I suspect Q13 could be my issue because source pins are pretty much dead, unlike Q14 under it. Also, next to Q13 there is this cap (C70), which is shorted. And down the board, other Mosfet chips are shorted on drain pins and the second source pin (painted in red) is also shorted. Not to mention some resistors (highlighted in pink) are also shorted.
Could anyone help me out? I’m willing to remove Q13 from the board, but I really need some advice from somebody who’s a pro on gpu card repair.
Hi, and thanks for answering!
Well, as you can see, there is no 12V connector for this card, so it only takes power from the PCI-E (I did measure the first three fingers from the left though and they output a positive resistance number, which climbs up to 5K Ohms). As for 3.3V, it’s good (4.433 Ohms) measuring the 4th pin from the right of the connector (including missing fingers), but that’s not the end of the story. I also measured my inductors (the three biggest ones between mosfets and capacitors) and they’re all low resistance. Each one outputs 1.7 Ohm when checked to ground BUT if I measure for resistance between the 3.3V to each inductor, I get 4K Ohms. So I’m not sure whether it means my mosfets are bad or if I should keep looking for the source of the short.
This is an EVGA GTX 1050, 2GB of GDDR5 RAM.
P/N: 02g-p4-6150-kr
Ok good info. i have a lot more experience with gpus than I do consoles so maybe i can help.
The 3 mosfets and 3 large inductors you mention are for the vcore 12 volt rail to the gpu chip.
The mosfets and smaller inductor at the top should be your memory power vmem.
It would be smart to locate the other voltage rails & regs like 1.8 volt and 5 volt and such but my guess is that just the 12 volts is shorted based on your numbers. 5 volt can be checked on that big 3 leg regulator behind the vcore mosfets near the ports.
Can you post a pic of the back lf the board as well?
Also post the markings or pic for those vcore mosfets and the 8 pin ic behind the top vcore mosfet.
I can tell u a lot more with those. Just don’t plug the card in to the pc again as it could send 12v to the chip if ur power supply has slow short protection.
I have a EVGA 1050ti with the exact same problem as described above. Anyone have any advice on what part I need to replace? My memory mosfets are fine but it seems all 3 vCore mosfets are bad. I’m not entirely sure. Anything helps.
See, this is a ‘regular’ card, meaning it has no GDDRX and the GPU is a standard one, so no FTW, overclocked edition or whatever. That being said, I’m not sure there is a 12V rail on this card (that is also why you do not see an additional power plug at the top of the board).
Good to know. And what voltage do those deal with? Is it 3.3 or 1.8V?
Noob level here, I don’t know what you mean by ‘regs’. Care to elaborate?
I’ve tested this 3-leg voltage regulator, and these are the results:
Hey, been a while haha. Again, good info and gove me a little bit to go throigh all of that and reply.
A few quick clarifications to make things clear:
All GPUs have a 12v rail. Yours just comes from PCIe alone. It can also pull 3.3v and then break the other needed voltages down from there. So 12v will be massive to establish and then see if it’s shorted. The rest are just as important but if 12v is shorted then there are potential issues.
Finally, by GDDRX I simply meant the RAM with the X being 5 or 6. It wound be 5 here but I use it so often that I have a habit of writing GDDRX when talking about the VMEM rail.
But let me take a look and get back to you. I may need more measurements so I’ll let you know. I have not been on this forum in a while but noticed your post in my inbox so thought I’d see if I could still lend a hand. Talk soon.
Quick addition and edit:
Also, while you wait let me answer some of the basic questions and/or explain how a GPU works in order to mayhe help focus your own searches:
A GPU needs certsin voltages to work. These are usually 12v, 5v, 3.3v and 1.8v. These come on in a certsin order and that order can change based on the card. However, they ALL have to be present for the card to work and the tricky thing is that some voltages are derived from others. So your 5v can come from 12v and 1.8v can come from any of them (depending on the card).
The reason for these voltsges is that certain components and parts of the card can only safely work on certain voltages. For instance, your PEX rail tends to be very low. The GPU chip (die) feeds off the 12v rail but the chip iself but anything more than 1.4v will kill it. RAM has it’s own voltage and so on.
So in diagnosing a card you have to first establish:
are there any blatant shorts on any of the main voltage rails. This requires finding where they are at and some are standard across all (12/3.3v on PCIe or Power plug) while others like 5v and 1.8v can be placed anywhere by the manufacturer…however it’s usually the same basic places.
So all shorts must be ruled out on these rails, especially 12v! If 12v is shorted it’s either an easy fix or a dead card, lol.
If no shorts are found, you have to power thr card on and see if the voltages are present where they should be on those rails. If the 5v location is missing 5v we know where in that boot sequence the card is failing. 5v isn’t being made/turned on…so then you figure out why. Etc.
Edit to the edit: It was so long ago that I forgot you made some basic measurements, haha. Talk soon.
Ok, first we’ll do the quick and easy testing. I had to read your OP again to refresh what’s going on and you may be right about the mem fets but let’s ignore that and start from scratch. Quick and easy.
The beauty of these cards is how simple they are compared to larger ones. Pretty bare bones and the PWM chips tend to fail on them a lot due to that. So, do this test and report back every reading in Ohms. No power at all to the card, from the same ground point:
With these two images, get a resistance reading on every components and/or pin I marked in red. Ignore the red you used in the first photo.
It’s basically all the coils and the larger of the voltage regulators (those fat 3 legged components). Some are ground and will read as such but mark those, too.
Then on the front of the card I want to focus on resistance readings at pin 1 on the VRM fets. So on Q11, 9, & 7 take resistance readings on the first pin (the one nearest the dots, next to the readings you took in the first post). Let me know those readings, too.
Finally, when that’s all done let me know if you have a way to power thia card safely. If we have no shorts we have to move on to testing voltages with power to it. If the PC won’t turn on at all while the card is plugged in you have a 12v problem and the card may just be dead. Bad fets at best, probably…but we’ll worry about that then.
Hello. If your Vcore fets are shorted you likely have a 12v issue. Have you checked the actual Vcore coils? If those are shorted to ground your issue is likely one lf those fets. If so, it may have damaged the GPU but you may have gotten lucky. Worth testing.
Let me know the part number of the vcore fets so I can see the pinout. Once you do that we can see if all test the same (some pins are exclusive to each and can give clues to which is bad if isn’t visually obvious).
