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

[WBF1]

Dogie- any chance for an explanation of the somewhat cryptic options towards the bottom of the settings page?

Specifically

RollNtime[ s]
WakeUp period[ s]
Normal/Smart[0/1]

[1715023023]

1715023023

[1715023023]

1715023023

[R-T-B]

If you really think GHash.io is going to go down (along with 30% of the network), buy a $20 Pi and use BFGMiner as a passthrough proxy. You can set up as many redundant pools as you want for as many miners as you want if you set miner difficulty properly.

While the first part of your comment may be valid (it is unlikely GHash.io is going to go down), I refuse to use GHash.io.  I suppose you could argue that's me creating a problem, but the suggested solution of BFGMiner is simply not a solution.  BFGMiner consistently crashes my controller on a NIGHTLY basis.

[-ck]

While the first part of your comment may be valid (it is unlikely GHash.io is going to go down), I refuse to use GHash.io.  

ckproxy (linux only) can go some way to sanitising these connections. It will allow you to connect to btcguild and slush stratum based pools like slush and ozcoin that I know of but it wont talk to eligius and emc because of the part of the enonce2 that these devices expect to be working on (it's coincidence that ckproxy can sanitise the other pools).

https://bitbucket.org/ckolivas/ckpool

Run it in standalone proxy mode (with the options -p -A)

[Scripts]

Flawless Guide. Awesome work Dogie. This will definitely help others.

[dogie]

i can confirm this
thoose fans are very weak...
but they are cooling dragon lketc with no issues.

coolermaster jetflo are much better choice, but i think not enough too

i will test new fans hope they arrive today.

No sensible noise level 25mm fan will get close to a 37.5mm high static fan with out voodoo.

The corsairs do work as (expensive) second fans though.

[dogie]

While the first part of your comment may be valid (it is unlikely GHash.io is going to go down), I refuse to use GHash.io.  

ckproxy (linux only) can go some way to sanitising these connections. It will allow you to connect to btcguild and slush stratum based pools like slush and ozcoin that I know of but it wont talk to eligius and emc because of the part of the enonce2 that these devices expect to be working on (it's coincidence that ckproxy can sanitise the other pools).

https://bitbucket.org/ckolivas/ckpool

Run it in standalone proxy mode (with the options -p -A)

BTCGuild can be used directly with stratum-lb-usa48.btcguild.com.

[dogie]

Dogie- any chance for an explanation of the somewhat cryptic options towards the bottom of the settings page?

Specifically

RollNtime[ s]
WakeUp period[ s]
Normal/Smart[0/1]

RollNtime is something you will never change, and determines when the pool's packet should expire. I'm not sure why its manually set as I thought its sent as a bit with the work packet. https://en.bitcoin.it/wiki/Getwork#rollntime
Don't know the other two.

[old nice guy]

Great guide Dogie the pictures are amazing as always. Will you provide information on over clocking I see you have said it will achieve 920gh/s over clocked how stable is that?

Sometimes a good reboot is all that's needed.. Oh and a nice drop in ambient temperature.

https://dl.dropboxusercontent.com/s/d12qutzdhnho8k0/RECKONER2_2.png?dl=0

thats the highest GH ive seen so far and one board is having problems lol

Same here Steve , i just set up my units  can you tell me how i do the reboot on these Tubes?

[dogie]

Great guide Dogie the pictures are amazing as always. Will you provide information on over clocking I see you have said it will achieve 920gh/s over clocked how stable is that?

Not at this second, no. There are too many cases of unstable units at stock as it is, or damaged units with unknown cause. Most of these are caused by improper power supplies being used, and/or PCI-E cabling, and/or cooling. Once that settles down I will. The other problem is that the OC pushes the unit over the 1KW @ PSU threshold, so all sensible powering solutions go out the window. You end up having to go 2x CX750s per Tube or 4x RM1000s per 3 Tubes.

[dogie]

Same here Steve , i just set up my units  can you tell me how i do the reboot on these Tubes?

Stop quoting the picture, its not real stats [Only 7 minutes of stats with super high difficulty = random variance]. You can reset your stats with ReSession or restart your unit with Settings -> Update/Restart.

[helipotte]

Dogie- any chance for an explanation of the somewhat cryptic options towards the bottom of the settings page?

Specifically

RollNtime[ s]
WakeUp period[ s]
Normal/Smart[0/1]

RollNtime is something you will never change, and determines when the pool's packet should expire. I'm not sure why its manually set as I thought its sent as a bit with the work packet. https://en.bitcoin.it/wiki/Getwork#rollntime
Don't know the other two.

I changed the NRolltime to 150 from 120 and it has reduced the "duplicate share" notifications I was receiving in Bfgminer (proxy).  I was averaging

around 5% rejects due to dups over 24Hr period.  When I changed it to 150, this dropped to about 1.5%. Just an observation.

[WBF1]

Dogie- any chance for an explanation of the somewhat cryptic options towards the bottom of the settings page?

Specifically

RollNtime[ s]
WakeUp period[ s]
Normal/Smart[0/1]

RollNtime is something you will never change, and determines when the pool's packet should expire. I'm not sure why its manually set as I thought its sent as a bit with the work packet. https://en.bitcoin.it/wiki/Getwork#rollntime
Don't know the other two.

I changed the NRolltime to 150 from 120 and it has reduced the "duplicate share" notifications I was receiving in Bfgminer (proxy).  I was averaging

around 5% rejects due to dups over 24Hr period.  When I changed it to 150, this dropped to about 1.5%. Just an observation.

Had a similar issue at some point. Make sure you set diff=n in bfgminer, where n is a max difficulty presented to the miner. From whatbi can tell, without setting it the miner->proxy difficulty stays at 1 even when the proxy->pool difficulty is higher. Bfgminer seems to not pass on the pool difficulty unless you set a max in bfgminer, even if that max is super high. After I did this, my reject rate was comparable to without the proxy. I think the controller or the boards get tripped up at difficulty 1 after some period of time. But I have no hard proof of this, just an observation.

[helipotte]

I did.  Set Bfgminer to present 128 diff to the tubes.  Poolside is set to 1024.

[jmumich]

Has anyone had a problem where one board shows all "-" dashes in the Test Status page, and doesn't hash at all, while the other three are all working?  I've tried putting it in all four possible positions in the chain, and it seems one board (the same one) always comes up like this:  "Board 04: |---|---|---|---|---|---|---|---| - V:1.41"  

Since it is coming up I assume it is not an issue with how it is assembled (and I even plugged the fan into that board to see if it works, and it did) - could this just be a bad board, or am I missing something?  

[bbxx]

doa board.

but sometimes i got same result after longer time of hashing
software reboot wake them up.

[dogie]

Has anyone had a problem where one board shows all "-" dashes in the Test Status page, and doesn't hash at all, while the other three are all working?  I've tried putting it in all four possible positions in the chain, and it seems one board (the same one) always comes up like this:  "Board 04: |---|---|---|---|---|---|---|---| - V:1.41"  

Since it is coming up I assume it is not an issue with how it is assembled (and I even plugged the fan into that board to see if it works, and it did) - could this just be a bad board, or am I missing something?  

It will come up regardless of how well you assembled it (unless you sawed it in half). If its been bent or damaged then it will display this.

Try swapping the power round from another board.

[jmumich]

doa board.

but sometimes i got same result after longer time of hashing
software reboot wake them up.

It will come up regardless of how well you assembled it (unless you sawed it in half). If its been bent or damaged then it will display this.

Try swapping the power round from another board.

Thanks, looks like its a dead board, it never came alive so rebooting doesn't help, and I get the same result no matter where it is powered (1 or 2 PCI-E).  There's no obvious physical damage.  Is there any reason why I couldn't run 3 together and try to RMA the single board?  Seems like it should work, and I'd rather have 3 going than nothing at all.  And it should be easier to OC 3 with a 1000W power supply.

[dogie]

It will come up regardless of how well you assembled it (unless you sawed it in half). If its been bent or damaged then it will display this.

Try swapping the power round from another board.

Thanks, looks like its a dead board, it never came alive so rebooting doesn't help, and I get the same result no matter where it is powered (1 or 2 PCI-E).  There's no obvious physical damage.  Is there any reason why I couldn't run 3 together and try to RMA the single board?  Seems like it should work, and I'd rather have 3 going than nothing at all.  And it should be easier to OC 3 with a 1000W power supply.

You can send one back, yes.

[yochdog]

assuming 3 tubes can be plugged into one controller......does each tube need a individual IP, or just one for the controller? 

[WBF1]

assuming 3 tubes can be plugged into one controller......does each tube need a individual IP, or just one for the controller? 

One for the controller