Overclock.net banner
1 - 20 of 36 Posts

·
Registered
Joined
·
37 Posts
Discussion Starter · #1 · (Edited)
Hi there,

I have two GPUs that require knowing the the Ohms of a resistor and then farads of a capacitor.

MSI 980Ti Gaming:
C622 = blown smd capacitor (I think I can match the size from a donor board...but it would be nice to know what microfarads this is)
C619 = blown smd resistor (no idea what value this is supposed to be...just simply shorted open)


Gigabyte 980 Ti Windforce:
C187 = lost smd capacitor (it's tiny...real tiny. Need to know microfarads)
R122 = lost smd resistor (it measured 6.3 KOhms before I lost it when I pumped my parts tray. I think that is not a good number and it's either lower or higher. Many pull-up resistors are closer to 10Kohms but not sure what this one is.)

If any of you have these boards and wouldn't mind measuring these components values that would be awesome. Maybe I can pay it forward one day.

Fun facts for you guys who like to repair GPUs...
The MSI 980 Ti has no power to the Pex rail but memory and other voltages are there. VRAM and Core have the right voltages. No shorts, correct ohms...just nothing at the pex choke. Should read 1.1v but doesn't. Flip the card around, follow the traces, and 5V stops at blown reistor and capacitor from what looks like a small leak from a liquid cooler. The card powers up but no diplay signal, which is exactly what you'd expect in when missing the 1.1V Pex. Should be an easy fix...if I knew the values. Wish I had the Tech Cemetary guy on a hotsline..."Hey man...any gueses on these values...???"

The Gigabyte 980 Ti Windforce booted but had terrible artifacting and and failed under load. Clearly a memory problem. I tested the memory with MATS and lots of errors. So many I decided to change the entire memory power phase...and all the VRAM. The card was in otherwise flawless condition so having bad VRAM meant thermal stress or memory power phase IC (which you can see in the photo) was out of spec. This could cause overvoltage or even the mosfets could have been failing, causing voltage spikes. This is the card I will try to turn into a Titan but that requires getting the dang memory power phase working again. I think have to install all new memory. Doable...just more time than I have.


Thanks!
 

Attachments

·
Registered
Joined
·
2,616 Posts
The 2nd chip looks like this
I'm not familiar enough with buck converters to guess at the R value, there's no high voltage, so it's probably not huge, it's just for some RC filter. I don't see it used in the datasheet.

What's the number on the 1st chip, and where is that on the PCB ? I only found the 2nd chip by guessing it was a buck controller.
 

·
Registered
Joined
·
1,101 Posts
If any of you have these boards and wouldn't mind measuring these components values that would be awesome. Maybe I can pay it forward one day.
You are out of luck:
a) Only the 0.001% of logged in people, including my self, we own a worth getting LCR.
b) Only the 0.00001% of logged in people, will measure and keep a log of components values, when their VGA this is still working.

Have a search by the use of several search engines, prefer results of pictures, use search keys as take-apart.
Good luck.
 

·
Registered
Joined
·
37 Posts
Discussion Starter · #4 ·
I really appeciate you both responding. I think I may have lucked out, at least on one of them. On the MSI 980TI I looked at another photo zoomed in from a different angle and both components are capacitors. Those I can match based on size on color (for the most part). I think I want to replace the chip too (as it looks corroded on one side) but I cant figure out what a it is. It's marked U502 with a 3A 2H label on top. I may have to watch a couple Tech Cemetary videos where he covers these Pex circuits. It's probably a buck controller as it takes down 5 or 3.3V to 1.1V.
 

·
Registered
Joined
·
37 Posts
Discussion Starter · #5 ·
The 2nd chip looks like this
I'm not familiar enough with buck converters to guess at the R value, there's no high voltage, so it's probably not huge, it's just for some RC filter. I don't see it used in the datasheet.

What's the number on the 1st chip, and where is that on the PCB ? I only found the 2nd chip by guessing it was a buck controller.

The fist chip is labeled 3A 2H and it may also be a buck converter for making the 1.1V Pex voltage. I really haven't been able to figure it out. I'm going to change out the two shorted capacitors and see if it works. Not too confident as it also looks like it got wet....but crossing my fingers.
 

·
Registered
Joined
·
1,101 Posts
I'm going to change out the two shorted capacitors and see if it works.
Caps were shorted due over-voltage, power circuitry this is blown too.
According my crystal ball, occurrence of over-voltage it takes with it the memory chips or the entire GPU.
This is why 95% of VGA cards its not worth repairing.
 

·
Registered
Joined
·
37 Posts
Discussion Starter · #7 ·
I’ve been lucky at 50% repairable but many fried gpu cores and memory listed in the “for parts only” section of eBay. If GPUs weren’t so overpriced I’d never bother to repair. Until recently, I scoffed at it. Best case scenario is bad VRAM or shorted low side mosfet that shorts to ground.
 

·
Registered
Joined
·
37 Posts
Discussion Starter · #8 ·
I’ve been lucky at 50% repairable but many fried gpu cores and memory listed in the “for parts only” section of eBay. If GPUs weren’t so overpriced I’d never bother to repair. Until recently, I scoffed at it. Best case scenario is bad VRAM or shorted low side mosfet that shorts to ground.
Forgot to ask, any idea what circuit that is for pex? Like you said, it’s probably dead also. Seems like it’s a buck circuit of some sort. Part number on the chip is not easy to identify as an actual part number
 

·
Registered
Joined
·
2,830 Posts
Forgot to ask, any idea what circuit that is for pex? Like you said, it’s probably dead also. Seems like it’s a buck circuit of some sort. Part number on the chip is not easy to identify as an actual part number
Business is slow for him he's gonna want payment
 

·
Registered
Joined
·
37 Posts
Discussion Starter · #13 ·
That's a good idea. I'll also try Buildzoid on discord and the vmod section. Anecdotally....I'm not a young guy and have a lot of life experience as they say. What I've learned is that there are engineers and there are fixers. Sometimes a person can be both but more often the former struggles to be good at the latter because of overthinking. I'm not blaming anyone here of this by the way. Anyway, I've found that the good fixers are good diagnosers with a healthy dose of optimism. The best fixers function more like artists, equal parts sense and science. Consistant GPU repair requires being at the very least, a good fixer. To get to the next level requires help from others on these forums. I think as more of this dialogue happens, the more cards we can keep out of the dumpster.
 

·
do not touch my butt
Joined
·
263 Posts
Unless you can read whats on it or have a capacitor tester its futile. The manufacturers wont get off the information.
if it is a reference design you can ring it out and likely match it to the schematic. for the caps it's pretty easy, you can PROBABLY just use 0.01uF 25v X5R or X7R rated. the size is probably 603/402. the board could possibly work without the capacitor at all. The big problem he pointed out is the failure mode... a direct short.

the large cap C622 is probably 805 size 1uF or 0.1uF 25v X5R/X7R

The resistor is connected to what is quite obviously NCP81172 buck regulator. if you look at the simple datasheet you can clearly see it is connected between VREF and REFIN. he may have to hunt around for the resistor connected to TSNS, but it is quick to eliminate the resistor below the highlighted one if one pin touches ground...

most likely there is another 81172 on the board and it will have the exact same layout as these are usually copy/paste and autorouted by lazy EEs
easy solution is to take off the matching resistor and measure it with RDC meter everyone should have laying around
 

·
Registered
Joined
·
37 Posts
Discussion Starter · #16 ·
It turns out I was wrong about those caps on the MSI 980 ti being shorted open open. The buck regulator seems to be what's dead. I actually found the same exact circuit on my Gigabyte 980 TI but I'll hunt for another on the same card as totally agree with copy and paste EE. The resistor on the Gig 980 ti is a zero ohm resistor btw. Regarding the buck converter, it's not a 81172 as there are only 10 pins, e.g. 5 one top and bottom. I've seen the buck converter you decribe used on a lot of VRAM power phases but this is for PEX. It's different...and I can't figure out. I do apprecate the help though.
 

·
Registered
Joined
·
1,101 Posts
The best fixers function more like artists, equal parts sense and science. Consistant GPU repair requires being at the very least, a good fixer.
What are you aware of good fixers behaviour on-line? The ones invested over 50000 euro dollars in tools?
The ones invested over ten years of their life so to graduate and earn working experience, prior starting a business?

My sensors give me now alerts, that I am having an conversation with a pensioner, whom disrespect the active hard working people.
By the same way that I do ignore the wannabe PC doctor (a farmer), I think that its time so I to act alike and for you too.

VGA cards repair business, this is not a business, no one true professional will engage repairing something which comes with out availability of key spare parts.
NVIDIA its smart, does not hand over GPU to any one else that their own approved partners AIB.

Industry does not allow trespassers to enter at a single factory, or them getting access to by law protected intellectual property, as is VGA cards circuits.
Keep playing with electronics, its also a useful sector for obtaining experiences, but stop using cheap excuses when you are a trespasser by choice.
 

·
Registered
Joined
·
37 Posts
Discussion Starter · #18 ·
What are you aware of good fixers behaviour on-line? The ones invested over 50000 euro dollars in tools?
The ones invested over ten years of their life so to graduate and earn working experience, prior starting a business?

My sensors give me now alerts, that I am having an conversation with a pensioner, whom disrespect the active hard working people.
By the same way that I do ignore the wannabe PC doctor (a farmer), I think that its time so I to act alike and for you too.

VGA cards repair business, this is not a business, no one true professional will engage repairing something which comes with out availability of key spare parts.
NVIDIA its smart, does not hand over GPU to any one else that their own approved partners AIB.

Industry does not allow trespassers to enter at a single factory, or them getting access to by law protected intellectual property, as is VGA cards circuits.
Keep playing with electronics, its also a useful sector for obtaining experiences, but stop using cheap excuses when you are a trespasser by choice.
My comments weren’t directed at you nor meant to offend anyone. I’m a hack if not based only on my limited tools but good enough to fix a gpu here and there. Few professional shops try to fix these as they are not intended be fixed by outsiders or at all really.

I don’t make excuses for my failures in card repair. Sometimes I try something and it doesn’t work. Sometimes it even makes it worse. Sometimes I right on and rescue a card from the dumpster. It’s just a hobby that keeps me from drinking at night. Or sometimes I do both
 

·
Registered
Joined
·
37 Posts
Discussion Starter · #19 ·
My comments weren’t directed at you nor meant to offend anyone. I’m a hack if not based only on my limited tools but good enough to fix a gpu here and there. Few professional shops try to fix these as they are not intended be fixed by outsiders or at all really.

I don’t make excuses for my failures in card repair. Sometimes I try something and it doesn’t work. Sometimes it even makes it worse. Sometimes I right on and rescue a card from the dumpster. It’s just a hobby that keeps me from drinking at night. Or sometimes I do both
By the way, I'm not retired. I'm still very much in my working years and work as consultant for top BioTech companies, leading a team of engineers for advanced control systems. I got in fixing GPUs because I'm into Crypto and through my Crypto network I get occasional access to bundles of broken mining cards. The last batch I had were a bunch of MSI RX 570s Gaming cards that all had bad power phase mosfets. After I fixed about half of those, I got the bug in me. Anyway, what we do is NOT easy and we need to support each other. If you have ideas or knowledge, please share. Listen, I even pick up the phone and call Alex from NorthridgeFix. He's a professional like you and very busy runngn his business and making YouTube videos. Again, we need to support each other.
 

·
Old and Decrepit
Joined
·
2,505 Posts
What are you aware of good fixers behaviour on-line? The ones invested over 50000 euro dollars in tools?
The ones invested over ten years of their life so to graduate and earn working experience, prior starting a business?

My sensors give me now alerts, that I am having an conversation with a pensioner, whom disrespect the active hard working people.
By the same way that I do ignore the wannabe PC doctor (a farmer), I think that its time so I to act alike and for you too.

VGA cards repair business, this is not a business, no one true professional will engage repairing something which comes with out availability of key spare parts.
NVIDIA its smart, does not hand over GPU to any one else that their own approved partners AIB.

Industry does not allow trespassers to enter at a single factory, or them getting access to by law protected intellectual property, as is VGA cards circuits.
Keep playing with electronics, its also a useful sector for obtaining experiences, but stop using cheap excuses when you are a trespasser by choice.
While your points on protection of IP are solid the remainder of your flaming post is harsh. If someone wishes to attempt a repair on anything there are no ethics violations or laws prohibiting such. Is is more difficult without schematics and parts nomenclature, of course it is. I’ve walked that road many times both professionally and privately. Every manufacturer plays their cards close to their chest. One of my professional roles within the company I work for is training. It often involves training 3rd party companies employees, some of which are “trusted” partners. The training and materials provided differ greatly between actual factory representatives and all others. The 3rd parties do not have access to proprietary software, schematics, field bulletins, procedures etc. Those courses are limited to operation and limited maintenance requirements. That being said there is nothing that prevents anyone who is educated enough to be able to trace and figure out circuit layouts from doing so and making repairs. Power components are generally simple and common across many industries. Where the level of difficulty comes in is the control circuits that are all but impossible to reverse engineer regardless of the level of education or how many thousands of $$ of equipment you have.
 
1 - 20 of 36 Posts
Top