Topic: Butterfly Labs - Bitforce Single and Mini Rig Box - page 23. (Read 186929 times)

Will do Garr...

They won't let me video the "fun stuff" I would assume...

Haha, get some footage of their place.

I have already advised BFL I will be picking up my MR in KC 

The Single has little weight, posting something like that it pretty easy, MiniRig will be the real challenge.
I hope they ship a test MiniRig around US before shipping it back to themselfs to see what weaknesses the packaging might have.

I've talked with someone who worked at UPS a while back... same sort of deal.  They would laugh at packages that said "fragile" on them, and simply toss them even harder.

Really, that's why you need so much foam and bubble wrap to ship anything safely.  I'd say that the BFL units are wrapped pretty well though.

http://img137.imageshack.us/img137/6628/capturevl.jpg

I've dealt with surface mount before, made sure the polarity was correct and have a working led again.

Yes, regarding above post, I worked for Securicor and DHL and saw how things are loaded onto trailers, sometimes thrown 2-3 meters forward for it to land hard on the floor, after that half a dozen other heavy boxed thrown on top of it. Only exception is the very top of the range service, for which a great premium is charged but the boxes are handles with more care.
I hope the Mini-Rig has some serious packaging.

Yeah... the LED is SMT, so it probably didn't fall off, but the little plastic thing easily could have if the package was thrown across the room or tossed onto a truck carelessly (not infeasible). Watching a large package sorting operation a la USPS, UPS, FedEx, etc, is a hair-raising experience.

The polarity of the LED is important. If attached in wrong polarity, it will never turn on.


Regards,
BF Labs Inc.

Are you saying the actual led actually fell off or are you talking about the light guide? I have had a light guide fall off on one of my units before. If its the actual led how did you reattach it?

Is overheating of either/both of the two fpga's the only thing that can cause throttling?  Or are there other factors that can cause a single to throttle?  I removed the heat pipe, the old paste was uniformily spread on both chips, cleaned up the old and reapplied some new stuff making sure it too was uniformily spread.  So far the highest I can go is with the 808 firmware and that might throttle a handful of times an hour.  Temperatures consistantly reported between 44-46C.  Any firmware above with throttle every 3-4 minutes.  The only thing that I forgot to check for was that both chips are sitting in the same plane, but by looking at the even distribution of the thermal paste, I would guess they are.

Also, as an FYI in case the front led stops working for any one else (or falls out like mine), the two leds on the right side will go out for about 12 seconds if the unit throttles.  Although, I did go out on a limb and was able to reattach the led that fell off.

OK please let us know how it goes. Throttling a lot usually means heatsink detachment or FAN inactivity.


Regards,
BF Labs Inc.

Room is temperature controlled and sits between 68-70F (20-22C).  There is nothing exhausting hot air and the singles sit about 8 inches apart and should be well vemtilated.  I have tried power supply, usb cables, host ports from working units and no change in operation.

I had a box of 3 and a box of 2, just wondering if both units I have issues with were from the box of two.  At least from appearance, no damage to either packaging.

I have tried the 832 and 800 firmware, both exhibit the same behavior in cgminer (after a minute or so, stuck at 600 +- 100).  I did a Light Diagnostic in easyminer at 832, Throttled 22 time before reaching 100 in about 600 seconds.

All the square heatsinks appear firmly in place as well as the small heatsink/fan on the botton, I don't want to put any pressure to them.  The main heatpipe had both clips seated correctly and I can see white thermal paste along the edges of the chips.  It also appears flat and not out of place.  Verified the top and bottom fans are both operational.  Top fan seems to be pushing the same as the other units, tested by checking deflection of a piece of paper over top vent.  I tried putting a powered 80mm fan on the button pushing into the unit, still problematic.

I have the ability as well as the usual white thermal paste.  But I will save removing the heat pipe, cleaning and reapplying as a last resort.

edit 1)  I am now trying the 768 firmware to see what happens, figure 768 is better than 600 or below.

edit 2) 768 firmware seemed solid, could not detect any throttling, ran at 45C for 2 hours with hm/s avg 754/762.  Tried the 792 firmware, ran at 127-200mh/s at 37C.  Downloaded and flashed it twice to make sure the flash wasn't bad.  Gave up on 792.  Now trying 800 firmware.  Will post results after a few hours.

It seems one of your units is throttling. There are two ways to know whats happening

1) Verify the Heatsink is firmly attached.
2) Try a lower bitstream like 800 or 768, etc. If the hash-rate increased, it means the unit was throttling.

It is odd that the unit throttles if your room temperature is 22C up to 25C and the units are not taking
exhaust air from another hot component (PC or PSU, etc). They are specifically tested for this situation
prior to shipping.

You can also verify the Silicon Paste under the heatsinks (if you have the utility to do so).

Please let us know how it goes. You can post or send PM.


Regards,
BF Labs Inc.

At one point it did light up, although it was dimmer when compared to the others.  The only thing I did was move it to another computer.  I was trying to figure out why it was getting low performance.

After 8 hours of runtime

http://img13.imageshack.us/img13/2936/bflrig.jpg

I was wondering, how did the LED fall out? Was it hit by something? Before it fell, did it light-up and blink
when the unit was powered on?

Regards,
BF Labs Inc.

tried the unit on a power supply from one that is working well, started mining and after about a minute it just stay between 500-700mh/s with temps at 45-46c.  Ran the light diagnostic and after getting to 100 had 22 Throttles in about 600 seconds run time.  Hard to tell if it is throttling since the led fell out.

I was amazed that I got a calling so quickly on a Friday evening after the email about the first one.  So at least it looks like they are standing behind their product, just hope I end up satisfied with the issues relating to the first and now the second problem I have had.

That is very bad luck rkozola.

I've got a total of 14 Singles from BFL. Not one has had a problem. 4 of them arrived this week. Reading through the forums there is some talk of the heat sinks coming loose in shipping but problems like yours are far from the norm.

Being DOA or nearly DOA like yours were I'd at least ask Sonny if BFL can cover your shipping costs.

Out of curiosity, have you tried another power supply on the single with the light out? The power supplies have been a problem occasionally. One of my power supplies died and behaved like yours. Singles themselves have been solid though.

and then there was 3.  So I noticed 1 of the 4 remaining singles was constantly running at 700mh/s with 45-46c temps.  I go to take a look to see if the front is blinking and the front led is no longer lit.  Great.  I go to move it to another pc too see what is going on and I hear a very slight tink, as I move it I see a very tiny led fall out.  So far went from happy to a little annoyed and now to kind of pissed.  I already had to pay to ship 1 single back, and I am guessing soon I will be paying for another.   

Got my single today.  I've been experimenting with firmwares since I got off work.  Noticed some odd dips in mh/s with the stock firmware.  No flashing throttle light, but happened in a noticeable interval.  Flashed the 864 and 872 firmwares.  Again noticed some dips in performance, but this time intermittent throttling light.  Pulled the cover off.  Immediately saw a good amount of white thermal paste on a sliver of one chip that isn't covered by the heatsink base.  Took the heatsink assembly off the board (FU push pins).  Confirmed my suspicion that the thermal paste used is not good(very runny white paste and applied copiously).  Cleaned it off and put some MX-4 on both chips, got a good press seal between the chips and heatsink, then reassembled.  My temps are a solid 3-4C lower with the 864 firmware.  No more throttling or dips in mh/s.  Mh/s stays pegged..  Haven't tried the 872, but suspect I should have good enough thermals to run it.  Just an FYI for anyone seeing something similar.