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

[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.

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.

Yep, mysterious network topology . Sometimes its the switches playing bad and not liking the frequency they get spammed with getwork packets, regardless of how small.

[dmcdad]

I just had the 30 amp fuse blow on my Cube.

Faulty fuse, faulty cube, or does the fuse just need to be higher rated than 30A for high clock setting?

[Swimmer63]

I just had the 30 amp fuse blow on my Cube.

Faulty fuse, faulty cube, or does the fuse just need to be higher rated than 30A for high clock setting?

I would bet first on faulty fuse.  I have run mine overclocked for 3 days with no issues.  Plus, putting a higher fuse in could damage the miner.

[dmcdad]

I just had the 30 amp fuse blow on my Cube.

Faulty fuse, faulty cube, or does the fuse just need to be higher rated than 30A for high clock setting?

I would bet first on faulty fuse.  I have run mine overclocked for 3 days with no issues.  Plus, putting a higher fuse in could damage the miner.

Note to folks: the fuse burned out very slowly which caused the arc to melt the fuse casing which slightly welded the fuse into the fuse holder. I got it out, but it was a pain.

I think I'm going to replace it with an ACO (closed on bottom versus open) fast fuse instead:

www.amazon.com/Bussmann-VP-ATC-30-Acting-Blade/dp/B000AZ7SW8/

[bcollan]

I have one cube hashing 32ghs on low  and one cube which worked at first now showing all x's and blinking an error code on the red led. Both are powered from a Corsair TX850M.
Any ideas on reviving the dead one?

[sidehack]

The fuse just blew in one of mine yesterday. There was actually an internal issue where the freewheel diode on the internal relay's coil failed (possibly what took out the fuse, not sure) but without it the relay coil arced over and shorted some turns - still clicked the relay, but allowed about 12 times the normal current through the relay drive transistor which fried it. Back up and running now, only slightly jerry-rigged. So I'd recommend before replacing the fuse, to give the board a once-over and make sure none of your parts have craters in them just in case.

[edgie13]

So, I've noticed something else happening to my cube.  It reboots at exactly 2 minutes.   It is also constantly switch between pool A and pool B(same address).  I will post my config below:

Cube config (only one mini blade is plugged in right now):



proxy showing work requested and even some shares accepted:



proxy's ip:



accepted share on pool stats page:



I'm sure it would be fine if it wasn't rebooting exactly every 2 minutes.  Does anybody have a user's manual for the cube?  Maybe something that would explain the 2 minute reboot.  I'm going to buy the psu that the op recommended to see if that fixes this. Thanks.

[CanaryInTheMine]

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.

Yep, mysterious network topology . Sometimes its the switches playing bad and not liking the frequency they get spammed with getwork packets, regardless of how small.

afaik the network interface on blades and cubes is 10 MB.  some network equipment may not deal with this properly.  if it's a managed switch, try setting the port config to 10MB.  however, a good switch should autosense to 10 MB.  I do not know if they are half or full duplex.

[CanaryInTheMine]

dogie,

what do you think about modifying the directions to show how to add a secondary IP to ones computer that would participate on the 192.168.1.x /24 network instead of folks having to re-IP their whole setup?
one can add a secondary IP to their NIC and then be able to reach a blade or a cube and then they can change miner's addy to join an existing network...
this may be easier?

edit: I meant to say 192.168.1.x, not 192.168.254.x

[dogie]

So, I've noticed something else happening to my cube.  It reboots at exactly 2 minutes.   It is also constantly switch between pool A and pool B(same address).  I will post my config below:

I'm sure it would be fine if it wasn't rebooting exactly every 2 minutes.  Does anybody have a user's manual for the cube?  Maybe something that would explain the 2 minute reboot.  I'm going to buy the psu that the op recommended to see if that fixes this. Thanks.

Your blade and stratum proxy aren't playing nice, hence 'long poll inactive'.

dogie,

what do you think about modifying the directions to show how to add a secondary IP to ones computer that would participate on the 192.168.254.x /24 network instead of folks having to re-IP their whole setup?
one can add a secondary IP to their NIC and then be able to reach a blade or a cube and then they can change miner's addy to join an existing network...
this may be easier?

I specifically avoid things like this in guides as the people who need the guide are looking for the simple solution. In this case that is changing router IP which they are relatively familiar with. Most people run on 192.168.1 as default anyway.

[lajz99]

My problem was using an Apple Airport Extreme router with an Airport Express...once I switched to a Netgear router all was well!  Happily hashing around 37-38gh/s on each of the cubes running through bfgminer/cgwatcher.

[CanaryInTheMine]

So, I've noticed something else happening to my cube.  It reboots at exactly 2 minutes.   It is also constantly switch between pool A and pool B(same address).  I will post my config below:

I'm sure it would be fine if it wasn't rebooting exactly every 2 minutes.  Does anybody have a user's manual for the cube?  Maybe something that would explain the 2 minute reboot.  I'm going to buy the psu that the op recommended to see if that fixes this. Thanks.

Your blade and stratum proxy aren't playing nice, hence 'long poll inactive'.

dogie,

what do you think about modifying the directions to show how to add a secondary IP to ones computer that would participate on the 192.168.254.x /24 network instead of folks having to re-IP their whole setup?
one can add a secondary IP to their NIC and then be able to reach a blade or a cube and then they can change miner's addy to join an existing network...
this may be easier?

I specifically avoid things like this in guides as the people who need the guide are looking for the simple solution. In this case that is changing router IP which they are relatively familiar with. Most people run on 192.168.1 as default anyway.

most people don't run that by default actually... depends on equipment and the ISP.  maybe it's more geographic.  but if anyone gets an issue with their internet service in the US for ex. (can't speak for the rest of the world however), one of the provider's favorite thing to do is to instruct the user on how to reset the router (after rebooting doesn't work). so that would be back to original IP config on the router/modem...

either approach works though.  thanks for making these guides!

[dogie]

So, I've noticed something else happening to my cube.  It reboots at exactly 2 minutes.   It is also constantly switch between pool A and pool B(same address).  I will post my config below:

I'm sure it would be fine if it wasn't rebooting exactly every 2 minutes.  Does anybody have a user's manual for the cube?  Maybe something that would explain the 2 minute reboot.  I'm going to buy the psu that the op recommended to see if that fixes this. Thanks.

Your blade and stratum proxy aren't playing nice, hence 'long poll inactive'.

dogie,

what do you think about modifying the directions to show how to add a secondary IP to ones computer that would participate on the 192.168.254.x /24 network instead of folks having to re-IP their whole setup?
one can add a secondary IP to their NIC and then be able to reach a blade or a cube and then they can change miner's addy to join an existing network...
this may be easier?

I specifically avoid things like this in guides as the people who need the guide are looking for the simple solution. In this case that is changing router IP which they are relatively familiar with. Most people run on 192.168.1 as default anyway.

most people don't run that by default actually... depends on equipment and the ISP.  maybe it's more geographic.  but if anyone gets an issue with their internet service in the US for ex. (can't speak for the rest of the world however), one of the provider's favorite thing to do is to instruct the user on how to reset the router (after rebooting doesn't work). so that would be back to original IP config on the router/modem...

either approach works though.  thanks for making these guides!

Thanks for your thanks. People only need to change their network stuff ONCE to get access to the blade to be able to change the IP and subnet of the blade to whatever they choose so either way its only about 5 minutes work

[edgie13]

So, I've noticed something else happening to my cube.  It reboots at exactly 2 minutes.   It is also constantly switch between pool A and pool B(same address).  I will post my config below:

I'm sure it would be fine if it wasn't rebooting exactly every 2 minutes.  Does anybody have a user's manual for the cube?  Maybe something that would explain the 2 minute reboot.  I'm going to buy the psu that the op recommended to see if that fixes this. Thanks.

Your blade and stratum proxy aren't playing nice, hence 'long poll inactive'.

I've tried 3 different computers with three different operating systems.  I've tried slush's proxy and bfgminer on 2 of those and only bfgminer on the third.  I've even taken everything off the network switch and connected it to a hub.  My thinking is that the hub being "dumb", that it will let any traffic pass.  ...still nothing.

Why would the stratum proxy cause the cube to reboot at exactly 2 minutes?  Thank you for all your help!!

[bcollan]

I have one cube hashing 32ghs on low  and one cube which worked at first now showing all x's and blinking an error code on the red led. Both are powered from a Corsair TX850M.
Any ideas on reviving the dead one?

Any suggestions on this one, or does it need to go back to canary?

[glendall]

Haven't had any luck getting my Canary cube going as well

I have no 'x' s, chips all o's

My stratum proxy seems fine but the cube can't process any work. I get endless "[date time here] INFO proxy client_service.handle_event # New job 1386190 629 15775 for prevhash 142637ca, clean_jobs=False" messages from my stratum proxy.

Very occasionally it'll end instead with clean_jobs=True.

Any ideas or suggestions? I've tried running another stratum proxy on a different computer on my network and same thing, the stratum proxy connects to the pool but never any mention of work /hashes being done.  I don't think it's a matter of incorrect pool credentials either because the second pool I set up is with Eligius which accepts any name password combo.

I turned off my pc software fire wall, didn't seem to help. Could it be a router based firewall problem maybe?  
The cube was a bit beat up, some of the boards where not lined up in the slots, (like the PCBs were out of the grooves likely indicating it was tossed around a bit) but opened it up and put the boards in the proper place and there is no physical damage besides that anyways.

[edgie13]

Haven't had any luck getting my Canary cube going as well

I have no 'x' s, chips all o's

My stratum proxy seems fine but the cube can't process any work. I get endless "[date time here] INFO proxy client_service.handle_event # New job 1386190 629 15775 for prevhash 142637ca, clean_jobs=True" messages from my stratum proxy.

Very occasionally it'll end instead with clean_jobs=True.

Any ideas or suggestions? I've tried running another stratum proxy on a different computer on my network and same thing, the stratum proxy connects to the pool but never any mention of work /hashes being done.  I don't think it's a matter of incorrect pool credentials either because the second pool I set up is with Eligius which accepts any name password combo.

I turned off my pc software fire wall, didn't seem to help. Could it be a router based firewall problem maybe?  
The cube was a beat beat up, some of the boards where not lined up in my slot, but opened it up and put the boards in the proper place and there is no physical damage besides that anyways.

I'm having that same issue.  My cube arrived with 2 screws completely out of the heatsinks.  The heatsink was kinda of flapping around. I had to gently tighten all the heatsink screws as they were all some level of loose. It did have a long journey from where it originally came from so I really wasn't too suprised to see it this way.  I just wish that it would hash. 

[glendall]

Well I wish you luck Edge...hopefully we can get these going. I'm pretty poor these days so if I lost a cube that'd really suck. Especially after being losing money in the WeExchange scam. Anyways though, optimistic can get this going.

BTW for your screen shot you show the ports as 8338. Should these not be 8333 ? You prob know what you are doing here and I don't think their would be any response at all if the ports were wrong but thought I'd mention it.


edit: WORKING  ! Praise the cyber gods.  I have no idea what happened. It was like the 100th try reset, refed all the same values in for the 20th try, and just on the last restart i'm getting hashes. Happy if somewhat confused.

[edgie13]

Well I wish you luck Edge...hopefully we can get these going. I'm pretty poor these days so if I lost a cube that'd really suck. Especially after being losing money in the WeExchange scam. Anyways though, optimistic can get this going.

BTW for your screen shot you show the ports as 8338. Should these not be 8333 ? You prob know what you are doing here and I don't think their would be any response at all if the ports were wrong but thought I'd mention it.

I changed it on both the cube and the proxy.  I had already tried the default a few times.

[glendall]

Well I wish you luck Edge...hopefully we can get these going. I'm pretty poor these days so if I lost a cube that'd really suck. Especially after being losing money in the WeExchange scam. Anyways though, optimistic can get this going.

BTW for your screen shot you show the ports as 8338. Should these not be 8333 ? You prob know what you are doing here and I don't think their would be any response at all if the ports were wrong but thought I'd mention it.

I changed it on both the cube and the proxy.  I had already tried the default a few times.

As said dunno what triggered it working now instead of the 100 other times I tried. Maybe might want to try starting/restarting the miners at different times then you start your mining proxy program.  Like I for example I was trying the cube first and then relaunching the stratum mining proxy, whereas before I was doing this in reverse. Again logically I doubt this would make a difference, maybe it was pure luck,  but all I can say is now a hundred false starts later and it is hashing.