1080ti stucked 139mhz and 405mhz

[alesiomaso]

PCB looks fine. No signs of modding from what i can tell?

No signs, its too weird, maybe some resistor or capacitor is causing me to suck 400w out of the pci port and it put the gpu in a kind of safe mode, but neither seems to be damaged. Any way to check this?

 

[80251]

Too bad you can't get a circuit schematic for the power sensing circuit, although how you would even troubleshoot something like that is beyond me.

 

[Toothless]

Is this one of those "1080ti but runs like a 1050" posts on ebay? Possibly a damaged power delivery or messed up bios from mining.

 

[alesiomaso]

Is this one of those "1080ti but runs like a 1050" posts on ebay? Possibly a damaged power delivery or messed up bios from mining.

Its a real 1080ti, but yes i think that de trobble es in the zone of "PCIe slot". GPU-z says that is consuming 390w and that its imposible. I have flashed the bios but its still the same.

 

[agent_x007]

[EDIT from future self : Don't assume PCB layout from placement of ref model, what I write below is wrong]
Well that explains it... you probably don't HAVE PCI-e shunt resistor (element with "R005" text)

It could fall off, after metal paste mod was used too for long time.

Either way, card won't work right in this state.
You need to solder new one.

 

[alesiomaso]

Well that explains it... you probably don't HAVE PCI-e shunt resistor (element with "R005" text)

View attachment 199378

Look at this video:

Are the same graphics and there are nothing too (or maybe I can't see well). In this video the graphics fault by other thing, so thats no the problem

 

[oldwalltree]

Well that explains it... you probably don't HAVE PCI-e shunt resistor (element with "R005" text)

View attachment 199378
It could fall off, after metal paste mod was used too for long time.

Either way, card won't work right in this state.
You need to solder new one.

Good catch!

 

[Anoniem]

Well that explains it... you probably don't HAVE PCI-e shunt resistor (element with "R005" text)

View attachment 199378
It could fall off, after metal paste mod was used too for long time.

Either way, card won't work right in this state.
You need to solder new one.

Nah, that's okay, just look at the link (TPU review of this card). I'm not sure if that ITE chip on photo 7 is in contact with a thermal pad (ergo silicone oil remnants) but it reminds me of SuperIO chips that crapped themself.

 

[StefanM]

Does following command at least find the higher clocks?

nvidia-smi.exe -i 0 -q -d SUPPORTED_CLOCKS

 

[agent_x007]

Nah, that's okay, just look at the link (TPU review of this card). I'm not sure if that ITE chip on photo 7 is in contact with a thermal pad (ergo silicone oil remnants) but it reminds me of SuperIO chips that crapped themself.

You are correct. PCI-e shunt for ASUS card is on the back of card (I assumed is on the same side as rest because of ref model - this is my mistake).
I still think shunt is the culprit though (or monitoring circuit for it is damaged in some way).
OP's card does have PCI-e shunt, and is attached (properly ?).

 

[Anoniem]

You are correct. PCI-e shunt for ASUS card is on the back of card (I assumed is on the same side as rest because of ref model - this is my mistake).
I still think shunt is the culprit though (or monitoring circuit for it is damaged in some way).
Card in question DOES have it attached.

Well you can measure a shunt And soldering something like a shunt isn't that hard but you need some equipment.

Oh and ignore my comment about the ITE 8915FN-56 chip, that regulates and monitors fans. If the card has massive GPU sag it COULD affect the solder joints, my RTX 2080 Trio from MSI had that issue. I'm no expert, but I have seen a lot of PCB's and SMD stuff, I can't find any fault in the pics. What doesn't help is that GPU PCB's are insanely complicated (four layer PCB anyone?).

 

[80251]

It might be worthwhile to contact Asus and see if they would be willing to repair your card for a fee or could maybe refer you to a company that could do so.

You might consider contacting der8auer too on facebook or instagram. He might have some ideas/connections.
https://www.youtube.com/c/der8auer/about

 

[alesiomaso]

It might be worthwhile to contact Asus and see if they would be willing to repair your card for a fee or could maybe refer you to a company that could do so.

You might consider contacting der8auer too on facebook or instagram. He might have some ideas/connections.
https://www.youtube.com/c/der8auer/about

The problem is that I am Argentine, here we do not have Asus service, much less could I contact Der8auer

Hey guys, I found this but I'm not sure if it's the problem. In the first photo we see a photo taken from the link provided by @Anoniem , in the second we see a capture of the gpu of the video that I passed previously, and in the third my gpu.

In the video the problem is not that and in the end the gpu works, I think it comes from the factory like that, maybe in the link that Anoniem provided is not the exactly model.

What do yo think?

 

[Anoniem]

The problem is that I am Argentine, here we do not have Asus service, much less could I contact Der8auer

Hey guys, I found this but I'm not sure if it's the problem. In the first photo we see a photo taken from the link provided by @Anoniem , in the second we see a capture of the gpu of the video that I passed previously, and in the third my gpu.

In the video the problem is not that and in the end the gpu works, I think it comes from the factory like that, maybe in the link that Anoniem provided is not the exactly model.

What do yo think?

You're right! Because of the stupidly long name I made an error, it's the Strix OC that I posted, not your Strix Gaming OC. There might be some minor variations in the design but nothing major. Anyhooo, those caps don't really matter if I'm right, they're probably output filtering for the memory VRM.

And it sucks that Roman isn't an option, he's a really nice dude Might want to check if there is any reputable shop that does SMD repair and all. Ask them what their hourly rate is and try to make a deal about how much time they're allowed to spend on it. It's a really tricky situation since getting a new card might be a bit of a challenge and these repairs can be quite expensive.

 

[80251]

I recently had a Gigabyte 1080ti Gaming OC repaired, but for mine, the problem was obvious, a capacitor was missing and a MOSFET had clearly become desoldered. The electronics technician who did the repair said that there were some components on the videocard that he couldn't source and that if they had also failed there would be nothing he could do about it -- even with his surface mount soldering station.
Anoniem might have the only solution and since the card is partly functional (my 1080ti was completely non-functional) at least there's some hope in that direction.

 

[alesiomaso]

I recently had a Gigabyte 1080ti Gaming OC repaired, but for mine, the problem was obvious, a capacitor was missing and a MOSFET had clearly become desoldered. The electronics technician who did the repair said that there were some components on the videocard that he couldn't source and that if they had also failed there would be nothing he could do about it -- even with his surface mount soldering station.
Anoniem might have the only solution and since the card is partly functional (my 1080ti was completely non-functional) at least there's some hope in that direction.

IS VERY CRAZY!!! If there don't seem to be any PCB faults, and flashing BIOS doesn't fix it either, what could it be???

 

[Falkentyne]

The problem is that I am Argentine, here we do not have Asus service, much less could I contact Der8auer

Hey guys, I found this but I'm not sure if it's the problem. In the first photo we see a photo taken from the link provided by @Anoniem , in the second we see a capture of the gpu of the video that I passed previously, and in the third my gpu.

In the video the problem is not that and in the end the gpu works, I think it comes from the factory like that, maybe in the link that Anoniem provided is not the exactly model.

What do yo think?

I can't even find any review anywhere that shows PCB pictures of a "Strix 1080 Ti Gaming OC" anywhere.
All i can find is the regular "Strix 1080 Ti OC".

And the shunts seem to be different on them. The alignment of the writing is different.

https://www.techpowerup.com/review/asus-gtx-1080-ti-strix-oc/images/front_full.jpg

https://www.techpowerup.com/review/asus-gtx-1080-ti-strix-oc/images/back_full.jpg

But from the picture you posted of your own card, I don't see any problem with the shunts, besides this issue.
I was just trying to determine if the shunts were original or not. But the soldering is way too clean for it to have been installed by an end user.

so I have no idea what's wrong with the card.

 

[80251]

IS VERY CRAZY!!! If there don't seem to be any PCB faults, and flashing BIOS doesn't fix it either, what could it be???

I'd figure a failed component on the board and maybe not something obvious or simple like a capacitor or resistor or inductor but an actual IC. I'd think the first circuit(s) to troubleshoot would be whatever monitors the power delivery to the card. Maybe a voltage comparator (if that's what they even use) has failed in the power monitoring circuit, but who knows.

 

[alesiomaso]

I can't even find any review anywhere that shows PCB pictures of a "Strix 1080 Ti Gaming OC" anywhere.
All i can find is the regular "Strix 1080 Ti OC".

And the shunts seem to be different on them. The alignment of the writing is different.

https://www.techpowerup.com/review/asus-gtx-1080-ti-strix-oc/images/front_full.jpg

https://www.techpowerup.com/review/asus-gtx-1080-ti-strix-oc/images/back_full.jpg

But from the picture you posted of your own card, I don't see any problem with the shunts, besides this issue.
I was just trying to determine if the shunts were original or not. But the soldering is way too clean for it to have been installed by an end user.

so I have no idea what's wrong with the card.

Do you think that changing a chip like the bios chip or another chip that is failing can solve the problem? Or is a software failure?

 

[Anoniem]

IS VERY CRAZY!!! If there don't seem to be any PCB faults, and flashing BIOS doesn't fix it either, what could it be???

Well that's the tricky part, visually it could be fine but you can't know for certain. Example: The current sensing circuit needs to know how much power is being drawn from the PEG cables (PCIe cables from the PSU) and PCIe slot. If something interferes with that and the controller doesn't get what it wants it'll block the card from running higher current/voltage. Why? To prevent damage or tampering. That is why some overclockers solder another shunt on top of an existing shunt to fool the controller in thinking it uses less power. That way it'll allow you to use more in reality, giving you more headroom to overclock.

That hunk of semi dead PCB is extremely complicated, bending it for a long time might fudge something up like a trace or solder joint. Could you post a picture of the card laying on it's side? Like how it's normally oriented in a PC.

PS. All the marks on the PCB that don't have components but do have solder is normal. There's no way any sane company would make a separate assembly line for a slightly different card.

 

[80251]

Do you think that changing a chip like the bios chip or another chip that is failing can solve the problem? Or is a software failure?

I'd think so. For example, if the power monitoring circuit uses a voltage comparator and that failed, it might throw off the power readings.

 

[alesiomaso]

Well that's the tricky part, visually it could be fine but you can't know for certain. Example: The current sensing circuit needs to know how much power is being drawn from the PEG cables (PCIe cables from the PSU) and PCIe slot. If something interferes with that and the controller doesn't get what it wants it'll block the card from running higher current/voltage. Why? To prevent damage or tampering. That is why some overclockers solder another shunt on top of an existing shunt to fool the controller in thinking it uses less power. That way it'll allow you to use more in reality, giving you more headroom to overclock.

That hunk of semi dead PCB is extremely complicated, bending it for a long time might fudge something up like a trace or solder joint. Could you post a picture of the card laying on it's side? Like how it's normally oriented in a PC.

PS. All the marks on the PCB that don't have components but do have solder is normal. There's no way any sane company would make a separate assembly line for a slightly different card.

I'd think so. For example, if the power monitoring circuit uses a voltage comparator and that failed, it might throw off the power readings.

How can I know if the monitoring circuit fails? With a multimeter?

 

[Anoniem]

How can I know if the monitoring circuit fails? With a multimeter?

That'd be a great start BTW, thanks for the photos, the card looks straight as an arrow so that makes it a little better

 

[Gungar]

I think that maybe with an volimeter we can find the problem and replace

Yeah probably but i don't know what i am looking for, it would be a lot easier for someone that knows this stuff to find what's not working properly.

 

[lowrider_05]

maybe this is a dumb idea of mine but have you tried flashing the XOC Bios on your Card, it has no Power restrictions and may ignore the fault? But it also could fry the card if it goes wrong.

here a link to tutorial: https://www.overclock.net/threads/how-to-flash-a-different-bios-on-your-1080-ti.1627212/

-Please be careful, I am not to blame if you blow up your card -