must be worth 90 BTC 

read the quoted text above your post. get your system working with the default v2 image (only for a v2 m-board), then run the commands from the RPi terminal

in my experience a little bit of airflow does a lot. these things are stable even when hot though so dont be too worried. alternatively, attach some small heatsinks to either the stock plate, or atop the asic chip

5 block erupters & my 36.5 Ghash/s Bitfury starter kit. Next step is 3.8 Ghash if all remains stable for 24hrs.

And the setup is semi-temporary. I hope to get a better spacing on the fans and wires when i get the chance to beautify it. Till then, heatsinks make good sturdy bases for fans and securing them

ooh, not being a linux guy that still looks like a headache! (massive respect for the method/result though!). I can see 1.5hr averages at bitminter that work okay, but a longer-spanning solution would be nice. maybe the next sd update ill try to get the logger set up.

right now, im seeing a very steady 36.5 Ghash at the pool over the last 3 hours, with the bitfury page suggesting it might be closer to 37

EDIT: after about 16hrs, looks like 36.8Ghash is the average hashrate, with the last 3hrs above 37Ghash
EDIT 2: some new tweaking, and it looks like i am averaging 37.7Ghash

I could offer pickup, but only if an overnight/2day xpresspost isnt a well-priced, easier method. (queen and lansdowne area)

from what I'm looking at:

20 units (If i can arrange some form of groupbuy/sub-groupbuy for at least 10, I can undertake some risk for my resale and own use purposes) : 0.82  BTC each
Shipping to Toronto (average per unit): ~0.04 BTC
Personal risk/time consideration per unit: ~0.08BTC

So approximately 0.94 BTC per unit, plus:
Priority/Xpresspost reshipping within Ontario (per sub-order): ~0.12BTC (plus maybe an extra 0.02 per extra unit if package size/weight is increased )

If this seems reasonable to a few people, I would be willing to go down this path. I have no experience with escrow, so feel free to either vouch confidence in my word, or point me in the direction of setting up escrow (John K.?) I imagine escrow would add a nominal fee, but may also make my up-front payment harder to acrue (who has 16BTC lying around? - not me [yet])

I have 1 in my order #13X and another in #8XX. That will satisfy me until whatever comes next (28nm? price drops? stock OC'd cards/more reliable chip?)

from 3 cards I expect to run 105 Ghash/s

what other chips can overclock 30% or more without extreme cooling mods (just airflow)? 

you can make the sd card at a windows system, but to change the ip settings you need to either use a monitor/KB with the bitfury, or use linux to change the network interfaces file.

The chainminer in the image works, but not optimally. once the sd is in the rpi with the right ip settings, do this:


It is much better version, and will cut your error rates massively. I went from 10% errors to 3%, and still have some tuning to do that should bring it under 2%

terrific, so can i expect them to return to my sending address, or do i need to email/PM/website-form you the address and request?

You have handled this very well, its too bad avalon screwed up so badly. If there are any pcb/partial boards already i'd still love to have one, even without chips. otherwise, I hope that an avalon gen2 or bitfury design will turn out better for you and those involved.

how are you making that sort of log? Im moving to manual tuning now, since most chips are stable at 54/55 with 1% errors and 36Ghash/37Ghash-nonce.

if it runs well for the next few hours i will consider moving the resistance from 1.178 to 1.165 (before it gave me issues, but that was prior to chainminer update)

The calculator is good for a 2-4 month idea of increase. After that i think we will become much more linear

This just stabilised my error rates ennormously. I had 2-15% errors per chip (10% average). Now its under 2% average (one chip is at 6%, thats the only exception)
35Ghash noncerate and climbing as the autotune does its thing. 37.5Ghash seems like a real possibility

UPDATE: after a little tuning to 54/55 as needed, 36.5 Ghash is my new [very stable] record.

just ran the above chainminer gitpull, up and hashing again now, at auto 54 and decent rates (still climbing)

Previously, i had about 34ghash average poolside using 2 workers at difficulty 16, autotuning, and pencil mod of 1.178k. The actual hardware hashrate was closer to 37.5 but 2-15% error rates on the chips. aftermarket cooling includes a fan on each face of the board, as well as multiple heatsinks (tiny one on each chip, and then 5 small heatsinks on the backside, 1 opposite each capacitor cluster)

EDIT #1 (5min): 1 worker, diff=16 (bitminter), autotuning is 54 across the board, hashrate: 35.86Ghash, noncerate: 35.03Ghash.  This is looking really good with the chainminer update!
EDIT #2 (15min): hashrate: 36.5Gh, Noncerate: 35.0Gh,   autotuning is moving most chips up to 55 now
EDIT #3 (30min): hashrate: 37.4Gh, Noncerate: 36.3Gh,   autotuning is about 50/50 between 54 and 55 on all chips. Several chips running at 54 have no errors, and a chip at 55 is 7% errors (hopefully fixed by next autotune)

not at the moment, except perhaps from people who bought multiple starter kits that were in the delay. Wait a few weeks and MBP should get a v2 board in your hands. (whatever 'loss' you experience now will surely be remedied by having an extra m-board (with arguable resale value around 500-800$)

I have tiny heatsinks on the face of each chip (except 2 of them), 5 small heatsinks on the back (centered opposite the capacitor clusters), a 120mm fan blowing directly at the back and a 60mm and 80mm blowing from opposite sides on the front face. This seems excessive, and nothing on the board gets past "warm". even the Dc converter (i assume thats what the cube is) has a tiny heatsink to help shed its warmth (it was the hottest component i think)

I was using 2 workers at difficulty 8 and getting hashrate fluctuating between 30-36, with 34 being about the average over the last 12 hours. I just changed to 1 worker at diff=16 and will see how that works. 2 workers seemed to be a bit more stable than 3 workers.

I have yet to do any tuning, will wait for new difficulty to kick in and see what the .stat.log looks like. Right now all 16 chips are auto at 55 and running 1.9-2.35 ghash each, with error rates between 2-15%  (good:2317 errors:221 spi-err:1 miso-err:6)

edit: at first 5min glance, 1 worker at diff:16 seems to only result in <26Ghash at the pool. Im switching back to 2 workers at 16

Where do you ship from? The shipping price to canada you just gave me seems excessive

International Express Courier (1-5 working days) - Tracked, Signature required    $426.74   
International Economy Courier (2-6 working days) - Tracked, Signature required    $290.02   

I am down to help organise a small ontario (or even just toronto) buy, and have enough experience to make reshipping quick and painless (local 1-2 day delivery would be in the range of $10-$15 via xpresspost). However, I have no experience with using escrow (and dont have the BTC wallet to support purchase of more then 4 units right now). If anyone wants to make this happen, ive requested a quote for 10x units and 20x units, of which i would probably take 3 for myself

7x the hashrate, 7x the price, 1/7th the ports.

Still wont make more then 0.4-0.6BTC by the time difficulty makes them obsolete

you should remove the FPGA chips (curious - what sort of monetary value is in them now?), then send chip-less units for the museum