gingernuts
Member
![*](https://bitcointalk.org/Themes/custom1/images/star.gif)
Offline
Activity: 89
Merit: 10
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
October 07, 2013, 07:43:22 PM |
|
Hi Guys, Nice to be here again, Can anybody tell me which cgminer to use with single and/or chained bitfury chips on raspberrypi gpio ports ? My boards are finished everything works fine but when i solder one chip on board and start chainminer it detects 256 chips and nothing happens. cgminer 3.5.0 dont detect chip at all ... Here are some pics of finished 2 layer boards ( for testing only ) final will be 4 layer ... ![](https://ip.bitcointalk.org/?u=http%3A%2F%2Fimgur.com%2Fa%2FoC6lr%230&t=665&c=3NlSZ_BuK_0XHA) Cheers, Any reason you're going to swap to 4-layers once you've proved that your 2-layer prototypes work? I'm looking at about a 70% price hike to go to 4-layer rather than 2 for my boards...
|
|
|
|
DrZeck
Member
![*](https://bitcointalk.org/Themes/custom1/images/star.gif)
Offline
Activity: 66
Merit: 10
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
October 07, 2013, 09:25:57 PM |
|
Hi Guys, Nice to be here again, Can anybody tell me which cgminer to use with single and/or chained bitfury chips on raspberrypi gpio ports ? My boards are finished everything works fine but when i solder one chip on board and start chainminer it detects 256 chips and nothing happens. cgminer 3.5.0 dont detect chip at all ... Here are some pics of finished 2 layer boards ( for testing only ) final will be 4 layer ... ![](https://ip.bitcointalk.org/?u=http%3A%2F%2Fimgur.com%2Fa%2FoC6lr%230&t=665&c=3NlSZ_BuK_0XHA) Cheers, Any reason you're going to swap to 4-layers once you've proved that your 2-layer prototypes work? I'm looking at about a 70% price hike to go to 4-layer rather than 2 for my boards... I am building custom heatsink too, so i want to avoid any traces on bottom layer, even vias will end at mid layer 2. that is the only reason i will jump to 4 layer board. and my prototype not working yet, i am not sure why, chainminer detect chip (i think) ![Cheesy](https://bitcointalk.org/Smileys/default/cheesy.gif) , when i start chainminer with 1 chip on module i get message 256 chips detected but not hashing or anything, when i start it without module i get 0 chips detected ... And at the moment having troubles with legkodymov cgminer. i just cant compile it. i am using same method as for ckolivas cgminer but with no success Ckolivas cgminer is working with just: clone git go to cgminer folder ./autogen.sh --enable-bitfury make ![Undecided](https://bitcointalk.org/Smileys/default/undecided.gif)
|
|
|
|
vs3
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
October 07, 2013, 10:45:19 PM |
|
Does anyone have a multi-chip board working stably on a 2-layer PCB?
I'm just about done on my 8 chip PCB design, but I'm wondering whether I really need to add a couple more layers for VDD & GND. At the moment I have >90% planes on the top and back sides, so I'm tempted to just try 2-layer, but the cost and turn around time of getting it wrong and having to respin are making me tempted to chicken out and go for the more expensive 4-layer option straight away!
Ginger - I got the NanoFury on a single layer ![Smiley](https://bitcointalk.org/Smileys/default/smiley.gif) The entire back side is just GND. I am dealing with one chip though. With some manual routing I think I could fit a bunch of chips the same way and keep the back clean from any tracks. Maybe that will cost a few 0 ohm resistors to do act as a "bridge", but that's very likely going to be cheaper than a 4-layer board. Thanks - well, that's one vote for 2-layer ![Smiley](https://bitcointalk.org/Smileys/default/smiley.gif) Here is how my board looks like: ![](https://ip.bitcointalk.org/?u=https%3A%2F%2Fi.imgur.com%2FzLk7aYyl.png&t=665&c=GnFrPeNQhaYBfA) (ignore the 4 pins that appear to stick out - they're not there. I couldn't find a proper 3D model for a surface-mount connector and used the through-hole one)
|
|
|
|
gingernuts
Member
![*](https://bitcointalk.org/Themes/custom1/images/star.gif)
Offline
Activity: 89
Merit: 10
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
October 08, 2013, 07:32:41 AM |
|
Any reason you're going to swap to 4-layers once you've proved that your 2-layer prototypes work? I'm looking at about a 70% price hike to go to 4-layer rather than 2 for my boards...
I am building custom heatsink too, so i want to avoid any traces on bottom layer, even vias will end at mid layer 2. that is the only reason i will jump to 4 layer board. Hmm, so you're going for 4-layers with blind vias? Who's fabbing your boards, and what's the cost premium on that? Wouldn't it be simpler/cheaper to just do all over backside solder mask, including the vias? ![Huh](https://bitcointalk.org/Smileys/default/huh.gif) Good luck with debugging your prototype - have you got the rest of the chain where your other chips would usually be tied off properly - I think you probably need to tie off OUTMOSI (or OUTMISO - whichever is the INPUT) so that you don't get crap injected into your SPI chain.
|
|
|
|
gingernuts
Member
![*](https://bitcointalk.org/Themes/custom1/images/star.gif)
Offline
Activity: 89
Merit: 10
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
October 08, 2013, 07:33:57 AM |
|
Does anyone have a multi-chip board working stably on a 2-layer PCB?
I'm just about done on my 8 chip PCB design, but I'm wondering whether I really need to add a couple more layers for VDD & GND. At the moment I have >90% planes on the top and back sides, so I'm tempted to just try 2-layer, but the cost and turn around time of getting it wrong and having to respin are making me tempted to chicken out and go for the more expensive 4-layer option straight away!
Ginger - I got the NanoFury on a single layer ![Smiley](https://bitcointalk.org/Smileys/default/smiley.gif) The entire back side is just GND. I am dealing with one chip though. With some manual routing I think I could fit a bunch of chips the same way and keep the back clean from any tracks. Maybe that will cost a few 0 ohm resistors to do act as a "bridge", but that's very likely going to be cheaper than a 4-layer board. Thanks - well, that's one vote for 2-layer ![Smiley](https://bitcointalk.org/Smileys/default/smiley.gif) Here is how my board looks like: ![](https://ip.bitcointalk.org/?u=https%3A%2F%2Fi.imgur.com%2FzLk7aYyl.png&t=665&c=GnFrPeNQhaYBfA) (ignore the 4 pins that appear to stick out - they're not there. I couldn't find a proper 3D model for a surface-mount connector and used the through-hole one) Very nice and clean - these really are simple chips to work with - thanks Bitfury! ![Smiley](https://bitcointalk.org/Smileys/default/smiley.gif)
|
|
|
|
vs3
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
October 08, 2013, 09:00:02 AM |
|
Does anyone have a multi-chip board working stably on a 2-layer PCB?
I'm just about done on my 8 chip PCB design, but I'm wondering whether I really need to add a couple more layers for VDD & GND. At the moment I have >90% planes on the top and back sides, so I'm tempted to just try 2-layer, but the cost and turn around time of getting it wrong and having to respin are making me tempted to chicken out and go for the more expensive 4-layer option straight away!
Ginger - I got the NanoFury on a single layer ![Smiley](https://bitcointalk.org/Smileys/default/smiley.gif) The entire back side is just GND. I am dealing with one chip though. With some manual routing I think I could fit a bunch of chips the same way and keep the back clean from any tracks. Maybe that will cost a few 0 ohm resistors to do act as a "bridge", but that's very likely going to be cheaper than a 4-layer board. Thanks - well, that's one vote for 2-layer ![Smiley](https://bitcointalk.org/Smileys/default/smiley.gif) Here is how my board looks like: ![](https://ip.bitcointalk.org/?u=https%3A%2F%2Fi.imgur.com%2FzLk7aYyl.png&t=665&c=GnFrPeNQhaYBfA) (ignore the 4 pins that appear to stick out - they're not there. I couldn't find a proper 3D model for a surface-mount connector and used the through-hole one) Very nice and clean - these really are simple chips to work with - thanks Bitfury! ![Smiley](https://bitcointalk.org/Smileys/default/smiley.gif) Very nice chips indeed! And incredibly hard to kill too! Truly awesome design! And you could only imagine how many months of hard work we could've saved if there was also just any kind of documentation ... Oh well, I guess we can't have everything ![Smiley](https://bitcointalk.org/Smileys/default/smiley.gif)
|
|
|
|
DrZeck
Member
![*](https://bitcointalk.org/Themes/custom1/images/star.gif)
Offline
Activity: 66
Merit: 10
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
October 09, 2013, 01:08:06 AM |
|
Hey,
Can someone explain this:
pi@bitfury ~ $ sudo /opt/bitfury/start-miner-console.sh INIT: 256 chips detected chip 2 programmed JOB 1 PROCESSED 0.001 sec [nonces:0, errors:0, spi:1 miso:3] (queue:0) chip 3 programmed JOB 2 PROCESSED 0.001 sec [nonces:0, errors:0, spi:130 miso:21] (queue:0) chip 4 programmed JOB 3 PROCESSED 0.009 sec [nonces:0, errors:1359, spi:1 miso:16] (queue:0) chip 5 programmed JOB 4 PROCESSED 0.007 sec [nonces:0, errors:1337, spi:1 miso:39] (queue:0) chip 6 programmed JOB 5 PROCESSED 0.007 sec [nonces:0, errors:1309, spi:1 miso:50] (queue:0) chip 7 programmed JOB 6 PROCESSED 0.007 sec [nonces:0, errors:1291, spi:1 miso:55] (queue:0) chip 8 programmed JOB 7 PROCESSED 0.008 sec [nonces:0, errors:1434, spi:1 miso:47] (queue:0) chip 9 programmed JOB 8 PROCESSED 0.008 sec [nonces:0, errors:1464, spi:1 miso:48] (queue:0) chip 10 programmed JOB 9 PROCESSED 0.007 sec [nonces:0, errors:1319, spi:1 miso:58] (queue:0) chip 11 programmed JOB 10 PROCESSED 0.008 sec [nonces:0, errors:1424, spi:1 miso:49] (queue:0) chip 12 programmed JOB 11 PROCESSED 0.009 sec [nonces:0, errors:1319, spi:1 miso:58] (queue:0) chip 13 programmed JOB 12 PROCESSED 0.009 sec [nonces:0, errors:1521, spi:1 miso:51] (queue:0) chip 14 programmed JOB 13 PROCESSED 0.006 sec [nonces:0, errors:1221, spi:1 miso:50] (queue:0) chip 15 programmed JOB 14 PROCESSED 0.007 sec [nonces:0, errors:1345, spi:1 miso:47] (queue:0) chip 16 programmed JOB 15 PROCESSED 0.007 sec [nonces:0, errors:1427, spi:1 miso:44] (queue:0) chip 17 programmed JOB 16 PROCESSED 0.008 sec [nonces:0, errors:1378, spi:1 miso:47] (queue:0) chip 18 programmed JOB 17 PROCESSED 0.010 sec [nonces:0, errors:1409, spi:1 miso:54] (queue:0)
It's one chip mounted on board ... i don't get any hashes etc ...
There is no resistor on miso to gnd atm
core voltage 0.79v
iovdd 1.82v
...
|
|
|
|
gingernuts
Member
![*](https://bitcointalk.org/Themes/custom1/images/star.gif)
Offline
Activity: 89
Merit: 10
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
October 09, 2013, 08:32:27 AM |
|
Hey,
Can someone explain this:
pi@bitfury ~ $ sudo /opt/bitfury/start-miner-console.sh INIT: 256 chips detected chip 2 programmed JOB 1 PROCESSED 0.001 sec [nonces:0, errors:0, spi:1 miso:3] (queue:0) chip 3 programmed JOB 2 PROCESSED 0.001 sec [nonces:0, errors:0, spi:130 miso:21] (queue:0) chip 4 programmed JOB 3 PROCESSED 0.009 sec [nonces:0, errors:1359, spi:1 miso:16] (queue:0) chip 5 programmed JOB 4 PROCESSED 0.007 sec [nonces:0, errors:1337, spi:1 miso:39] (queue:0) chip 6 programmed JOB 5 PROCESSED 0.007 sec [nonces:0, errors:1309, spi:1 miso:50] (queue:0) chip 7 programmed JOB 6 PROCESSED 0.007 sec [nonces:0, errors:1291, spi:1 miso:55] (queue:0) chip 8 programmed JOB 7 PROCESSED 0.008 sec [nonces:0, errors:1434, spi:1 miso:47] (queue:0) chip 9 programmed JOB 8 PROCESSED 0.008 sec [nonces:0, errors:1464, spi:1 miso:48] (queue:0) chip 10 programmed JOB 9 PROCESSED 0.007 sec [nonces:0, errors:1319, spi:1 miso:58] (queue:0) chip 11 programmed JOB 10 PROCESSED 0.008 sec [nonces:0, errors:1424, spi:1 miso:49] (queue:0) chip 12 programmed JOB 11 PROCESSED 0.009 sec [nonces:0, errors:1319, spi:1 miso:58] (queue:0) chip 13 programmed JOB 12 PROCESSED 0.009 sec [nonces:0, errors:1521, spi:1 miso:51] (queue:0) chip 14 programmed JOB 13 PROCESSED 0.006 sec [nonces:0, errors:1221, spi:1 miso:50] (queue:0) chip 15 programmed JOB 14 PROCESSED 0.007 sec [nonces:0, errors:1345, spi:1 miso:47] (queue:0) chip 16 programmed JOB 15 PROCESSED 0.007 sec [nonces:0, errors:1427, spi:1 miso:44] (queue:0) chip 17 programmed JOB 16 PROCESSED 0.008 sec [nonces:0, errors:1378, spi:1 miso:47] (queue:0) chip 18 programmed JOB 17 PROCESSED 0.010 sec [nonces:0, errors:1409, spi:1 miso:54] (queue:0)
It's one chip mounted on board ... i don't get any hashes etc ...
There is no resistor on miso to gnd atm
core voltage 0.79v
iovdd 1.82v
...
I could be wrong, but as I understand it, the SPI chain is one big shift register. The S/W does some magic to see how many chips are there, presumably by trying to talk to the next chip along, and as you have no termination on the end of the chain, it will read some random garbage. It then seems to be happy that this garbage is a real chip, so configures your chain for multiple chips... I also believe that the S/W expects the SPI response to come back out of the 1st chip in the chain, which means that all the other chips are routing their SPI data all the way up the chain, back around in the last chip of the chain and all the way back to the first chip, so if the S/W doesn't correctly detect how many chips there are in the chain, you are toast! If I've got this all upside down, will someone please shoot that full of holes before I send my PCB's off tonight!
|
|
|
|
KNK
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
October 09, 2013, 11:12:14 AM |
|
It is recommended not to leave floating inputs, so you should add a resistor on MISO, but this is not your (only) problem.
For chainmainer there are some definitions in miner.h which you should set according to your board or it will cycle for MAXCHIPS which is 256 by default
You may also try Legkodymov's cgminer fork, but there are also some settings to configure (in butfury-config.h) before compiling
|
|
|
|
DrZeck
Member
![*](https://bitcointalk.org/Themes/custom1/images/star.gif)
Offline
Activity: 66
Merit: 10
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
October 09, 2013, 12:02:45 PM |
|
It is recommended not to leave floating inputs, so you should add a resistor on MISO, but this is not your (only) problem.
For chainmainer there are some definitions in miner.h which you should set according to your board or it will cycle for MAXCHIPS which is 256 by default
You may also try Legkodymov's cgminer fork, but there are also some settings to configure (in butfury-config.h) before compiling
Hi, i added resistor now, on my design i have 10 chip modules and 12 modules goes on 1 "M" board. so its a 120 chips total. But now i am trying to hash with 1-2 chips only. i am using V1 image for raspi, so if i put 16 chip on board ( one full board and second board with just 6 chips ) that should work with default settings in miner.h ? for cgminer do you know what need to be changed in bitfury-config.h ( software is my bad side ![Grin](https://bitcointalk.org/Smileys/default/grin.gif) ) Cheers,
|
|
|
|
KNK
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
October 09, 2013, 03:04:09 PM |
|
I haven't looked thoroughly the chainminer code neither i have looked at your board design, so may be wrong, but it seems chainminer uses GPIO pins to select banks (spi_reset in spidev.c). so if your design uses different pins you need to change them. The cgminer code uses either just 1 bank (with BITFURY_NEEDBMW_NOMUX defined) or I2C multiplexers (if BITFURY_METABANK is defined) to select banks. I am working on adding GPIO bank selection in my fork and when done the changes will be probably merged back to Legkodymov's cgminer, so you may use my fork or Legkodymov's with BITFURY_NEEDBMW_NOMUX defined for a single bank, single chip. From your output it seems like there is no bank selection made and your chip is detected multiple times, so either you don't have that on your board or it is not working properly i.e. your bank is always selected
|
|
|
|
gingernuts
Member
![*](https://bitcointalk.org/Themes/custom1/images/star.gif)
Offline
Activity: 89
Merit: 10
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
October 26, 2013, 05:33:00 PM |
|
Has anyone got spitest working on a beaglebone black?
|
|
|
|
ashg1990
Newbie
Offline
Activity: 30
Merit: 0
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
October 26, 2013, 11:52:59 PM |
|
I would Like to Apply to Be a Tester I have experience with C coding I have experience with raspberry pi I have soldering qualification j standard recognized worldwide + over 5 years experience soldering I have a good knowledge on hacking I have got all the supplies and tools to create a test jig and then develop my own boards I have access to industrial reflow machines and a huge amount of other tools used in the electrical industry I have sufficient knowledge to debug and produce a quality asic board I would love to have the opportunity to test a chip before I proceed to plunge money into bitfury asic and also release open source designs
|
|
|
|
KNK
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
October 27, 2013, 08:59:08 AM |
|
You are about 4 months late. There are no more free chip samples - you buy chips from the store now, in order to test your design.
|
|
|
|
vulgartrendkill
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
October 27, 2013, 09:04:10 AM |
|
hi guys, I`ve just got a bitfury based miner, and one ofthe chips only hashs at .5 to 1gh. the manufacturer says i can send back for a replacement, but is there anything I can do to increase the hash rate that isn`t too complicated just increase that one chips rate?
Thanks
|
|
|
|
KNK
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
October 27, 2013, 09:07:46 AM |
|
Some chips have bad hashing cores and can't perform any better than that no matter what you do.
|
|
|
|
vulgartrendkill
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
October 27, 2013, 09:17:46 AM |
|
Some chips have bad hashing cores and can't perform any better than that no matter what you do.
Thanks, I guess its RMA for me then...
|
|
|
|
gingernuts
Member
![*](https://bitcointalk.org/Themes/custom1/images/star.gif)
Offline
Activity: 89
Merit: 10
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
October 27, 2013, 10:47:22 AM |
|
Hi guys, I am trying to bring up my own board design, I have 0.8VDD, 1.8V IOVDD and 0.9V IOREF, but I just seem to get all '1's out of the SPI.
Do you have any troubleshooting suggestions?
Cheers.
|
|
|
|
vs3
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
October 27, 2013, 10:55:32 AM |
|
Hi guys, I am trying to bring up my own board design, I have 0.8VDD, 1.8V IOVDD and 0.9V IOREF, but I just seem to get all '1's out of the SPI.
Do you have any troubleshooting suggestions?
Cheers.
Start by sending the RESET sequence - you should see it also on the OUT pins (as it is being relayed by the chips).
|
|
|
|
gingernuts
Member
![*](https://bitcointalk.org/Themes/custom1/images/star.gif)
Offline
Activity: 89
Merit: 10
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
October 27, 2013, 11:02:13 AM |
|
Hi guys, I am trying to bring up my own board design, I have 0.8VDD, 1.8V IOVDD and 0.9V IOREF, but I just seem to get all '1's out of the SPI.
Do you have any troubleshooting suggestions?
Cheers.
Start by sending the RESET sequence - you should see it also on the OUT pins (as it is being relayed by the chips). Thanks - I'll give that a go ![Smiley](https://bitcointalk.org/Smileys/default/smiley.gif)
|
|
|
|
|