Sapphire RX 480 - 8go no power - 0 volts à +VDDC - Help to find why

[fab13]

Sapphire RX 480 - 8GB no power - 0 volts to +VDC

Hello to all =D
I would like to start this card again for my little brother.

Symptoms :
No power on , no rgb led , black screen , no rotating fan.

What I found before disassembly a 22uF capacitor on the +VDDCI line was loose. Before powering up and checking if the component was missing, the capacitor was replaced.

Then powering up
+12V ok
+5 Volts ok
+1.5 V ok
+ 0.90 ok

What I found abnormal :
-No 0 Volt on the +VDDC line of the CPU, value found 0 Ohm, diode mode 0mv and finally 0 Volts on the whole phase row (Which can be seen in grey metal capsules and I circled in yellow.
The +VDDC seems to be controlled by the power regulator if I am not mistaken.


What I did:
1) Removed the synchronous buck gate (IR3553 chip) driver phase 6 and phase 1 and still 0ohm and diode mode 0.01 mV on the phases
2) Removed then one by one the 4C10N mosfets but before removing everything I found 12 Volts drain on each but still 0 Ohm on the phases.

Can you look at the diagram I am attaching on page 15 of 26 all the phase lines and the +vddc line and I am attaching the other photos for your convenience.
and the Buck controller datasheet

To tell the truth, I may be going the wrong way, no?

Thank you very much in advance team =D =D

[DrasticRepairs]

It's normal for vcore vddc to be 0 ohms. The only way to test that is with voltage injection. If it draws more than 5A at 0.8V, the core is probably shorted.
Don't worry about that, though.
If a MOSFET is bad, it shows up as a short on the 12V rails.
You'll want to check enable on the vddc controller. If you don't have EN, then trace that back to it's source. You can also check resistance and voltage on perst_buf. It should have 3.3V and nearly oL resistance.

[dicky96]

The resistance of Vcore should not be 0 ohms! Don't try to read it in diode mode you need to read it in resistance mode.

Some GPU have a very low Vcore resistance around 0.6 ohms or less - I measured an RTX 3080 at 0.2 ohms and that powered up though it crashes after drivers load. That is the lowest one I ever found, but even so it should not be 0.00 ohms - that is a shorted GPU or some other component (MLCC capacitor) that is short

[DynaxSC]

Hi, there are special so called milliohm meters, which can measure such resistances very accurately (they use a bridge architecture). I use YT2050, you can buy it on a famous chinese portal. It's cheap and quite accurate. And these resistances are even lower, for GP104-200-A1 (GTX1070) I measure it as ca 0,14 ohms. Everything under 0,1 ohms is already a suspect to be damaged, but 0,14 ohms is already OK.

[fab13]

I would like to thank you for your help really for the moment it is complicated to find the problem

So to answer you DrasticRepairs
I injected 0.8 V at +VDDC and consumed 2.40 Ampere and as you said if the core is shorted it should overconsume then about 5 A and nothing heats up on the thermal camera.
In addition I measured in Ohm to rectify what I said above, I am at 00.4Ohm (Same value when I touch the red and ground on my voltmeter) =S


For the mosfet IR3550 on VIN on phase 1 to 5 = 0.670Kohm and VIN phase 6 (Located the lowest) measured at 6.01 Komh it seems to me quite high compared to others.perhaps following my withdrawal.
A VCC = measured all phases at 14Kohm

What does it mean when a value is high in kOhm on a line for example?


I can't find what you indicate =D (prest_buf) , do you mean I have to check on the buck controller IR3567B on ENABLE line 21?.
What I get is 10.36 Kohm and seems to be VDDC_VDDCI_OE_VR which is the 0.8V_PGOOD
The 0.8V_PGOOD goes to the GS9238-ATQ chip what should I measure on it?
Am I doing it wrong?
IN VCC = 0.410kOhm measured (I think I saw 3.3 Volts before removing the mosfets)

The mosfets 4C01N (which are circled in white and removed) the Q4609 on drain = 6Kohm
Q4610 AND Q4611 = 17 Ohm

SEE THE PICTURES TO UNDERSTAND BETTER

Sorry if I don't know the terms, I'm a beginner since 1 month =D

Hello dicky96 =D
Thanks for the little tip really but what does MLCC mean? I can see it on the +PW_MVDD_PHASE line going to GS7256 (a single buck controller)
On phase 8 (+PW_MVDD_PHASE) = 20 Ohm

What else should I check because I don't know what to do ^^
and then I'll reassemble everything I have desoldered.

[dicky96]

Sorry for using terminology - it means Multi Layer Ceramic Capacitor

I explain them, and how/why they fail, in this beginners video

https://www.youtube.com/watch?v=R9slJUqT6Ls

Also check this out if you didn't already see it - it has 30K views, so I must have done something right. It might help you understand things.

https://www.youtube.com/watch?v=SmVdl8kew_4

[fab13]

Hello Dicky96

You did a great job and I understand better the way a graphic card works and I'm less blocked on some points
However I will have to look again and take notes on the points I don't understand.
I'm looking forward to follow other videos on the repair of graphics cards or motherboards.

For information no short circuit on the VIN of IR3553 but on SW which has two mosfets internally I still fall to 0.06 Ohm despite removal of all controllers and obviously to the coils

At the end of the video you insist on the subject of PCI-Express not short circuit.

For the moment I do not know yet where is the problem =S