I switched the crystal to the stock frequency and now efficiency is back up to
95% and rising.
But just before I did the switch, I noticed that the OTHER 3 blades (stock
configuration) were resetting every couple of minutes (WTF)? But since I
changed the clock back on the first, no more resets, no hashing
slowdowns, and the board I was fiddling with now acts as it did before,
in spec.
I use the same proxy for all boards, and that has been incredibly stable for
a week plus. Up time for all boards was over a week.
So I am back to where I started, at least. The board still has the
power supply voltage boost, so it's running a little hotter than before.
Don't know if I will fiddle around with it again, but I wish I could figureout
what is going on.
Topic: Low efficiency on a newly overclocked ASICMiner blade (Read 755 times)
I just got my overclock components for my ASICMiner blades, and started modifying
them. I have a few crystals to try, but all I've tried exhibit the same problem: efficiency
varies widely but is typically in the 25% range. Hashing is as expected at the
overclock rate (e.g. 13108 GH with a 15.538 clock). I have done the PS resistor mod to
all 8 regulators. I have no X's, ever. Not a thermal issue because the problem is
exhibited at temp and stone cold. I have 3 more blades to mod, but I don't want to
do more harm than good. Any clues?
them. I have a few crystals to try, but all I've tried exhibit the same problem: efficiency
varies widely but is typically in the 25% range. Hashing is as expected at the
overclock rate (e.g. 13108 GH with a 15.538 clock). I have done the PS resistor mod to
all 8 regulators. I have no X's, ever. Not a thermal issue because the problem is
exhibited at temp and stone cold. I have 3 more blades to mod, but I don't want to
do more harm than good. Any clues?
Jump to: