with or without my CPU overclocked, or XMP enabled. When I was debugging the code 43 I noticed some times the GPU only worked with my XMP and CPU/cache overclocks enabled.

I have a Corsair RM850x [850 Watt Gold] power supply.

Images from OCCT monitor tab; the high values are from trying to play Call of Duty Black Ops 6. There was heavy artifacting and I was certainly going to crash but I managed to close it down in time to also get these readings.

Edit:

Because I'm not allowed to insert photos here some possible answers to questions:

Max Temp / HotSpot Temp: 49.34C / 60.35C
GPU Mem Temp: 56C
GPU Max Power: 187.81Watt
GPU Max Clock: 1980MHz
GPU Max Memory Clock: 1187.5MHz
GPU Video Clock: 1740MHz

id you make sure all the pads for the vrm etc are intact its possible that they are overheating or not making contact.

 

[DasReelGood]

id you make sure all the pads for the vrm etc are intact its possible that they are overheating or not making contact.

I opened it back up and there are light indents in the thermal pads; leading me to believe they are making adequate contact.

However, I can not find any thermal readings for them. Only the card's junction, hotspot and memory.

 

[JayGau]

Hello,

I recently repadded my EVGA XC3 10GB LHR 3080. I used Kritical Pads from their website. I followed their instructions from their website on where each pad goes and applied PTM7950 on my GPU's core.

I experienced a number of issues after repasting my GPU.

Code 43:
I managed to fix this by using DDU and reinstalling a graphics driver using NVCleanstall. I also removed and then re-enabled XMP on my RAM, and lowered my cache overclock from 4.8 to 4.4.

Now my current issue is that my PC will blackscreen and crash entirely when there is a heavy load from gaming and some benchmarks on it. I cannot figure why exactly this is.

For some reason MSI Afterburner is allowing my card to use voltage up 1.094v. This is insanely high from my understanding and I've tried to undervolt it to 1950MHz at 0.950v. This undervolt was stable beforehand. It never crashed during countless hours of gaming and withstood OCCT and Unigine Heaven [I know Unigine Heaven is not a good test for stability but I noticed it does generate a LOT of heat].

These crashes happen with or without my CPU overclocked, or XMP enabled. When I was debugging the code 43 I noticed some times the GPU only worked with my XMP and CPU/cache overclocks enabled.

I have a Corsair RM850x [850 Watt Gold] power supply.

Images from OCCT monitor tab; the high values are from trying to play Call of Duty Black Ops 6. There was heavy artifacting and I was certainly going to crash but I managed to close it down in time to also get these readings.

Edit:

Because I'm not allowed to insert photos here some possible answers to questions:

Max Temp / HotSpot Temp: 49.34C / 60.35C
GPU Mem Temp: 56C
GPU Max Power: 187.81Watt
GPU Max Clock: 1980MHz
GPU Max Memory Clock: 1187.5MHz
GPU Video Clock: 1740MHz

Are those temps while gaming? If they do it looks extremely low. Hot spot and VRAM should be much higher than that.

 

[beyondlogic]

I opened it back up and there are light indents in the thermal pads; leading me to believe they are making adequate contact.

However, I can not find any thermal readings for them. Only the card's junction, hotspot and memory.

did you use the correct size of thermal pads. even if temps look fine it might not be dissipating the heat properly.

 

[DasReelGood]

did you use the correct size of thermal pads. even if temps look fine it might not be dissipating the heat properly.

The company I bought them from pre-cuts and sizes them. I checked online for negative reviews but couldn't find any.

Are those temps while gaming? If they do it looks extremely low. Hot spot and VRAM should be much higher than that.

Yeah, I was on the verge of crashing nearly instantly and these are the readings just before I almost crashed.

The power draw was only 187 watts and I know this card can easily do 310 watts.

That's why I'm so confused.

 

[Gaidax]

By size, you need to actually have them at proper thickness, the same as were on stock pads for that very specific card model.

Examples of such disasters:

View: https://youtu.be/wi7MkHv6TKI

View: https://www.youtube.com/watch?v=nIzgQhRoJS8

 

[DasReelGood]

By size, you need to actually have them at proper thickness, the same as were on stock pads for that very specific card model.

Examples of such disasters:

View: https://youtu.be/wi7MkHv6TKI

View: https://www.youtube.com/watch?v=nIzgQhRoJS8

I understand this. The company selling the pads pre-cut the sizes required. 1mm, 2.6mm, 2.8mm and 2.75mm. They all seem to make proper contact.

I can see the chips' indents in the pads when I opened the card. I do not see any light between the pads and anywhere they're supposed to be touching.

The only thing I can't find is vrm temps or anything that could make it easier to be certain that nothing is over heating.

 

[Gaidax]

I understand this. The company selling the pads pre-cut the sizes required. 1mm, 2.6mm, 2.8mm and 2.75mm. They all seem to make proper contact.

I can see the chips' indents in the pads when I opened the card. I do not see any light between the pads and anywhere they're supposed to be touching.

The only thing I can't find is vrm temps or anything that could make it easier to be certain that nothing is over heating.

The problem may be, just like in a first video, that the whole tampering ended up damaging memory or core contacts.

You have very similar problem with artifacts, just like in that vid. It is possible while doing replacement you ended up displacing or damaging component connection.

Can try to do the same or similar memory test that guy did to begin with to see if errors are returned. If they are, then you'd need a repair. Artifacts are a strong indicator of either memory or core contact issue.