The memory was overclocked to 1500 but not the gpu itself. Considering this is the "OC Edition" of the card the software out of the box also allows that to happen. I was considering that may be part of it though. Perhaps overclocking the memory threw off some of the voltage needs and overloaded the motherboard 24-pin connector?
I dunno. The cards are fine afterwards, the psu's don't pop and continue to power up peripherals. I guess I suspect there's an issue on the board itself when running 2 cards near load for a sustained period.
Topic: GPU Mining: Motherboard & PSU issues question (Read 668 times)
Yes PCIE do pull some power but its only 75 watt, so as long as you have 3 or less graphic card you don't need molex. The weird part is you only have 2 graphic card is this thing happened? Do you overclock?
Just trying to find out how common this is.
System: Asrock 990FX Killer mobo, MSI R9270 (non-x), CorSair Vengeance case with lots of air cooling
So initially, I had the setup running for over a month with no issues. Card was hashing along at ~400kH with a temp between 67-70. I decided to add a 2nd card to the case and double output, or at least increase it without causing the heat to go up too much. Corsair HX850 power supply utilized, (2) R9 270's in crossfire configuration - within a week I started getting random shutoffs so I figured there was a short somewhere (the PSU would power up, I'd hear the HDD's spinning, but nothing powered by the mobo directly would power). Found that pins 10 & 11 had been charred and wires exposed. OK, no biggie, swap in a new power supply and RMA the PSU.
Swapped in a Corsair 750 (can't recall which one but 80+ Gold of course), same thing within a week power loss. But this time, the wires weren't exposed just slightly charred. I ran the cards at less intensity/kH just to keep the temps below 70.
I found an old post on here where somebody said the power draw from the PCI-e slots was too much which caused the 12V rails (pins 10 & 11) to burn up. I'm just not sure how common this is. It's a new board designed for crossfire/SLI configurations for up to 3 cards. The gold rated power supply should be sufficient (right around 65-70% wattage need). But it happened twice. Any ideas? Bad PSU's, bad motherboard, bad setup?
Waiting now to see if Corsair will honor an RMA since the PSU's are still under manufacturer warranty but outside the MicroCenter exchange timeframe. The mobo is covered by an extended Micro Center warranty so at least I can get that replaced easy enough.
System: Asrock 990FX Killer mobo, MSI R9270 (non-x), CorSair Vengeance case with lots of air cooling
So initially, I had the setup running for over a month with no issues. Card was hashing along at ~400kH with a temp between 67-70. I decided to add a 2nd card to the case and double output, or at least increase it without causing the heat to go up too much. Corsair HX850 power supply utilized, (2) R9 270's in crossfire configuration - within a week I started getting random shutoffs so I figured there was a short somewhere (the PSU would power up, I'd hear the HDD's spinning, but nothing powered by the mobo directly would power). Found that pins 10 & 11 had been charred and wires exposed. OK, no biggie, swap in a new power supply and RMA the PSU.
Swapped in a Corsair 750 (can't recall which one but 80+ Gold of course), same thing within a week power loss. But this time, the wires weren't exposed just slightly charred. I ran the cards at less intensity/kH just to keep the temps below 70.
I found an old post on here where somebody said the power draw from the PCI-e slots was too much which caused the 12V rails (pins 10 & 11) to burn up. I'm just not sure how common this is. It's a new board designed for crossfire/SLI configurations for up to 3 cards. The gold rated power supply should be sufficient (right around 65-70% wattage need). But it happened twice. Any ideas? Bad PSU's, bad motherboard, bad setup?
Waiting now to see if Corsair will honor an RMA since the PSU's are still under manufacturer warranty but outside the MicroCenter exchange timeframe. The mobo is covered by an extended Micro Center warranty so at least I can get that replaced easy enough.
Jump to: