Topic: [Guide] Dogie's Comprehensive ASICMiner Blade Setup - page 9. (Read 580771 times)

I had intermittent problems using wifi with my Blades.  I bought a used switch and hard wired, instant 10.8ghs and runs for days and days with out problem.  Using Pi with bfgminer proxy.

From another node on your network, you could attempt a TCP connection to 192.168.1.101:8332 just to make sure that nothing on either your network or OS or whatever is blocking incoming connections to your proxy. Otherwise, I can't see any reason, many people reported that Blades are *very* sensitive to network issues that are generally not causing problems to other devices.

I swapped the psu with my desktop psu, i got the status OOO but now it just restarts itself after ~2 minutes of doing nothing. The mining proxy seems to cause my OS to freeze every now and then

http://s14.postimg.org/cl9mpqqgv/Untitled.png

Any ideas? 

I think confusions are arising because of differences between V1 and V2 blades.

My previous post about polarity was based on my V1 blades, here's how they look from under the PCB:



As you can see, the + is on the left and the - on the right. It's possible that it was actually reversed on V2 blades. If so, it's very fortunate that you're pointing this out - always use polarity according to the PCB, since guides (for the other version) could be misleading.

Then about the fuse, V1 Blades required the owner to plug in a 10A fuse for operation, the kind that is also used in automobiles. As far as I know, V2 Blades use SMD style fuses (smaller package, soldered in), rated 8A, and these seem to be more problematic. However, you don't need to plug them in on setup - they're already onboard.

I'm afraid I can't help you with the rest though. If you ever get the chance to try another PSU, that's a possible cause that you could rule out for sure. If that fails and you still only get X's, I would go for RMA or try to arrange something with the seller maybe?

About the polairty, isnt it the opposite? Because minus (-) is drawn on the left side on the back side of the blade, while + on the right one. Just opened a brand new blade, same situation status XXX. The first time i powered the other blade i remeber the psu's fan starting to move and the heatsink on the blade got warm, but i had issues with my router then. Ever since it just gets powered and the psu's fan is silent.Theres something else i remeber about the intial powering of the blade i remeber i didnt mount the screws, i just stuck the cables in (they seemed loosely), is it possible that i have messed up something with the cables 

My source seemed reliable, i hope they are ok.

Theres something about the OP i couldnt understand "Remember to plug the 10A fuse [red disk] in, either way round will do." any clarification on that?

Sometimes if a cellphone is  hooked to your network/Router there will be sporadic connection.

Glad you found the bad actor on the network.  Mine was a bad wifi router.  Another network-woe-sensitivity data point.

My problem turned out to be related to a wifi repeater. After unplugging it, the blade picked right up. Thanks to all for the suggestions.

My problem with the all O's but not hashing:

There was network traffic, and work received sometimes when i put a plain tcpdump on it with headers.  On visual inspection of the board the pic pads were off on 3 sides, I figured couldn't hurt, and fixed it.  On first restart same thing.  I was working against getwork wallets and a slush proxy.  After dicking with it for awhile I put -A option in tcpdump and noticed json returning "Bad credentials".  I sucked it up and loaded bfgminer with the stratum proxy, and have had no problems.  I get the feel the miner is not updating the userpass because the auth hash never changed no matter what i did.  No problems besides killing a rail on a cheapo 220w psu.

HELP?

http://www.coolermaster.com/powersupply/gxii-series/gxii-650w/ <-- can this cube overclock to HIGH MODE?

http://www.coolermaster.com/images/upload/images/psu/gxii-650-ac.jpg

I can't think of anything that's just started happening the last 3 nights. But it's down again for the night... The first night I tried a getwork pool and I tried running a stratum proxy on a different machine. It made no difference. Yet my USB sticks have been rock steady for months.

Indeed, I've seen something like that before with my blades - turned out it was my network.  I know I keep harping on that, but it was the root of so many of my and others' problems.  Anyway, what do you do at those times, or what automatic things happen at those times (keep in mind the time shown might not match your local time)?  Does it do this if you've got nothing else running/hooked-into your network (like cellphones on wifi, streaming video, etc)?

Boobs

I see an interesting pattern emerging:

Yeah you Blade look like he's on strike - normal idle proxy output indeed.

Maybe network sniffing / packet inspection (tcpdump or wireshark) can help, at this stage nothing else would come to my mind...

It looks like the proxy is still connected to the pool:


2014-01-24 06:45:34,548 INFO proxy getwork_listener._on_authorized # Worker 'worker.1' asks for new work
2014-01-24 06:45:34,600 INFO proxy getwork_listener._on_authorized # Worker 'worker.1' asks for new work
2014-01-24 06:45:34,653 INFO proxy getwork_listener._on_authorized # Worker 'worker.1' asks for new work
2014-01-24 06:45:34,705 INFO proxy getwork_listener._on_authorized # Worker 'worker.1' asks for new work
2014-01-24 06:45:34,757 INFO proxy getwork_listener._on_authorized # Worker 'worker.1' asks for new work
2014-01-24 06:45:34,810 INFO proxy getwork_listener._on_authorized # Worker 'worker.1' asks for new work
2014-01-24 06:45:36,363 INFO proxy client_service.handle_event # New job 7af38b9 for prevhash 06de5ca8, clean_jobs=False
2014-01-24 06:45:44,955 INFO proxy getwork_listener._on_authorized # Worker 'worker.1' asks for new work
2014-01-24 06:45:56,288 INFO proxy client_service.handle_event # New job 7af38ba for prevhash 06de5ca8, clean_jobs=False
2014-01-24 06:46:06,368 INFO proxy client_service.handle_event # New job 7af38bb for prevhash 06de5ca8, clean_jobs=False
2014-01-24 06:46:16,817 INFO proxy client_service.handle_event # New job 7af38bc for prevhash 06de5ca8, clean_jobs=False


This is what it looks like on the blade side:

Total MHS:   00000
Received:   0000000009
Accepted:   0000000000
Per Minute:   000.00
Efficiency:   000.00%
Up Time:   0d,00h,01m,31s

Current Server: 192.168.0.253:8332
Chip: OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO


I'm currently connected to Ghash, but I was having the same trouble with BTCGuild.

If you have a way to log the output from your stratum mining proxy, you should check for unusual output at the moment when it stopped hashing. It should be at least a starting point to debug.

Sometimes, my stratum proxy doesn't reconnect on disconnection from pool, or takes ages before it reconnects.

OK, this is weird. After being down overnight for about 11 hours, my blade spontaneously started working normally for about 11 hours, and it has now stopped again. 

I'm having similar problems to others I've seen on this thread, which never seem to have any resolution that I can see. My blade has been working great since November, but then in the last couple weeks the hashrate would occasionally fall for a few hours, and then go back to normal. I thought maybe my ISP was throttling me, but my USB BEs are chugging along normally. I switched pools and it seemed to improve things. But then last night I noticed it had pretty much stopped altogether. I tried different pools, brought up a stratum proxy on my PC (I usually use one on a Raspberry Pi), and even tried a getwork pool. I then power cycling and a hard reset. I left it on overnight, and there was no improvement this morning. I turned it off while I got ready for work, but when I turned it back on it still wasn't hashing. So I left it and went to work. I now see that it finally started working again on its own during the day.

During this whole time, the status has been all Os (no Xs). Oh, and I happened to change ISPs last week (due to unrelated things), so I know it's not a throttling problem - plus, as I said, the USB BEs are still doing just fine.

Facing the screws, the two left screws are GND, the two right screws are +12V

You don't need to connect all 4 terminals, you can use only one of each (as long as you have enough wires for 10A per Blade)


This PSU should be enough for up to 4 Blades


Yes. I don't know what's wrong if not the PSU. All the doc I've seen mentions either a faulty PSU or something awful when you only get x's instead of O's and it persists over cold restarts. Are you sure that they were OK in the first place?