BFL board project COINTAMINATION - EU facility - ORDERS OPENED for Chili - page 12.

Lucko

hero member

Activity: 826

Merit: 1000

Quote from: asjfdlksfd on January 06, 2014, 09:03:32 AM

Another problem is, after rebooting the chillies will not further work mostly because all engines mostly are enabled at 0 MHz. bfgminer 3.9.0 will not find the board or cannot refind it. Only to disconnect power brings them back online.

Cheers...

Is only 7 blinking? That means that it is cooling... I have same problem. If I don't use a pipe to get cooled air to it it will not start. They are at 43C and you need to get them to about 38C(a guess) or so to start startup... I guess it is too hot...

Quote from: Mudbankkeith on January 06, 2014, 10:00:45 AM

Hi Lucko

Now that I have unpacked these boards,

The delivery has been rough on them.

9 boards have bent pins on the fan connectors, I am now checking each board for other damage.

With all that bubble raping? Well as long as it is only bent pins it is not a big deal...

EDIT: now that I think of... Was box damaged? I think it could even be me packing it. I might pressed bubble raping in to hard...

ronin4bits

newbie

Activity: 59

Merit: 0

Hi Lucko. Excellent work - and thanks to MrTeal for assisting.

Please proceed with making my boards, as well.

Thanks!

Mudbankkeith

hero member

Activity: 868

Merit: 1000

Hi Lucko

Now that I have unpacked these boards,

The delivery has been rough on them.

9 boards have bent pins on the fan connectors, I am now checking each board for other damage.

asjfdlksfd

full member

Activity: 128

Merit: 100

Hello,

yesterday I had find a solution of some of my miners by replacing the Enermax 1500 W Platinum (supports 6 rails a 30 A so I tried to use on three different rails one chilly/rail) by an older style Enermax 1350 W Revo power supply. This helps me so the chilly will not restart below 1.160 V.
I find this issue because my old style 700 W Enermax Noisetaker (2 rails with 18A) mostly find not a problem.

But in fact, I'm still not cooling my fets because if the voltage will increase to max short time later every chilly I'm cooling also the chilly will reboot. So my chillies will work ~30 GH/s which is not perfect but ok as long I find a suitable xp version which i can use to flash it. But still it would be welcome if anybody find a solution to flash with linux. I can use also my rasberry pi to program directly with jtag (in same way I do that with jally

In fact, Mr. Teal, could you add an command to limit the voltage manually e.g. after restarting?
Also can you add in firmware a way so we can disable each chip?
Also it would be fine can we have a command to recheck/reset every chip by hand?

Another problem is, after rebooting the chillies will not further work mostly because all engines mostly are enabled at 0 MHz. bfgminer 3.9.0 will not find the board or cannot refind it. Only to disconnect power brings them back online.

Cheers...

Mudbankkeith

hero member

Activity: 868

Merit: 1000

Boards are here

https://www.dropbox.com/s/58hnjh3wnijcr2m/2014-01-06%2012.48.40.jpg

Heatsinks are here

https://www.dropbox.com/s/noj49yxyw3k5cbv/2014-01-06%2012.48.52.jpg

Even the nuts and bolts

https://www.dropbox.com/s/0zasjaemgckqkul/2014-01-06%2012.49.25.jpg

Now it gets serious

bcollan

member

Activity: 69

Merit: 10

Please make mine too Lucko.
I'm not getting any younger... Grin

bx8389

member

Activity: 80

Merit: 10

Quote from: MrTeal on January 05, 2014, 04:38:45 PM

Quote from: asjfdlksfd on January 05, 2014, 09:53:20 AM

Quote from: Lucko on January 04, 2014, 12:39:32 PM

Yes that is same thing... It can go over 20A per chips at this speed... And power supply is only 160A with cooling...

I guess your boards have better copper plates or so. All first 4 boards I've assembled now have the same issue. It is not relevant how many engines could be started. My board with 126 engines have the same problem as my board with only 109 engines. It looks like not only the problem comes from higher voltages > 1.1. One board with 109 engines allready resets at 1.0V. I can start it only by stopping the cpu fan so the blf will arrive the 70° limit earlier. I will check later.

All these are now limit to the power fets. Maybe there can be replaced by ones with 60 A support?

Cheers...

Restarting at @1V does not have much to do with the number of engines. I only have one board here so it is difficult to diagnose whether all the boards with those issues are due to the same root cause, but the one I have is due to an actual hardware error. Lucko is looking into whether those other ones exhibit the same symptoms.
As for replacing the FETs, the ones rated to 50A available in stock now, so if you wanted to replace them on some boards you could do that. You'd also need to make changes to the controller itself though as it is what limits the current to 160A. The fets would definitely need more cooling though; the 96370 @ 40A will dissipate ~7W each if you keep the junction temperature to 75C. The 96371 @ 50A will dissipate ~9.5W each at the same conditions, and keeping the junction temperature down will be harder.

100A isn't the recommended for the long run, it's just the current at the rated Ploss spec. The fets will run fine at the 40A rating continuously as long as they're cooled, but efficiency is lower than if they were run at 25A per phase.

I don't believe it is likely a component issue; we have made almost 1000 of these and none have experienced (that I know of) this particular problem. I also replaced the affected phases with parts out of my stock and it did not solve the issue. I will try to find some time this week to feed the control and current feedback signal on the bad phase outside the board to see that can solve the problem.

Many thanks for your attention to this problem and explanations MrTeal.

MrTeal

legendary

Activity: 1274

Merit: 1004

Quote from: asjfdlksfd on January 05, 2014, 09:53:20 AM

Quote from: Lucko on January 04, 2014, 12:39:32 PM

Yes that is same thing... It can go over 20A per chips at this speed... And power supply is only 160A with cooling...

I guess your boards have better copper plates or so. All first 4 boards I've assembled now have the same issue. It is not relevant how many engines could be started. My board with 126 engines have the same problem as my board with only 109 engines. It looks like not only the problem comes from higher voltages > 1.1. One board with 109 engines allready resets at 1.0V. I can start it only by stopping the cpu fan so the blf will arrive the 70° limit earlier. I will check later.

All these are now limit to the power fets. Maybe there can be replaced by ones with 60 A support?

Cheers...

Restarting at @1V does not have much to do with the number of engines. I only have one board here so it is difficult to diagnose whether all the boards with those issues are due to the same root cause, but the one I have is due to an actual hardware error. Lucko is looking into whether those other ones exhibit the same symptoms.
As for replacing the FETs, the ones rated to 50A available in stock now, so if you wanted to replace them on some boards you could do that. You'd also need to make changes to the controller itself though as it is what limits the current to 160A. The fets would definitely need more cooling though; the 96370 @ 40A will dissipate ~7W each if you keep the junction temperature to 75C. The 96371 @ 50A will dissipate ~9.5W each at the same conditions, and keeping the junction temperature down will be harder.

Quote from: bx8389 on January 05, 2014, 03:32:47 PM

You can check on TI site Fet Specifications http://www.ti.com/product/csd96370q5m
They have also a version w/ silighty more output. Chilli have 4 so 40A x 4 = 160 A max, 100 A recomended in long run

For me, don't seem a problem of intensity output, this is rather a symptom. I haven't checked yet ( have a daytime work ) but I belive more on things like saturation on inductors or capacitor rds due the random nature with same components. For instance a saturated inductor on this kind of power supply acts like a short so this can reset a controler.

But your sugestion to have some firmware configuration is good: these units can hash at 23-25 Mhs so if can program firmware to stop at 2 VRM stage and don't got 3th stage VRM at least we can have at least some return during problem investigation time. As is, on this moment, these boards are useless

Regards

100A isn't the recommended for the long run, it's just the current at the rated Ploss spec. The fets will run fine at the 40A rating continuously as long as they're cooled, but efficiency is lower than if they were run at 25A per phase.

I don't believe it is likely a component issue; we have made almost 1000 of these and none have experienced (that I know of) this particular problem. I also replaced the affected phases with parts out of my stock and it did not solve the issue. I will try to find some time this week to feed the control and current feedback signal on the bad phase outside the board to see that can solve the problem.

bx8389

member

Activity: 80

Merit: 10

Quote from: asjfdlksfd on January 05, 2014, 01:46:25 PM

Quote from: Lucko on January 05, 2014, 12:07:26 PM

I tested every board with 2 restarts. But I did figure out that they act batter if they have there own unshare power supply... If it manage to get to speed(when speeds start drooping) I declare it without problems. Also if it crashed above 38GH. If it chresh but was able to get it going with restarts it was also send out. So that is why your 109 is restarting because of unknown problem that MrTeal and I are looking. You did opted to have them made anyway... Only if I could not make it run I didn't send it out(OK I did one by mistake)...

And I'm not sure what you mean by 60A support...

Replace all 40A psr chips by 60A types? I've not checked that they exists. I guess it's a problem with limit of the psr.

It would be fine we can disable chips manually by Z-Commands before starting bfgminer.
And it would be fine we can limit manually by Z-Commands the voltage and also maybe the voltage-ramp.

I think the fw is not open source to extend them by myself? This is a question which can only answered by mr. teal.

Cheers....

You can check on TI site Fet Specifications http://www.ti.com/product/csd96370q5m
They have also a version w/ silighty more output. Chilli have 4 so 40A x 4 = 160 A max, 100 A recomended in long run

For me, don't seem a problem of intensity output, this is rather a symptom. I haven't checked yet ( have a daytime work ) but I belive more on things like saturation on inductors or capacitor rds due the random nature with same components. For instance a saturated inductor on this kind of power supply acts like a short so this can reset a controler.

But your sugestion to have some firmware configuration is good: these units can hash at 23-25 Mhs so if can program firmware to stop at 2 VRM stage and don't got 3th stage VRM at least we can have at least some return during problem investigation time. As is, on this moment, these boards are useless

Regards

asjfdlksfd

full member

Activity: 128

Merit: 100

Quote from: Lucko on January 05, 2014, 12:07:26 PM

I tested every board with 2 restarts. But I did figure out that they act batter if they have there own unshare power supply... If it manage to get to speed(when speeds start drooping) I declare it without problems. Also if it crashed above 38GH. If it chresh but was able to get it going with restarts it was also send out. So that is why your 109 is restarting because of unknown problem that MrTeal and I are looking. You did opted to have them made anyway... Only if I could not make it run I didn't send it out(OK I did one by mistake)...

And I'm not sure what you mean by 60A support...

Replace all 40A psr chips by 60A types? I've not checked that they exists. I guess it's a problem with limit of the psr.

It would be fine we can disable chips manually by Z-Commands before starting bfgminer.
And it would be fine we can limit manually by Z-Commands the voltage and also maybe the voltage-ramp.

I think the fw is not open source to extend them by myself? This is a question which can only answered by mr. teal.

Cheers....

Lucko

hero member

Activity: 826

Merit: 1000

Quote from: asjfdlksfd on January 05, 2014, 09:53:20 AM

Quote from: Lucko on January 04, 2014, 12:39:32 PM

Yes that is same thing... It can go over 20A per chips at this speed... And power supply is only 160A with cooling...

I guess your boards have better copper plates or so. All first 4 boards I've assembled now have the same issue. It is not relevant how many engines could be started. My board with 126 engines have the same problem as my board with only 109 engines. It looks like not only the problem comes from higher voltages > 1.1. One board with 109 engines allready resets at 1.0V. I can start it only by stopping the cpu fan so the blf will arrive the 70° limit earlier. I will check later.

All these are now limit to the power fets. Maybe there can be replaced by ones with 60 A support?

Cheers...

I tested every board with 2 restarts. But I did figure out that they act batter if they have there own unshare power supply... If it manage to get to speed(when speeds start drooping) I declare it without problems. Also if it crashed above 38GH. If it chresh but was able to get it going with restarts it was also send out. So that is why your 109 is restarting because of unknown problem that MrTeal and I are looking. You did opted to have them made anyway... Only if I could not make it run I didn't send it out(OK I did one by mistake)...

And I'm not sure what you mean by 60A support...

GrapeApe

sr. member

Activity: 476

Merit: 250

Quote from: asjfdlksfd on January 05, 2014, 09:19:59 AM

Does anybody knows how I can flash with linux without any windows installing? I use kvm but in real usb is not working from it.

Cheers..

I just went through this myself. You need Virtual Box (VB) It has settings for usb support but you have to add your user to the vb group for the usb to work. Another member helped me a little bit. here.

Quote from: twib2 on January 02, 2014, 07:11:44 PM

No worries, you just go down to the bottom right of the Virtual Box in which XP is running and click on the USB icon. Then choose the one that corresponds to the Chili. I'm at work, so I don't remember which one that is, but you can lsusb on the linux side or just unplug and see where the difference is in that popup menu choice. You have to have set up the VirtualBox for XP to allow USB passthrough, but that's an easy configuration choice. Depending on your port, you might also have to have installed the extension pack to allow USB 3.0 if it's one of those.
.

asjfdlksfd

full member

Activity: 128

Merit: 100

Quote from: Lucko on January 04, 2014, 12:39:32 PM

Yes that is same thing... It can go over 20A per chips at this speed... And power supply is only 160A with cooling...

I guess your boards have better copper plates or so. All first 4 boards I've assembled now have the same issue. It is not relevant how many engines could be started. My board with 126 engines have the same problem as my board with only 109 engines. It looks like not only the problem comes from higher voltages > 1.1. One board with 109 engines allready resets at 1.0V. I can start it only by stopping the cpu fan so the blf will arrive the 70° limit earlier. I will check later.

All these are now limit to the power fets. Maybe there can be replaced by ones with 60 A support?

Cheers...

asjfdlksfd

full member

Activity: 128

Merit: 100

Quote from: GrapeApe on January 04, 2014, 11:21:53 AM

The link for the modified firmware is up now. He has a link to it on the first page of his thread with the other firmware links.

https://www.dropbox.com/s/rfa4kuigov99sg8/Chili14e1v1.hex
works again it was down

Does anybody knows how I can flash with linux without any windows installing? I use kvm but in real usb is not working from it.

Cheers..

Lucko

hero member

Activity: 826

Merit: 1000

Quote from: asjfdlksfd on January 04, 2014, 06:06:10 AM

Quote from: bx8389 on January 03, 2014, 07:48:00 PM

Also received today mines, thanks Lucko,
At this moment i try to put first one at work but resets at 27 GHs and 32 ºC. Bfgminner have restarted board many times but until now without results. Tomorrow i will install remaining boards and see what are results.

Ok, I have the same issue here now with the first 3 boards. The others are not completed, yet because of missing coolers, which should arrives over the day.

The messages looks like the same as is described on Mr. Teals forum a a result of high performers.
With 119, 121 they are not really high performers, the third with 126 looks like more that.

Code:

[2014-01-04 09:52:11] Devices detected:
[2014-01-04 09:52:11] BitFORCE SHA256 SC by MrTeal and ChipGeek (driver=bitforce_queue; procs=1; serial=FTX30WIZ; path=/dev/ttyUSB0)
1 devices listed
l))
[2014-01-04 09:52:11] Acquired exclusive advisory lock on /dev/ttyUSB0
[2014-01-04 09:52:11] BFL: Attempting to open /dev/ttyUSB0
[2014-01-04 09:52:11] Found BitForce device on /dev/ttyUSB0
[2014-01-04 09:52:11] DEVICE: Chili SC
[2014-01-04 09:52:11] MANUFACTURER: MrTeal and ChipGeek
[2014-01-04 09:52:11] FIRMWARE: 1.2.14e
[2014-01-04 09:52:11] CHIP PARALLELIZATION: NO
[2014-01-04 09:52:11] QUEUE DEPTH:40
[2014-01-04 09:52:11] PROCESSOR 0: 16 engines @ 181 MHz -- MAP: FFFF
[2014-01-04 09:52:11] PROCESSOR 1: 16 engines @ 181 MHz -- MAP: FFFF
[2014-01-04 09:52:11] PROCESSOR 2: 16 engines @ 180 MHz -- MAP: FFFF
[2014-01-04 09:52:11] PROCESSOR 3: 16 engines @ 178 MHz -- MAP: FFFF
[2014-01-04 09:52:11] PROCESSOR 4: 16 engines @ 200 MHz -- MAP: FFFF
[2014-01-04 09:52:11] PROCESSOR 5: 16 engines @ 189 MHz -- MAP: FFFF
[2014-01-04 09:52:11] PROCESSOR 6: 16 engines @ 186 MHz -- MAP: FFFF
[2014-01-04 09:52:11] PROCESSOR 7: 14 engines @ 182 MHz -- MAP: DFF7
[2014-01-04 09:52:11] THEORETICAL MAX: 23.28 GH/s
[2014-01-04 09:52:11] ENGINES: 126
[2014-01-04 09:52:11] FREQUENCY: 184 MHz
[2014-01-04 09:52:11] CRITICAL TEMPERATURE: 0
[2014-01-04 09:52:11] TOTAL THERMAL CYCLES: 0
[2014-01-04 09:52:11] XLINK MODE: MASTER
[2014-01-04 09:52:11] XLINK PRESENT: NO
[2014-01-04 09:52:11] Devices detected:
[2014-01-04 09:52:11] BitFORCE SHA256 SC by MrTeal and ChipGeek (driver=bitforce_queue; procs=1; serial=FTX30WIZ; path=/dev/ttyUSB0)

The miners have near by the same problem. They start and bfgminer test shows more or less:

Code:

bfgminer version 3.9.0 - Started: [2014-01-04 09:53:08] - [ 0 days 00:02:14]
[M]anage devices [P]ool management [S]ettings [D]isplay options [H]elp [Q]uit
Connected to removed.btcguild.com diff 8 with stratum as user removed
Block: ...6cb9f38a #278547 Diff:1.42G (10.15Ph/s) Started: [09:53:08]
ST:7 F:0 NB:1 AS:1 BW:[163/192 B/s] E:36.16 I: 399uBTC/hr BS:2.92k
1 62.0C | 30.36/33.02/30.15Gh/s | A:163 R:2+1(.97%) HW:148/ 13%
--------------------------------------------------------------------------------
BFL 0 : 62.0C | 40.75/32.03/29.25Gh/s | A:163 R:2+1(.97%) HW:156/ 13%
--------------------------------------------------------------------------------

Select processor to manage using up/down arrow keys
BFL 0 : 65.0C | 40.75/34.93/31.72Gh/s | A:161 R:2+1(.99%) HW:134/ 13%
BitFORCE SHA256 SC from MrTeal and ChipGeek
Serial: FTX30WIZ
Kernel: bulk queue
Temperatures: 43.0C 65.0C
Voltages: 3.290 / 1.157 / 11.781

[D]isable [I]dentify [F]an control
[Slash] Find processor [Plus] Add device(s) [Enter] Close device manager
[2014-01-04 09:55:12] BFL 0: Received unexpected queue result response:
[2014-01-04 09:55:13] BFL 0: Received unexpected queue result response:
[2014-01-04 09:55:14] BFL 0: Received unexpected queue result response:
[2014-01-04 09:55:14] BFL 0: Error: Get temp returned empty string/timed out
[2014-01-04 09:55:14] BFL 0: Received unexpected queue result response:

Here the LED 2 an 4 are on, 7 is blinking ~4Hz. After a while 7and 8 are blinking where 7 but now 2Hz by both going on same time but first 7 goes off (1:1) then 8 (1:2-3).:

7 011001100110011001100...
8 011101110111011101110...

After a while more left 6 are blinking randomly for a second and then all are going off. Now it looks like the board has to been reset.

Then I can restart bfgminer. It looks like really the problems comes from the power rails. I didn't cool the FETS active by a fan any more at the moment so the fet temp rises 45 °C which limits the vcore <1.1 V. So flashing to 1.1V limit could be solve the problem.

Code:

Block: ...6cb9f38a #278547 Diff:1.42G (10.15Ph/s) Started: [10:02:38]
ST:7 F:0 NB:1 AS:0 BW:[101/ 81 B/s] E:95.64 I: 541uBTC/hr BS:4.75k
1 59.0C | 36.78/36.78/34.57Gh/s | A:266 R:5+1(.68%) HW:389/8.0%
--------------------------------------------------------------------------------
BFL 0 : 59.0C | 37.13/36.78/34.57Gh/s | A:266 R:5+1(.68%) HW:389/8.0%
--------------------------------------------------------------------------------

Select processor to manage using up/down arrow keys
BFL 0 : 59.0C | 37.12/36.78/34.35Gh/s | A:260 R:5+1(.74%) HW:368/7.9%
BitFORCE SHA256 SC from MrTeal and ChipGeek
Serial: FTX30WIZ
Kernel: bulk queue
Temperatures: 45.0C 59.0C
Voltages: 3.290 / 1.097 / 11.906

At the moment so I guess do not blow the power fets. This will limit also the voltage and is a work around. Also the other one could I fix by removing the fan to blow the fets, too :-/

Cheers...

You have too good chips. You need limiter firmware... I would put it on but it was not available at the time. You can upload that with https://www.dropbox.com/s/xpccbhfkbpinov8/ChiliFlash.exe https://www.dropbox.com/s/rfa4kuigov99sg8/Chili14e1v1.hex

Blinking 7 light means that board is cooling off and waiting to start selftest. Blinking 7 and 8 means selftest...

Quote from: bx8389 on January 04, 2014, 11:52:36 AM

Quote from: GrapeApe on January 03, 2014, 11:45:24 PM

Quote from: Swimmer63 on January 03, 2014, 11:14:25 PM

Quote from: Bogart on January 03, 2014, 09:01:19 PM

Quote from: Lucko on January 03, 2014, 03:55:07 PM

if anyone else would like me to make his boards let me know...

Go ahead and make mine please.

Thanks.

Go ahead and make mine Lucko.

Go ahead and make mine as well. I'm curious if you ever tried the modified firmware that limits the voltage to 1.1v, if so what were your results?

One of my received board acheived 39.2 during 1 minute then stop responding: I flashed 1.1V firmware and now got 36GHs stable board

Regards

Yes that is same thing... It can go over 20A per chips at this speed... And power supply is only 160A with cooling...

bx8389

member

Activity: 80

Merit: 10

Quote from: GrapeApe on January 03, 2014, 11:45:24 PM

Quote from: Swimmer63 on January 03, 2014, 11:14:25 PM

Quote from: Bogart on January 03, 2014, 09:01:19 PM

Quote from: Lucko on January 03, 2014, 03:55:07 PM

if anyone else would like me to make his boards let me know...

Go ahead and make mine please.

Thanks.

Go ahead and make mine Lucko.

Go ahead and make mine as well. I'm curious if you ever tried the modified firmware that limits the voltage to 1.1v, if so what were your results?

One of my received board acheived 39.2 during 1 minute then stop responding: I flashed 1.1V firmware and now got 36GHs stable board

Regards

bx8389

member

Activity: 80

Merit: 10

Quote from: Lucko on January 04, 2014, 04:30:34 AM

Quote from: bx8389 on January 03, 2014, 07:48:00 PM

Also received today mines, thanks Lucko,
At this moment i try to put first one at work but resets at 27 GHs and 32 ºC. Bfgminner have restarted board many times but until now without results. Tomorrow i will install remaining boards and see what are results.

Regards

I was sure I didn't send any of thus out but it looks like I did a mistake because I'm missing one... I will send you another can you send this one back...

I probably made a mistake when randomizing boards for sending...

Thanks by prompt support, I sent PM about.

af_newbie

legendary

Activity: 2688

Merit: 1468

Quote from: Bogart on January 03, 2014, 09:01:19 PM

Quote from: Lucko on January 03, 2014, 03:55:07 PM

if anyone else would like me to make his boards let me know...

Go ahead and make mine please.

Thanks.

Lucko,

Make mine too.

Thanks,

GrapeApe

sr. member

Activity: 476

Merit: 250

Quote from: Lucko on January 04, 2014, 04:30:34 AM

Quote from: GrapeApe on January 03, 2014, 11:45:24 PM

Go ahead and make mine as well. I'm curious if you ever tried the modified firmware that limits the voltage to 1.1v, if so what were your results?

It was not on dropbox at that time so no... Link didn't work...

The link for the modified firmware is up now. He has a link to it on the first page of his thread with the other firmware links.

https://www.dropbox.com/s/rfa4kuigov99sg8/Chili14e1v1.hex
works again it was down

asjfdlksfd

full member

Activity: 128

Merit: 100

Quote from: bx8389 on January 03, 2014, 07:48:00 PM

Also received today mines, thanks Lucko,
At this moment i try to put first one at work but resets at 27 GHs and 32 ºC. Bfgminner have restarted board many times but until now without results. Tomorrow i will install remaining boards and see what are results.

Ok, I have the same issue here now with the first 3 boards. The others are not completed, yet because of missing coolers, which should arrives over the day.

The messages looks like the same as is described on Mr. Teals forum a a result of high performers.
With 119, 121 they are not really high performers, the third with 126 looks like more that.

Code:

[2014-01-04 09:52:11] Devices detected:
[2014-01-04 09:52:11] BitFORCE SHA256 SC by MrTeal and ChipGeek (driver=bitforce_queue; procs=1; serial=FTX30WIZ; path=/dev/ttyUSB0)
1 devices listed
l))
[2014-01-04 09:52:11] Acquired exclusive advisory lock on /dev/ttyUSB0
[2014-01-04 09:52:11] BFL: Attempting to open /dev/ttyUSB0
[2014-01-04 09:52:11] Found BitForce device on /dev/ttyUSB0
[2014-01-04 09:52:11] DEVICE: Chili SC
[2014-01-04 09:52:11] MANUFACTURER: MrTeal and ChipGeek
[2014-01-04 09:52:11] FIRMWARE: 1.2.14e
[2014-01-04 09:52:11] CHIP PARALLELIZATION: NO
[2014-01-04 09:52:11] QUEUE DEPTH:40
[2014-01-04 09:52:11] PROCESSOR 0: 16 engines @ 181 MHz -- MAP: FFFF
[2014-01-04 09:52:11] PROCESSOR 1: 16 engines @ 181 MHz -- MAP: FFFF
[2014-01-04 09:52:11] PROCESSOR 2: 16 engines @ 180 MHz -- MAP: FFFF
[2014-01-04 09:52:11] PROCESSOR 3: 16 engines @ 178 MHz -- MAP: FFFF
[2014-01-04 09:52:11] PROCESSOR 4: 16 engines @ 200 MHz -- MAP: FFFF
[2014-01-04 09:52:11] PROCESSOR 5: 16 engines @ 189 MHz -- MAP: FFFF
[2014-01-04 09:52:11] PROCESSOR 6: 16 engines @ 186 MHz -- MAP: FFFF
[2014-01-04 09:52:11] PROCESSOR 7: 14 engines @ 182 MHz -- MAP: DFF7
[2014-01-04 09:52:11] THEORETICAL MAX: 23.28 GH/s
[2014-01-04 09:52:11] ENGINES: 126
[2014-01-04 09:52:11] FREQUENCY: 184 MHz
[2014-01-04 09:52:11] CRITICAL TEMPERATURE: 0
[2014-01-04 09:52:11] TOTAL THERMAL CYCLES: 0
[2014-01-04 09:52:11] XLINK MODE: MASTER
[2014-01-04 09:52:11] XLINK PRESENT: NO
[2014-01-04 09:52:11] Devices detected:
[2014-01-04 09:52:11] BitFORCE SHA256 SC by MrTeal and ChipGeek (driver=bitforce_queue; procs=1; serial=FTX30WIZ; path=/dev/ttyUSB0)

The miners have near by the same problem. They start and bfgminer test shows more or less:

Code:

bfgminer version 3.9.0 - Started: [2014-01-04 09:53:08] - [ 0 days 00:02:14]
[M]anage devices [P]ool management [S]ettings [D]isplay options [H]elp [Q]uit
Connected to removed.btcguild.com diff 8 with stratum as user removed
Block: ...6cb9f38a #278547 Diff:1.42G (10.15Ph/s) Started: [09:53:08]
ST:7 F:0 NB:1 AS:1 BW:[163/192 B/s] E:36.16 I: 399uBTC/hr BS:2.92k
1 62.0C | 30.36/33.02/30.15Gh/s | A:163 R:2+1(.97%) HW:148/ 13%
--------------------------------------------------------------------------------
BFL 0 : 62.0C | 40.75/32.03/29.25Gh/s | A:163 R:2+1(.97%) HW:156/ 13%
--------------------------------------------------------------------------------

Select processor to manage using up/down arrow keys
BFL 0 : 65.0C | 40.75/34.93/31.72Gh/s | A:161 R:2+1(.99%) HW:134/ 13%
BitFORCE SHA256 SC from MrTeal and ChipGeek
Serial: FTX30WIZ
Kernel: bulk queue
Temperatures: 43.0C 65.0C
Voltages: 3.290 / 1.157 / 11.781

[D]isable [I]dentify [F]an control
[Slash] Find processor [Plus] Add device(s) [Enter] Close device manager
[2014-01-04 09:55:12] BFL 0: Received unexpected queue result response:
[2014-01-04 09:55:13] BFL 0: Received unexpected queue result response:
[2014-01-04 09:55:14] BFL 0: Received unexpected queue result response:
[2014-01-04 09:55:14] BFL 0: Error: Get temp returned empty string/timed out
[2014-01-04 09:55:14] BFL 0: Received unexpected queue result response:

Here the LED 2 an 4 are on, 7 is blinking ~4Hz. After a while 7and 8 are blinking where 7 but now 2Hz by both going on same time but first 7 goes off (1:1) then 8 (1:2-3).:

7 011001100110011001100...
8 011101110111011101110...

After a while more left 6 are blinking randomly for a second and then all are going off. Now it looks like the board has to been reset.

Then I can restart bfgminer. It looks like really the problems comes from the power rails. I didn't cool the FETS active by a fan any more at the moment so the fet temp rises 45 °C which limits the vcore <1.1 V. So flashing to 1.1V limit could be solve the problem.

Code:

Block: ...6cb9f38a #278547 Diff:1.42G (10.15Ph/s) Started: [10:02:38]
ST:7 F:0 NB:1 AS:0 BW:[101/ 81 B/s] E:95.64 I: 541uBTC/hr BS:4.75k
1 59.0C | 36.78/36.78/34.57Gh/s | A:266 R:5+1(.68%) HW:389/8.0%
--------------------------------------------------------------------------------
BFL 0 : 59.0C | 37.13/36.78/34.57Gh/s | A:266 R:5+1(.68%) HW:389/8.0%
--------------------------------------------------------------------------------

Select processor to manage using up/down arrow keys
BFL 0 : 59.0C | 37.12/36.78/34.35Gh/s | A:260 R:5+1(.74%) HW:368/7.9%
BitFORCE SHA256 SC from MrTeal and ChipGeek
Serial: FTX30WIZ
Kernel: bulk queue
Temperatures: 45.0C 59.0C
Voltages: 3.290 / 1.097 / 11.906

At the moment so I guess do not blow the power fets. This will limit also the voltage and is a work around. Also the other one could I fix by removing the fan to blow the fets, too :-/

Cheers...

Topic: BFL board project COINTAMINATION - EU facility - ORDERS OPENED for Chili - page 12. (Read 73031 times)