freshjuice wrote:
I started suddenly having serious problems with my AGP NX6600GT card. In "desktop mode it was fine, but the moment I tried ANYTHING 3D, it would fail. Major artifacts, stprites, games crashing, you name it. It did it even with DXDiag. It happened out of the blue.
SO... after what would be a very long story to tell (and coming pretty clsoe to an RMA), I came around to replacing the VM-101 I had installed on my card with the orignal stock fan.
And the problems are now gone!
What baffles me is that why it would start doing this all of a sudden when previously it had seemed to run stable and flawless
It leaves me to believe one of two things:
1) The VM-101 does not actually dissipate heat enough to work properly over the long haul. (I.e. wear and tear finally took it's toll? kind of odd...)
2) Possibly the VM-101 got bumped and was touching the heatpipes of the XP-90 (but I would think either this would short the system altogether or at least the video would not work!)
This is a mystery to me. I might do a little more investigation, but I'd like to know your thoughts.
Thanks!
I had previously mentioned that I felt the actual mounting of the unit isn't very (i.e. sufficiently) secure. If it was sliding out of position slowly over time, it might have gotten to a point where contact was poor, resulting in insufficient thermal transfer--or, it slid enough that the TIM you applied was no longer fully effective. This is far too difficult to arrive at a solid result for, though. Since the card works fine again with the stock cooler, I hesitate to believe that the chip has suffered extended duration heat damage (not impossible, just highly unlikely; if it were really slow heat damaged, no amount of cooling should fix it, in most cases).
Contact with the XP-90 shouldn't prove to be an issue, I believe. In fact, my VM-101 touches part of the case, which is grounded (or earthed, depending on where you're from), and that hasn't affected operation at all.
Perhaps if you get the time, try to investigate a bit more? I wonder if there's a possbility the pipes on your unit have a slow leak, thus reducing the effectiveness of the pipes over time (no more convection action)?
But how would one test for that?
-Ed