i would buy these if it makes sense to compared with a 16-chip board.

just tweaked mine today.

BEFORE: 1.130K at cold state = 0.788 V    ------> unknown hot state resistance = 0.808 V     ---> 39.5GH average
NOW:     1.100K at cold state = 0.815 V   ------> unknown hot state resistance = 0.834 V     ----> 42.0GH average (based on 2.5hours)

I am using small chip-sized heatsink on 14/16 chips (i need more lol - and its been suggested that backside cooling may be optimal) and have 5 1"x1" heatsinks on the back, centered over the 4 capacitor clusters and in the center of the board. a 120mm fan is aimed at the back and the front is cooled by a 60mm and an 80mm fan on opposite sides

I don't intend to push further until I know that:
a) the chips can handle a certain amount of heat (if they can handle 70% of what asicminer does, then ive got a lot of headroom)
b) what the 30A converter can handle and its failure mode(s) (intron suggested that they simply switch off at overvolt/overheat)

The heat issue seems minimal so long as the chips can handle some warth and voltage. The issue lies in the 30A limit. If my 'math' is right, the chips use around 0.8-0.9w/GH at 2.65GH/s * 16 chips * 0.835 V = 28.3-31.8 A. The other components may draw additional power on top of this...?

will it blend?

well as i said, theres no way i set 136 as the timing value. I was changing it from 55 to 54, so 554 or 545 are the conceivable saved numbers. how it uncapitalised the aifdso command i have not a clue - failsafe perhaps?

^see my edit to the post. turns out whatever switched it from 54[5] -> 136  also changed the aIfDSo -> aifdso

i returned it to the capitalised form and its active again     waiting for hashrate to climb and stabilize right now

hmm so why would it show as 136 in .stat.log? like i said, its now no longer hashing. would using autotune or a set value other then 54 get the chip back up to operation?

EDIT: well shit, apparently whatever happened it changed "aIDSo" to "aidso" which would effectively have caused the issue - i just changed it back and am restarting the system now.

i just made a mistake with my h-board best.cnf file, and i may have saved the first chip to "545" instead of "54" because on reboot it read as 136 and 0.000 hashrate.

i fixed it to 54 and rebooted but the chip still appears unresponsive. any advice or tips? looks like voltage is still going to the chip, so did i ruin the chip or is it salvagable?

I would hate to think i accidentally crapped out the chip with this tuning attempt

*smirk* +1

AFAIK its 0% fees

Its 0.26BTC/GHS not vice versa. managment fees seem to be limited to electricity cost only with no pool fees - i assume this will change in the coming week as trhe site ramps to full use.

buying cex.io GHS gives you access to the ghash pool so that you can watch your worker (which is set to as many GHS as purchased). It also gives instructions for pointing your own units there, (although im not entirely sure if this was on purpose or accidental)

worth noting that i moved to ghash.io and now have a steady 39.5-40 GH/s from my starter kit. Difficulty 32, 1 worker, 0% pool fees.

oh, and cex.io is partnered with them, so you can buy hosted GH/s (bitfury based) instantly for about 0.23BTC/GH - great if you dont want to upkeep hardware or preorder anything

i usually check voltage across the caps, seems pretty reliable, esp if you want to average the reading of a few caps (usually all the same of course)

i havent gotten a free GH/s yet, but when you withdraw, they send you a confirmation email. check your junk bin

edit: got my bitfury pointed at ghash.io now, and it is working amazingly

i just moved to another pool that chose a difficulty for me (not sure what) that forced my hashrate (both at pool and via webgui) down from 38.5->27 GH/s. I changed it to 16 manually and im all good again.

what difficulty best suits a 35-40GH/s miner? 8,16, or 32?

edit: bumped to 32 difficulty, and it seems to be a little better for my miner. went from about 38.5GH to 39.5GH

just withdrew 0.065 BTC without a hitch. 8GH mining there now, and i just pointed my 38.5GH/s bitfury system at my new ghash.io account. - all looks good, seems to have picked it up pretty quick, but failed to properly set a difficulty (it mustve used 4 or 32, because i was getting 70% hashrate until i manually set it at 16)

cex.io - for the price of a block erupter you can have 1GH hosted by ghash.io

I think they are a bit longer it looks like, but similar width. however, the heatsink is larger, and presumably only about 5 fit in a 10-port hub for this reason. (although, the heatsinks can likely be smaller with time, since they are dissipating just as much heat as a block erupter)

looks really interesting - i will be testing with 3-4 GH/s i think

is there an easy way to build/'gitpull' this, or does it involve imaging the SD or otherwise turning off the miner?

im hashing quite happily at the moment, and SSH is okay as long as i maintain >38 GH/s