Check if the damage is minor and consider repair options. If the issue is severe, it might be best to replace the card.
Check if the damage is minor and consider repair options. If the issue is severe, it might be best to replace the card.
Hello, I'm facing some problems with the graphics card I bought recently. It came with an unknown condition, and the seller had issues getting it to function properly. During repairs, they replaced some components, but it didn't work in my PC. When I sent it back for another check, they confirmed it was operational, so they decided to sell it. After installing it in my computer, I removed all previous drivers and reinstalled the correct ones. Everything ran smoothly until I tried a stress test. Within minutes, my display shut down and there was a burning odor. I quickly powered off the power supply, took out the GPU, and examined it. A small chip on the board was nearly melted. I thought it might be dead, so I put it back in to see if it would power on. Although there was no display output, I heard the Windows boot sound. Because the chip was close to all outputs, I assumed the display issue was likely from the connections. I bought this card with the hope of using it in a server environment, possibly for running a parsec VM. Do you think it's still feasible? Would I be taking a fire risk if I tried to fix it myself? Also, do you know where to find a guide with module names so I could attempt soldering a chip and potentially restore display functionality?
Hello, I'm facing some problems with the graphics card I bought recently. It came with an unknown condition, and the seller had issues getting it to function properly. During repairs, they replaced some components, but it didn't work in my PC. After a return, they confirmed it was working, so they decided to sell it. When I installed it, I removed the old drivers and reinstalled the correct ones. Everything ran smoothly until I tried a stress test. Within minutes, my display shut down and there was a burning odor. I turned off the power supply, removed the GPU, and examined the board. A small chip was almost completely melted. After putting it back in, I noticed it wouldn't power on properly—only the Windows boot sound came through. Since the chip was near all outputs, I thought the display connections might be damaged. I considered using this card for a server with a parsec VM, wondering if it would still work without display outputs and whether it posed a fire risk. Also, I'm curious if anyone knows where to find a guide with module names to help me solder a chip and potentially restore display functionality.
The fried chip AON6414A is essential for power delivery; without it the GPU likely won't function properly or at all, possibly leading to premature failure. It can be replaced relatively simply by someone skilled in surface soldering, though confirming its presence is crucial before removal. If you need repair work, a board specialist may be necessary, but the cost might exceed the value of the GPU itself.
For future guidance, avoid this when encountering a dead VRM part. Repeated power cycles increase the risk of damaging the silicon. This isn’t a high-side MOSFET failure; it could be linked to memory issues or an iGPU output from the board’s audio module. The device might still produce some sound, or it could be using an iGPU that’s feeding audio through the motherboard’s speakers. In any case, it won’t run games properly now. Check TechPowerUp for PCB images to identify the part number—likely a 6414A. Finding GPU schematics is nearly impossible without insider knowledge from engineering staff. You might find luck, but don’t expect success.
You can swap the damaged MOSFET for another unit with the same part number (6414A) next to it. Test to determine if it functions properly.
It would make sense if the chip included other components of the GPU, as the side logo seemed to have built-in RGB. I only realized this after the GPU started flickering. My CPU doesn’t have integrated graphics, which is why the issue became apparent. I took the motherboard out to fit two GPUs and check if Windows still recognized it—unfortunately, it didn’t. It might have been a coincidence. I noticed the 6414A module is present at every HDMI port, with a chip in front, while the display port has a different module. Perhaps it’s just a coincidence. The board appears damaged overall, but closer inspection didn’t reveal any further issues. If I replace that chip, it could work again—provided it didn’t short out nearby components. Since the warranty expired about a month ago, it likely won’t cover this. However, because the GPU was returned multiple times for repair and stored for a while, I’m hoping they’ll consider a replacement or fix.
If the PCB's base material remains intact, it can likely be fixed, though there are situations where the device turns into just components or waste.
This situation seems quite random. The best images come from your GPUs (Gigabyte GTX 1650 Gaming OC 4G). Most of what you see between the three blue capacitors on the left and the seven grey ones is the VRM. It looks like fixing things here might be simpler because the parts are more accessible and don’t require a lot of special tools or expertise. But once something breaks, related issues can pop up easily. You don’t need to see obvious damage; even small shorts or malfunctions can stress other components, causing more problems. You mentioned the GPU has been fixed several times before, so it likely has multiple issues and might need a full inspection. If you manage to get it to a service center or warranty program, do it. Otherwise, it’s probably time to let it go and replace it. The GTX 1650 price makes rebuilding the VRM more costly than buying a new card.
Based on the extent of the damage, fixing it might not be worthwhile. The cost could exceed the card's value.