DFI-SLi killing eVGA 6800U's?
Hello Everyone, having experienced sudden death on my secondary eVGA 6800U on the DFi-SLI-DR board, and knowing of at least 3 other local benchers with the same issues, I posted a thread over at dfi-street.com, but as yet none of the DFI people have answered - see here: http://www.dfi-street.com/forum/showthread.php?t=18416
Check it out, please, and let me know if you guys have experienced something similar. At least on my system, and another one, they were running default speeds when this happened, both FX55 rigs, with eVGA 6800U's in SLi config, 6693 drivers.
Your feedback would be appreciated!
Here is an excerp from the post:
Quote:
Hello DFI,
I am seeing a disturbing trend with the Sudden Death Syndrome of the "Slave" card, specifically 6800Ultra eVGA in SLI config, on the DFI-SLi mobo. 3 Cases involved 6800U eVGA's, one case XFX 7800GTX.
Under normal running conditions, no overclocking, and proper watercooling, myself, and 3 other Locals (South Africa) have lost the secondary card on our very similar DFI-SLi/eVGA 6800U rigs.
In nearly all the cases, a small hexfet on the hsf-side of the GFX card popped, close to the edge of the pcb in the back center. Under normal running conditions, these hexfets stay very cool, and myself being an Electronics Engineer, is quite mistified by this behaviour.
My setup is:
DFI SLI mobo
FX55 Clawhammer
2x eVGA 6800U cards
2x Mushkin 256 Black Level2 BH5 modules
Enermax 600W 24Pin psu
2x WD Raptor Harddrives
HP Lightscribe DVD
Cooling is via Maze4 GPU blocks (2x), Frozen CPU copper mem sinks, Modified Antarctica (to 1/2") cpu block, Coolwave 3x 120mm radiator, and Eheim 1260/DD5 Laing. No heat issues.
One thing that I did, was to measure all PSU voltage rails with an Oscilloscope, and they are consistently flat, ripple well within specs. As it should be on a non-overclocked stock running setup.
Under similar circumstances, default clocks, we lost in total 4 6800U cards in the 2nd PCI-eX xlot.
Is their a possible issue we should be aware off, can you have a chat to Oscar concerning this? My thinking is thermal runaway in a vreg circuit, but yeah, it is pure conjecture, but it is suspect that we should loose secondary GFX card, under the same conditions?
All of us used different memory brands, different PSU's, Enermax600W, OCZ 600W, Thermaltake680W, Mushkin BH5, Mushkin TCCD, Kingston HyperX, and decent cooling.
Would like some feedback, please?
Kind regards,
MrBean.