Official Avalon Technical Support Thread

[Humax]

Did you try 05-19 firmware. I remember I too tried the latest (at least latest on the wiki) and it completely borked the unit where it stopped hashing after 1 min (this was on Batch 1 unit but that should not matter).

Another thing you can look into is the F1 or F2 fuse issue (if your unit was early batch 2). Sorry I can't give you specifics which fuse it is as the avalon wiki is offline at the moment, but I do remember that if you do have the fuse in place you need to disable it.

Thank you allinvain.

It's batch 2 and I got it yesterday. I haven't tried 05-19 firmware and I'll certainly going to look into the F1 and F2 fuse issue. I am also thinking of changing the psu. Those things can also cause lots of trouble when they're not working perfectly. And I'd like to have more room to spare. It came with a 750W Enermax psu and I'd like to try to swap it for a 1000W+ Gold psu, leaving room for 4th unit and/or overclocking.
My stable unit already draws 733W at frequency of 336.

Wow shmadz! You've been busy!  
I keep my fingers crossed for you and hope it will fix your problem.
To be honest, I don't get that much error like you

edit: I think you don't need to worry about the box below. That's just log for accessing your avalon.

[allinvain]

Did you try 05-19 firmware. I remember I too tried the latest (at least latest on the wiki) and it completely borked the unit where it stopped hashing after 1 min (this was on Batch 1 unit but that should not matter).

Another thing you can look into is the F1 or F2 fuse issue (if your unit was early batch 2). Sorry I can't give you specifics which fuse it is as the avalon wiki is offline at the moment, but I do remember that if you do have the fuse in place you need to disable it.

Thank you allinvain.

It's batch 2 and I got it yesterday. I haven't tried 05-19 firmware and I'll certainly going to look into the F1 and F2 fuse issue. I am also thinking of changing the psu. Those things can also cause lots of trouble when they're not working perfectly. And I'd like to have more room to spare. It came with a 750W Enermax psu and I'd like to try to swap it for a 1000W+ Gold psu, leaving room for 4th unit and/or overclocking.
My stable unit already draws 733W at frequency of 336.

Wow shmadz! You've been busy!  
I keep my fingers crossed for you and hope it will fix your problem.
To be honest, I don't get that much error like you

edit: I think you don't need to worry about the box below. That's just log for accessing your avalon.

At least give 05-19 firmware a try first before anything. Then the next step I would look into is the fuse issue. Batch 2 units shipped after a certain date should already have that fuse removed (check the wiki).

Hmm, 750 watt eh. I only opened one of my units and it comes with a 850 watt PSU with some Chinese characters on the side but I think it's still an Enermax unit. I am tempted to swap all the power supplies with Corsair AX860 as I don't intend to overclock or to expand to 4 modules (existing power setup won't handle it and definitely not the cooling). One of my PSUs is running hotter than the rest but that may be due to it's placement but who knows it may be a slightly weaker one.

[-ck]

I've just tried to re-flash the 0627 firmware with "keep-settings" disabled - turned off the WIFI (removed the interface in the GUI) and I tried running with Con's donation pool settings (maybe there's something wrong with the way I set up my pools?) -- he deserves it anyways. -- been running half an hour now, got my fingers crossed!

:UPDATE: running over an hour now, Humax's comment below gave me the idea to try turning system log to "warning" to remove the news.info logs, didn't work, maybe needs reboot to apply?
 --- new update, just restarted again, right after I added my bitminter pool... I'm going to try to run it to ozcoin with similar settings to what Con was using and see how that goes 

Thank you very much. I definitely saw a spike in my donation hashrate for that hour, so it should be mining fine for you now.

[shmadz]

I've just tried to re-flash the 0627 firmware with "keep-settings" disabled - turned off the WIFI (removed the interface in the GUI) and I tried running with Con's donation pool settings (maybe there's something wrong with the way I set up my pools?) -- he deserves it anyways. -- been running half an hour now, got my fingers crossed!

:UPDATE: running over an hour now, Humax's comment below gave me the idea to try turning system log to "warning" to remove the news.info logs, didn't work, maybe needs reboot to apply?
 --- new update, just restarted again, right after I added my bitminter pool... I'm going to try to run it to ozcoin with similar settings to what Con was using and see how that goes 

Thank you very much. I definitely saw a spike in my donation hashrate for that hour, so it should be mining fine for you now.

No, seriously, THANK YOU!

I mimicked the settings you had in the defaults (including the 56 LSD setting on my ozcoin account) and it's been almost 2 hours without a restart *and I have set to 300 now (was on 282 while trying to troubleshoot)



On another note, is it possible to change the way balancing works? For example, if you could set 3 or 4 different pools, set it to balance, and then set an arbitrary limit to one or more of those pools... this would allow me to send 1 or 2 GH/s your way, while still mining at 25-30 or so on my own pools...

Just a suggestion   

[Humax]

F1 fuse is not on the board! So much for that little chance of fixing the problem.
Flashing to 05-19 firmware has made it impossible to log on. I can see Lucy for half a second about 1 of 10 attempts to login.
Before, I could login about 50% of the times, but now not even once and I've been trying for over half hour!

I'm going to try different psu in a few days. All stores are closed now for the weekend.

If new psu won't help, then I will have to send it back to China. Really not looking forward to that.... imagine how long it will take to get another unit....  

...and I will also try Con's donation pool settings 

[shmadz]

F1 fuse is not on the board! So much for that little chance of fixing the problem.
Flashing to 05-19 firmware has made it impossible to log on. I can see Lucy for half a second about 1 of 10 attempts to login.
Before, I could login about 50% of the times, but now not even once and I've been trying for over half hour!

I'm going to try different psu in a few days. All stores are closed now for the weekend.

If new psu won't help, then I will have to send it back to China. Really not looking forward to that.... imagine how long it will take to get another unit....  

...and I will also try Con's donation pool settings  

Hi Humax, just to clarify. I understand you have 2 avalons (I'm assuming both are the black model with the F1 fuse removed) -- one works fine right out of the box, and the other was restarting all the time? please let me know if I'm wrong in my assumptions.

I have seen the Lucy stuff when trying to connect on the previous IP address after the device defaulted back to 192.168.0.100

One thing to check, turn the device on, try to ping the original address as well as the default address and see if you get a response.. if you do get a response from some address, unpulg the avalon and try to ping that same address, if you still get a response, you may have a network conflict.

-edit- to be certain if this is the case, connect to the device with a laptop via a crossover cable and try to ping or connect...

(*my apologies if this is too basic and you've already tried*)

[ProfMac]

One thing to check, turn the device on, try to ping the original address as well as the default address and see if you get a response.. if you do get a response from some address, unpulg the avalon and try to ping that same address, if you still get a response, you may have a network conflict.

(*my apologies if this is too basic and you've already tried*)

arp  <-- also a useful utility on both windows and linux

[allinvain]

F1 fuse is not on the board! So much for that little chance of fixing the problem.
Flashing to 05-19 firmware has made it impossible to log on. I can see Lucy for half a second about 1 of 10 attempts to login.
Before, I could login about 50% of the times, but now not even once and I've been trying for over half hour!

I'm going to try different psu in a few days. All stores are closed now for the weekend.

If new psu won't help, then I will have to send it back to China. Really not looking forward to that.... imagine how long it will take to get another unit....  

...and I will also try Con's donation pool settings  

Maybe you forgot to checkmark the "keep settings" when you flashed to 05-19? If so it reverts to default settings which as another poster mentioned is IP address 192.168.0.100. The firmware update did complete successfully though, right? I hope it did not take a long time and you power cycled the unit like I did, cause if you did I'm afraid you may have bricked the TP-LINK router. Don't mean to scare you though.

It could be the PSU as well. Go for a high quality unit with at least 800 watts and one single 12V rail.

Edit: I forgot to mention, can you try to ping your avalon unit so see if there is any packet loss happening. If so replace the ethernet cord first. If persists, perhaps the port on your router/switch is bad. If problem persists then maybe the tp-link router's ethernet port is flaky. You could try mining with it directly attached to a computer or raspberry pi if you're into having some fun. Careful though as you have to remove the J1 jumper on the control board (fgpa control board) if you plan to connect it directly to a computer/other device.

[Humax]

My greatest gratitude to everybody trying to help!

I appreciate all your input.
Sofar, I have already done all the things you guys mentioned here and almost everything I could come up with.
I want to wait for a new psu before plunging myself into solving this problem again.
Meanwhile, I try to clear my head and take a break.
Sometimes, things just miraculously will be fine after a break. 

[cypherdoc]

once i configure cgminer, save & apply, the wiki says to restart cgminer to initiate mining.  how do i do that?

[Humax]

Since you have one working unit there is a way to connect the modules of your non-working unit to the working unit so that all the modules are hashing through 1 unit.   I have not done it myself so I am not sure how it is done.  I would make the working unit a 4-module unit and then run the other 2 in the other box if you can.  I believe it is done by using a USB splitter to bypass the TPlink card.  One guy reported he got it to work but the fans would not spin so the fans have to powered directly.

EDIT:  Also, there is a guy selling TP Link cards burned for Avalon.  (I think it is in this thread somewhere).  I would try hooking things up this way and order a new card rather than send the whole thing back to China just for a TP Link card.  Or you can do the Raspberry Pi thing that some people have done.

Thanks. I'll keep this in mind.

Meanwhile, I'm going to re-arrange my whole network. Last time I also couldn't mine efficiently because I got too many devices hooked up.
I'm going to balance the load and hope it works like last time. 

[fpgaminer]

Humax:  Try connecting your sick Avalon directly to a PC.  That means:

1) Switch the unit's power off, and unplug from wall.
2) Open the case.
3) Remove the "P5 jumper" (labelled J1 on the board) located immediately behind the USB port on the FPGA controller board.  See: https://en.bitcoin.it/wiki/Avalon#FPGA_controller
4) Either a) disconnect the USB cable from the TP-Link and connect to a PC, or b) disconnect the USB cable from the FPGA controller and use your own USB cable.
5) Power on the unit.
6) Make sure your PC recognizes the Avalon (it comes up as a serial port).
7) Get/compile a copy of cgminer for your PC; make sure it includes Avalon support.
Run cgminer (e.g. "cgminer -S /dev/ttyUSB0 -o stratum.btcguild.com:3333 -O xxxx:yyyy -D --verbose --avalon-options 115200:24:10:45:282")


Of course, you'll need to adjust how you run cgminer depending on your PC's OS and what port your OS assigned the Avalon to).

In case you're wondering, I had to do the same thing for one of my Avalons.  It constantly restarted, due to the dreaded -71 errors.  That particular Avalon has been running with no problems ever since I hooked it up to a Raspberry Pi instead.

NOTE: If you choose to reconnect the Avalon back to its TP-Link, DO NOT forget to reattach the P5 jumper.  If you don't, it definitely won't run

[shmadz]

<snip>
It constantly restarted, due to the dreaded -71 errors.  
</snip>

Thanks for the workaround, good to know there is a way to bypass the TP-Link.

just a note about these -71 errors, I was getting the same thing (see   https://bitcointalk.org/index.php?topic=156323.msg2611536#msg2611536  )

ckolivas' new firmware 0627 seems to have fixed this somehow, I've been over 5 hours now without a restart, no more -71 errors, running at 345 mhz - average speed reported 81660 - pool (bitminter) has been guesstimating my speed between 70 and 90+ so it seems to be working quite well,

just wanted to note this in case anyone else is having constant restarts due to -71 errors.

[Sitarow]

<snip>
It constantly restarted, due to the dreaded -71 errors.  
</snip>

Thanks for the workaround, good to know there is a way to bypass the TP-Link.

just a note about these -71 errors, I was getting the same thing (see   https://bitcointalk.org/index.php?topic=156323.msg2611536#msg2611536  )

ckolivas' new firmware 0627 seems to have fixed this somehow, I've been over 5 hours now without a restart, no more -71 errors, running at 345 mhz - average speed reported 81660 - pool (bitminter) has been guesstimating my speed between 70 and 90+ so it seems to be working quite well,

just wanted to note this in case anyone else is having constant restarts due to -71 errors.

what options are you using?

[-ck]

<snip>
It constantly restarted, due to the dreaded -71 errors.  
</snip>

Thanks for the workaround, good to know there is a way to bypass the TP-Link.

just a note about these -71 errors, I was getting the same thing (see   https://bitcointalk.org/index.php?topic=156323.msg2611536#msg2611536  )

ckolivas' new firmware 0627 seems to have fixed this somehow, I've been over 5 hours now without a restart, no more -71 errors, running at 345 mhz - average speed reported 81660 - pool (bitminter) has been guesstimating my speed between 70 and 90+ so it seems to be working quite well,

just wanted to note this in case anyone else is having constant restarts due to -71 errors.

Well, the -71 errors were something like:

Code:

[29560.890000] ftdi_sio ttyUSB0: failed to get modem status: -71
[29560.900000] usb 1-1: clear tt 1 (8090) error -71

And the updated cgminer code bypasses the ftdi_sio driver entirely now since it uses direct USB, so there is one less potential point of failure (but of course that doesn't make it a fix for all potential USB issues!)

[shmadz]

what options are you using?

right now, my options are --quiet --avalon-temp 0 --avalon-cutoff 45  ... fans are reporting 3840 rpm (not actually that loud, unless you're a sissy  ) --- temp 1 is at 22 and temp 2 and 3 are at 37 (AC pointing directly at the intake)

Well, the -71 errors were something like:

Code:

[29560.890000] ftdi_sio ttyUSB0: failed to get modem status: -71
[29560.900000] usb 1-1: clear tt 1 (8090) error -71

And the updated cgminer code bypasses the ftdi_sio driver entirely now since it uses direct USB, so there is one less potential point of failure (but of course that doesn't make it a fix for all potential USB issues!)

whatever you did, it worked like magic!

Code:

Elapsed        MHSav     FoundBlocks    Getworks   Accepted  Rejected   HW   Utility   Discarded   Stale	
8h 56m 11s   81655.25    0              1094         10024       26    8566  18.69       2108         0

Thanks again CK!

[Humax]

Thank you everybody! my issue has been solved!

The miracle I was hoping for happened! 

by doing these weird things: (in case you want to know..)
- play with the cables inside the unit
- shake the unit
- flashing to newest firmware again. (without keeping setting)
- rearranged network
- moved to my other switch.
- revoked all other devices/miners from direct access to internet
- mining for CK at default settings for an hour while I left the room and watched some tv
- changed settings, powered off, powered on and let it run while going to bed with positive high hopes and faith......


....and it has been mining non stop! 

So which step solved the problem? shaking, rearranging network or donating? I don't know and honestly, right now, I don't even care.....

[cypherdoc]

can someone tell me how to delete the blockchain in ubuntu and start a re-download?

unless you can tell me how to get out of this error:

Code:

: Error opening block database.
Do you want to rebuild the block database now?

[allinvain]

Thank you everybody! my issue has been solved!

The miracle I was hoping for happened! 

by doing these weird things: (in case you want to know..)
- play with the cables inside the unit
- shake the unit
- flashing to newest firmware again. (without keeping setting)
- rearranged network
- moved to my other switch.
- revoked all other devices/miners from direct access to internet
- mining for CK at default settings for an hour while I left the room and watched some tv
- changed settings, powered off, powered on and let it run while going to bed with positive high hopes and faith......


....and it has been mining non stop! 

So which step solved the problem? shaking, rearranging network or donating? I don't know and honestly, right now, I don't even care.....

Congrats!

I'm telling you man, there is magic in them darn boxes!

Happy hashing!

[Bogart]

I just solved a very difficult problem.

My daughters laptop stopped connecting to the internet.  I won't go into the details but the solution was turning off WiFi and disabling DHCP on my batch 1 Avalon.

This.

I had the same problem when the DHCP service got turned on again somehow on my Batch 1.

Having a rogue DHCP server on your network will cause all kinds of connectivity problems like Humax describes, and their cause is usually not obvious.

In the Avalon Web UI, navigate to Network --> Interfaces, click the Edit button on the interface(s), and select the "Disable DHCP for this interface" checkbox.  Then "Save and Apply".