Zotac 1070 Ti AMP Extreme with no vcore

Thieri

n00b
Joined
May 19, 2021
Messages
2
Hi there!

the card from my father-in-law suddenly went dead while on desktop. I tried a diagnosis, but found a dead end. Maybe you can help:

At first i measured all main parts. So no shorts on:

- VRMs (GPU nor RAM)
- GPU (as far as i can tell with such low resistance)
- Main power rails (12V from both peg and from pcie slot)

Then i tried measuring voltages, but there were no voltages from gpu or ram VRMs.

Next i looked at the pwm controller (uP9511P) and sure enough there is no EN signal to the chip. According to the datasheet it should be >0.4V but it stays 0V from start.
I have to admit that i also don't fully understand the EN signal line. The datasheet shows this sample circuit:


Bildschirmfoto 2021-05-19 um 20.05.40.png


So EN is connected to GND via the resistor and the MOSFET (if voltage is supplied at the gate). But where is any Voltage coming from for the EN signal? As said it's my fault, that i don't understand the basics here.

Where is the EN signal coming from on the pcb specifically? I found the corresponding mosfet which has 5V at the gate when running. But where are the 5V coming from?
Here again the datasheet states, that the gate has to be pulled low to get the EN signal:


Bildschirmfoto 2021-05-19 um 20.10.12.png


But what should pull the gate low?


Thanks in advance! Sorry for the missing knowledge regarding the basics. And let me know if you need more information.
 

RazorWind

2[H]4U
Joined
Feb 11, 2001
Messages
3,969
I think Razorwind ought to start charging people for advice/help.
The thought had occurred to me, but using [H] for the purpose would be super scummy, IMHO.

Related: Should we have a subforum or something for these threads, so we don't clog up the main graphics card forum? When I started posting about working on my own cards, it didn't occur to me that we'd get this many new folks showing up just to ask for help.

I think the short answer to the OP's question is there's probably a minor rail on the board that isn't working, and thus the enable signal isn't being generated, but as we saw in another thread just today, it may be that the core VRM is starting up, failing to hit the targeted voltage, and then shutting down to avoid damage due a short.

Op: What is the exact resistance to ground you measure through the GPU core and memory?
 
Joined
Jan 16, 2013
Messages
2,317
The thought had occurred to me, but using [H] for the purpose would be super scummy, IMHO.

Related: Should we have a subforum or something for these threads, so we don't clog up the main graphics card forum? When I started posting about working on my own cards, it didn't occur to me that we'd get this many new folks showing up just to ask for help.

I think the short answer to the OP's question is there's probably a minor rail on the board that isn't working, and thus the enable signal isn't being generated, but as we saw in another thread just today, it may be that the core VRM is starting up, failing to hit the targeted voltage, and then shutting down to avoid damage due a short.

Op: What is the exact resistance to ground you measure through the GPU core and memory?
Haha it's not scummy. You're providing a service here. We might need a subforum, I have noticed a lot more of these new accounts posting similar threads.
 

Thieri

n00b
Joined
May 19, 2021
Messages
2
Haha it's not scummy. You're providing a service here. We might need a subforum, I have noticed a lot more of these new accounts posting similar threads.

Well with the current prices for graphics hardware how is anyone to blame who is seeking help in repairing their cards? ;-)
Even repair services for these cards are hard to get by in my area due to shortages in part supply chains.

Op: What is the exact resistance to ground you measure through the GPU core and memory?

For GPU core i get something around 0.4 to 0.6 ohms. It's hard to say since my multimeter isn't the best.
For Memory i am a bit stumbled. I thought i had measured something around 100 ohms but i did it again just now. The reading is also below 1 ohm now. Am i measuring wrong? I am testing between the output side of the choke against ground.

If it says under 1 ohm now i think that's no good news. :-/
 

RazorWind

2[H]4U
Joined
Feb 11, 2001
Messages
3,969
Well with the current prices for graphics hardware how is anyone to blame who is seeking help in repairing their cards? ;-)
Even repair services for these cards are hard to get by in my area due to shortages in part supply chains.



For GPU core i get something around 0.4 to 0.6 ohms. It's hard to say since my multimeter isn't the best.
For Memory i am a bit stumbled. I thought i had measured something around 100 ohms but i did it again just now. The reading is also below 1 ohm now. Am i measuring wrong? I am testing between the output side of the choke against ground.

If it says under 1 ohm now i think that's no good news. :-/
it sounds like you're taking the measurement correctly. If you really have 1 ohm to ground on the memory circuit, that would explain your problem.

Unfortunately, my experience with diagnosing a short on the memory rail is not good. It's a pretty common issue, but I've never succeeded in figuring out where the short occurs, which strongly implies that the short is inside one of the large BGA components - either the memory ICs themselves or the memory controller on the GPU.
 
Top