Ok, left bags covering most of the heatsinks overnight and stats are still better. The temp sensor must be in the chip because otherwise how would I have shown it being hotter with an open case? Edit: used an IR temp sensor and highest temp I can find anywhere is 130F, so even though the temp sensor may not be in the chip, I am still below the temp of 70C that many are finding ideal. 130F is only 55C. So this method is only if your unit is in a cold basement like mine.
hmmm my unit is in a cold bsmt 41c on 2 units and 27c and 32c on other two...the bsmt is 62F...i'm running .95 at 555gh stable (it drifts up 10-15gh and back down) and 547.7ave ...i'm doing like 537.7 stable on slush and I have the case open and 2 fans both 3" both in front right and left blowing crosswise (x pattern) across the miner
from posts here seems my unit works better then most...(or am I in denial) my impression ...and thus fence sitting on going to .98
also do the 'correct' error formula (not using knc cgminer hw errors alone) and I get a 5.5% or so error rate
the only other thing is once in a great while i'll have 1 core 4-86? enbable on and then off ..but prob only every 3hrs or so at most is my guess
so if your unit is ying my unit is yang
ying/yang
go figure..unless I'm missing something oh...my work unit is 7767.6 or some such....again it is stock miner .95 came in a destroyed box filled with rainwater ..I'm still going with the theory that the reason it works as well (or at all) is they had to hand carry the box/unit with 2 hands everywhere or it would have fallen thru the soggy box that was ripped in half...there fore it never "took flight" like I suspect other units did in such a large if light box....or I'm full of it..but anyway
whatever ...I usually walk by the unit and avert my eyes...I'm not worthy....its the boss
anywhoo....from where I see it right now correct me if I'm full of it (likely)
Searing
Yep, you are doing pretty good. No reason to change anything if you are happy and don't want to risk it.
