Pages:
Author

Topic: Avalon ASIC users thread - page 33. (Read 438714 times)

newbie
Activity: 41
Merit: 0
January 05, 2014, 11:49:55 PM
I know some people have mentioned this in the past, but I can't seem to find it any more.

Is there a command or cgminer setup flag through the web UI to restart cgminer automatically after a certain interval? I find using the default router as a host I get quite a few instances (a couple times a day) where cgminer will restart but won't detect the Avalon modules, and has to restarted manually. Power cycling isn't necessary, I can just hit Save&Apply in the config tab to quickly restart cgminer again, and it picks up the controller and starts back up.

Under System|Scheduled Tasks is the Crontab. I found this doing a search... Hope it works

59 2, 15 * * * /sbin/shutdown -r +5

it should reboot at 2:59am and 3:59pm every day.
legendary
Activity: 1274
Merit: 1004
January 05, 2014, 10:49:00 PM
I know some people have mentioned this in the past, but I can't seem to find it any more.

Is there a command or cgminer setup flag through the web UI to restart cgminer automatically after a certain interval? I find using the default router as a host I get quite a few instances (a couple times a day) where cgminer will restart but won't detect the Avalon modules, and has to restarted manually. Power cycling isn't necessary, I can just hit Save&Apply in the config tab to quickly restart cgminer again, and it picks up the controller and starts back up.
sr. member
Activity: 378
Merit: 250
January 05, 2014, 09:23:38 PM

You would need to replace the entire module, the sub boards are fabricated at one large board which = 1 module.

I am in the same boat as you, my 29 sub board is dead for some reason even though there is nothing burnt on it.

There is a score between pairs of sub modules on some board versions.  I have used the heat sink as a support, and snapped the board at the score to remove a dead sub module pair.


Do you think you could have used a dermal tool to cut along the score. Snapping seems a little risky.
 

I love Dremel tools. 

I was very concerned how it would turn out.  That is why I placed it on the heat sink, as it is very flat and had a sharp machined edge at exactly the right place to match the score.  My board snapped completely cleanly before I had depressed the unsupported end by a finger width. 


The reason I ask is because I just bought a broken module from a fellow forum member.
I do not have a controller to use with the module.
I am going to try and separate the module into smaller units and build some Klondike controllers to interface the modules to USB.
sr. member
Activity: 403
Merit: 250
January 05, 2014, 09:18:44 PM
smoothrunnings,

Yep ... Tried everything you mentioned and nothing works.  I believe I "bricked" the TL-WR703N module.  The reason I think this is that is I observed the TL-WR703N board during power up the blue LED no longer comes on signifying an attempt at Ethernet physical layer connection (i.e OSI layer 1 attempt).  This previously was not the case as the blue LED ALWAYS came on and provided expected blink sequences during start-up and operation.  Oh THE HUMANITY!!!!  Do you (or anyone else for that matter) have any more ideas I might try?

H@shKraker
legendary
Activity: 1246
Merit: 1002
January 05, 2014, 09:16:23 PM

You would need to replace the entire module, the sub boards are fabricated at one large board which = 1 module.

I am in the same boat as you, my 29 sub board is dead for some reason even though there is nothing burnt on it.

There is a score between pairs of sub modules on some board versions.  I have used the heat sink as a support, and snapped the board at the score to remove a dead sub module pair.



Now that is what I like to hear. ProfMac, since you are talking about pairs, I take it there would be 4 subs per module each containing 2 subs. So my problem would be sub 3 and mod 4 correct? I guess I have to buy someone's bad board, break it down and start swapping them out to find the good subs on the purchased "bad board".

I think the modules are numbered from left to right, when the heatsink is away from you.  By that designation, your sub-modules 5 & 6 are the pair.  I do notice that no "diamond" pettern of connectors exists between your sub-modules 1&2.  On my boards, those conductors are there, and I did not want to cut them.  I can't tell from your photograph if your modules are scored between each sub-module, or between each pair.

Also, on my module, the holes in the board and the heat sink only matched up if I moved the PCB 2-sub-modules width.  Check your situation to see if you can get away with only dropping 1 submodule.

After this is all done, move the usable sub-modules all the way to the low numbered position, and re-arrange your modules so that the under-spec module is the nighest number (3 or 4).  Then, change the count in the configuration panel to 23 or 31, whichever is 1 smaller than now.

legendary
Activity: 1246
Merit: 1002
January 05, 2014, 09:06:44 PM

You would need to replace the entire module, the sub boards are fabricated at one large board which = 1 module.

I am in the same boat as you, my 29 sub board is dead for some reason even though there is nothing burnt on it.

There is a score between pairs of sub modules on some board versions.  I have used the heat sink as a support, and snapped the board at the score to remove a dead sub module pair.


Do you think you could have used a dermal tool to cut along the score. Snapping seems a little risky.
 

I love Dremel tools. 

I was very concerned how it would turn out.  That is why I placed it on the heat sink, as it is very flat and had a sharp machined edge at exactly the right place to match the score.  My board snapped completely cleanly before I had depressed the unsupported end by a finger width. 

hero member
Activity: 630
Merit: 501
January 05, 2014, 08:16:44 PM

You would need to replace the entire module, the sub boards are fabricated at one large board which = 1 module.

I am in the same boat as you, my 29 sub board is dead for some reason even though there is nothing burnt on it.

There is a score between pairs of sub modules on some board versions.  I have used the heat sink as a support, and snapped the board at the score to remove a dead sub module pair.


Do you think you could have used a dermal tool to cut along the score. Snapping seems a little risky.
 

Just take a chain saw to it! Wink

Kidding.
hero member
Activity: 630
Merit: 501
January 05, 2014, 08:15:40 PM
smoothrunnings,

I was upgrading FROM the 519 firmware to the most recent as indicated by Yifu's post.  I did try the reset button thing but that hasn't seemed to help.  The wiki doesn't indicate how long one should hold said button.  On the IP stuff No Wifi active on the Laptop and the laptop's IP was set to 192.168.0.22/24.  That should mean I can talk to a device listening on 192.168.0.100/24.  I just found the OpenWRT "Trigger fail-safe mode with button push during TL-WR703N startup" Wiki.  It says to hold the reset button during device startup.  Do you think that's the proper route to take to get back to factory default so I can then attempt to load the most recent firmware?

H@shKraker

Yes. I would have to used the latest firmware from CK, most Avy owners have had a better experience with CK's firmware.
sr. member
Activity: 378
Merit: 250
January 05, 2014, 08:02:23 PM

You would need to replace the entire module, the sub boards are fabricated at one large board which = 1 module.

I am in the same boat as you, my 29 sub board is dead for some reason even though there is nothing burnt on it.

There is a score between pairs of sub modules on some board versions.  I have used the heat sink as a support, and snapped the board at the score to remove a dead sub module pair.


Do you think you could have used a dermal tool to cut along the score. Snapping seems a little risky.
 
sr. member
Activity: 403
Merit: 250
January 05, 2014, 07:25:27 PM
smoothrunnings,

I was upgrading FROM the 519 firmware to the most recent as indicated by Yifu's post.  I did try the reset button thing but that hasn't seemed to help.  The wiki doesn't indicate how long one should hold said button.  On the IP stuff No Wifi active on the Laptop and the laptop's IP was set to 192.168.0.22/24.  That should mean I can talk to a device listening on 192.168.0.100/24.  I just found the OpenWRT "Trigger fail-safe mode with button push during TL-WR703N startup" Wiki.  It says to hold the reset button during device startup.  Do you think that's the proper route to take to get back to factory default so I can then attempt to load the most recent firmware?

H@shKraker
hero member
Activity: 630
Merit: 501
January 05, 2014, 07:04:31 PM

You would need to replace the entire module, the sub boards are fabricated at one large board which = 1 module.

I am in the same boat as you, my 29 sub board is dead for some reason even though there is nothing burnt on it.

There is a score between pairs of sub modules on some board versions.  I have used the heat sink as a support, and snapped the board at the score to remove a dead sub module pair.



The point is you can't simply buy a sub board for Avalon, you can only buy the entire module. Smiley And I doubt everyone here is as good as you when it comes to repairing electronics. In North America people throw away their electronics when they break and buy new ones, so getting them repaired cheaply is very hard. It would be different if we all lived in Colombia, they use and reuse electronics until they turn to dust! Smiley
newbie
Activity: 41
Merit: 0
January 05, 2014, 07:04:11 PM

You would need to replace the entire module, the sub boards are fabricated at one large board which = 1 module.

I am in the same boat as you, my 29 sub board is dead for some reason even though there is nothing burnt on it.

There is a score between pairs of sub modules on some board versions.  I have used the heat sink as a support, and snapped the board at the score to remove a dead sub module pair.



Now that is what I like to hear. ProfMac, since you are talking about pairs, I take it there would be 4 subs per module each containing 2 subs. So my problem would be sub 3 and mod 4 correct? I guess I have to buy someone's bad board, break it down and start swapping them out to find the good subs on the purchased "bad board".
hero member
Activity: 630
Merit: 501
January 05, 2014, 07:01:02 PM
Sooo .... I tried to upgrade the firmware on my Gen 1, 4module Avalon yesterday and I fear I may have bricked it.  During the upgrade the web page just sat there telling me i twas updating the firmware but it never sent me back to the login page.  After 10 or so minutes I power cycled the machine and now I can't connect to it (web nor ssh) on its assigned IP of 192.168.120.23.  I thought maybe the machine has reset itself to factory defaults so I reset my IP to 192.168.0.22, plugged the Avalon into my configuration port (where nothing else runs), plugged in my laptop to that same hub and tried to ssh and/or http to 192.168.0.100.  The machine did not reply to that either.  Does anyone know how I might determine A) if the Avalon's Ethernet port is listening and if so B) what IP it has chosen.

H@shKraker

Press the rest button on the back, you will see a hole on the back to the right of the RJ45 connector, if you have the lid off on it still you can see button on the network card and the hole.

You didn't mention which firmware you use to upgrade your avay and if you un-selected the "keep settings" or not. If you did un-select keep settings after you hit confirm you need to wait first and for most for the avy to reboot, you should be able to tell this by now as everything on it will go quiet for a couple of seconds before the fans start back up.

Then you need to change the IP address on your computer to 192.168.0.10/sub: 255.255.255.0, GW and DNS blank, and connect to the Avy's RJ45 port and surf to 192.168.0.100, if you can resolved it, open a command prompt and type: ping 192.168.0.100 if you don't get a response check your machines IP address, type ipconfig and check to see its IP address is 192.168.0.10.

If you are using a laptop to connect and you have wifi in your house or building, disable the wifi on your laptop before you attemp to connect to the Avy through the lan cable or your computer will use the wifi as the default connection for all traffic.

sr. member
Activity: 403
Merit: 250
January 05, 2014, 06:47:33 PM
Sooo .... I tried to upgrade the firmware on my Gen 1, 4module Avalon yesterday and I fear I may have bricked it.  During the upgrade the web page just sat there telling me i twas updating the firmware but it never sent me back to the login page.  After 10 or so minutes I power cycled the machine and now I can't connect to it (web nor ssh) on its assigned IP of 192.168.120.23.  I thought maybe the machine has reset itself to factory defaults so I reset my IP to 192.168.0.22, plugged the Avalon into my configuration port (where nothing else runs), plugged in my laptop to that same hub and tried to ssh and/or http to 192.168.0.100.  The machine did not reply to that either.  Does anyone know how I might determine A) if the Avalon's Ethernet port is listening and if so B) what IP it has chosen.

H@shKraker
legendary
Activity: 1246
Merit: 1002
January 05, 2014, 06:08:23 PM

You would need to replace the entire module, the sub boards are fabricated at one large board which = 1 module.

I am in the same boat as you, my 29 sub board is dead for some reason even though there is nothing burnt on it.

There is a score between pairs of sub modules on some board versions.  I have used the heat sink as a support, and snapped the board at the score to remove a dead sub module pair.

member
Activity: 77
Merit: 10
January 05, 2014, 05:46:55 PM
pleas help. All I did was move my avalon mini down stairs to the basement and now the ftdi device discconects itself from the ttyusb port. see log below. How can i fix this? it was working fine and now not =( anyone can help me to get this thing back up? several reboots and this is how the log looks everytime now.

   11.700000] ftdi_sio 1-1.1:1.0: FTDI USB Serial Device converter detected
[   11.710000] usb 1-1.1: Detected FT232RL
[   11.710000] usb 1-1.1: Number of endpoints 2
[   11.720000] usb 1-1.1: Endpoint 1 MaxPacketSize 16384
[   11.720000] usb 1-1.1: Endpoint 2 MaxPacketSize 16384
[   11.730000] usb 1-1.1: Setting MaxPacketSize 64
[   11.730000] usb 1-1.1: FTDI USB Serial Device converter now attached to ttyUSB0
[   15.050000] device eth0 entered promiscuous mode
[   15.200000] eth0: link up (100Mbps/Full duplex)
[   15.210000] br-lan: port 1(eth0) entered forwarding state
[   15.210000] br-lan: port 1(eth0) entered forwarding state
[   17.210000] br-lan: port 1(eth0) entered forwarding state
[   22.260000] wlan0: authenticate with 08:60:6e:20:65:f0
[   22.300000] wlan0: send auth to 08:60:6e:20:65:f0 (try 1/3)
[   22.300000] wlan0: authenticated
[   22.320000] wlan0: associate with 08:60:6e:20:65:f0 (try 1/3)
[   22.320000] br-lan: received packet on eth0 with own address as source address
[   22.330000] wlan0: RX AssocResp from 08:60:6e:20:65:f0 (capab=0x411 status=0 aid=1)
[   22.330000] wlan0: associated
[   22.710000] br-lan: received packet on eth0 with own address as source address
[   22.720000] br-lan: received packet on eth0 with own address as source address
[   29.190000] ftdi_sio ttyUSB0: FTDI USB Serial Device converter now disconnected from ttyUSB0
[   29.210000] ftdi_sio 1-1.1:1.0: device disconnected
newbie
Activity: 41
Merit: 0
January 05, 2014, 05:39:20 PM
Got a weird issue...maybe. Yesterday I woke up and noticed the the miner had failed over and hung so I did a hard bounce on the box and it came up clean. But when I looked at the cgminer api log, I noticed this and figured that a couple of chips failed because it never changed. The box had run for 3 days straight flawlessly before that. A search turned up that these get flaky if they are not rebooted regularly. So I did another search for reboots and found that it is done using cron. So i did some more searching and setup a job to bounce every 12hrs.  This morning cgminer status shows the reboot occurred on time but now the 2 failed chips show some work...all be it way behind the others and every once in awhile they will increase buy a few points. Any one that knows these boxes well know what is going on? Are the chip eventually going to completely fail? Can the cards be replaced? My hash rate has dropped by 10G.

[match_work_count27] => 18
[match_work_count28] => 18
[match_work_count29] => 0
[match_work_count30] => 0
[match_work_count31] => 18
[match_work_count32] => 18

[match_work_count28] => 19143
[match_work_count29] => 444
[match_work_count30] => 381
[match_work_count31] => 19006

Also see this...is this a good reading?

[match_work_count32] => 24067
[Controller Version] => 2000000000

Thanks ahead for looking.

I see from from posts in hear that modules do drop out and it looks like workers 29 and 30 are the common failed modules. So I guess I can replace them if I find working mods for sale. I also noticed that Utility is only at 7...I take it that is not right?

Elapsed      10h 34m 12s
MHSav      105,267
MHS5sec      106,038
FoundBlocks   0   
Getworks      2,830
Accepted      4,664
Rejected      13
HW              28,669
Utility              7
Discarded      5,621
Stale              0
GetFailures           0
LocalWork           1,023,906
RemoteFailures   0
NetworkBlocks   89   
TotalMH      4.005594e+09
WU              1,470
DiffA              917,144
DiffR              2,328
DiffS              0
BestShare           366,995

I could really use some help...I hope I haven't broken some kind of protocol by asking these questions.

You would need to replace the entire module, the sub boards are fabricated at one large board which = 1 module.

I am in the same boat as you, my 29 sub board is dead for some reason even though there is nothing burnt on it.

Thank you for the reply. I made a mistake of not reporting that I have a Avalon 1 Batch3/4 miner. From the pics I have seen, each mod looks to comprise 8 subs. I thought I might be able to replace the 2 bad sub-modules.

http://thegenesisblock.com/avalon-shipping-39th-of-bitcoin-miners-600-units-in-batch-3-begin-arriving/#!prettyPhoto/0/


Your picture shows 1 module with 8 subs. The subs are manufactured as 1 board and are not cut into 8 separate boards. So if sub 29 and 30 are damaged you will have to replaced the entire module.


Thanks for the info, I never looked inside. As soon as I got it, it went right to work and the pic looks like it is modular. Poor design imo...if they were modular, you could add extra income to your product line by sell the sub-modules. Also if the system could be designed plug and play, you could start swapping out the bad units with upgraded tech...extending the life of the miner.

Oh well, thanks for taking the time to educate this 'ol dog.
sr. member
Activity: 351
Merit: 250
January 05, 2014, 05:28:07 PM
I'm Back Avalon took a shit any ideas?

Sat Jan  4 21:02:15 2014 auth.emerg kernel: [10065.420000] usb 1-1: clear tt 1 (9031) error -71

The tplink router in the avalons is really flaky with wifi enabled and this is the symptom of it. Plug it in via ethernet only and disable all wifi.

I turned it off and on last night. It would last a few hours and die again. Just disabled wifi and flashed to the latest firmware which is something I recognize the original must have been ancient 20130723. Will leave it on eligius so I can see if holds together. Thanks for the suggestions will post in a few days if it worked, sooner if it doesn't.
hero member
Activity: 630
Merit: 501
January 05, 2014, 05:23:13 PM
Got a weird issue...maybe. Yesterday I woke up and noticed the the miner had failed over and hung so I did a hard bounce on the box and it came up clean. But when I looked at the cgminer api log, I noticed this and figured that a couple of chips failed because it never changed. The box had run for 3 days straight flawlessly before that. A search turned up that these get flaky if they are not rebooted regularly. So I did another search for reboots and found that it is done using cron. So i did some more searching and setup a job to bounce every 12hrs.  This morning cgminer status shows the reboot occurred on time but now the 2 failed chips show some work...all be it way behind the others and every once in awhile they will increase buy a few points. Any one that knows these boxes well know what is going on? Are the chip eventually going to completely fail? Can the cards be replaced? My hash rate has dropped by 10G.

[match_work_count27] => 18
[match_work_count28] => 18
[match_work_count29] => 0
[match_work_count30] => 0
[match_work_count31] => 18
[match_work_count32] => 18

[match_work_count28] => 19143
[match_work_count29] => 444
[match_work_count30] => 381
[match_work_count31] => 19006

Also see this...is this a good reading?

[match_work_count32] => 24067
[Controller Version] => 2000000000

Thanks ahead for looking.

I see from from posts in hear that modules do drop out and it looks like workers 29 and 30 are the common failed modules. So I guess I can replace them if I find working mods for sale. I also noticed that Utility is only at 7...I take it that is not right?

Elapsed      10h 34m 12s
MHSav      105,267
MHS5sec      106,038
FoundBlocks   0   
Getworks      2,830
Accepted      4,664
Rejected      13
HW              28,669
Utility              7
Discarded      5,621
Stale              0
GetFailures           0
LocalWork           1,023,906
RemoteFailures   0
NetworkBlocks   89   
TotalMH      4.005594e+09
WU              1,470
DiffA              917,144
DiffR              2,328
DiffS              0
BestShare           366,995

I could really use some help...I hope I haven't broken some kind of protocol by asking these questions.

You would need to replace the entire module, the sub boards are fabricated at one large board which = 1 module.

I am in the same boat as you, my 29 sub board is dead for some reason even though there is nothing burnt on it.

Thank you for the reply. I made a mistake of not reporting that I have a Avalon 1 Batch3/4 miner. From the pics I have seen, each mod looks to comprise 8 subs. I thought I might be able to replace the 2 bad sub-modules.

http://thegenesisblock.com/avalon-shipping-39th-of-bitcoin-miners-600-units-in-batch-3-begin-arriving/#!prettyPhoto/0/


Your picture shows 1 module with 8 subs. The subs are manufactured as 1 board and are not cut into 8 separate boards. So if sub 29 and 30 are damaged you will have to replaced the entire module.
newbie
Activity: 41
Merit: 0
January 05, 2014, 05:17:46 PM
Got a weird issue...maybe. Yesterday I woke up and noticed the the miner had failed over and hung so I did a hard bounce on the box and it came up clean. But when I looked at the cgminer api log, I noticed this and figured that a couple of chips failed because it never changed. The box had run for 3 days straight flawlessly before that. A search turned up that these get flaky if they are not rebooted regularly. So I did another search for reboots and found that it is done using cron. So i did some more searching and setup a job to bounce every 12hrs.  This morning cgminer status shows the reboot occurred on time but now the 2 failed chips show some work...all be it way behind the others and every once in awhile they will increase buy a few points. Any one that knows these boxes well know what is going on? Are the chip eventually going to completely fail? Can the cards be replaced? My hash rate has dropped by 10G.

[match_work_count27] => 18
[match_work_count28] => 18
[match_work_count29] => 0
[match_work_count30] => 0
[match_work_count31] => 18
[match_work_count32] => 18

[match_work_count28] => 19143
[match_work_count29] => 444
[match_work_count30] => 381
[match_work_count31] => 19006

Also see this...is this a good reading?

[match_work_count32] => 24067
[Controller Version] => 2000000000

Thanks ahead for looking.

I see from from posts in hear that modules do drop out and it looks like workers 29 and 30 are the common failed modules. So I guess I can replace them if I find working mods for sale. I also noticed that Utility is only at 7...I take it that is not right?

Elapsed      10h 34m 12s
MHSav      105,267
MHS5sec      106,038
FoundBlocks   0   
Getworks      2,830
Accepted      4,664
Rejected      13
HW              28,669
Utility              7
Discarded      5,621
Stale              0
GetFailures           0
LocalWork           1,023,906
RemoteFailures   0
NetworkBlocks   89   
TotalMH      4.005594e+09
WU              1,470
DiffA              917,144
DiffR              2,328
DiffS              0
BestShare           366,995

I could really use some help...I hope I haven't broken some kind of protocol by asking these questions.

You would need to replace the entire module, the sub boards are fabricated at one large board which = 1 module.

I am in the same boat as you, my 29 sub board is dead for some reason even though there is nothing burnt on it.

Thank you for the reply. I made a mistake of not reporting that I have a Avalon 1 Batch3/4 miner. From the pics I have seen, each mod looks to comprise 8 subs. I thought I might be able to replace the 2 bad sub-modules.

http://thegenesisblock.com/avalon-shipping-39th-of-bitcoin-miners-600-units-in-batch-3-begin-arriving/#!prettyPhoto/0/
Pages:
Jump to: