[GUIDE] BitFury Miner Support/Tuning

[daddyfatsax]

running cgminer 4.3.3 and compiling with --enable-bab i get BaB0: Delay getting work reply ...

chips are detected at startup and connected to pool, can any cgminer users help?

I always had better luck with bfgminer than with cgminer. Over time I have switched both of my rigs back to chainminer, which I did not want to do.

How did you get BFGminer to recognize Blackarrow bitfury boards?  I never have been able to.  I am running minepeon.

I compiled bfgminer with --enable-bfsb. Won't work on V1 board, but I have a v2.3 board coming to replace my broken V1. So I'll see if I can get the black arrow driver to work.

[arorts]

For one of m-boards, the cables that connect the PSU with the rPI seem to be getting hotter and hotter over time. In fact, the white connectors are starting to get brownish.  I have not done any over clocking.

Another board with even more h-cards is actually cooler than the one I mentioned before. The PSU supports more power than what the set of m-boards consume.

Does anyone know why this might be happening? Any workarounds? :-/

[Trupik]

For one of m-boards, the cables that connect the PSU with the rPI seem to be getting hotter and hotter over time. In fact, the white connectors are starting to get brownish.  I have not done any over clocking.

Another board with even more h-cards is actually cooler than the one I mentioned before. The PSU supports more power than what the set of m-boards consume.

Does anyone know why this might be happening? Any workarounds? :-/

Sounds like there is a thinner amount of conducting metal somewhere along the path of electricity. The problem may be in the cable, the connectors, the soldering, or the PCB. Hard to tell. Is the cable hotter on the side of a PSU, or on the side of M-board? Either way I would try to replace the PSU if possible.

[terrapinflyer]

running cgminer 4.3.3 and compiling with --enable-bab i get BaB0: Delay getting work reply ...

chips are detected at startup and connected to pool, can any cgminer users help?

Did you find a way to get them to work? I am having the same issue.

[littleai]

Dear,
I want to buy one 400G rig, anyone is planing to sell your used rig. PM me.

Thank you!

[warrensgun]

anyone have a link to the pi image?  or can i just use a standard raspbian image and then copy over the chain miner etc.

[Micky25]

My sd card got corrupted so I imaged it again not realizing I used an old version of the image file from megabigpower. Nothing but low hash rates and 0 hash rates for 2 days.

When I realized this I put a new image on the sd card from megabig, the one linked in their help file pdf.

Its been running rocksolid for 2 days at full hash rate, about 30-34 ghash per card on an 8 card setup.

so, make sure you are running the most current image file as it makes a huge difference

Make sure you are using the correct image.  On the M-Board it will say what version the board is.  You can fit the V1 and V2 image on a 4GB SD card, however you will need an 8GB SD card for the V3 image. 

V1 (3.64GB)
https://mega.co.nz/#!GZEwQRiJ!TqG52DBj07C_zGO38dhXuyeNhwZ42VB2qk2ErfsrUZ8
 

V2 (3.64GB)
https://mega.co.nz/#!iVEwFKpT!TqRndnWurfcQ0V2gJb2sQlNcSGoMF1B7YpUoGDvVebI​

V3  (3.69GB)
https://mega.co.nz/#!7AFVhLwQ!BmgAIx32aQQxyadT1WKdJzNOUGQAQcQgYemfSJ6kMNM

Mirror: https://www.dropbox.com/s/skr9vitueeyxwmw/miner-v3-image-11-15-2013-customer.img

[AMD FTW]

so one of my boards recently stopped working (slot 3, I have other boards in slots 1,2 5,6 & 9,10) - I tried reseating it several times and tried it in other slots (slot 6 and slot 13) with no effect. no hashrate, no heat, no detection in the stats.log file.

looking at it theres no visible problem. a bit of dust had accumulated between the inductor and the TPS53355 possibly (but not likely) shorting 2-3 of the pins ont he upper side of the chip. I dont think this is the problem, but its theonly lead i have

any easy troubleshooting techniques?  I have a spare TPS53355 chip I could actually drop in if that is the issue, but i dont have the fine-point soldering skils to do that myself. (contracting out that task might work, but not sure what to look for in the yellowpages (Toronto, Ontario) to find someone capable of doing it.)

any ideas? I did not have a miultimeter handy at the time so no voltage info yet. I might have to wire up a 12V/GND to the board to check at home

Did you ever figure out what was wrong with your H-Card. I own a lot of these cards and I have at least 16 dead H-cards between all my rigs that I'd like to get them up and running. I did once observe one H-Card pop and I saw some smoke and definitely could identify the electrical burn smell. I determine the 30 amp fuse on the board was no longer allowing current to go through so I desoldered it and replaced it with another fuse. Unfortunately that board no longer works and the board looks completely fine. The other 15 or so defective boards between my other units also look good but don't hash.

[goxed]

so one of my boards recently stopped working (slot 3, I have other boards in slots 1,2 5,6 & 9,10) - I tried reseating it several times and tried it in other slots (slot 6 and slot 13) with no effect. no hashrate, no heat, no detection in the stats.log file.

looking at it theres no visible problem. a bit of dust had accumulated between the inductor and the TPS53355 possibly (but not likely) shorting 2-3 of the pins ont he upper side of the chip. I dont think this is the problem, but its theonly lead i have

any easy troubleshooting techniques?  I have a spare TPS53355 chip I could actually drop in if that is the issue, but i dont have the fine-point soldering skils to do that myself. (contracting out that task might work, but not sure what to look for in the yellowpages (Toronto, Ontario) to find someone capable of doing it.)

any ideas? I did not have a miultimeter handy at the time so no voltage info yet. I might have to wire up a 12V/GND to the board to check at home

Did you ever figure out what was wrong with your H-Card. I own a lot of these cards and I have at least 16 dead H-cards between all my rigs that I'd like to get them up and running. I did once observe one H-Card pop and I saw some smoke and definitely could identify the electrical burn smell. I determine the 30 amp fuse on the board was no longer allowing current to go through so I desoldered it and replaced it with another fuse. Unfortunately that board no longer works and the board looks completely fine. The other 15 or so defective boards between my other units also look good but don't hash.

Did you hot plug your cards? If yes the first and last chips comms are destroyed. you need to bypass them and your cards should hash again with 14chips.

[klondike_bar]

so one of my boards recently stopped working (slot 3, I have other boards in slots 1,2 5,6 & 9,10) - I tried reseating it several times and tried it in other slots (slot 6 and slot 13) with no effect. no hashrate, no heat, no detection in the stats.log file.

looking at it theres no visible problem. a bit of dust had accumulated between the inductor and the TPS53355 possibly (but not likely) shorting 2-3 of the pins ont he upper side of the chip. I dont think this is the problem, but its theonly lead i have

any easy troubleshooting techniques?  I have a spare TPS53355 chip I could actually drop in if that is the issue, but i dont have the fine-point soldering skils to do that myself. (contracting out that task might work, but not sure what to look for in the yellowpages (Toronto, Ontario) to find someone capable of doing it.)

any ideas? I did not have a miultimeter handy at the time so no voltage info yet. I might have to wire up a 12V/GND to the board to check at home

Did you ever figure out what was wrong with your H-Card. I own a lot of these cards and I have at least 16 dead H-cards between all my rigs that I'd like to get them up and running. I did once observe one H-Card pop and I saw some smoke and definitely could identify the electrical burn smell. I determine the 30 amp fuse on the board was no longer allowing current to go through so I desoldered it and replaced it with another fuse. Unfortunately that board no longer works and the board looks completely fine. The other 15 or so defective boards between my other units also look good but don't hash.

Did you hot plug your cards? If yes the first and last chips comms are destroyed. you need to bypass them and your cards should hash again with 14chips.

i dont think i ever hotplugged, but i havent tried putting the bad card back in for months now since its at home and the bitfury is a 40min drive away, for only ~33GH if it worked magically.

there was no visible issue on the board, but i alcohol cleaned some of the dusty and greasy components

[daddyfatsax]

Even sometimes just wiggling around in the slots will cause a small short. I had a box fan pointing at one of my rigs and the constant wobble eventually caused a slot on the M-board to not work. M-boards are really picky, at least the 3 I have had and now I have 10 cards and no board to run them on.

[adib]

Here is how i solved the mboard connection problem:


ill post some more with an actual mboard with these on, self made of course

Edit: daddyfatsax care to sell those unused bitfury boards?

[adib]

so one of my boards recently stopped working (slot 3, I have other boards in slots 1,2 5,6 & 9,10) - I tried reseating it several times and tried it in other slots (slot 6 and slot 13) with no effect. no hashrate, no heat, no detection in the stats.log file.

looking at it theres no visible problem. a bit of dust had accumulated between the inductor and the TPS53355 possibly (but not likely) shorting 2-3 of the pins ont he upper side of the chip. I dont think this is the problem, but its theonly lead i have

any easy troubleshooting techniques?  I have a spare TPS53355 chip I could actually drop in if that is the issue, but i dont have the fine-point soldering skils to do that myself. (contracting out that task might work, but not sure what to look for in the yellowpages (Toronto, Ontario) to find someone capable of doing it.)

any ideas? I did not have a miultimeter handy at the time so no voltage info yet. I might have to wire up a 12V/GND to the board to check at home

Did you ever figure out what was wrong with your H-Card. I own a lot of these cards and I have at least 16 dead H-cards between all my rigs that I'd like to get them up and running. I did once observe one H-Card pop and I saw some smoke and definitely could identify the electrical burn smell. I determine the 30 amp fuse on the board was no longer allowing current to go through so I desoldered it and replaced it with another fuse. Unfortunately that board no longer works and the board looks completely fine. The other 15 or so defective boards between my other units also look good but don't hash.

Check your voltage with a meter, if it says 0.6...0.75 youre good, just the first chip if youre lucky is gone.
Sometimes some sort of short in the spi chain can go wrong and it kills the whole h-card, finding it is quite challanging but doable.
As i have posted earlier, the 2.2 mboard can be modded with other type of connector. the one i posted sit on 2 mboards they are gold plated and tight
the rigs are running for months and are imune to power outs as the bfsb image sits on an usb stick and imune to airflow as the connectors are doing their job.
I could try o repair them or buy them from you faulty.

[AMD FTW]

Here is how i solved the mboard connection problem:


ill post some more with an actual mboard with these on, self made of course

Edit: daddyfatsax care to sell those unused bitfury boards?

If you had any pics of this setup, I'd be interested in seeing them.

[adib]

Here you go:
v1.1, v1.2 and v2.2 moddable


v1 chips, 12 chips/board(overclocked), 250ghs, 300W at the wall, 100 chips
same setup, v2 chips, 290ghs, 250W at the wall, 96 chips

[AMD FTW]

Here you go:
v1.1, v1.2 and v2.2 moddable


v1 chips, 12 chips/board(overclocked), 250ghs, 300W at the wall, 100 chips
same setup, v2 chips, 290ghs, 250W at the wall, 96 chips

Thanks for the pix, interesting setup. I just purchased a hot air rework station and figured this would be a great project to learn on and possibly salvage some boards. I'm pretty good with a soldering iron so this shouldn't be too bad. I'm looking now for an oscilloscope that looks decent and would be great for this project along with other projects I might tinker with in the future. My parts should arrive next week some time, so the salvaging of the boards begins next weekend.

Hot air station I got
http://www.amazon.com/gp/product/B00A2VQA24/ref=oh_aui_detailpage_o00_s01?ie=UTF8&psc=1

[adib]

first check if you have voltage,
if you do, you are good, the chances are the first chip is gone
if its not that, the pcb traces are gone, you have to verify which one
a meter should help also

Have fun

[AMD FTW]

I checked 5 cards tonight in over 20 dead cards in my defective bitfury card stack and I've found that all of them have no voltage going to the bitfury chips. I checked this by using a multimeter and saw 12 volts coming out of the fuse. Then I checked the inductor and used the bottom of the voltage regulator pad as the ground and saw no voltage. So what are the odds that the voltage regulators are all toast? I have no experience with the durability of these voltage regulators and if they were pushed to far maybe they finally just kicked the bucket. I bought an intersil 0-6v 30 amp board below to verify if the chips do in fact work. I should have it in a few days to verify if the bitfury chips are indeed good. If the h-boards traces are gone, and replacement pcb's aren't available, maybe I'll try to design a board with all these chips in series so I don't need to use a vreg which should make it more reliable. If I run them in series of 16 bitfury chips with 12 volts, then they'd only get .75 volts or so which will keep them running efficient and cool. I checked the voltage on a few good H-cards and they are taking .865 volts, and they've been this way since day 1 as I've never adjusted the potentiometer. Anyone know the rest of the h-board pinouts on the pictures I posted.

Power Source
http://www.mouser.com/Search/ProductDetail.aspx?R=ISL8225MEVAL3Zvirtualkey57760000virtualkey968-ISL8225MEVAL3Z

[adib]

I think you were better off, buying the power TPS buck.
Also check the inductor to see if it has voltage before and after it.

Good luck

[AMD FTW]

I think you were better off, buying the power TPS buck.
Also check the inductor to see if it has voltage before and after it.

Good luck

I'll hopefully will have the Intersil 30 amp development board tomorrow. I had a chance to use my new hot air station. I removed the old TI buck converter on the H-Board and replaced it with a new one. Unfortunately I didn't have any voltage before the inductor prior to removal of the old part and also after I installed the new one. I tried another new TI buck converter and had no luck here either. This was the first time I've ever used a hot air rework station, so I decided to just practice on one of my h-boards removing capacitors, resistors, buck converter, and inductor. I got the hang of using the hot air station, so I figured I'd try it on the solid aluminum capacitor. I removed the capacitor and then I started replacing the components that I removed. I reinstalled everything with no issues until I got to the aluminum capacitor. I guess I heated it up to long and it blew up. Lucky it didn't hit me in the eye as at this moment I had placed my saftey glasses on the work bench and forgot to put them back on. I'll have some more time tomorrow to mess around with the boards. I'll post the TI stock h-board buck converter diagram here.