MSI RTX3060TI VENTUS 3X OC issues

[Foodog]

Hello, I have 2 mining rigs with a mix of MSI RTX 3060TI's VENTUS OC 2X and VENTUS OC 3X (they are all LHR's). These rigs have been mining for about 5 months now. And I am starting to have issues with the VENTUS 3X (although it runs "cooler" than the 2X model).  Here are the symptoms : The card will run without problems with some overclocking settings for a few months (mining ETH in this case) then it will start to crash and the OC settings have to be reduced, after the changes it will work fine for a few days/weeks and it will start to crash again, and OC settings have to be reduced again up to the point where they have to be set to 0 (in HiveOS) when it finally crashes all the time. This problems seems to affect all the VENTUS 3X that i have (in 2 different rigs).
At first i suspected the thermal pads (on the memory chips) because after disassembling the card. The pads were really oily however it is difficult to determine the actual thickness of the pads, although i used a caliper (they are very soft and quite squashed) but they appear to be in the 2.5mm-3mm. I have tried 3mm Gelid Extreme (are quite stiff although the Ultimate is stiffer)   did not work at all (the processor was not making contact with the heatsink) i tried 2mm and 2.5mm Maxtor (which are softer) and the best result i got was a "running" card with 0 Memclock applied in HiveOS for 2 weeks after it started to crash.
We are in winter and the cards are running in the low 50's with hotspot (Junc. Temp low 60's) so overheating is not the issue.
It is now the 4th GPU (out of 9) affected with this and probable it will be all of them in the near future. And i cannot believe that it is a harware failure.
Has anybody encountered this problem ?? Has anybody changed the thermal pads on this card  ?? if so what is the thickness and the model of the pads used ??
Any solution or advice would be greatly welcome :-) Thks in advance.           

[swogerino]

Hello, I have 2 mining rigs with a mix of MSI RTX 3060TI's VENTUS OC 2X and VENTUS OC 3X (they are all LHR's). These rigs have been mining for about 5 months now. And I am starting to have issues with the VENTUS 3X (although it runs "cooler" than the 2X model).  Here are the symptoms : The card will run without problems with some overclocking settings for a few months (mining ETH in this case) then it will start to crash and the OC settings have to be reduced, after the changes it will work fine for a few days/weeks and it will start to crash again, and OC settings have to be reduced again up to the point where they have to be set to 0 (in HiveOS) when it finally crashes all the time. This problems seems to affect all the VENTUS 3X that i have (in 2 different rigs).
At first i suspected the thermal pads (on the memory chips) because after disassembling the card. The pads were really oily however it is difficult to determine the actual thickness of the pads, although i used a caliper (they are very soft and quite squashed) but they appear to be in the 2.5mm-3mm. I have tried 3mm Gelid Extreme (are quite stiff although the Ultimate is stiffer)   did not work at all (the processor was not making contact with the heatsink) i tried 2mm and 2.5mm Maxtor (which are softer) and the best result i got was a "running" card with 0 Memclock applied in HiveOS for 2 weeks after it started to crash.
We are in winter and the cards are running in the low 50's with hotspot (Junc. Temp low 60's) so overheating is not the issue.
It is now the 4th GPU (out of 9) affected with this and probable it will be all of them in the near future. And i cannot believe that it is a harware failure.
Has anybody encountered this problem ?? Has anybody changed the thermal pads on this card  ?? if so what is the thickness and the model of the pads used ??
Any solution or advice would be greatly welcome :-) Thks in advance.           

I have a couple of the 2X OC Ventus from the 3060 ti and they are mining from nearly 1.5 years now,I am in the middle of the summer here,running the cards with 91% fan and the core temperature is 69 degree Celsius,because is hot in Europe at 40 degree Celsius daily temperature,with the correct Windows OC settings being -200 core clock and +1059-1200 Memory clock depending how much it can handle,mine max runs at 1068 without crashing.Not a single crash I have from this card and it run perfectly from the first day up to now.I use Windows 10 which gets all the updates and latest version of Lolminer.The card is in a mixed rig environment,meaning it run together with different other cards.

For the 3X OC I don't know but I doubt to be hardware failure because a couple of friends of mine have them and they are mining without problems in HiveOS and Windows.

[FP91G]

Any solution or advice would be greatly welcome :-) Thks in advance.            

Show your settings in Hive OS and write the configuration of the mining farm with the power supply.
What drivers do you have installed?
Have you tried changing the miner?
What memory do video cards have? Hynix or something else?
Need more information.

[Foodog]

Hi swogerino,
Thks for answering  my post, Regarding the gpu u have (ventus 2x) if you 've been mining with it for a year and half i presume you have the first version (which has samsung memory , the ventus 2X OC1 (actual model) and the ventus 3X OC have Hynix) it is also thicker than the actual model (it was derived fron the 3070 from what i read) and therefore has better cooling than the actual ventux 2x. Anyway i do not have the problem (yet) with the Ventus 2x.  I do not know if i explained my problem correctly : I have 20 cards (10 ventus 2x OC1 and 10 ventus 3x OC ) in 2 rigs . They are all about 6 months old (i will not dwell on the obvious warranty issue since i live in Argentina (that says it all)). I have some Ventus 3x on Rig1 and some on Rig2. I am running HiveOS (but its trivial on windows it's the same) I have run into the situation where all cards (ventus 3x) on both rigs are running perfectly... Then some (again ventus 3X) started to degrade (crashing) -> Reduced the OC settings (obviously on the crashing cards) -> to cannot mine anymore on these cards (17-21 Mhs). I have now 4 Ventus 3x in this state (which i took out from the rigs) . I figured i needed to change the thermal paste and pads and upon opening the GPU its OBVIOUS that there is a problem with the pads : they are squashed and dripping with oil (yes it's that bad) i will post some pics if possible. although it does not really show the oil on the pics.
Upon replacing the pads (on the VRAM and the VRM's) and the thermal paste : No improvements ( i cannot mine with the cards even without OC (51Mhs) it will crash). The pads i tried are 2mm , 2.5mm (Maxtor)  and 3mm (Gelid-Extreme) which are quality pads but stiffer than the "originals" which seem VERY soft (therefore VERY CHEAP...) and maybe that's the solution. I sent a ticket to MSI here are their answer :   "...The thickness of your thermal paste is 2.25~2.5mm. According to the company's requirements, I cannot provide the brand of thermal paste..." they probably meant PADS not too many details in their response. One guy replied to my post (exact same as this one) ont he MSI forum. He has a ventus 3x was mining on windows with a hashrate of ~57 Mh/s. Figured he could improve that by changing the thermal paste and pads and guess what he is mining at 17 Mh/s now (with a Gpu temp of 41C, at least he is running cool :-)) just like me the guy tried 2mm, 2.5mm and 3mm pads to no Avail....https://forum-en.msi.com/index.php?threads/msi-rtx-3060ti-ventus-oc-3x-issues.377118/
I would be really interested in hearing about somebody who replaced the pads or only the thermal paste on his Ventus 3x and got somethg running "normally"). I actually tried (on the last failing card) to only replace the thermal paste and clean (remove the oil on the chips and the pads themselves) without removing the pads. It DID NOT SOLVE the problem.
Now regarding ur friends using the ventus 3x , just like have now perfectly running ventus 3x (6 of them), until the problem arises and they will face the exact situation i am in Guaranteed : a GPU that you can use to browse the web and open a spreadsheet .... Again i feel really frustrated about the whole adventure but i really appreciate ur feedback :-)

[Foodog]

Hello FP91G,
No offense here, but it has nothing to do with HiveOS settings or OC settings if you read my posts the problem gradually worsens on some of the cards (either rig1 or 2) to a perfect running GPU ---> crashes all the time without OC. NO changes in hiveOS or any hardware. FYI the memory are Hynix (but again nothing to do with the problem i am facing or maybe if hynix are known to degrade really fast which i haven't heard of and again i don't think its the cause since the Ventus 2x has hynix and i haven't encountered the problem YET on the ventus 2x) check the link of the MSI post (the guy has the exact same problem running windows).
Another why hiveOS and the whole rig setup has nothing to do : is that i run various benchmark (FurMark, heaven) on my windows (windows 10 so you don't ask ;-) machine on the 4 failing cards and they all crashed.... :-( 
My belief is : The pads failed (but maybe somethg else which i don't know) but certainly the pads are in really bad shape i would like to post pics (says it all) but cant see how to do it 
Anyway thks for ur feedback and hope that someone can help me solve my problem....
   

[FP91G]

To post photos, you can use https://imgbb.com/
I've seen a lot of graphics cards leaking thermal pads, but it doesn't affect mining. Maybe this liquid got into the connectors on the motherboard or the connectors on the raiser, then it needs to be cleaned with a special liquid. I use medical alcohol  80%.
If the FurMark test crashes on a separate computer, then I'm looking for a different BIOS for this video card.
https://www.techpowerup.com/vgabios/

[Foodog]

Hi FP91G,
"....leaking thermal pads, doesn't affect mining..."my question here is : leaking thermal pads ----> defective thermal pads or at least less effective thermal pads. The problem with the RTX 3060Ti is that there is NO WAY to show the temp on the VRAM (unlike the RTX 3070Ti or 3080 etc....) I have watched several Youtube's https://www.youtube.com/watch?v=kPYpaROZ0ZE&t=243s about people complaining of high memory temps (rtx 3080's and the like) due to bad (and cheap) original thermal pads, the advantage of the 3070ti, 3080 etc... is you can see the temp on the VRAM (GPU-Z for ex.) therefore YOU KNOW if they are making contact with the heatsink. With the 3060TI there is just NO WAY of knowing.

"...a different BIOS for this video card..." unfortunately there are 4 of them (out of 9)!!!!  i think it would be really incredible that all 4 GPUs had their BIOS affected (on 2 different rigs)

when i say the FurMark benchmark crashes i mean it will run for sometime (+/- long depending on the GPUs)  and then crashes.

What worries me is that i managed to RMA one of the faulty Ventus3X under warranty and got an answer from MSI yesterday they will return me a Ventus 2X not a Ventus 3X....     

[FP91G]

Many older graphics cards didn't even have that option. Many video cards showed only the core temperature. You can only change the thermal pads and update the bios. I was helped by updating the BIOS on video cards 2xxx series. Further, only a guarantee or seek support on the official forum.