If you have stripped all the vram and the impedance is 0 ohm, it's best to throw it away because your gpu has a problem. Normally, if the gpu is healthy, remove the ram and it should be above 100 ohm. After installing the entire vram, the internal resistance drops to 40-48 ohm, which proves that the communication part to the ram is quite strong. If it's from 20 to 35 ohm, it's still acceptable. Usually, an aging gpu will have such internal resistance. For DDR5 only, DDR6 is not certain because it hasn't happened much.

Hi buddy! According to the picture you took and the area you circled, that is the power supply coil for vram. Did you measure it to see if the internal resistance is below 30 ohms? If it is shorted to 0 ohms, the entire ram set is likely shortened. You did not specify whether the ram you replaced was new or taken from another card! If you want to be clear, you should remove all the vram and measure again.

Usually these R9s are too old and the pcb is also degraded to 100 cards, 65 cards open gpu 20 cards vram errors the rest are faults and pcb burns, if u want learning repair graphic card you should buy dead card model begin 750Ti to 1050

The result you have after measure look like nice.
First, you need to recheck the 1.8V rail. After that, check the 0.9V rail from the GS9238 or APW8713A. If the voltage is sufficient and you detect a square wave before the inductor, then it's temporarily okay. The main IC is usually the least likely to fail, especially on this 500 series.
Next, measure the output pulse for each phase from the uP9505P or IR3567B to the phase controller IC. Do you see a square wave? Then, measure the gate pin of the high-side MOSFET; is there a square wave coming from the phase management IC? Check for...

Thanks! but it's long times ago. Again thanks you very much

this original dual mosfet...

i was repair case like this...

u can replace this by EMF20B02V