I have a GA-965-S3 board, used to run 3 cards with non-powered extenders with some overclocking on bamt but without overvolting. Cards were 5830 and 2x 5870. Later I changed 5830 with a 5850, 2 cards were on PCIe x1-x1 12cm extender, one was in PCIe x16 slot. Few days ago I added added a 5830 so there were 4 GPUs in total, one for each slot (3x PCIe x1, 1x PCIe x16).
PSU is a Tagan mid-range 1kW unit.
Yesterday or so, bamt started to lock up as soon as mining started, I checked the ram and cpu on another mobo and they were fine, also, they aren't overheating. Without a card in that slot the rig mines without any problems, I tried a few GPUs in it and I can hear BIOS beep indicating the mobo starts but no display output whatsoever ('no signal detected'). If I remember correctly the mobo would even load an OS with that card and I could connect via SSH but the rig crashed or didn't detect any card in that slot. Now nothing will happen, even the BIOS won't POST, I can use use other slots though and connect a monitor to any of the cards, all three PCIe x1 are fine for both display output and mining. I looked carefully at the board, the are no burn marks or leaking (or 'fat') capacitors, tomorrow I'll give a look at the back of the board. I think it is possible to (de)solder those slots but I don't know how difficult it is to desolder the old one and solder in a new one (2*82 pin with an ejector:
Or how much time does it take, is it worth paying someone more experienced (I'm clumsy at soldering).
I'm not entirely sure the 4th GPU should do that (yeah, no powered extenders but 4 isn't much), on the other hand, there was only a 4 pin additional power connector on that mobo.TL;DR
Any advice on replacing a possibly dead PCIe x16 164pin slot? Or maybe there is something else I could look at, I am not 100% sure the problem doesn't lie somewhere else within the motherboard, maybe it's the mobo's VRM? Anything to measure? Advice is welcome here.