Topic: **US** BitFury Setup Guide - page 8. (Read 53224 times)

hmm. few ideas:
1) take a voltmeter to the board and check voltage across the pads that [used to] have capacitors. - you should be able to see ~0.79V, but IME not each pad set has voltage on my unit, so test a few to make sure
2) check your ip - is it actually on your home network with correct gateway setup so it can get online
3) check your pool info - if this was the issue, youd likely see your chips all at 0GH
4) check the orientation of everything. if a card or the rpi is backwards you could either cause damage or simply not see anything
5) make sure you are filling the banks as advised (slots 1, 5, 9 in your case)
6) try with only 1 card at a time and/or rearranging the order - maybe an spi/miso issue somewhere is causing the issue

Hello BFS friends,

I recieved the BFS rig today. v2 m-board and 3 x h-boards . The problem is I cannot see anything running.
I tried lot of things. Boot screen says something like "miner: no process found"
the /run/shm/ is quite empty (no logs)
can somebody help me, please thx

I have V2 M-boards from Punin. I have been fiddling with board placement a lot and have noticed that the 4th bank is not very useful for me. I get the best result when I am running the 2 M-boards with an empty 4th bank(2+2+2 and 3+2+2 cards). The difference was over 30GH/s between using and not  using the 4th bank.

YMMV, but give it a try if you have boards in the 4th bank doing 15GH/s or less.

  Your sd FS bust it.

I have 5 cards running  34.53 Gh/s average each.  Only cooling is a $16 box fan from walmart, and they are stacked on top of a 4 module Avalon unit which gets pretty warm.  So yes, good airflow is all you need to reach 34-38gh/s.  I had one doing 43+gh/s, but that was when I put the voltage up significantly.

Success!  Thank you all. 

the sd card image file should have the newest chainminer on it. (if you, you update later once the rig is running)

SD card can be imaged from whatever your daily deaktop/laptop is

Thanks!  Can I reimage the new card on my laptop or does it all have to be done on the pi?
Also, what's the difference between the image file and the chainminer files?

EDIT: Anyone?

SD Card went bad.  Had the same issue on 2 of 3 of mine.  The SD cards these were shipped with appear to be of low quality.

Go out and buy a new SD card and then image it from the V1 image in this thread.  That solved my problem.

I restarted my miner for the first time in 2-3 weeks and can't get it going again.  I have a starter kit plus another h - card and two on the way. V1.0.
Here are some screen shots.  Can anyone diagnose what's up?  I am thinking the SD card went bad?  I have never had a problem with it but everyone else seems to be writing about it.
http://i.imgur.com/6GtLrDY.jpg
http://i.imgur.com/Zke9YYV.jpg
http://i.imgur.com/dZsOaBn.jpg
http://i.imgur.com/hhO5Bw1.jpg

  Why dont you do RMA for exchange hboard?

I can't really see or smell anything wrong. I used my multi-meter to test the continuity of components but everything is at least coming back with some sort of reading. The C04 spot may look a little different but it's too difficult to tell. I'm not an expert in this area to begin with...

Here are some pictures of what I see. http://imgur.com/a/1ozpu#4

I almost rather just sell the card as is for $300 with free US shipping or $20 international. I don't have much free time to deal with it right now.

the 'pulse' cube is an inductor if im correct. i put a tiny heatsink on mine to cool it but its probably unneccessary. the fuse is actually a 'polyfuse' and is the big flat resistor-like item marked F300L or something similar. it is resettable and will work again after a failure if the power is turned off. (intron has a post about it somewhere in the forums)

if its tripping the psu, thats not good, and is quite strange. it would indicate that either the psu is having far too much load taken across its connection (what brand/capacity is yours, and how is it wired to the m-board?) or that there is some sort of backfeed from the board. (less likely)

I would suggest looking at the small chip located near the inductor/pulse cube. it is the regulator and gets very hot during operation over 0.8V or 32GH. it should be unable to blow or fail (presumably), but from your issue it seems like a potential place to check first for any discolouration or damage. youll know the component by its ~12 pins and the thermal vias on the backside of the board

OK I'm back up and running again, but one H-board I think needs to be looked at. I checked each of the new H-board individually, and one of them doesn't want to play anymore... Every time I use it either in a group with other cards or by itself, it seems to trip up my PSU and everything goes dead.

What can I check with my multi-meter to see if anything is damaged? I checked the big thing with the word "Pulse" on it, since I heard it was a fuse. I do get a little bit of a reading with my multi-meter, so I assume that isn't blown.

I wish MBP would have sent an identical board like what was sent in August. Those seem to be a little more robust.

at 35GH you need to have heatsinks and fans. IME:

25GH : no/minimal airflow needed, no heatsinks needed
30GH : airflow needed
35GH : airflow + some heatsinking required
40GH : airflow and decent heatsinking required (I use 4 * 35x35x6mm heatsinks, plus tiny heatsinks on the face of each chip [probably unneccessary to cool from the front], plus a few tiny heatsinks on the inductor and the backside of the regulator.
45GH : I would imagine you need a more intensive method such as copper heatsinks or liquid/immersion cooling

In your setup i cannot imagine pushing more than 32-35GH per card without problems.Its also worth noting that when a pencil mod is in use, it tends to creep a little bit as the board gets warm or ambient temps rise. I often see a pencil mod board start up at xxGH, after an hour its xx+1GH, after 6hrs its xx+3GH, and at the warm points in the day is can be up to xx+4GH

needless to say, this means that the pencil mod sometimes nees to be repeted a few times, with 3-12hr sampling time to see what sort of speed you get. if the components get hot, the hashrate can get wonky or the chips/regulator will switch off.

btw, wasnt dave supposed to roll out some v2 m-boards to those still using the v1? by being able to put cards in each bank it gives much more space in an unfilled unit to get heatsinks or airflow. I actually found that the gap between slots 1 and 5 is perfect for leaning an 80mm fan at a 45degree angle against the card in slot 5, so that the slot1 card has sufficient airflow

ps: if you have another fan or case fan, aim it directly at the backside of card #4. it will cool it much better then the fan at the side is currently doing, since its airflow is making little/no contact with the backside of the board, which is the hottest region to cool

Get an IR thermometer and make sure the regulators aren't running too hot. I've had boards go dark like that, and usually a full poweroff is required to fix the situation.

Also, this might be related: https://bitcointalk.org/index.php?topic=250249.msg3099676;topicseen#msg3099676

Well I woke up all three of my new boards not hashing anymore... I think I want to vomit. Any suggestions to try to with my multi-meter?

I have been abusing the SD on one of my PIs, I mean turn off power without shutdown, etc etc and it has never died since the day I got the M board (Sep).

 The SD on the second PI on the other hand just dies every other day.

I think it has something to do with the SD. I have re-imaged a third SD and will check if SD is indeed the issue.

win32 disk imager?

  My sd FS dead again and I am using usb flash for FS + sd card for boot up.


ps: happened yesterday morning and I got it up less than 5 minutes.