The new pads are less hard than the originals.
Thank you for your thoughts. There is a pad flow in the gap between the chips indeed. But honestly I don't see a problem in that. Here are two pictures from the predecessor card, same modell (please ignore the two small pad pieces missing on the block). Looks like fine contact to me. Please note: this one has shown no VRAM issues after installing the block.
View attachment 2635238
after a couple of days' use:
View attachment 2635239
card with VRAM problems, current assignment, same pads, before assembly:
View attachment 2635240
I don't think the pads are responsible for the current Mem OC problems. I agree that something bent the PCB and made the BGAs lose contact, probably myself while disassembling the air cooler. This will hopefully be fixed by reflowing as reported before.
I mean the spacer rings that come with the block and which Bykski recommends (orange) to place between the cooler and the PCB.
View attachment 2635241
View attachment 2635242
I followed this instruction in one of the several assemblies I did while trying to solve the VRAM issue on the current card. The temperatures with the rings installed were awful, about 90 °C hot spot temperature. They are just fine without them, almost too good to be true actually:
Take a look at a few
driver comparisons I did yesterday if you like. Tell me what you think about the GPU, Mem and Hotspot temps with several 500W+ stock Time Spy runs. HWiNFO columns from left to right: current, minimum, maximum. Translations (bottom of the screenshots): Wasser = water; Luft = case temp; Raum = ambient temp. Water flow was about 135 l/h.