j.c.f.'s Overclocked Athlon
Project
And what do I get when I fire up jig #2? Identical, I
mean i@#$%ingdentical results as the last jig. So I sit
down and try all the combinations and find that BP_FID2 and BP_FID3 are reversed. Either
I'm suddenly suffering from some amazing malady only affecting those two connections and
am unable to distinguish just those pins or Tom's schematic is wrong. I don't care, I got
it going at last.
So how fast does it go now? Not a whole lot faster,
750MHz @ 1.6V is about the limit of my setup, 800 results in a badly corrupted POST,
graphics and characters all over the place, urgh. So I figure, maybe the 235W POWERMAN
supply is finally running up against it's limits so I slap the L&C 300W supply in
there.
Well, that was a mistake, that supply is a piece of
junk. You can't run TECs between the 12V and 5V rails, so it's either 5V or 12V and at 12V
even that 120cfm fan has trouble dumping the heat. Not only that, but I
can't even POST at 750MHz (meaning it's not likely to be a heat issue) which confirms my
theory about the supply being taxed. Here's a clear case of one power supply making 750MHz
and the other not. So out it came and back in went the 235W POWERMAN FPS235-60GT, maybe
because I can boost the current of the 5V rail or what I dunno, but at least I can hit a
rock solid 750MHz with the limiting resistor on the fan bypassed (summer of course will be
another thing but that's months away).
All attempts to up the voltage lead to more complete
failures (no video, nada) so more research is probably in order there to find out if the
settings are doing what they should be or whether the supply is crapping out or what, but
in any event given the clear thermal situation I can't be expecting to up many settings
and it's difficult to be unhappy with the golden 50% overclock even if it is technically
only 15%.
Perhaps the TECs are at their limit, my calculations
certainly indicate that I am pushing them so if I get really enthusiastic I'll try just a
plain ol' heatsink, but then I loose the current shunt and I should probably get a
temperature probe so I can add a bit of science to this whole affair.
Ah, the life of the overclocker is never complete...
Should you really want to communicate you can eMail to: jforster@u.arizona.edu
[Page 1] [Page 2] [Page 3] [Page
4] [page 5] [Page 6] [Page 7] [Home]
|
