Topic: AVALON 6 CURRENTLY - page 2. (Read 4165 times)

Ok. In which case I'll take off the left side and see what's going on.

C

Really through eBay and PayPal protection, you aren't getting refunded for DOA unit. That's crazy. Sorry to hear about that crapshoot of an experience.

no money back on either sadly... no justice so your my only hope lol

Maybe....

Did some research here. For purposes of keeping things baseline and cool, I'm running all tests at speed 100. That should give a nice 400-500mh per blade with little temp difference. My new unit here works fine; two blades run great at up to 3.5th at speed 350. Room to grow there, but work well.

Test 1: Run with one blade. Works, flashes red, blade 1 is down, error is 18440, Loop 1 failed PG1 failed, ADC1_err. Temps are 32,47,273. Ok, fair enough. One important thing, the plug on the blade side was exceptionally tight. Pull hard and you *will* rip off the socket. I think they glued them, keep this in mind.

Test 2: Run with second blade on controller 2 (stux's controller from second Avalon). Goal is to establish a working MCU if possible. Plugged it into my blade, his cable, his controller, my auc, my Pi. Works fine, with same errors and 400-500gh hashing rate. Ergo his second MCU is good. No need to test third MCU, leave in box.

Test 3: Put his MCU on blade 1 of his Avalon. Fails with both loops down. Ergo that blade is bad, problem is in blade.

It is also to note that the ripped off plug matches the fact that the seller was retarded and probably thought he could pull the good blade and put it in another box. So he ripped off the plug, replacing the plug (bitch on wheels) could get things to work on that blade. Will think about that.

Meantime will take apart blade 1 from unit 1 and see if there are any solder bridges that can be fixed. Leaving unit 2 alone unit stux says he isn't getting his money back evah....

Sweet.

With BTC up, it will be nice to see those babies mining lol

Burp. Just got it last night, along with a nice reference A6 that I need to test. Also pulled the microcontroller from the first MCU board that was on the first A6 (labelled "A") the chip was running super hot which indicates shorts. Will try a new chip once I verify location of pin 1.

maybe some Turkey gravy will do the trick LOL

Sounds good, will be able to work on it over T-Day. Also just bought another A6 to use as a test platform, gotta spend the money for the tools. :-)

C

looks like the art will arrive to you tomorrow

second was ebay, I'm looking into that now.

I already printed a label for the MCU to be sent, will go out tomorrow am.

Meantime I took one chip off the bad board to try and get the good board to pass loop tests. Still not coming up, it's possible the problem is in the MCU; if you have a spare you would like to send over I can do a cross-check.

I'll post a few pictures from the second board; there are at least 5 chips on there that have solder balls or shorts on them, what is making me wonder is how a unit could operate long enough with a single short in order to burn up additional chips. I wonder if the unit only self tests the loops on startup, and not while running. If so there may be a flag or option I can set in cgminer to ignore bad chips and continue, then I could figure out which chips are good/bad.

Researching..... In the meantime didn't see them on Amazon, but Amazon may have a return policy for the second unit.

was on their amazon store

That's crazy I'm sorry to hear about that, no one wants to be sold bunk equipment. I'm kinda floored that they didn't get back to you through their company website. Also wondering how did you find their used equipment for sale?

How did you pay for the units? if it was credit card at least they will offer purchase protection. Hope so, but at least you found them on the forum. Best of luck.

gonna be filing a scam report if I don't hear back by monday since they are on these forums

Right. That is what I am calling the second unit you sent over. I'm not fiddling with that at all in case you want to return it. The board work is being done on the first unit.

C

Likewise you should see the other side of the first unit. At least 4 chips have either bridges or solder extrusions, my guess is these were run way too hot and eventually blew up. When blew up they were sold. This side shorted +12 over to the signal lines blowing the MCU.

Fix.... Hm. Have other things to do so will put this aside (I have not touched unit 2, but I'm expecting similar things). But anyway that's one way to blow a board.

C

second one you mean? with the pins ripped?

If you bought it from Ebay, then do a return request. I'm going to take off the second board, anyone know of a source for these chips?

C

ohhhhhh wowwwwwwwwwww.

Since they never responded I guess I'll name the company that sold me that DOA one also, it came from oregon mines. (thats the first one)

And in the meantime I've solved the mystery of your other unit's side 0. As mentioned it comes up with a Loop failure, which probably matches your description of "it would work sometimes then drop out". Taking the board apart and looking at it with the 15x loupes reveals the problem:



Anyway the issue is that one of the chips has a solder bridge along several of the connection points. Avalons do the same thing as the Antminers in that the power for the chip comes in on the top and bottom sides with signals coming in on the sides. In this case the chip overheated and melted the solder enough to make a nice little 3 wide bridge.

I'm guessing that is the problem. Now to see if I can fix it. Hm.....

C