KNC on fire !

[rammy2k2]

This is a repost from KNC forum, i was thinking it should be posted here too.

Unit's catching fire!!!
6th October 2013, 09:35 PM
Just caught this on irc #kncminer

http://imgur.com/2UZuUzH, http://imgur.com/a/hyA3a


Apparently modules are blowing up what appears to be capacitors, WTF!?!?

[dogie]

... and I was called an armchair engineer.... Sorry to those with financial losses, and to those who's houses/families are at risk from these units.

Edit: If the poster is looking at what I'm looking at, those are resistors.
Edit2: I can see the dark silhouettes of capacitors too. Could be any of it that started.

[Zeek_W]

Whoops! Any word from KNC on this?

[CMMPro]

Yes, the word is NOT to plug / unplug your miner power cables while it is powered up or it will surge the boards and kill them.

(Bitcoinorama just stated this was the cause.)

[zedicus]

Sorry to hear about your issue.

Look at the capacitor and see what the rating on it is and swap it out. You have a soldering gun? Sending it out to knc is going to be weeks worth of down time vs replacing a simple capacitor.

Just look at he microfarad and voltage rating and grab a capacitor at an electronics store or radioshack and call it a day.

Hopefully your handy with a soldering gun.


 

[Littleshop]

Sorry to hear about your issue.

Look at the capacitor and see if what the rating on it is and swap it out. You have a soldering gun? Sending it out to knc is going to be weeks worth of down time vs replacing a simple capacitor.

Just looks at he microfarad and voltage rating and grab a capacitor at an electronics store or radioshack and call it a day.

Hopefully your handy with a soldering gun.


 

It looks worse then one capacitor.  I would guess the PCB itself is damaged and things would need to be re-routed on the surface for a repair.   

[zedicus]

Seems you are correct, I hadnt looked at the pics, i thought it was just a bad cap.


Might as well get the snail mail going asap! Unless they want to send a new board.

[Tehfiend]

... and I was called an armchair engineer.... Sorry to those with financial losses, and to those who's houses/families are at risk from these units.

Yes because you were absolutely completely wrong in your constant trolling in the KNC thread. For example:
https://bitcointalk.org/index.php?topic=273584.msg2937636#msg2937636

LOL at the desperation in this thread, people going to be sad when KNC runs away.

It was a larger model than I thought as I just skimmed over the pic, but just try running that at 320W 24/7 LOL.

LOL indeed as I'm looking at those coolers working perfectly with my Jupiter as I type this. Don't be bitter just because you were 100% wrong with all your arm-chair engineering and I'm sure you will find some defective units in every single model of ASIC miners out there. The fact is that KNC brought an ASIC miner to market in around 4 months and while not 100% perfect, they are currently at the top of my list.

Yes, the word is NOT to plug / unplug your miner power cables while it is powered up or it will surge the boards and kill them.

(Bitcoinorama just stated this was the cause.)

I initially had some issues with my PSU and this is exactly how I just powered mine on so I don't know if that is the case. While trouble shooting I plugged the controller in first while PSU was already on and then plugged in each module one at a time. It ended up being a bad cable and is now working perfectly. Hopefully this is an isolated incident as I have't heard of any other problems like this.

[frankenmint]

Okay so for those of us who still have stuff on order or have recently received anything - IS THIS A PROBLEM?  I saw a guy on IRC who said "use VRM heatsinks"  that will prevent the problem altogether by reducing heat output ~20%.  Is that true?  Will this postpone orders going forward?  I'd be willing to wait if it means being sure my house will not catch on fire.  I'm not saying it will or that this is bound to happen (I'm still hoping that this was a freak occurence Human error based incident)  [not human error as a result of bad engineering]. 

Just looking for answers.

[tacotime]

Okay so for those of us who still have stuff on order or have recently received anything - IS THIS A PROBLEM?  I saw a guy on IRC who said "use VRM heatsinks"  that will prevent the problem altogether by reducing heat output ~20%.  Is that true?  Will this postpone orders going forward?  I'd be willing to wait if it means being sure my house will not catch on fire.  I'm not saying it will or that this is bound to happen (I'm still hoping that this was a freak occurence Human error based incident)  [not human error as a result of bad engineering].  

Just looking for answers.

It's an extra $5-10, there's no reason not to

If your MOSFETs are running hot it means fluctuations in voltage over time to the chip and the rest of the hardware, which will eventually destroy it.  Gigabyte's old 7950s had insufficient VRM cooling and a 100% failure rate for me within 12 months (mining LTC 24/7, with the VRMs running 80-100C); the GPU cores themselves were running in the 60-70s

[calichomp]

I'm trying to get a refund now.  These are barely on the edge of ROI for me.  It isn't worth burning the apartment building down for a few extra bucks.  Will let you all know if I have any luck.

[tacotime]

Just an FYI, these are the things you'll wanna slap a heatsink on:

You can tell them by the three prongs at one even if you don't know anything about PC hardware

Each of these chips is pulling 180 W from my quick and dirty calculation, it's likely that the MOSFETs NEED to be cooled at that level of draw

Can someone who owns one of these please tell me how many phases they are using?  They should be using no less than 4 (you can tell by the number of MOSFETs/chokes), with a more solid design being 6-8

[dropt]

... and I was called an armchair engineer....

That's because you *are* an armchair engineer.  Are you *sure* those are resistors?

[Paladin69]

I'm trying to get a refund now.  These are barely on the edge of ROI for me.  It isn't worth burning the apartment building down for a few extra bucks.  Will let you all know if I have any luck.

At the rate of difficulty increase, by December of 2014, a 70TH machine (if one will exist) would only have a shelf life of one month.

ASIC is really destroying mining.  If code could be implemented to exclude ASIC it may need to be done in 2015 or sooner.

[MoneyMorpheus]

I'm trying to get a refund now.  These are barely on the edge of ROI for me.  It isn't worth burning the apartment building down for a few extra bucks.  Will let you all know if I have any luck.

At the rate of difficulty increase, by December of 2014, a 70TH machine (if one will exist) would only have a shelf life of one month.

ASIC is really destroying mining.  If code could be implemented to exclude ASIC it may need to be done in 2015 or sooner.

If the asics where excluded the network would be left unprotected for a 51% attack. It needs the asics to get big, what's killing mining are irrational buyers...

[Paladin69]

I'm trying to get a refund now.  These are barely on the edge of ROI for me.  It isn't worth burning the apartment building down for a few extra bucks.  Will let you all know if I have any luck.

At the rate of difficulty increase, by December of 2014, a 70TH machine (if one will exist) would only have a shelf life of one month.

ASIC is really destroying mining.  If code could be implemented to exclude ASIC it may need to be done in 2015 or sooner.

If the asics where excluded the network would be left unprotected for a 51% attack. It needs the asics to get big, what's killing mining are irrational buyers...

I would think faster machines make it more vulnerable to a 51% attack, not slower machines.  People would spread their GPU's and FPGA's out again if ASIC was excluded.

[MoneyMorpheus]

I'm trying to get a refund now.  These are barely on the edge of ROI for me.  It isn't worth burning the apartment building down for a few extra bucks.  Will let you all know if I have any luck.

At the rate of difficulty increase, by December of 2014, a 70TH machine (if one will exist) would only have a shelf life of one month.

ASIC is really destroying mining.  If code could be implemented to exclude ASIC it may need to be done in 2015 or sooner.

If the asics where excluded the network would be left unprotected for a 51% attack. It needs the asics to get big, what's killing mining are irrational buyers...

I would think faster machines make it more vulnerable to a 51% attack, not slower machines.  People would spread their GPU's and FPGA's out again if ASIC was excluded.

It is more likely to get the network overpowered by someone with access to a few supercomputers than a lot of asics, because if the latter, he would be stuck with a pile of junk and the first one doesn't plus he gets whatever he made while ruining btcs...

If the time comes that the algorithm needs to be changed it will be due to security reasons and not because of unprofitable mining.

[kmtan]

too bad...

[tolip_wen]

... and I was called an armchair engineer....

That's because you *are* an armchair engineer.  Are you *sure* those are resistors?

+1

[Fefox]

Yes, the word is NOT to plug / unplug your miner power cables while it is powered up or it will surge the boards and kill them.

(Bitcoinorama just stated this was the cause.)

ummmm, so you have a power failure, and when the power comes back on the unit will die?


ok never mind, I just got it, don't unplug the psu from the miner.  hahaha well who would do that?