Nope, not the power supply. Nope, not the proxy being used. Nope, not the pool being used. Nope, not the color tshirt I was wearing. Nor the phase of the moon.

It was the damned fan.

Even while it seemed to be moving plenty of air, it appears that the auto-shutoff circuit in the cube is based on how much current the fan is pulling. This is why when the fan was just starting to go bad, the stupid baffle trick worked. I should have thought of replacing the fan then, it seems so obvious now... ah well.

I replaced the fan with a 120mm 2200RPM 80CFM unit (cost me $5), and suddenly everything works just fine again.

http://www.tigerdirect.com/applications/searchtools/item-details.asp?EdpNo=8208359

My beleaguered cube is now happily hashing away in the high clock speed (where before it would immediately shut off if I even attempted to overclock it).

In summary; if you have a working cube that suddenly starts turning itself off, replace the fan -- even if the fan appears to be moving plenty of air at first glance.

I switched to the slush proxy, and it stays up for between one and three days at a time now. But then it will still *click* and shut itself off, requiring a manual toggling of the power.

[edit] I'm using the slush proxy with the -rt option to relay the vardiff changes along to the cube directly. I'm not sure if that may be one of the reasons it stays up longer this way; back with the bfgminer proxy it would do the diff filtering on the miner software rather than the cube (I think... if I am wrong here please anyone feel free to correct me)

I gave up on the baffle idea, it was garbage anyway. There's some slight possibility that the fan draw influences the issue, but I'm just discarding that line of thinking.

I had been using a DYNEX DX-520WPS. I spoke to someone from WTCR and they suggested that it may be underpowered on the 12v line (20A vs a recommended 30A). I switched on their recommendation of a Corsair CX500.

It fell over again afterwards, on the new power supply. Oh well, that was $60 on the wind

It's currently running after yet another restart. It's been up for 22 minutes. All ASIC are "o", and their hash numbers seem to be in line with what I'd expect. It's telling me:

Jobs:0000010115 Accepted:0000009662 Rejected:0000000285 (5:0) F1 F2 F3
MHS:31264 Utility:427 Efficieny:095.52%

Which seem... okay? I don't have a good baseline to compare that against.

I fully expect it to fall over again. Before the current restart it ran for maybe an hour? Before that it wouldn't stay up for more than two minutes for a few of restarts. I need to start taking better notes on how long it lasts for every single attempt...

I really, really wish that the mystery "debugging USB" port had some kind of documentation! I would love to be able to get SOME kind of feedback beyond the web page, like an error message prior to shutdown!

I want to reiterate to those playing along at home that it will shut itself off after an apparently random amount of time whether it's connected to my proxy or not! But with that said, I'm running it against bfgminer (and have been since I first got it).

Maybe if I don't touch ANYTHING it'll stay alive this time... hahaha.

20 hours or so of uptime and it fell over.

Its current status is back to what it had been doing; start (flash red activity light). Finish booting (light goes solid). Five to ten seconds later, click, and it's off.

I can't help but wonder if the auto-shutoff mechanism isn't handled by the draw of the fan?

Well, after I thought I had it solved (it seemed the left-most two slots would cause it to flounder, regardless of which card was plugged in) it fell over again.

I put all the cards back in again, and ... had a weird thought. I put a small baffle in front of the intake fan, about halfway covering it (limiting the air flow into the case for cooling).

Without the baffle, it'll shut off in around 2 minutes. With the the baffle completely covering the fan, immediate shut-off. Half-covered... it's been running for two and a half hours on Low clock, no issues.

I'm... flabberghasted. I have to admit that the baffle idea came out of nowhere -- I thought that maybe letting it heat up a bit might help get it running reliably... pure voodoo on my part, I readily admit that. I don't like voodoo in electronics, so if this does indeed keep the silly thing stable, I'd like to figure out WHY. Is it because the fan has resistance and has to work harder? Is it because the cards get a chance to heat up a bit? Is something more or less conductive at the higher temperature, and thus "happier" from the perspective of a working system??

I'll keep monitoring and report back.

So, perhaps I can get this post moved from the newbies forum over to the appropriate one at some point. But in the interim, I'd like to post some of the issues that I've been experiencing with the Block Erupter Cube.

For the last couple of weeks, the cube had been operating just like it was supposed to. I was channeling it through a bfgminer proxy, and all was well. I started noticing that it was resetting itself to the "Low" clock speed, and after attempting to set it "High" and not having it stick for more than a few hours, I just let it continue running on "Low"

Another week passes, and I notice one morning that the cube isn't running. The power supply that feeds it is happily spinning its fan, and the "SB PWR" light is lit on the cube, but nothing else seems to be. I toggle the power off and on, and cube boots up and begins running again, hashing at the normal rate just like it should. But around five hours later, it's off again.

I read on here that some of the blades of the cube might be misaligned and indeed they did appear to be nudged outside of their intended grooves in the case. So I took everything out (including the "Block Erupter" sticker which had for some reason fallen inside) made sure nothing looked burnt, replugged all the cards, and carefully slid it all back into its case. I swapped from the old fuse to the other one (nice of them to send along two), and powered it up.

Five minutes later, it's off. It never even got warm!

So I begin the task of trying to debug the thing further. I perform a factory reset (just hold down the reset button while it's powered on and wait until it flashes) and let it sit running but idle on my desk, just to observe it.

After around 60 seconds, I notice the activity light start flashing, and then there is an audible "click" and the unit powers off. The case is cold to the touch (since it never even started hashing, I didn't have it plugged into the network). The big fan is still running, the "SB PWR" light is on, but for all other intents and purposes the unit lies dead. I repeat this a few times just to verify it's going to continue shutting itself off cold, and indeed it does.

So I've ruled out overheating. The power supply is good (but just in case I did indeed try another one temporarily). Cooling is adequate. The bloody thing just won't stay running!

Now, there's a USB port on the back of this thing which the promotional material label as "for debugging and firmware updates". However, I notice there is a distinct lack of information on just what "debugging" entails.

Has anyone gotten some sort of console or memory dump via USB? What did you do?

I see that some other people are experiencing the "click of death" on their cubes as well. Many are reporting "X" on the status interface. In the short amount of time that I can actually access the status page, I don't see anything marked as "X".

I'm going to try removing all of the blades and replacing them one at a time to see if I can narrow it down to either an individual blade, or an individual slot that causes the failure.

The software that is actually running on the cube SEEMS to be shutting it down deliberately (at least that's what I take the flashing of the status light to mean, before it clicks and shuts down). I would love to be able to debug a REASON from the machine. And the only method of debugging is that USB port and its mysterious uses!