Topic: Avalon ASIC users thread - page 203. (Read 438602 times)

I believe you have to make the ip addresses unique.  Is the OpenWRT web interface installed on the systems?

Your calculations are wrong it is capable of 650 watt Output which means 650x 1.2 roughly 780 wats at the wall. However if the psu is crapy it might be issue also not providing enough amps and making whole system to bug because of that
there are two 12V rails up to 30 Amps. According to specs it shall be OK unless PDU board does not spread the load. If three ASIC units are powered from one 12 V rail it would be a problem, but i am sure that nghzang took care of that:)

Highlighted, italicize, underlined, glowing key word.

Edit: not much of a glow.

Asic-J wrote "Avalon comes with Antec 650 watt bronze PS. It is only capable of 82% efficiency. (533 watts)." Could that be the problem causing the errors, if possible it is, have you considered trying a beefier psu in it?

Thats not as much as i thought. So Asicminer is still the biggest with 3.6TH at the moment. But what i wonder is what Asicminer is working on. It looks like they can build 0.2TH per day and bring it online. But avalon claimed that they can prepare 12 units a day, together 0.78TH. And Avalon has to create a full standalone thing while asicminer only has to put in one board in the rack.

So what i think about is what friedcat and co are working on that takes that much time. Are they sitting with a soldering bolt at a PCB and working to get the asics on the plate or what is the timeeater? I mean they dont have the 12TH online till now and in the next days 50TH will come shipped to them because the fab was faster than they thought. I see there is a problem but i dont know what it is that is eating the time. Even though they work around the clock they claim.
I already suggested that even when they hire professional electricians the salary wouldnt be high in comparision to the mined bitcoins with the units. But i would like to know what steps they have to do now that is eating the time. Does someone know? Friedcat doesnt answer now that he is busy so much. But i still think the speed now is a problem and some solution must be found. And i dont mean that friedcat or co has to work more. They should get help. But without knowing wheres the problem no one can suggest a solution really.

Any way dude:) The Ip address would be the last "issue" for us. Besides the static ip address  is assigned by default unit gets ip address via dhcp when told so:). Just use a computer with crossover cable(just in case), plug it in avalon Ethernet port and configure it

Make sure you PC LAN has STATIC address/mask assigned matching the sub net where avalon default ip is:)

You have to think about how to know when avlon stops hashing as reported here and to be able to power off/on REMOUTELY. That will be the main issue until there is a fix for it. Think about that not the ip:)
And that process shall happen automatically:). I do have solution for that (snmp managed power switch) and when avalon comes in my hands i will test it and publish it here if needed and hang issue is not fixed. Moreover that solution shall inspect avalon temp (via separate temp sensor) and it can power off the unit remotely/automatically in case of overheating:)



 

Edited my post to add router log. Much more efficient. What happen if you connect 2+ units they all get the same static IP?

Ip is well known already:)
Besides nmap is good if you know or have an idea what you are looking for:) What about scanning whole B or A network:)

Anyone who need to find the IP address of their unit could use Nmap or look in the router log.

This model looks fairly accurate atm:

5 x 66 Gh/s have been confirmed, maybe 6. So optimistically 400 Gh/s. One unit had 4 instead of 3 modules, so it's a bit faster.

The bulk of the built machine have probably only shipped out after CNY and are either currently in transit or stuck at customs... hard to tell without any updates or tracking being available.

As all of the first batch #1 of 300 units will be shipped out by the end of Feb, by mid March you will notice about 20 Th/s come online. Perhaps by then it will only be background noise compared to the 50 Th/s ASICminer will release. And the 2 test machines BFL will put online :-)

Obviously this model did not work out due to CNY and shipping problems:

http://organofcorti.blogspot.be/2013/01/96-asic-choices-wwje-update.html

edit:
(I am referring to the arrival part, the update #3 of the model is indeed more realistic)

I would really like to know how many avalon-units are mining right now. I heard they were able to ship 12 units a day with 0.065TH each. And it looks like many of you are mining at ozcoin. The hashrate of the network grew around 10TH in the last 1.5month i have read and asicminer is at 4.2TH. I doubt the rest is avalon isnt it?

So how many avalons are mining now at what hashrate?

Avalon miner fan behavior is as follows:
   - fans stay at ~0/840/840 until temp3 reaches ~55
   - fans accelerate to much higher speed (0/3480/3240)
   - fans never slow down, even though temp3 decreases to ~43

It would be nice if the fans slowed down.  Much noise with new firmware!

It's better to run cgminer in an infinite bash loop (with a 5s sleep after it exits to avoid a hard loop) and send it a 'quit'
Although 'restart' is usually pretty simple in linux, it can mess up on rare occasions

They have a support site at http://support.avalon-asic.com/

I am trying to contact Avalon with questions about delivery etc. I emailed Yofu, but I didn;t receive a reply after 48 hours. Does anyone know the best person to contact there? Do they have a support email address?

Is this helping out (working) to automate the process of restarting?

https://bitcointalksearch.org/topic/my-impression-of-avalon-machine-cooling-info-updated-145885
------------------
6 there is some monitoring logic in cron, but it does NOT prevent the machine to get wonky (jgrazik & me both have experienced, crying~~).
Here is a snippet might be useful (default port is 4028):
RESTART_MONITOR="(echo -n 'devs' | nc ${IP} ${PORT} | grep 'MHS 5s=0.0') && (echo -n 'restart' | nc ${IP} ${PORT})"
----------------

The miner was power-cycled.  (updated original note to indicate this)

Just cgminer? Or all of OpenWrt? Or the entire machine?

I wonder what the reason is .. hopefully bugs in firmware or cgminer, but wonder if unit could be defective somehow =(