Need help to repair a RX580 Sapphire Pulse

[mengpo]

Hi everyone,

I've got a rx580 Sapphire Pulse stopped working while I was playing a game.
Since then, when I turn on my computer, the fans are spinning but the card is not recognized (no display neither). The GPU core is heating when turned on. I'm a noob in electronic and I followed some youtube stuff but no luck for now.

I wrote down what I found (see the pictures) and I got a strange behavior on the APW8722A, on the top left. The VCC pin (the 5th) has 0V and 0 Ohm when the card is powered up.
According to the APW8722A datasheet, from a second APW8722A found at the rear on the board and from a japanese guy (from this site : https://nekosoftpc.blogspot.com/2019...-rx470-8g.html and see the picture too), I should have found 12V.

I thought that the IC was faulty so I removed it. But when I re-measured the voltage and the resistance (without the chip) and I still got : 0V and 0 Ohm. I ordered another APW8722A but now I'm pretty sure that it's a short before this chip...

Since I cannot find any RX470/RX480/RX570/RX580 Sapphire schematics, I don't know what to do next...

Can someone help me further with my diagnosis?

Thank you very much (in advance) !

[fab13]

Re: Need help to repair a RX580 Sapphire Pulse

Hello

Effectively it would come from higher up before this regulator but without a diagram it is always more complex =S

However the 5V arrives well at the transistor level which is supplied in 12 V.

I found these schematics maybe that can help

[mengpo]

Re: Need help to repair a RX580 Sapphire Pulse

Hi there,

Thanks very much for the schematics. I already have them all and the MSIs are the best lead here. Nevertheless, it seems that sapphire adds some voltage regulation circuit that MSI has not...

As you said my problem is more complex without these diagrams... The strangest thing here (and you pointed it with the 5V) is that the 12V PEX doesn't seem to be impacted by the short...

I'll try to look further into the MSI schematics... Thank you again !

[foshland]

Re: Need help to repair a RX580 Sapphire Pulse

Hi, I have a Sapphire pulse RX570 that looks identical, but doesnt work either, maybe I can help meassuring any point of your interest in resistance mode, just PM me and we can work together to fix them. (We will post the info we get, of course, here). I stopped working on it because I lost a component (Q210) during the repair, if you could take a detailed photo of that zone so I can identify it, ill be very happy and could continue working on it. I post 2 photo to show the missing component, thanks in advance.

Now you know, write here, PM me or whatever you want, lets fix them!!!!

[erminba]

Re: Need help to repair a RX580 Sapphire Pulse

Removing APW driver and powering up the card is very risky. Those mosfets gates are floating and can open any time to fry memory controller.

[mengpo]

Re: Need help to repair a RX580 Sapphire Pulse

Hi, I have a Sapphire pulse RX570 that looks identical, but doesnt work either, maybe I can help meassuring any point of your interest in resistance mode, just PM me and we can work together to fix them. (We will post the info we get, of course, here). I stopped working on it because I lost a component (Q210) during the repair, if you could take a detailed photo of that zone so I can identify it, ill be very happy and could continue working on it. I post 2 photo to show the missing component, thanks in advance.

Now you know, write here, PM me or whatever you want, lets fix them!!!!

Hi there,

Thanks for your message. Well, I'd like to tell you what measuring points I need but I don't
Simply because I don't what to do with these measures
That's why I asked some guidance here. But I appreciate your help
For your missing part, here is a picture of the zone (sorry the picture was taken with my phone)

Just fyi, the U300 is a GS7103 and the Q210's name is : S1A. After searching it, I found the ref in a russian forum (I can send you the link if you want) and it seems that the S1A is interchangeable with the K6N. The model they bought (and worked at the end) is this one : aliexpress

Here is the Datasheet

The package used on our cards is : SOT-363

Removing APW driver and powering up the card is very risky. Those mosfets gates are floating and can open any time to fry memory controller.

I should have know better
I thought that the mosfets would have stayed closed if they did not receive any voltage from the APW... Thank you very much for the warning... I hope I did not fry anything with my tests...

[foshland]

Re: Need help to repair a RX580 Sapphire Pulse

Thanks a lot MENGPO, very apreciated, if you can please send me the link to the russian page, my google translator is very good at russian,

Thanks again, I hope to continue repairing that card and if I do, maybe I can help more with yours.

It was the first GPU that I tried to repair and since I lost the component, I let it there just getting dust, so maybe now I have learned some more, I can try to fix it.

[fab13]

Re: Need help to repair a RX580 Sapphire Pulse

Indeed, if this little mosfet is missing, we'll have to see what happens afterwards and I hope that the power has not been turned on despite everything?

So the part in question is on order?
And to see what it gives once in place at APW8722A but you indicate 0 Ohm on pin 5 of this chip, possible that there is a short circuit on the VCC line with possible resistance HS.
Because on the datasheet we see clearly that pin 5 is connected to a resistor.
But the value I don't know.

See here to find in beep mode the resistor which must be close.

After maybe I'm wrong



Translated with www.DeepL.com/Translator (free version)

[mengpo]

Re: Need help to repair a RX580 Sapphire Pulse

Hi fab13,

I finally found out what was wrong in the card. I'm joining two schematics extracts you gave me (the MSI one).
The pin 5 of the U2 chip was at 0V. As I previously said, I though (wrongly) that the responsible of this was the U2 itself. After remove it, the pin 5 was still directly to the ground when I powered up the card (short notice : as Erminba said, don't do this it could fry memory controller). So I bought another APW8722A, put it in place and I tested the other components from the +VDDCI_VCC line.
The PC610 condensator was the faulty one (smallest component I've seen in my life). I replaced it by another condensator (same characteristics : 100nF / 16V) and I got 6V at the pin 5. Again it was problematic since I was expected 12V. I continued my investigation and I also found out that the PR637 seemed damaged. I removed it and tested it and I measured 100 ohms instead of 2.2 ohms.
Finally I changed everything once again with "stronger" components : PR637 (2.2ohms 1% 125mW), PC610 (100nF 25V) and also the C610 (100nf 25V)

And now, my card is working again

I've got some temperature issue but I think the heater does not touch perfectly the GPU core... But it's for another day

Thanks again fab13 for the schematics and your help !

[fab13]

Re: Need help to repair a RX580 Sapphire Pulse

Wow that's great really that's what it seemed to me a resistance Hs in the lot or low value as you indicated which is not normal My diagnosis seemed good I imagine I'm happy for you that it works again as on the first day and if we can help each other between technicians. And then with diagram it is always easier. Small question how was resistance physically? If we can help with a next card that would be cool. For the GPU heating problem at this level I don't really know

[mengpo]

Re: Need help to repair a RX580 Sapphire Pulse

Wow that's great really that's what it seemed to me a resistance Hs in the lot or low value as you indicated which is not normal My diagnosis seemed good I imagine I'm happy for you that it works again as on the first day and if we can help each other between technicians. And then with diagram it is always easier. Small question how was resistance physically? If we can help with a next card that would be cool. For the GPU heating problem at this level I don't really know

Hi fab13,

Thanks again for your help ! The resistance seemed a bit burnt on the side but nothing bad (no strong decoloration or no deformation here). ... That's why I did not decide to change it at first, but I guess I was wrong...

Since the value of this resistance is very low, maybe any damage (even the smallest) can change it... I don't know, my knowledge here is very poor...

[sakai]

Re: Need help to repair a RX580 Sapphire Pulse

Hi, I have a similar card but no voltage on core. I hope someone here can guide me on what should I check.

[Bozog]

Re: Need help to repair a RX580 Sapphire Pulse

Hi, I have a similar card but no voltage on core. I hope someone here can guide me on what should I check.

Well according to the schematic IR3567B VCC comes direct from +3.3V_BUS (PCIe edge connector), so I'd check that for shorts or find out why it isn't 3V3

You should probably start a new thread rather than hijack someone elses...