I have built myself a cheap 7 GPU rig to screw around with, it is completely within TDP and running stable at the moment albeit overclocked and belching heat out all day, it has been a wild ride but have found a lot of valuable information.
I have 4 old Antminers in my laundry room, and have been in the process of tinkering with them and have noticed something that is driving me nuts:
Each hashboard has three 75W (max) rated 6 pin pci connectors…
Same with the GS mini doge, one single 75W (max) rated power connector.
One congruency I have seen is that this seems to be the standard:
One 75W (max) 6pin pci power connector from the psu to the unit being requested to drive upwards of 150W each.
The miners then go on to troubleshoot every possible problem involved with power and heat, nobody is addressing that these things come from the factory as FIRE HAZARDS.
So does anyone have any valid opinion on this? Or am I just getting too high off hot GPU plastic and missed some sort of massive engineering shift to 150W(LOL) rated 6 pin connectors?
Im failing to understand why anyone would have problems troubleshooting a rig thats burning up or not running for long periods of time when the ONE connector on the back is fried. Thats the problem, its right in front of you; the PSU connectors arent rated for the power they have been DESIGNED to run.
To put things into perspective:
GTX 1060s have one connector on the back of the cards, this is because the TDP is 120W and the PCI-E slot provides 75W. So the rest of the juice is supplied through the 6 pin on the card itself.
Asus ROG strix 1060 has an 8-pin, apparently rated to 150W because they doubled the grounding capability of the 6 pin.
This is a disgusting breach of safety in my eyes by almost every ASIC manufacturer I’ve seen now, but i could be dead wrong and would like someone else’s opinion.