My 1080 Ti completely stopped working mid game. Please help me!

ShiniX

n00b
Joined
Jun 23, 2021
Messages
5
Hello everyone,
My GPU (1080 TI) stopped working mid game for no reason. The screens went black and the GPU fans stopped spinning.
Did a lot of troubleshooting as mentioned below and the result is always no signal output, fans don't spin and rgp don't turn on.
I tried the following:
Plug and unplug the GPU and the 8+8 power connector
Plug it in the secondary PCI-E
Tried to put the GPU in a completely different old system, no luck.
Tried a different GPU in my system and it worked (an old ATI HD 5550)

The built-in Intel GPU is working
The GPU is Gigabyte Aorus 1080 TI Extreme Edition.
I would really appreciate any help, Is there anything I can do or try to fix it?
Especially with the current state of the market getting a replacement of any sort is out of the question 😞
 
Last edited:
Is there anything I can do or try to fix it?

Serious, not-mocking-you question: Do you have a hot air rework station?

It appears that this card has a fuse on the 12V input rail from the slot. I betcha that fuse is blown. What this usually means is that something else on the card failed, and you now have a short to ground. So, what you need to do next is take the heatsink off, and measure the resistance to the ground plane between each side of that fuse and ground. You should have a few tens of Ks from both sides. I bet what you have instead is infinity on one side, and near zero on the other.

Confirm that, and we can maybe come up with a plan to repair it.
 
Serious, not-mocking-you question: Do you have a hot air rework station?

It appears that this card has a fuse on the 12V input rail from the slot. I betcha that fuse is blown. What this usually means is that something else on the card failed, and you now have a short to ground. So, what you need to do next is take the heatsink off, and measure the resistance to the ground plane between each side of that fuse and ground. You should have a few tens of Ks from both sides. I bet what you have instead is infinity on one side, and near zero on the other.

Confirm that, and we can maybe come up with a plan to repair it.
Too much technical details there :)
I had the card since late 2017 plugged in to the PSU with no problem. Do you mean the fuse on the card popped that caused a short?
I'm not very technical but please bear with me :)
So few questions, what tool do I need to use to measure the resistance?
Once I get it, how do I locate the fuse and how to connect the other side to ground?
Thanks for responding :)
 
Too much technical details there :)
I had the card since late 2017 plugged in to the PSU with no problem. Do you mean the fuse on the card popped that caused a short?
It's the other way around. A short circuit developed somewhere on the board, and that allowed too much current to flow through the fuse, causing it to burn out. This is the intended behavior of the fuse - it sacrifices itself to prevent further damage to the system. With the proper tools, replacing the fuse is very easy, but it won't help if you don't also fix whatever created the short circuit first.

I'm not very technical but please bear with me :)
So few questions, what tool do I need to use to measure the resistance?
Oh boy... I'm going to sound super condescending when I say this, but there is no avoiding the fact that this will be a very technical operation. You will need to arm yourself with at least a basic knowledge of how the circuits on the board work in order to successfully repair a graphics card that's experienced a hardware failure.

Electrical resistance is measured in units called ohms. The tool you use to measure this is a digital multimeter, with a resistance function, like the one linked below.

Electricity always attempts to flow from regions of high voltage (the 12V wires on the power supply) to regions of low voltage (the ground plane, which is 0V). The amount of electricity that flows is called current, measured in amperes, or amps. The amount of resistance to this flow is measured in ohms, symbolized with the greek Omega. The omega symbol on the meter denotes the resistance measurement mode.

https://www.amazon.com/AstroAI-Mult...jbGlja1JlZGlyZWN0JmRvTm90TG9nQ2xpY2s9dHJ1ZQ==

All materials are electrically conductive to some degree. Some, like copper, are very conductive, and will have almost no resistance to electricity. Other materials, like air, are very poor conductors, and have really high resistance. You can measure this with the multimeter. Other materials are called semiconductors, and have variable resistance, depending on factors such as temperature, the presence of a magnetic field, or a voltage being applied to them.

What's likely happened is that a piece of semiconductor has basically melted, allowing the copper traces at each end of it to touch. Normally, the semiconductor would have a high resistance when the board is switched off, but because it's melted, its resistance is near zero, and allows a large amount of current to flow through it. When this happened, it exceeded the limits of the fuse, which burned out and stopped the flow of current.

The fuse is the white component near the left side of the slot connector. Click to embiggen; I've circled it for you.
Gigabyte-Aorus-GTX-1080Ti-Review-KitGuru-Teardown-PCB.jpg

Image shamefully stolen from KitGuru.net.

To confirm that this is what's happened, you need to take a resistance measurement between each end of the fuse, and a known ground on the board. The shell of the DVI connector is a good ground to use. On a healthy board, you'd have a few K (that is, thousand) ohms to ground on both ends of the fuse, and zero ohms when measuring from one end to the other. On your board, I suspect you'll find that you have infinity (the meter will say "OL") on the left, and 0 on the right.
 
As an Amazon Associate, HardForum may earn from qualifying purchases.
Also when in 2017 did you get since if you registered it within 30 days could be under warranty with the 4 year warranty that creates other wise 3 year
 
It's the other way around. A short circuit developed somewhere on the board, and that allowed too much current to flow through the fuse, causing it to burn out. This is the intended behavior of the fuse - it sacrifices itself to prevent further damage to the system. With the proper tools, replacing the fuse is very easy, but it won't help if you don't also fix whatever created the short circuit first.


Oh boy... I'm going to sound super condescending when I say this, but there is no avoiding the fact that this will be a very technical operation. You will need to arm yourself with at least a basic knowledge of how the circuits on the board work in order to successfully repair a graphics card that's experienced a hardware failure.

Electrical resistance is measured in units called ohms. The tool you use to measure this is a digital multimeter, with a resistance function, like the one linked below.

Electricity always attempts to flow from regions of high voltage (the 12V wires on the power supply) to regions of low voltage (the ground plane, which is 0V). The amount of electricity that flows is called current, measured in amperes, or amps. The amount of resistance to this flow is measured in ohms, symbolized with the greek Omega. The omega symbol on the meter denotes the resistance measurement mode.

https://www.amazon.com/AstroAI-Mult...jbGlja1JlZGlyZWN0JmRvTm90TG9nQ2xpY2s9dHJ1ZQ==

All materials are electrically conductive to some degree. Some, like copper, are very conductive, and will have almost no resistance to electricity. Other materials, like air, are very poor conductors, and have really high resistance. You can measure this with the multimeter. Other materials are called semiconductors, and have variable resistance, depending on factors such as temperature, the presence of a magnetic field, or a voltage being applied to them.

What's likely happened is that a piece of semiconductor has basically melted, allowing the copper traces at each end of it to touch. Normally, the semiconductor would have a high resistance when the board is switched off, but because it's melted, its resistance is near zero, and allows a large amount of current to flow through it. When this happened, it exceeded the limits of the fuse, which burned out and stopped the flow of current.

The fuse is the white component near the left side of the slot connector. Click to embiggen; I've circled it for you.
View attachment 368753

Image shamefully stolen from KitGuru.net.

To confirm that this is what's happened, you need to take a resistance measurement between each end of the fuse, and a known ground on the board. The shell of the DVI connector is a good ground to use. On a healthy board, you'd have a few K (that is, thousand) ohms to ground on both ends of the fuse, and zero ohms when measuring from one end to the other. On your board, I suspect you'll find that you have infinity (the meter will say "OL") on the left, and 0 on the right.

I really appreciate your detailed reply.

It's the other way around. A short circuit developed somewhere on the board,
By board you mean the motherboard?
But connecting an old GPU in the same PCI-E didn't cause any problems, or that might be just a random short?

I understood must of your explanation about circuits except the last paragraph when it comes to using the tool. I will follow up on your solution once I get the multimeter.

Another member below mentioned that I might still have that 4 year warranty if I registered my card within 30 days which I didn't unfortunately. Nonetheless, I wrote a ticket to Gigabyte support (no telephone for EU/Sweden support) and will see if I can still claim the 4 years warranty. Removing the heatsink now might void the warranty ofc :)

I will post an update once I get an answer from Gigabyte. Again, I'm really thankful for your help :)
 
As an Amazon Associate, HardForum may earn from qualifying purchases.
Also when in 2017 did you get since if you registered it within 30 days could be under warranty with the 4 year warranty that creates other wise 3 year
Nov 2017 but I didn't register my card in the first 30 days for the extended 4 years warranty. Anyway, I submitted a support ticket to Gigabyte to see what happens.
 
I really appreciate your detailed reply.


By board you mean the motherboard?

"Board," in this context, means the circuit board portion of the graphics card assembly.

But connecting an old GPU in the same PCI-E didn't cause any problems, or that might be just a random short?

The short circuit is on the graphics card. When you swapped out the bad graphics card with a good one, the system worked because you had removed the faulty component and replaced it with a non-faulty one.

I understood must of your explanation about circuits except the last paragraph when it comes to using the tool. I will follow up on your solution once I get the multimeter.

Another member below mentioned that I might still have that 4 year warranty if I registered my card within 30 days which I didn't unfortunately. Nonetheless, I wrote a ticket to Gigabyte support (no telephone for EU/Sweden support) and will see if I can still claim the 4 years warranty. Removing the heatsink now might void the warranty ofc :)

I will post an update once I get an answer from Gigabyte. Again, I'm really thankful for your help :)
 
Had exactly the same thing happen to my Gigabyte 1080 Ti about a week ago. Took a pic of the fuse:

1625184247598.png


I would say just from looking at it that it's fucked. Thankfully mine is still under warranty for a few more months so I'll be taking it back.
 
Had exactly the same thing happen to my Gigabyte 1080 Ti about a week ago. Took a pic of the fuse:

View attachment 371173

I would say just from looking at it that it's fucked. Thankfully mine is still under warranty for a few more months so I'll be taking it back.
Why would you say it's fucked just by looking at it? The oxidation on the solder joints is pretty normal. I'm guessing it's just a bit humid in your house.

You can check the state of the fuse with an ohmmeter. You should have zero ohms from one side to the other. If you have more than like half an ohm, it's blown, indicating a short elsewhere on the board. If you have zero ohms, then your problem lies elsewhere.

All of that being moot if you can just RMA it.
 
Why would you say it's fucked just by looking at it? The oxidation on the solder joints is pretty normal. I'm guessing it's just a bit humid in your house.

You can check the state of the fuse with an ohmmeter. You should have zero ohms from one side to the other. If you have more than like half an ohm, it's blown, indicating a short elsewhere on the board. If you have zero ohms, then your problem lies elsewhere.

All of that being moot if you can just RMA it.
I'm not exactly an electronics expert - it just looks bad to me. But if that's normal then no worries. I've tested it in two different PCs and have the same issue. Will be RMA-ing it this weekend.
 
RazorWind I have recently been watching YouTube videos on this and you explained this great. I have become so interested in small electronics and love learning about it.

If you have no RMA with gigabyte ShiniX the person I watch on you tube is NorthRidge Fix. You could contract them if you want it repaired. He seems to do good work.
 
Quick update, after going back and forth with the retailer and Gigabyte they agreed to check the card even though I didn't register for the 4 years warranty at the time. After some time, they confirmed to me that the card is actually dead but didn't give me specific diagnosis. Nonetheless, they agreed to replace it with Asus Strix 3060 Ti. Although they sent me only the card, no box or accessories but at least I have a decent GPU now. Wish they sent me the dead 1080Ti as well to attempt to repair it myself :)
Anyway, thank you very much RazorWind for being here to help
 
Wanted to ask for help…
 

Attachments

  • 646DD7A3-0A4A-42A9-8CD7-58E742D9274A.jpeg
    646DD7A3-0A4A-42A9-8CD7-58E742D9274A.jpeg
    584.3 KB · Views: 0
  • CCC6C236-A527-4A26-B2AA-11A66F54E88B.jpeg
    CCC6C236-A527-4A26-B2AA-11A66F54E88B.jpeg
    622.2 KB · Views: 0
My issue with a 1080 TI Aorus (not the extreme or watercooled), i have a blown fuse. I traced my problem to the vram mosfet(short) top left, right next to the coil.Its teh same issue as seen in this video


View: https://www.youtube.com/watch?v=8C_j8DBRTAI , if i cant post that kind of link let me know i can remove it.


https://cdn.discordapp.com/attachme...143ddcde075aaa77ae6ed974db75dbb5e1461b7f8e40&

I removed the mosfet and no more short there. The resistance on the vram coil is only 4.1 though which is low.
I em not sure its the right procedure what i want to do next is(maybe RazorWind can help me here), bridge the blown fuse(question how thick or thin should the wire be, or what else should i use?) , connect the card with a riser als secondary.

Now i check the resistance on the vram coil and also if possible feel if any vram chip gets overly warm/hot.

If all seems ok, then test with mats mods.

Thanks for any help.
 
Last edited:
I hope this is something you're doing for fun :p

Also, the first few posts are pretty funny. I'd be surprised if the average person, even in a tech forum like this, knows what a reflow station is - let alone has access to one.
 
How did you determine which FET package you needed to remove?

Also, what type of memory chips does this particular board have? One thing that should probably be noted when troubleshooting memory issues is that not every brand of memory chips has the same resistance, so the "known good" value can vary when looking at cards with one brand, versus another.
 
My issue with a 1080 TI Aorus (not the extreme or watercooled), i have a blown fuse. I traced my problem to the vram mosfet(short) top left, right next to the coil.Its teh same issue as seen in this video


View: https://www.youtube.com/watch?v=8C_j8DBRTAI , if i cant post that kind of link let me know i can remove it.


https://cdn.discordapp.com/attachme...143ddcde075aaa77ae6ed974db75dbb5e1461b7f8e40&

I removed the mosfet and no more short there. The resistance on the vram coil is only 4.1 though which is low.
I em not sure its the right procedure what i want to do next is(maybe RazorWind can help me here), bridge the blown fuse(question how thick or thin should the wire be, or what else should i use?) , connect the card with a riser als secondary.

Now i check the resistance on the vram coil and also if possible feel if any vram chip gets overly warm/hot.

If all seems ok, then test with mats mods.

Thanks for any help.

I take it you could see damage to the mosfet in question? If not, you measured a short across the part, as compared to the others?

Low resistance on the coil is normal, it's a solid copper wire wrapped around a core.

If you've managed to safely remove the damaged part, and the blown fuse, just replace it with the same part. Be a good idea to ensure there are no other shorts, find a ground point on the card, then measure resistance to each of the pads for the removed mosfet. One should lead to ground, i.e. register as a short, but the rest of the pads should not. If this is what you measure, it's likely replacing that part will fix it. Wherever you can find that sells the part (try digikey.com) likely also will have the fuse.
Too bad the video is in German, I don't understand him. Keep researching, the amp value of that fuse shouldn't be too hard to find. Email Gigabyte support with your model and serial, let them know which mosfet is blown and that the fuse is blown, and ask for the part numbers of each. Once you have that finding the parts on digikey should be easy. Order 2 or 3 of the mosfets, just in case you make a mistake. It's likely just a dollar or 2 per part, and digikey has a minimum order amount anyway. Get a few extra fuses too.

Install the new mosfet and fuse, then take some resistance measurements as shown here: in the first few minutes, if everything reads as expected, there is likely no other shorts.

View: https://www.youtube.com/watch?v=-Wz_h0kwMjg
 
Last edited:
I take it you could see damage to the mosfet in question? If not, you measured a short across the part, as compared to the others?

Low resistance on the coil is normal, it's a solid copper wire wrapped around a core.

If you've managed to safely remove the damaged part, and the blown fuse, just replace it with the same part. Be a good idea to ensure there are no other shorts, find a ground point on the card, then measure resistance to each of the pads for the removed mosfet. One should lead to ground, i.e. register as a short, but the rest of the pads should not. If this is what you measure, it's likely replacing that part will fix it. Wherever you can find that sells the part (try digikey.com) likely also will have the fuse.
Too bad the video is in German, I don't understand him. Keep researching, the amp value of that fuse shouldn't be too hard to find. Email Gigabyte support with your model and serial, let them know which mosfet is blown and that the fuse is blown, and ask for the part numbers of each. Once you have that finding the parts on digikey should be easy. Order 2 or 3 of the mosfets, just in case you make a mistake. It's likely just a dollar or 2 per part, and digikey has a minimum order amount anyway. Get a few extra fuses too.

Install the new mosfet and fuse, then take some resistance measurements as shown here: in the first few minutes, if everything reads as expected, there is likely no other shorts.

View: https://www.youtube.com/watch?v=-Wz_h0kwMjg

Ηι there and thank you for taking time to answer and help.

The video is in german but you click CC and it ads subtitles you can choose english.

Brand of the memory chips should be MIcron thats what i gather form this code on the chips ,7TA77 D9VRL .
Guy on the second video has like 35Ohm on the vram coils, which in my case is 4.2, that is low compare to it.

There was no damage visible but the resistance on that mosfet was like 20-30Ohm, the other three were around 1500 Ohm on the same spot.

I measured after removing the mosfet, which has 4 feet top and bottom. The values on all 4 of them now are almost identical.


Another guy on the german video said, it could possible for a very short ammount of time till the fuse is blown that 12v got through the ram or memory controller on the chip which will mean isnt worth to repair. He suggested also to try and see if any chip get really warm etc maybe one or more are shorted too.
Also i em in Europe so digikey is no option, parts are more costly here . Mosfet is Aon6994(https://www.gpufix.de/en/aon6994.html) i would need two of these since they work together, and a 10A fuse(https://www.gpufix.de/en/sf-1206hh10m-2.html) Think this one is suitable .That alone will cost me around 25usd include shipping. It is some cost especailly if the card/chip is dead.

That the reason i would want to bridge the fuse and connect on a pc on a riser, take some measurements, check if vram chip gets hot etc..
 
Last edited:
When you say "resistance on that mosfet," which terminal are you measuring? It matters. A lot. The only terminals on the FET packages where you can expect to see different resistance values (edit: from one phase to another) are the gates. The sources and drains are all wired up together.

Assuming one or more of the FET packages failed in the typical way, where it becomes a short between the source ("output" to compute logic and then ground) and the drain (the 12V power supply), then yes, the person talking about how the logic circuit probably received 12V for some period of time is correct. This can lead to irreparable damage to the GPU, but it doesn't necessarily mean that it actually did. Only way to find out is to clear the short, and then try to run the board and see if it works. Given that you have less than the expected resistance to ground on the output of the memory VRM, chances are, you have not yet cleared the short, which may be either inside the board or inside the GPU silicon. You may need to do a current injection test with freeze spray or a thermal camera to find it, if you run out of clues.

What resistance to ground do you have on the GPU power circuit? Measure from the coil terminals to the shell of one of the display outputs.

Do not bypass a fuse unless you are absolutely sure that you have fixed whatever the original problem was. If you do that, you could very easily do serious damage to the board that you definitely won't be able to repair.
 
Last edited:
When you say "resistance on that mosfet," which terminal are you measuring? It matters. A lot. The only terminals on the FET packages where you can expect to see different resistance values (edit: from one phase to another) are the gates. The sources and drains are all wired up together.

Assuming one or more of the FET packages failed in the typical way, where it becomes a short between the source ("output" to compute logic and then ground) and the drain (the 12V power supply), then yes, the person talking about how the logic circuit probably received 12V for some period of time is correct. This can lead to irreparable damage to the GPU, but it doesn't necessarily mean that it actually did. Only way to find out is to clear the short, and then try to run the board and see if it works. Given that you have less than the expected resistance to ground on the output of the memory VRM, chances are, you have not yet cleared the short, which may be either inside the board or inside the GPU silicon. You may need to do a current injection test with freeze spray or a thermal camera to find it, if you run out of clues.

What resistance to ground do you have on the GPU power circuit? Measure from the coil terminals to the shell of one of the display outputs.

Do not bypass a fuse unless you are absolutely sure that you have fixed whatever the original problem was. If you do that, you could very easily do serious damage to the board that you definitely won't be able to repair.
Mosfet has 4 feet i measure the resistance where the arrow is (first bottom left )which right now after removing the mosfet is 1210Ohm and on the feet directly over it (first top left) which is 1520Ohm.

with that " Given that you have less than the expected resistance to ground on the output of the memory VRM " , you mean the resistance of the vrm coil (Μagic R15) left of the mosfet which is only 4.2Ohm for me? As i see online this should
be a value of 30-40Ohm and up to 70-80Ohm depends on the memory brand.


with that "What resistance to ground do you have on the GPU power circuit? Measure from the coil terminals to the shell of one of the display outputs " you mean the Magic R22, if so:

PLacing black probe on DVI shell and red probe on either side of that coil i have almost no resistance at all, its 00.2 Ohm.
 

Attachments

  • 1704735938209.jpg
    1704735938209.jpg
    350.1 KB · Views: 0
  • 1704736746949.jpg
    1704736746949.jpg
    461.4 KB · Views: 0
Last edited:
RazorWind I have recently been watching YouTube videos on this and you explained this great. I have become so interested in small electronics and love learning about it.

If you have no RMA with gigabyte ShiniX the person I watch on you tube is NorthRidge Fix. You could contract them if you want it repaired. He seems to do good work.
A second for Northridge Fix. I've been subbed to him for a while. Do know that unless you pay $95 for expedited service you'll be waiting an average of 4-8 weeks for a repair.

EDIT: Eek... Didn't notice this was a 2+ year old post
 
Last edited:
Mosfet has 4 feet i measure the resistance where the arrow is (first bottom left )which right now after removing the mosfet is 1210Ohm and on the feet directly over it (first top left) which is 1520Ohm.

with that " Given that you have less than the expected resistance to ground on the output of the memory VRM " , you mean the resistance of the vrm coil (Μagic R15) left of the mosfet which is only 4.2Ohm for me? As i see online this should
be a value of 30-40Ohm and up to 70-80Ohm depends on the memory brand.

Yes, that's what I meant. 4.2 ohms is probably low, as you suspect.

You could try removing the low side FET from the affected phase. If that doesn't work, try removing the FETs from the other phase. If that still doesn't clear the short, solder a wire the coil terminals of the memory, and inject 1.0V at up to say, 10 amps, and see if you have any components that get warm. A thermal camera is the best way to do this, but you can use freeze spray or isopropanol to detect warm components too. Just be aware that isopropanol is flammable, so take the appropriate precautions if you do that.

with that "What resistance to ground do you have on the GPU power circuit? Measure from the coil terminals to the shell of one of the display outputs " you mean the Magic R22, if so:

PLacing black probe on DVI shell and red probe on either side of that coil i have almost no resistance at all, its 00.2 Ohm.
It's been a while since I looked at a 1080 Ti, but I believe ~200 milliohms is in the realm of normal, so that's good.

What is the resistance you have from each of the fuses to ground? It'll obviously be infinity on the positive side on the blown ones, but what do you have on the negative side?
 
Yes, that's what I meant. 4.2 ohms is probably low, as you suspect.

You could try removing the low side FET from the affected phase. If that doesn't work, try removing the FETs from the other phase. If that still doesn't clear the short, solder a wire the coil terminals of the memory, and inject 1.0V at up to say, 10 amps, and see if you have any components that get warm. A thermal camera is the best way to do this, but you can use freeze spray or isopropanol to detect warm components too. Just be aware that isopropanol is flammable, so take the appropriate precautions if you do that.


It's been a while since I looked at a 1080 Ti, but I believe ~200 milliohms is in the realm of normal, so that's good.

What is the resistance you have from each of the fuses to ground? It'll obviously be infinity on the positive side on the blown ones, but what do you have on the negative side?
Aright so you want me to remove the mosfet right next to it, and if still there remove the other 2 mosfets right under it. So despite the resistance on all mosfets are within expected range(also on the removed one on the pcb) , i should remove them in the steps you suggest?

200 milliohms is in the realm of normal - thats correct according to other sources.

The fuse that blown is directly over the slot. You want me to measure the ones that under the 2 8pins , if so : black probe dvi shell, red probe either side outer 8pin fuse is 9.8KOhm , inner 8pin fuse is 4.1KOhm. Is this normal?
 
Aright so you want me to remove the mosfet right next to it, and if still there remove the other 2 mosfets right under it. So despite the resistance on all mosfets are within expected range(also on the removed one on the pcb) , i should remove them in the steps you suggest?
Yes, although now that I think about it, I think you should do the current injection test first, and resort to removing the FETs if you don't find anything that gets warm.

Keep in mind that the large BGA components (GPU and memory ICs) are supposed to get warm, so you'd really only be concerned about them if you see that there's one tiny spot that gets super hot. If they just get vaguely warm, that doesn't indicate a problem.
 
Yes, although now that I think about it, I think you should do the current injection test first, and resort to removing the FETs if you don't find anything that gets warm.

Keep in mind that the large BGA components (GPU and memory ICs) are supposed to get warm, so you'd really only be concerned about them if you see that there's one tiny spot that gets super hot. If they just get vaguely warm, that doesn't indicate a problem.
Injection test needs a specific powersupply which i dont have(i have to look for someone who has or a store),I em able to get a thermal camera though.. What temp and over it should be considered super hot?
 
Last edited:
Injection test needs a specific powersupply which i dont have, i have to look for someone who has or a store. What temp and over it should be considered super hot?
You're not looking for any particular temperature - just areas that get hotter than the rest. Assuming the board is at ambient temperature, you're looking for components or parts of the board that are detectably warmer than that, since that's where the energy from the current you inject is going.

This is super easy with a thermal camera, since it's capable of detecting even small differences in temperature. If you have to use freeze spray or isopropanol, a successful tes will show that the frost immediately disappears from some particular component as soon as you turn on the power supply. Similarly, if you use isopropanol, what you'll see in a successful test is that the alcohol is boiling off around some component. Usually, the component in question is the one you need to replace, but in some cases, it's not, and it's some other component that's in series with the indicated one that's actually the problem.
 
You're not looking for any particular temperature - just areas that get hotter than the rest. Assuming the board is at ambient temperature, you're looking for components or parts of the board that are detectably warmer than that, since that's where the energy from the current you inject is going.

This is super easy with a thermal camera, since it's capable of detecting even small differences in temperature. If you have to use freeze spray or isopropanol, a successful tes will show that the frost immediately disappears from some particular component as soon as you turn on the power supply. Similarly, if you use isopropanol, what you'll see in a successful test is that the alcohol is boiling off around some component. Usually, the component in question is the one you need to replace, but in some cases, it's not, and it's some other component that's in series with the indicated one that's actually the problem.
I see but in order to do this test a laboratory power supply is needed. Doing 1-1.35v on maximal 10A. If I can find something like this ok good, but if not do I have another option?
 
I see but in order to do this test a laboratory power supply is needed. Doing 1-1.35v on maximal 10A. If I can find something like this ok good, but if not do I have another option?
There are a few options:

* Perform a visual inspection and see if you can see physical damage to any components
* Guess what might be wrong and start removing parts until the short is cleared. Be aware that if you remove any semiconductors, you will likely need to replace them with new ones. FETs, especially, are pretty easy to damage with heat, especially when you're working on a board like a graphics card, which tends to have a ton of copper in it, so you have to use a ton of heat to remove them.
* X-ray the board and confirm that the problem isn't traces that are burned together
 
The blown fuse explanation makes the most sense given the fans stopped, ie. all power cut to card. Unless you’re an electrical engineer that card is probably done for good.

RIP 1080 Ti
 
As an Amazon Associate, HardForum may earn from qualifying purchases.
https://www.amazon.com/Adjustable-Switching-Regulated-Quick-Charge-Interface/dp/B08HYK2ZW3/

Could get away with something cheap like this. Good to have if you are diving into repairing electronics.
I giggled at 'Wanptek.' Sounds like it's named after the sad trombone riff.

He said he's in Europe, and that stuff there is a lot more expensive, so he may not have the same number of options we have in the states, but I have a similar unit that I've used with success to fix some graphics cards in the past. You don't necessarily need super expensive gear.
 
As an Amazon Associate, HardForum may earn from qualifying purchases.
As i understand, the best is to find such power supply and continue from there. Thanks for all your answers suggestions and help, if I make any progress ill post it here(might help others).
 
Hi , i kind of missed this here" solder a wire the coil terminals of the memory, and inject 1.0V at up to say, 10 amps" . So you dont want me to inject the voltage from the 12v contacts slot of the card. Solder wire to one coil or both,side etc.. Thanks.
 
Hi , i kind of missed this here" solder a wire the coil terminals of the memory, and inject 1.0V at up to say, 10 amps" . So you dont want me to inject the voltage from the 12v contacts slot of the card. Solder wire to one coil or both,side etc.. Thanks.
Correct. You want to inject current into the coil, basically replacing the functioning VRM with your lab power supply, so you can use whatever means you have available to see where all that energy is going.
 
Well how I exactly I do that? Do I have to solder a wire if so on which coil, which side, to the coil with the complete mosfet or the one I removed, or both? Thank you.
 
Back
Top