[BOUNTY] Sapphire Dual -x 7950 Problem - 3 Cards Dead - Help needed

[Prelektro]

FYI : I'm not talking about cgminer dead, cards literally went dead  

1.5 week ago I bought 4 sapphire 7950 Dual-x . They are installed on two different mainboards (asus z77a and amd m5a97). Two of them are powered with 750 watts cooler master bronze and other two are powered with 1050 watts cooler master silent pro gold (I was planning to add more cards).

First, one card which is installed at second pci-e slot (m5a97) went dead, before going dead, system restarted itself several times (I thought it was about electricity in my area) After that card went dead, I couldn't power the system without taking card out. Than, card is replaced by manufacturer. After 2-3 days, card which is installed at first pci-e slot went dead. This time I was able to power the system, but there were no display and fans were spinning at %100. I used 1100/1250 clocks. Temperatures never exceeded 80 celcius. (They were between 77-79 for first card and under 70 for second card)

Same thing happened to one of the cards installed on z77a - no display %100 fan.

Now I'm using two brand new sapphire 7950 on msi mpower board, I was afraid to overclock so I just oc'ed engine to 1000. Cards worked perfectly for 3 days and now again system started restarting itself and I'm afraid it looks like it will happen again if I continue mining.

Windows message contains these error codes :

 BCCode:   a0
  BCP1:   000000000000000B
  BCP2:   00000000BD606000
  BCP3:   0000000000000003
  BCP4:   00000000034A7000

I'm using win7 ultimate and 13.9 drivers - cgminer 3.7.2. I got other 7950 cards (asus) and they've been working perfectly for months.

If someone can help me to solve this problem, I will pay some bounty and I will be generous about it believe me. Thanks

[Wipeout2097]

You claim the cards didn't go above 80ºC but it is incorrect. What never went above 80ºC is the GPU core.

Load Gpu-z, look at the Sensors tab and see if it shows VRM temperature. If they go above 80ºC, you MUST undervolt your cards with VBE7. If it doesn't show anything, then it's even worse because one never knows. You might as well pray. I've seen 110ºC vrm on a 7870.

You should undervolt anyway, to protect the GPU, the vrm and the PSU, reduce heat and save power

Then you aren't suppose to overclock the core for getting hashrate. The ram first, to 1500, 1625 or 1750 Mhz, then the core only up to a point

The Cooler Master PSUs aren't that good either.

[Prelektro]

You claim the cards didn't go above 80ºC but it is incorrect. What never went above 80ºC is the GPU core.

Load Gpu-z, look at the Sensors tab and see if it shows VRM temperature. If they go above 80ºC, you MUST undervolt your cards with VBE7. If it doesn't show anything, then it's even worse because one never knows. You might as well pray. I've seen 110ºC vrm on a 7870.

You should undervolt anyway, to protect the GPU, the vrm and the PSU, reduce heat and save power

Then you aren't suppose to overclock the core for getting hashrate. The ram first, to 1500, 1625 or 1750 Mhz, then the core only up to a point

The Cooler Master PSUs aren't that good either.

Yes I thought about it too but problem is there are no sensors on vrm so gpu-z won't show the temperatures.

I discovered something strange today, cards came with bios 015.040.000.001.000000. When I search it on net, it shows its R9 280x bios ? Couldn't see that bios in techpowerup's bios list for 7950 sapphire. Any ideas?