GPU How standardized are Nvidia's PCB layouts?

Kmpkt

Innovation through Miniaturization
Original poster
KMPKT
Feb 1, 2016
3,382
5,936
Hey Guys,

Wondering if anyone here can comment as to whether Nvidia's AIB cards will follow spec placements for not just the GPU, but memory, MOFSETs, etc? I'm guessing by looking at the sheer number of water blocks that EKWB offers that these components vary wildly in their position on each board, but I am considering making a universal Nvidia contact plate that would at the very least cover the GPU and RAM. Any knowledgable input would be greatly appreciated. Cheers!
 
Last edited:

AleksandarK

/dev/null
May 14, 2017
703
774
Making a uneversial contact plate is next to impossible. You could only make one that would be compatible with Founders edition cards, but as far as AIB versions, they put fan connectors or power connectors near the memory and every manufacturer does it differently. On board layout, some put VRMs first, than DRAM, some go opposite. Some put DRAM 180 degrees arround the GPU core, some put it 270 degrees arround the GPU. It is impossible to predict.
 

Phuncz

Lord of the Boards
SFFn Staff
May 9, 2015
5,858
4,914
I'd look into the concept of modularising the VRM and VRAM part, since these are often the parts that differ a lot. VRAM is not really an issue of placement but of what taller components might be blocking something. In the end copper will conduct well enough and could potentially be made of (a series of) 2D parts, possibly with heat-conducting material in between layers:



Somewhat like the Bitspower monoblocks for ITX boards.