I bought a Sapphire 128MB 9500 to attempt the famous resistor mod. Well, once I pulled the hsf I realized what I was in for -- the resistors and solder points are tiny! Of course I had no luck even desoldering the little bastard. I dug up my half-used window defogger kit to try the cheap method, but the paint had gummed up beyond use. What the heck, why not give the old pencil trick a shot. I tried to connect points 1&2 (according to the xbitlabs diagram) with a mechanical pencil, but when I attempted the bios flash I had no luck. Big surprise. What's worse, 3dmark crapped out halfway through.
Oops! I figured my hamhanded handling of the board had fouled something up. Board not oc'ed yet btw. I didn't restore the 9500 bios at this point, because I had another thought in my head -- what if it's heat related? I opened the window to cool my room down and tried some more testing -- no problems. Then tonight, I ran the unit up to about 320/290 and settled in for some Madden 2003. Strange texture shimmering then odd shadows that should not be there ruined my experience.
Remembering the gummy yellow TIM I scraped off the core -- and hearing about people using shims, or even removing the bracket -- I pulled my card for some more abuse. Pulling the hsf, I noticed that there seemed to be no, or almost no, contact with the core. The only compound (AS3) on the hsf had balled up into a couple little droplets. I tried to yank the bracket ala digit-life (see the R300 roundup) but had no success (can't see how they managed it...).
Failing that, I grabbed a small piece of tin foil (got that idea from a post somewhere), folded it, folded it again, cut it to fit the core, and dropped it on top. Then I smeared a drop or two of AS3 on the hsf to cover the other side of the foil. Replaced the hsf, and, feeling confident, ran the core up to 340. Played a full game of Madden 2003 -- no artifacts whatsoever, in contrast to my previous experience.
In summary:
jaybee
Oops! I figured my hamhanded handling of the board had fouled something up. Board not oc'ed yet btw. I didn't restore the 9500 bios at this point, because I had another thought in my head -- what if it's heat related? I opened the window to cool my room down and tried some more testing -- no problems. Then tonight, I ran the unit up to about 320/290 and settled in for some Madden 2003. Strange texture shimmering then odd shadows that should not be there ruined my experience.
Remembering the gummy yellow TIM I scraped off the core -- and hearing about people using shims, or even removing the bracket -- I pulled my card for some more abuse. Pulling the hsf, I noticed that there seemed to be no, or almost no, contact with the core. The only compound (AS3) on the hsf had balled up into a couple little droplets. I tried to yank the bracket ala digit-life (see the R300 roundup) but had no success (can't see how they managed it...).
Failing that, I grabbed a small piece of tin foil (got that idea from a post somewhere), folded it, folded it again, cut it to fit the core, and dropped it on top. Then I smeared a drop or two of AS3 on the hsf to cover the other side of the foil. Replaced the hsf, and, feeling confident, ran the core up to 340. Played a full game of Madden 2003 -- no artifacts whatsoever, in contrast to my previous experience.
In summary:
- don't try to pull the resistor unless you are very skilled at soldering and have the right equipment.
- be careful removing the power plug for the hsf -- it's easy to slip & yank one of the wires out
- flashing the bios should allow a significant overclock
- once you scrape off the yellow TIM, find some way of ensuring good contact between the core and hsf!
jaybee
Facebook
Twitter