Topic: [CLOSED] - Stumptown Miners - Avalon PCB Assembly - West Coast USA - page 2. (Read 30763 times)

If they work with the existing board design for the gen 1 chips, then they're great stats.

That's some crappy stats

We are pleased to announce the launch of our next auction. BitSyncom is auctioning their Avalon ASIC A3255-Q48 chips: the second generation SHA256 Processor. BitSyncom will be auctioning off 20 reels of 500 chips per reel. Each chip outputs 1.5-1.6 GH/s at a standard 0.9V, with 2.05W consumption. The minimum bid for a single reel will be 24 BTC.

For more detailed information on the A3255-Q48 mining chip, please see Avalon’s Datasheet. This product has already cleared customs and is ready to ship upon the close of the auction.

Bidders who have deposited bitcoin in their Tradehill Auction accounts may begin bidding as of Friday November 1, 2013, 1:00 PM EST. The auction will close on Tuesday November 5, 2013, 2:00 PM EST.

All bidders that have submitted identity documents since the close of the prior auction should have received a response from Tradehill Auction Support either indicating that their account has been approved or detailing additional documentation needs.
For support related to your auction account, please contact [email protected].

They found the firmware bug!!  So far two people report 0 (Z.E.R.O.) HW errors and over 5gh/s.

It's really late, so I'll confirm mine in the morning.

I've got a few work items to deal with tomorrow am, but after that (and I confirm myself) I'll start running boards.  Also need to order the replacement parts.

It was a dumb bug.  Just a matter of time before it was found.

Ok.. Couldn't wait   I can confirm no errors on my unit.  We're golden.  It's hashing right now.

Any word? I have yet to receive anything in the mail nor a reply via pm since Colin's latest update via email that you posted here. Curious what's happened since.

...
Now for the parts, I'm thinking we may do better just building more K16s for other people with chips or re-purposing the parts on a Gen2 design.  Not sure.  Just that kitting them out in 'refunds' isn't really doable.  If I do get a Gen2 design going I may buy them myself, but that's really speculative at this point.

As for the K16, I'm not sure at this point.  They are hashing, but with ~25% HW that just doesn't seem to be getting fixed.  It looks like there is a problem with the PIC and USB and CGMINER, but I really can't tell if its fixable on the current PCB.  It is taking way longer then I figured it would which is a bad sign... The PIC seems to be underpowered for the task at hand.
...

...
I'm really sure the problem with the current K16 is firmware.  The group seems to be having a hard time as well with CGminer itself.  It all revolves around the USB being the trouble spot.  My hunch is the little PIC can't handle feeding the ASIC's and dealing with USB traffic at the same time.  Should be able to fix it, but who knows how experience the guys working on it are.  It isn't trivial code that's for sure.  USB is a dragon.

I've already written a simple stand alone stratum client.  I'll try to write K16 firmware from scratch that doesn't use USB or CGminer.  This may at least help the others find the issue.  Right now it seems a struggle to debug with CGminer/USB being so complex.  I want a really simple test setup.  This will lead right into doing my own Gen2 board as well.

I'll be back home later this week and should be able to get my test client running with the K16.  If that works, and I suspect it will, we'll be able to figure out some kind of solution.
...

Tolan, I'll send an email to Colin asking about the six chips. I'm sorry you haven't received them yet.

Is there any update on Gen 2 chips?

Thanks for all the feedback, guys. I'm going to go ahead with a group buy, then, as soon as Avalon's terms of sale are available. And again, if you're interested in being notified when that happens, PM me or watch this thread. I'll be announcing it here when it happens.
[/quote

Hi all,

Thought I'd send you all a quick update:  With the latest part swaps and firmware my one built unit has been hashing most of the day without a single issue.

The only thing now is the high reported error rate, but its really looking like a firmware issue.  A great deal of progress was made today by the dev group and I'm getting a good feeling they are converging on a solution.  I'm checking on the progress every couple hours.

Even with the HW errors its around 4Ghash/s.  Not bad.

You should see quite a bit of work submitted to the pool.  Perhaps that will boost spirits a little.  I'll leave it running from now on.  I may take it offline now and then to update or for phone calls in my office.  The fan I have on hand is noisy as hell.

I know several of you have been looking at refunds or putting the ASIC's on other boards.  Still an option of course, but things are progressing and Ryan and myself are doing what (little) we can.

I did want to stress that this board is NOT DIY friendly.  Unless you have another known to work board and a top-tier assembly contractor you'll be in for a bad time.  Most of the other boards assembled by the current dev team, some by contract assembly houses, have needed significant rework and some have failed outright. The only way to reliably assemble a board like this is with high end automated equipment and very experienced operators.

I also want to give my reassurance that the moment the dev team (and myself) are confident the design is good and working I will proceed to build the boards and get them shipped asap.  The parts are loaded in the machine and ready to rock.  I'll work with Ryan to expedite the logistics as much as possible.

I'll keep you all posted, and thank-you for your patience.  I do understand how trying this is.

Ryan, please forward this to anyone else you like (other then the Zefir batch 1 group who I've Bcc'd).

Have a good week all!

Colin Fitzgerald, Founder/CEO
Place Droid Inc.

Yes, that includes the boards. We can ship your components to you at this time. Colin has all of them at the moment, so I'll make sure he does that ASAP.

Sorry about that. I did get your message and thought I had already replied to it. On top of this, Colin emailed me later that day about a customer who'd been asking for their chips and for some reason I thought he was talking about you. So I figured the two of you had handled it. Looking back at the emails, though, I see now that he was *not* talking about you, but about another customer, and that I had *not* in fact replied to your message. Mea maxima culpa.

Last I heard from Colin, he and the other developers had discovered that some of the component values needed to be changed. He was going to try the new components on Friday, but I haven't heard anything from him about it yet. I just sent off another email asking for a status update. We are still planning on getting chips on boards, but I understand if you want out at this point. Please let me know what you decide. If you do want out, I'll go ahead and refund the cost of assembly plus my margin to the address you PMed me.

Edit: Have you gotten the six chips you requested earlier? They were supposed to have been sent around the end of August.

I have yet to hear from ryepdx since August 26th. He has my chips from batch 1 and I have no idea whether he's sending them to me or going to be able to build them into a miner. I'd like to hear from him when he gets a chance.

Hi Ryan, when you say you have the parts, does that include boards? We had previously discussed with my refund that I would like to have all of the parts I purchased shipped to me since, understandably, they are not refundable. Is that possible at this time? That is the last step my refunds with the project (you've refunded labor, steamboat has refunded chips, just need components now).

We have the parts