I have good heatsinks but perhaps could use slightly larger ones on the regulator/inductor. 0.865-0.87 is good for 40GH on my boards, but the pencil mod ramps up to maximum (or minimum?) resistance over a few days and its difficult to get the exact balance. once it climbs to >0.88v the boards become unstable and i see high error rates followed by entire boards or just some of the chips dropping to 0 every few hours and upkeep is a PITA.

i'm sure if i had cold air moving through the heatsinks (rather then the 21 C room temperature) I could reach 0.9V though, or at least achieve a stable 0.88v

work on your math. assuming the boards run at >30GH without overclocking, you should be able to see 2BTC return within a few months, and the bitfury chip is pretty future-stable (on par with many 28nm designs when underclocked)

impressive. I cant get past 0.88v (39-41GH per board) without having chips or boards drop to 0 hashrate or produce massive3 error rates time to time

Looking to buy a few more H-cards for my setup, for reasonable prices. All conditions considered (pencil mods, EOL boards with bad chips, heatsinks attached, etc). I am in Toronto, Ontario, Canada but am open to worldwide shipping.

I can offer to work with escrow if desired to ensure smooth and fast purchase. Feel free to contact me by PM, I'm looking to pay less than $350 or 1.75BTC each.

$75

hopefully the price will be set <2BTC at this point

ahahahahahahahah responding to an ignored post? You get the gold trollie award for that one

BFL deserves to be sued or have action brought against them by the FTC. The delays are unnacceptable and Inaba/Josh is the most abrasive personality on the forums, constantly mis-informing and insulting.

Im lucky paypal was able to refund me after 45 days (closer to 80 days) when i realised that the device i ordered in early june with a mid-september estimated delivery was actually coming no sooner then late october. (Now late-november seems to be the situation).

Id love to know how KNC shipped hundreds of units out in a few days (each one containing several of what is similar to the BFL board+heatsink assembly) and BFL seems incapable of averaging moe than a few dozen units a day.

If BFL had proper forum presence (ie: kicked inaba to the curb and put someone more positive in) I might have not demanded my refund, and been able to accept delays. but to have inaba call people idiots and ban CUSTOMERS from the BFL forums is insane.

a new one makes a good backup. 4GB is mine and makes keeping both with a working image much simpler if the issue happens again. you'll need an sd card reader, or a lot of linux skills. (I suggest a 2$ card reader :p)

not true. I think there will be another 1-2 months of 30% increases, then it will HAVE to slowly adjust downwards. Costs will slowly approach the price of manufacturing, in the case of bitfury it is unlikely they can make a board for any less than $200 (and continue to pay off nre costs) for some time. Other manufacture use more complex designs that may be even more difficult to bring, and machines like asicminer / BFL / avalon (first-gen anyways) will become inefficient far sooner than bitfury and other gen2 asics capable of <1w/GH

to double hashrate by Nov 24 (less than 3 difficulty changes at 30+%), each company that is now shipping would basically need to ship 70%+ of the amount already shipped (Basically asking that they have tightened their production speeds by 2-3x current.

Its you SD Card. Load up another one or re-image it to get going again. It seems like sometimes powering down causes the issue.

FOLLOW-UP QUESTION: what would happen with there being fans drawing power from the M-board? Is it possible that when the unit powers off, the fans' continued rotation is forcing a dropping (but available) voltage across the RPi damaging either the SD or causing partial operation/writes by the processor?

im curious whether the victims of SD cards or RPis dying commonly use more or bigger fans drawing from the board connectors? I was using a 3-pin 230mm and 2-wire 80mm the one (and only) time my SD card needed a swap. Im still using that since with no issues

I expect next week they will announce the november pricing, hopefully bringing these closer to parity

what about if all the chips had parallel power, and the m-board (or an external PSU) simply provided 0.85V?

any chance they will be made for sale, or the schematics made public? And what defines whether the cards run well vs turn into melted goo like seen in some of the other photos

from my research it appears the main issue is that to develop a board based on someone else's chips is usually an expensive venture unless you source it to china or north europe, and even then the asic manufactureres charge too much for thier chips on a reel.

for example lets take a bitfury 16-chip board:

PCB prototype cost (3-10 boards): ~$60-200 per board with varying timeframes
parts (not including bitfury chips): ~$15-30 per board depending if you want to upgrade components for better overclocking
assembly (in north america): ~$80-$200 per board depending how many and where you go. generally there is a first-time fee of >$200

total cost per board for small-run batch : $200-$350 approx
total cost per board for batch of 200 boards : $50-$100 approx
CHIP COST: currently the lowest ive seen is around $15 if ordering a reel. for 16-chip board : $240

this means that even with large production batches, we would be limited by what the chip designer wants for a reel. In the above example, a medium-sized run of boards and assembly would not be much less than direct purchase from MBP/BFSB (especially if they drop pricing to $350-$400 soon like i think they should)

Even with the promise of $5 chips in janruary, you would see costs of around $150 per board and it would be minimally cheaper then what they would likely charge at the time. The only way to step ahead of the curve is to design your own chips or have an open-source design.

*open-source cant be difficult, since most designs are based on FPGA code, right? I would imagine it cant be difficult to make chips for $1 a peice that are a basic design (exactly what avalon and asicminer did IMO)*

im guessing janruary we will be down to 60%, and march down to 25%

should we write on the side of food : "might make you fat. hoarding your food money will make you thinner and richer"?

(if you say yes, welcome to the nanny state for idiots)

i would beleive that you cant fill every slot on the board if using a heatsink taller than 8-10mm. i have some 35x35x6mm heatsinks on mine that easily allow airflow between every slot

you cant be serious? asicminer, bitfury, and avalon are all selling units that are available immediately.

i would guess that the regulator is half the size of a chip, but emits 3-4x the heat of a chip

1) create product [check]
2) bring to market [....uh....check?]
3) fulfill all pre-orders and sell from in-hand stock [i guess in two weeks...right now you are shipping preorders 2-4months later than initially 'promised'/suggested]
4) take preorders on next-gen product that is rediculously overpriced by today's standards, and promise to deliver it before the 110nm preorders are filled