[Guide] Dogie's Comprehensive ASICMiner Cube Setup [HD]

[dmcdad]

A couple of times now I've had my cubes reset from High clock to Low clock while unattended. Anyone else seen this? In High clock mode they seem to run fine (occasionally get 1 "x" on a chip, rejects and efficiency look great).

Related to this, is there any type of HTTP API to get the status of the cubes and/or control the clock rate? I'd like to setup a monitoring script that gets the reported hash rate (to log/graph) and automatically switches the clock back to High if it ends up on Low. The HTTP request that the "Clock" button uses toggles between Low and High, so I wonder if there is a better HTTP mechanism that allows discrete Low/High control (otherwise I'll have to screen scrape the current Clock state to make sure I don't toggle it to Low).

I just realized I probably didn't press the Update/Restart button after changing the clock to High. It looks like pressing that may be required for the High clock setting to stick? Note though I didn't restart the unit -- it spontaneously went to Low and/or restated itself.

[1714709688]

1714709688

[1714709688]

1714709688

[1714709688]

1714709688

[1714709688]

1714709688

[dogie]

A couple of times now I've had my cubes reset from High clock to Low clock while unattended. Anyone else seen this? In High clock mode they seem to run fine (occasionally get 1 "x" on a chip, rejects and efficiency look great).

Related to this, is there any type of HTTP API to get the status of the cubes and/or control the clock rate? I'd like to setup a monitoring script that gets the reported hash rate (to log/graph) and automatically switches the clock back to High if it ends up on Low. The HTTP request that the "Clock" button uses toggles between Low and High, so I wonder if there is a better HTTP mechanism that allows discrete Low/High control (otherwise I'll have to screen scrape the current Clock state to make sure I don't toggle it to Low).

I just realized I probably didn't press the Update/Restart button after changing the clock to High. It looks like pressing that may be required for the High clock setting to stick? Note though I didn't restart the unit -- it spontaneously went to Low and/or restated itself.

Just checked mine and it had gone back to low. Its possible there is a temperature based trigger we're not aware of, or that an interruption in pool traffic triggers the same. We will have to see.

There are api scripts people have been running for V1 and B2 blades to monitor lots of them at once, so they do exist.

[wiredmine]

A couple of times now I've had my cubes reset from High clock to Low clock while unattended. Anyone else seen this? In High clock mode they seem to run fine (occasionally get 1 "x" on a chip, rejects and efficiency look great).

Related to this, is there any type of HTTP API to get the status of the cubes and/or control the clock rate? I'd like to setup a monitoring script that gets the reported hash rate (to log/graph) and automatically switches the clock back to High if it ends up on Low. The HTTP request that the "Clock" button uses toggles between Low and High, so I wonder if there is a better HTTP mechanism that allows discrete Low/High control (otherwise I'll have to screen scrape the current Clock state to make sure I don't toggle it to Low).

I just realized I probably didn't press the Update/Restart button after changing the clock to High. It looks like pressing that may be required for the High clock setting to stick? Note though I didn't restart the unit -- it spontaneously went to Low and/or restated itself.

Yep thats exactly right. I've experienced the same problem, it will either randomly swap from pool A to pool B or high to low if you don't specifically click "Update/Restart" after any changes. Also you might notice that sometimes just clicking refresh will cause the settings to alter.

[dbzenhosting]

I just got my cube miner today. Its up and running, but it seems that one of the boards are bad. I hear some rattling noise inside when i move/shake the miner.

ASIC_01-16: O O O O O O O O O O O O O O O O
ASIC_17-32: O O O O O O O O O O O O O O O O
ASIC_33-48: O O O O O O O O O O O O O O O O
ASIC_49-64: x x x x x x x x x x x x x x x x
ASIC_65-80: O O O O O O O O O O O O O O O O
ASIC_81-96: O O O O O O O O O O O O O O O O


I am only getting 30-32GB/hash.
Is there any way to fix this problem? I haven't opened up the miner.
Any suggestions?

[dogie]

I just got my cube miner today. Its up and running, but it seems that one of the boards are bad. I hear some rattling noise inside when i move/shake the miner.

ASIC_01-16: O O O O O O O O O O O O O O O O
ASIC_17-32: O O O O O O O O O O O O O O O O
ASIC_33-48: O O O O O O O O O O O O O O O O
ASIC_49-64: x x x x x x x x x x x x x x x x
ASIC_65-80: O O O O O O O O O O O O O O O O
ASIC_81-96: O O O O O O O O O O O O O O O O


I am only getting 30-32GB/hash.
Is there any way to fix this problem? I haven't opened up the miner.
Any suggestions?

Open up the miner, 4 screws on the outside. Take a look how the blades are sitting in the rivets along the top of the case - you may need to jiggle them a bit to get them back in their slots. Then EVERYTHING, including the pcb will slide out in one go and you can take a look what is going on. You may be able to get that blade back in its slot without removing everything, but do take a look.

[sidehack]

If you need to pull the downed card for a once-over but don't want a lot of downtime, these guys will run without any problems with cards missing. Hopefully it's fairly obvious and fairly simple, like a fuse or bent pin.

[dogie]

If you need to pull the downed card for a once-over but don't want a lot of downtime, these guys will run without any problems with cards missing. Hopefully it's fairly obvious and fairly simple, like a fuse or bent pin.

Its also a lot easier to get a replacement, just a diddy little daughter board sent out rather than the block of aluminium.

[dogie]

AntMiner S1 setup thread is up.

[mchu168]

Hey dogie - maybe this has been covered, but if I have more than one Cube, do I simply change its IP address but leave other settings (port #, pool IP, etc) the same?

[dmcdad]

Hey dogie - maybe this has been covered, but if I have more than one Cube, do I simply change its IP address but leave other settings (port #, pool IP, etc) the same?

You'll change the IP address and you'll need a different username/password combo (presuming you are using slush).

[mchu168]

Hey dogie - maybe this has been covered, but if I have more than one Cube, do I simply change its IP address but leave other settings (port #, pool IP, etc) the same?

You'll change the IP address and you'll need a different username/password combo (presuming you are using slush).

Got makes sense.  I will probly use BFGMiner, so just change the IP address of the Cube? 

Another question for you guys, how many Cubes can I put on one 15A household breaker?  or 20A (assuming my wife let's me run an extension cord from the kitchen  )? 

I figure given they draw 280W in OC, I could run 5 or 6 max on the 15A or 7 or 8 max on the 20A (without toaster on).

[edgie13]

I had an issue with an old CoolerMaster 1000w PSU. Could not get the cubes to power up and the would click every 10-15 seconds. Hooked them up to a Corsair 1200 watt and everything works fine. I have 2 on the 1200, 1 on a Corsair 750, and the 4th cube is split between the 1200 and 750.  They are all hashing away at 36 gh/s. Thanks for the guide!!!

Which Corsair 750 are you using?  I have mine connected to a cx750.  I can get the the admin screen, but the damn thing won't hash!  I see it requesting work from the proxy, but not actually hashing the work.  I've tried Slush's proxy and bfgminer with the same results. I even tried disconnecting all but one mini blade, hoping that would give it enough power, but now luck.  I also tried an 500 watt cheap psu, but it just trips the psu as soon as I turn it on.  Any help would be GREATLY appreciated.  Thanks.

[mchu168]

I had an issue with an old CoolerMaster 1000w PSU. Could not get the cubes to power up and the would click every 10-15 seconds. Hooked them up to a Corsair 1200 watt and everything works fine. I have 2 on the 1200, 1 on a Corsair 750, and the 4th cube is split between the 1200 and 750.  They are all hashing away at 36 gh/s. Thanks for the guide!!!

Which Corsair 750 are you using?  I have mine connected to a cx750.  I can get the the admin screen, but the damn thing won't hash!  I see it requesting work from the proxy, but not actually hashing the work.  I've tried Slush's proxy and bfgminer with the same results. I even tried disconnecting all but one mini blade, hoping that would give it enough power, but now luck.  I also tried an 500 watt cheap psu, but it just trips the psu as soon as I turn it on.  Any help would be GREATLY appreciated.  Thanks.

I'm using the CX750 (black Friday special  ) and it's been hashing like a champ for a few days.  I've got it to work on a Win8 machine and a Minepeon (Raspberry Pi) using BFGMiner. 

If you are using Windows, make sure to install the 32bit version of BFGMiner, as the 64bit version doesn't support proxy mining (I tore my hair out for an hour before realizing this).

[hephaist0s]

I have 4 cubes running on 3 different power supplies, an Antec 1200, a Corsair CX750, and one on a Rocketfish 500W I had lying around. Right now, switching the clock to high makes no difference in the speed -- they all hover around 29.5GH/s, whether in high or low. When I set up the first one, it did overclock up into the mid 30's...  then I got distracted setting up the rest, and never really noticed when everything changed. Any ideas at all? I'm not sure where to start on this one.

[edgie13]

I'm not having any luck getting mine set up either.  They'll start up and accept a few shares then they basically stop.  I've even swapped out PSU's to make sure that they're not an issue.

...

Did you ever fix your issue?  I'm running into the exact same thing.

[dogie]

I'm not having any luck getting mine set up either.  They'll start up and accept a few shares then they basically stop.  I've even swapped out PSU's to make sure that they're not an issue.

...

Did you ever fix your issue?  I'm running into the exact same thing.

His login details were wrong.

[dogie]

I have 4 cubes running on 3 different power supplies, an Antec 1200, a Corsair CX750, and one on a Rocketfish 500W I had lying around. Right now, switching the clock to high makes no difference in the speed -- they all hover around 29.5GH/s, whether in high or low. When I set up the first one, it did overclock up into the mid 30's...  then I got distracted setting up the rest, and never really noticed when everything changed. Any ideas at all? I'm not sure where to start on this one.

Potentially might be limited by proxy or network topology limiting hash rate.

[hephaist0s]

I have 4 cubes running on 3 different power supplies, an Antec 1200, a Corsair CX750, and one on a Rocketfish 500W I had lying around. Right now, switching the clock to high makes no difference in the speed -- they all hover around 29.5GH/s, whether in high or low. When I set up the first one, it did overclock up into the mid 30's...  then I got distracted setting up the rest, and never really noticed when everything changed. Any ideas at all? I'm not sure where to start on this one.

Potentially might be limited by proxy or network topology limiting hash rate.

Ok! That's kind of the direction I was going, too. I did just add a switch in order to increase ports. But I'm lousy with networking, any further thoughts? Is having a switch bad? But I needed more ports.. ugh.

[edgie13]

I'm not having any luck getting mine set up either.  They'll start up and accept a few shares then they basically stop.  I've even swapped out PSU's to make sure that they're not an issue.

...

Did you ever fix your issue?  I'm running into the exact same thing.

His login details were wrong.

It was actually Chris_Sabian that had the login issues. lajz99 posted that his started working, but stop shortly after.

I have a bunch of v2 blades that are working just fine.  I even copy/pasted the config from a working blade to the cube, but still no dice. I've tried 3 different PSUs, different network cables, different network switches, bfgminer, slush's proxy, Windows 7, Ubuntu, and minepeon.

[hephaist0s]

I have 4 cubes running on 3 different power supplies, an Antec 1200, a Corsair CX750, and one on a Rocketfish 500W I had lying around. Right now, switching the clock to high makes no difference in the speed -- they all hover around 29.5GH/s, whether in high or low. When I set up the first one, it did overclock up into the mid 30's...  then I got distracted setting up the rest, and never really noticed when everything changed. Any ideas at all? I'm not sure where to start on this one.

Potentially might be limited by proxy or network topology limiting hash rate.

Ok! That's kind of the direction I was going, too. I did just add a switch in order to increase ports. But I'm lousy with networking, any further thoughts? Is having a switch bad? But I needed more ports.. ugh.

While I still don't know the exact cause of the network-wide slowdown, I moved the proxy server to another machine on the network and that fixed the problem. The cubes are running in the high 30s again.