ASUS RTX2070 Short Circuit Repair

ASUS RTX2070 shorted out, first fixed all the chips under the shorted 5V, and tested normal after replacing the missing parts one by one.

ASUS RTX2070 Short Circuit Repair

ASUS RTX2070 shorted out, first fixed all the chips under the shorted 5V, and tested normal after replacing the missing parts one by one.

Machine model: ASUS RTX2070

Fault phenomenon: short circuit

Repair process: An RTX2070, second repair card, arrived and was dumbfounded. Table test + burn-in revealed that almost all the chips were bad except for the core and video memory.
The initial failure of the card is a 5V FB alignment with a poor over-hole, resulting in the 5V output voltage cannot be fed back to the FB pin of MP1475, causing the output voltage to be too high, burning a total of UP9512R 45491*2 UP1666Q GS9216*2 GS7155, AMX114330 and 1666Q damaged when burned all the way to the memory power supply MOS tube (memory (Fortunately, the resistance value is normal).

ASUS RTX2070 Short Circuit Repair

ASUS RTX2070 short circuit repair Figure 1.

ASUS RTX2070 Short Circuit Repair

ASUS RTX2070 short circuit repair Figure 2.

First removed all the chips under the shorted 5V, then repaired 5V, and then replaced all the ICs under 5V, because there is no GS7155 on hand, so the chip by 5V step-down to 3.3V SEQ voltage, instead of resistive solder bits from the PCB with 0 ohms resistor directly connected to the gold finger 3.3V. after replacing the chip power found no EN, and then found that the card PCB left NVVDD EN-related signals Normal, but the output goes inside the board layer of the line to the right side is disconnected, because the card gold finger burned at the digging, may give digging broken line.

ASUS RTX2070 Short Circuit Repair

ASUS RTX2070 short circuit repair Figure 3.

These alignments run from the left side of the gold finger all the way to the right side, so whether burning the gold finger 12V or a cracked hook may cause some signal loss.

There is no way out, if you can fly the wires, if you can’t, you create a fake signal, and eventually the card lights up and tests OK, but it is not stable. Regular black screen fan spinning wildly. Once inadvertently touched the card’s mining board and found that it was caused by poor contact with the PCIE slot. I suspected that the 3V3 SEQ direct connection to Goldfinger 3V3 was not reliable because the 45491 VCC takes power from the 3V3 SEQ, and the 45491 is very sensitive to power supply voltage fluctuations. So it was better to order the GS7155 from Taobao, Thule the original card design, replace the parts and test it properly. The card was then truly repaired!