I just replaced the thermal pads on two of my Zotac 3080 Trinity OC. The reported temp is now 12-15
degrees cooler than they were before. Since changing those tapes, the hashrate on these two cards is
in the high 60’s. Previously both cards were at 95+MH. I have verified the tapes are in the right position,
and that the thermal paste for the CPU is making good contact. I do not want to tear anymore of these
cards down to replace the thermals if I am going to loose 30MH! Set up below.
MB: ASROCK H110
CPU: I5 6500
MEM: 16GB
Power: 2- EVGA 1600 and 1- EVGA 1300 all platinum models
GPU’s: 6 AMD 5700XT (hash 51-54MH, at 100W)
7 Zotac 3080 Trinity OC
Any suggestions would be greatly appreciated!!
Here is a picture of what is happening. GPU 0 and GPU 2 have had the thermal pads changed to new ones. GPU 1 is stock with no changes. The hasrates for GPU 0 and 2 were just as high prior to the thermal pad change.
well are your XT’s modded? that hashrate for those is down about 20mhs also. those should do 50 easy. why are you setting your fans at 100%? you can def see that the power is down on the 2 that are hashing low, and those were the same settings when you turned it off to replace the pads? how hot was it running when you shut it off to re-pad them? it almost seems like maybe your power supply’s are putting out enough? im just trying to think here because just changing pads unless you put them on wrong there is no reason for it to throttle, what kind of pad did you use? how thick did you use? one of them the mem is set at 200,
try core 200 mem 2200 power 240
Im sorry about what happened after your pad changed.
I actually have the same but worse problem.
Im actually running at 38 and it low to 33 MH with a 3080 FE.
I did all repadding over and over but can’t find any solution.
The stranger is that my card is actually working well in my gaming PC did all sorts of benchmarks and there is nothing wrong.
I just cant mine with it anymore.
If someone heard about that and can help I’ll be super glad
So it seems my rebuild left something to be desired. My cpu temp went way up at first.
After rebuilding these cards one more time, the temp is down but the hashrate is higher, 92MHS, but not the 98 I am used to seeing. I think that my “fixing” these cards caused all of my issues.
Chmmac, I would love for you to PM me. I would like to go over your 3080 settings and your replacement of the thermals. Thanks!!
You might have used incorrect thickness pads. Sometimes if the pads are thicker than the factory you might actually end up creating a little space between the cooler and the actual die of the GPU which could be leading to your issues.
I did the thermal pad mod for my Gigabyte Gaming OC 3080 and now I am sitting at 100mhs at only 90c memory junction temp.
Earlier I used to start throttling at stock settings and never cross 80mhs on my card.
Try to do the thermal mod again but this time check the thickness of your pads and match that to what your card needs.
My Gigabyte needed 2mm pads on the top of memory on the front. I also added thermal pads to the back of the card even though there is no memory there just so I could use the metal backplate as a passive heat sink.
I needed 3mm pads for the back of my card. So again, double check the thickness of your pads
do you have more 3080? what are the normal temperatures for them!? i have 3, with replaced thermopads, i try to keep them under 50c… what are acceptable/safe temperatures?ambient temp is 26-27 (…) i know but its cheapper to replace the fans then the GPU
For the 3080 core temperature does not matter. You need to check mem junction temp and unfortunately that cannot be seen in Linux. I checked my temps while mining in Windows after my mod and then I switched over to HiveOS.
Without the mod. I wouldn’t even be able to run at stock without throttling.
So core would be around 57-58c and mem junction would be close to 108c at stock.
After the mod core would be 57-58c and mem junction would be at 88c running at absolute clock 1060 for core and 2600 for mem in HiveOS hashing at 100mhs
Ok, thx
I run -200core 2960mem 238w 103,5mhs stabel
Now it gets warmer so i decresed to 2500 mem
But was wondering whaa are acceptable/safe temps (gpu)
Btw, if i set autofan to cpu 45c what it does? It turns on fans first and after it reach 99-100% of the fans them it starts to throttle? Is that correct?
Wow those are nice overclocks but I’d definitely recommend using an absolute clock for nvidia instead of a negative offset.
Autofan keeps your card at the temp you have specified. Fans will eventually go to 100% if the mem starts throttling. Usually autofan only tracks core temp and not mem junction.
Super, thx for all infos! I try with absolut values on 3080 it did not work for me before , i have all 3070 running “absolute”
Thx once more, happy mining especially today ;))
