The Chili – 30+GH/s BFL based Bitcoin Miner Assembly

[i3luefire]

I've got like 3 or 4 cards that all of the sudden are showing 250GH/100% hardware errors?

I have them off now and will let them cool completely down before turning them back on.. But reading on here that some people solved this with firmware reflash's.

Do these cards get in situations where they need to be re-flashed? Or can a reboot always get them "back"

I am currently running the 14e flash that has the 1.1v limit.

What were those cards achieving before they went to 250GH/100%HW ?

[Hands]

Low to mid 30's

I left them off for 2 hours then turned them back on and they are fine...

I guess for now the protocol is when the system goes south turn it off for a few hours then turn it back on.. .

[i3luefire]

the 1.1v limited FW is only for cards that max out the power supply's current limits and that only happens on the miners hitting close to 40GH. so you are probably limiting your voltage and hash rates for no reason if your Chili is only hitting mid 30's. the normal 14e FW should be able to get a higher hash rate out of boards in the mid 30's.

[ipxtreme]

How can I make mine go faster? They are about 64 degrees but only working at about 28gh each. Is there a switch I can use with these using either bfgminer (which I use now).

[i3luefire]

out of the 18 i have had here the only ones that have been below 32gh have always had bad contact for the cooler or no cooling on the backside of the board under the FETs. i have basically the whole backside of the board under the FETs covered with small heatsinks and a fan blowing on them. all my units run above 32gh. most of them are ~35gh

[ipxtreme]

out of the 18 i have had here the only ones that have been below 32gh have always had bad contact for the cooler or no cooling on the backside of the board under the FETs. i have basically the whole backside of the board under the FETs covered with small heatsinks and a fan blowing on them. all my units run above 32gh. most of them are ~35gh

thanks. I thought if the temp was under 70 it would hash faster until it hits 70 and maintain that speed. Since my temp is 64, it isnt running as fast as it can.

[Henchman24]

out of the 18 i have had here the only ones that have been below 32gh have always had bad contact for the cooler or no cooling on the backside of the board under the FETs. i have basically the whole backside of the board under the FETs covered with small heatsinks and a fan blowing on them. all my units run above 32gh. most of them are ~35gh

thanks. I thought if the temp was under 70 it would hash faster until it hits 70 and maintain that speed. Since my temp is 64, it isnt running as fast as it can.

Yeah, sounds like your asics are cool, but your FETs are hot.  Get some cooling on them and your hashrate and asic temp will both rise.

[ipxtreme]

out of the 18 i have had here the only ones that have been below 32gh have always had bad contact for the cooler or no cooling on the backside of the board under the FETs. i have basically the whole backside of the board under the FETs covered with small heatsinks and a fan blowing on them. all my units run above 32gh. most of them are ~35gh

thanks. I thought if the temp was under 70 it would hash faster until it hits 70 and maintain that speed. Since my temp is 64, it isnt running as fast as it can.

Yeah, sounds like your asics are cool, but your FETs are hot.  Get some cooling on them and your hashrate and asic temp will both rise.

ah. ok thanks.

[lightfoot]

Weird stuff. I added small heat sinks on the FETs and my performance went *down*. Tightened the water block a bit and temps spiked and performance *crashed*

Removed little sinks, loosened up block and snugged down gently and all normal. But man this thing is sensitive to fiddling; my guess is tightening caused the board to flex a hair enough to lift the sink off a chip. Result instant fail.

Back to 38gh. If I bolt a water block to the back of the board and run it with the one on the front I could probably get 40, but another thing I noticed is if you turn the block upside down the pump must cavitate or something, total fail.

Wow.

[twib2]

Mr Teal or anyone "in the know",

I start and stop my bfgminer quite a bit, but often one or the other of my Chilis doesn't come up again and I can't add it manually.  I have to do a hard reset or power on/off to get it to reboot and get seen again.  Restarting bfgminer or invoking it with --force-dev-init doesn't seem to do the trick.

My question is there any way I can force it to reset via a commandline so I don't have to wait until I get home from work or whatever?  I know the devices are on /dev/ttyUSB0 or /dev/ttyUSB1 and when they're working, I can see some output if I cat them directly, but I get nothing if it's hung.  Any ideas to force a reset remotely?  

Hmm, on a whim I just did:

Code:

echo "reset" >/dev/ttyUSB1

and now a few minutes later when I started bfgminer again, it CAME BACK!   It's never come back from a hangup before, so maybe that did it?  Or was this just a fluke?

Anyone?  Bueler?

edit: spoke too soon - it was recognized for a bit, but NOT hashing.  then trying to restart from within bfgminer just made it disappear again.  Argh.

[i3luefire]

Mr Teal or anyone "in the know",
I start and stop my bfgminer quite a bit, but often one or the other of my Chilis doesn't come up again and I can't add it manually.  I have to do a hard reset or power on/off to get it to reboot and get seen again.  Restarting bfgminer or invoking it with --force-dev-init doesn't seem to do the trick.
My question is there any way I can force it to reset via a commandline so I don't have to wait until I get home from work or whatever?  I know the devices are on /dev/ttyUSB0 or /dev/ttyUSB1 and when they're working, I can see some output if I cat them directly, but I get nothing if it's hung.  Any ideas to force a reset remotely?  
Hmm, on a whim I just did:

Code:

echo "reset" >/dev/ttyUSB1

and now a few minutes later when I started bfgminer again, it CAME BACK!   It's never come back from a hangup before, so maybe that did it?  Or was this just a fluke?
Anyone?  Bueler?
edit: spoke too soon - it was recognized for a bit, but NOT hashing.  then trying to restart from within bfgminer just made it disappear again.  Argh.

I have also been wanting a solution to this very same problem although I have been restarting bfgminer less often lately. the only reason I am restarting now is if there is a a power faliure or something else like that. I have also wondered if there was a way of making the miners come up on the same /dev/xxxx every time they or the computer is rebooted.
+1

[twib2]

Mr Teal or anyone "in the know",
I start and stop my bfgminer quite a bit, but often one or the other of my Chilis doesn't come up again and I can't add it manually.  I have to do a hard reset or power on/off to get it to reboot and get seen again.  Restarting bfgminer or invoking it with --force-dev-init doesn't seem to do the trick.
My question is there any way I can force it to reset via a commandline so I don't have to wait until I get home from work or whatever?  I know the devices are on /dev/ttyUSB0 or /dev/ttyUSB1 and when they're working, I can see some output if I cat them directly, but I get nothing if it's hung.  Any ideas to force a reset remotely?  
Hmm, on a whim I just did:

Code:

echo "reset" >/dev/ttyUSB1

and now a few minutes later when I started bfgminer again, it CAME BACK!   It's never come back from a hangup before, so maybe that did it?  Or was this just a fluke?
Anyone?  Bueler?
edit: spoke too soon - it was recognized for a bit, but NOT hashing.  then trying to restart from within bfgminer just made it disappear again.  Argh.

I have also been wanting a solution to this very same problem although I have been restarting bfgminer less often lately. the only reason I am restarting now is if there is a a power faliure or something else like that. I have also wondered if there was a way of making the miners come up on the same /dev/xxxx every time they or the computer is rebooted.
+1

Yep, I have some scripts I wrote that automatically shift mining to the most profitable coin every 15mins.  I have so many coins to choose from it's not handy to use the various pools and switch that way, so I just invoke bfgminer in different ways via scripts.

As for the same /dev/xxxx, I hava short line that just picks out the ones that have certain devices on them.  For example, to pick out the antminer ones:

Code:

usb="1:"$(./antminer -n | grep -e ea60 | sort -k 8 | cut -d ' ' -f 7 | tr '\n' ',1:' | sed 's/,$/\n/g' | sed 's/,/,1:/g')

puts the the devs that antminer (cgminer modded for antminer U1s) recognizes that are indeed U1s in a format that the --usb switch can recognize.  I set that up when I invoke antminer and that way whatever /dev/xxxx they are in I don't have to worry about and since I automatically tell it the right ones.  Fun stuff!

[i3luefire]

where did you get your U1?

[Mudbankkeith]

where did you get your U1?

https://bitcointalk.org/index.php?topic=382515.msg4110282#msg4110282
https://bitcointalk.org/index.php?topic=395506.msg4267487#msg4267487

[twib2]

i may just try and replace the ftdi chip

Before doing that, try shorting the center two pins on the unpopulated 0.1" connector labeled PROG. Pin 5 is MCU reset and pin 6 is ground, so it will hold the microcontroller in reset. If it's an MCU issue holding it in reset should cause the device to pop up on your computer as the FTDI chip will be able to connect. If it doesn't show up, it likely is the FTDI.

replacing the FTDI chip fixed it!!!!!!!!
 

You rock. I've replaced a few of those on the BFL devices; bitch in heat.
How did you do it on the Chili? And what is with those chips?
C

It's the first one on a Chili I've heard of failing, but they're actually not that hard to remove. Just heat it with air and pull it off.

I used an X-acto blade to cut the legs off then cleaned the pads off with solder wick and replaced the chip. no hot air gun available.

I think my shorting of pins 5 and 6 finally screwed up and hit an adjacent pin.  The Chili seems to power up fine (startup lights etc all look normal), but now it doesn't show up on the USB bus no matter what I do.  My other Chili is fine.  I haven't had a chance to look at it since I just had to reset it just before heading to work, but I expect I'll find a melted FTDI since the bootup process doesn't give any error lights.

So, in anticipation of that, what is the exact part number of the FTDI chip that I should order, and possibly from where?  I do have a hot air station, so that part will be easy, but what temperature should I set it at so as not to melt the chip body?

Thanks in advance!

[lightfoot]

So, in anticipation of that, what is the exact part number of the FTDI chip that I should order, and possibly from where?  I do have a hot air station, so that part will be easy, but what temperature should I set it at so as not to melt the chip body?

Thanks in advance!

For chips with legs, I'd go with 400c for 30 seconds tippy tops. See how long it takes to get them off, then use a bit less than that to get the new one on.

You don't need anywhere near as long heat for the chips with legs.

[twib2]

So, in anticipation of that, what is the exact part number of the FTDI chip that I should order, and possibly from where?  I do have a hot air station, so that part will be easy, but what temperature should I set it at so as not to melt the chip body?

Thanks in advance!

For chips with legs, I'd go with 400c for 30 seconds tippy tops. See how long it takes to get them off, then use a bit less than that to get the new one on.

You don't need anywhere near as long heat for the chips with legs.

Thanks much!   Part number?

[i3luefire]

So, in anticipation of that, what is the exact part number of the FTDI chip that I should order, and possibly from where?  I do have a hot air station, so that part will be easy, but what temperature should I set it at so as not to melt the chip body?

Thanks in advance!

For chips with legs, I'd go with 400c for 30 seconds tippy tops. See how long it takes to get them off, then use a bit less than that to get the new one on.

You don't need anywhere near as long heat for the chips with legs.

Thanks much!   Part number?

http://www.mouser.com/Search/ProductDetail.aspx?R=0virtualkey0virtualkey895-FT232HL-REEL

[twib2]

So, in anticipation of that, what is the exact part number of the FTDI chip that I should order, and possibly from where?  I do have a hot air station, so that part will be easy, but what temperature should I set it at so as not to melt the chip body?

Thanks in advance!

For chips with legs, I'd go with 400c for 30 seconds tippy tops. See how long it takes to get them off, then use a bit less than that to get the new one on.

You don't need anywhere near as long heat for the chips with legs.

Thanks much!   Part number?

http://www.mouser.com/Search/ProductDetail.aspx?R=0virtualkey0virtualkey895-FT232HL-REEL

Perfect - thanks!  I actually just ordered that just in case, but good to know I was on the right track! 

[Cascaders28]

Question -

What is the max amperage provided by the 3 pin fan power ports on these? I've got a pretty hefty fan I'd like to point along the edge to keep the FETs stuff cool. It is however a 2A fan and I don't want to overdraw anything. Thanks!