KNC Jupiter suddenly died. Anybody had similar issue?

[czz]

Hi everybody,

My Jupiter died today. It looks like either the beaglebone or the controller board is dead. This happened just today suddenly during normal operation. It worked well a week before that.

After I turn on the power, following happens:
1) the "power" LED turns on
2) in 3-4 seconds, User LED 0 turns on
3) in 1-2 seconds, all User LED's turn on
4) in a second, all LED's go black, and the white light on the controller board never turns on
5) in some time, this repeats

This happens with two different power supplies. Also this happens regardless if I have PCI-E connectors connected, or if I have them disconnected. I have also made sure that Beaglebone is connected firmly to the controller.

Video here:
https://dl.dropboxusercontent.com/u/4387628/2013-10-26%2015.57.12.mp4

Of course I contacted KNC but they will be out for weekend, so for now I'd just like to know if anybody had a similar issue. Also I have a slight hope there is something that can I can fix myself.

[ralree]

Have you tried the hard-reset procedure?

[czz]

Hard reset did not help

[DyslexicZombei]

Can you please tell us what firmware version that you're using?

You didn't mention this crucial piece of info. Thanks.

[czz]

DyslexicZombei,

Firmware 0.97

[DyslexicZombei]

DyslexicZombei,

Firmware 0.97

Oof. Have you had anything similar happen with previous firmware versions? I'd suggest that anyone not do any firmware upgrades to chase a few more GH/s without thoroughly checking out how well *other* canaries in the coal mine are doing. I've been in the line of work of US Govt. Test & Evaluation and Research & Development for about 10 years.

These miners, and every other miner, are essentially prototypes and would be rated at low TRL levels using the military's classification system for new tech.

Quite a few people have had major issues w/ 0.97. Hopefully, they'll cross ship you a replacement Jupiter without waiting for your Jupe to arrive.

[czz]

DyslexicZombei,

No I did not have this with any other firmware. It was also working well on 0.97 since the day it came out.
Anyways this more likes a hardware failure to me.

The beaglebone seems to work when disconnected from the controller board. It only behaves this way when plugged into the controller. So the controller might be dead.

[czz]

No way to reflash either. When attempting to boot from an sd card with a recovery image, it behaves exactly the same way.

[DyslexicZombei]

No way to reflash either. When attempting to boot from an sd card with a recovery image, it behaves exactly the same way.

Not a good sign. Honestly sounds like the magic smoke came out when you weren't looking.

Any fried marks on any PCBs or the smell of anything burnt?

[czz]

No, there is no visible damage. That would be too simple

[Dalkore]

My issue:    One unit does not get through the boot process.  The LEDs on the right side of the Ethernet port never turn on.  It stops at that point.   I have triple checked all connection and swapped a known good PSU.  I sent an email in without response.  At this point I think I have a bad controller board and just need them to send me a new one so I can mail this one back.  Anyone else have this issue?

-D

[bobsag3]

My issue:    One unit does not get through the boot process.  The LEDs on the right side of the Ethernet port never turn on.  It stops at that point.   I have triple checked all connection and swapped a known good PSU.  I sent an email in without response.  At this point I think I have a bad controller board and just need them to send me a new one so I can mail this one back.  Anyone else have this issue?

-D

Have you tried reseating the beagle board and or all the connector cables from the ASICs to the control board? Also what firmware were you on before the issue came up?

[Paladin69]

Try booting it up without the network cable plugged in and see if it stays on.

You're also going to have to test each board alone with the other PCI-E's disconnected.

[FeedbackLoop]

Try booting it up without the network cable plugged in and see if it stays on.

You're also going to have to test each board alone with the other PCI-E's disconnected.

In his video the network cable IS off and he stated the PCI-E's were all unplugged. 

[Dalkore]

My issue:    One unit does not get through the boot process.  The LEDs on the right side of the Ethernet port never turn on.  It stops at that point.   I have triple checked all connection and swapped a known good PSU.  I sent an email in without response.  At this point I think I have a bad controller board and just need them to send me a new one so I can mail this one back.  Anyone else have this issue?

-D

Have you tried reseating the beagle board and or all the connector cables from the ASICs to the control board? Also what firmware were you on before the issue came up?

I did but I will do it again to be absolutely sure.  It has the stock firmware which I believe 0.96.  Can I actually take the beagle board, which I assume as the network port on it, and reseat it?   I didn't know if it was secured on there or if it was removeable. 

-D

[Paladin69]

Try booting it up without the network cable plugged in and see if it stays on.

You're also going to have to test each board alone with the other PCI-E's disconnected.

In his video the network cable IS off and he stated the PCI-E's were all unplugged. 

oops...weary of clicking links.

Mine had a bad card and it took down the whole system after 10 minutes every time.  I couldn't get it to turn back on to test each board until the network cable was unplugged.  That was all I personally had to do.

RMA'd the bad board.  Then the next day they came out with the enablecores.bin, always wondered if that would've helped...

[bobsag3]

My issue:    One unit does not get through the boot process.  The LEDs on the right side of the Ethernet port never turn on.  It stops at that point.   I have triple checked all connection and swapped a known good PSU.  I sent an email in without response.  At this point I think I have a bad controller board and just need them to send me a new one so I can mail this one back.  Anyone else have this issue?

-D

Have you tried reseating the beagle board and or all the connector cables from the ASICs to the control board? Also what firmware were you on before the issue came up?

I did but I will do it again to be absolutely sure.  It has the stock firmware which I believe 0.96.  Can I actually take the beagle board, which I assume as the network port on it, and reseat it?   I didn't know if it was secured on there or if it was removeable. 

-D

I know its not secured well enough to survive shipping and not coming loose, so you should be able to remove it.