35w per K16, so 8 x 35w = 280w in total?
All 280W is on the 12V rail though so don't compare with the overall capacity of the supply.
|
|
|
Just doing my sanity probing on the K16. Found first error: Vcc power for PIC is not connected.
To me this is good news. When I move house, I always relax once the first thing breaks, something has to be broken, once you find it the tension disappears and you can get on with the work at hand (this is not rational). Thanks for your hard work BKK, we're a few weeks from chips arriving, certainly time for a re-roll or two. Let us know if you need funds.
|
|
|
Ouch I need full assembly then. I think my first two orders maybe didn't even include saying what i wanted done with the things but the last order I sent the form did include saying, and I foolishly put full kit not realising that would mean the chips could be bad.
Is there a procedure for updating that? As now i obviously am going to want full assembly at least, and if that doesn't include chi-checking then i guess I would need to full assembly plus testing.
All you've committed to is what you've paid for, buying some chips. As above SB is finalizing a shopping cart for chip buyers to nominate and, if necessary, pay for what they want done to their chips. Commitment comes with payment, you've only committed for what you've paid for. SB did send out a survey to batch 1 buyers, afaik this was just a survey to gauge interest.
|
|
|
Well, there goes my comfort level.
Did Avalong release their own firmware? Would it help or is this design completely different?
This is a completely different design. The only thing in common is they both have Avalon ASIC chips. The Avalon design does not have a micro-controller on board, my design does not have an FPGA. So their firmware is FPGA firmware and mine is PIC firmware. Due to this even the cgminer driver will be different. But this has all been known for 2 months now so it's not news to anyone who has been investing effort/time . Has anyone put a cro or logic analyzer on a real Avalon? Without any of their hdl it is tough to know what the comms looks like. Not releasing their code is annoying, they could release what they can and we could check out the comms protocol even if we can't build chips. In a few months their ip will be useless.
|
|
|
Holy crap, there's a photo of an assembled board on ebay and BKK is still driving to pick his parts. This is a dog eat dog business for sure. No mention of a heatsink.
|
|
|
BKK: I'm not sure if you saw it but Yifu posted on another thread (where someone got an Avalon clone board working with a sample chip) this: hint hint: the chip has a much higher clock cap. I don't think anyone knows details yet, so just an fyi. I hope he means capability and not capacitance! ps. lets close the thread above and keep this OT, we need to help BKK to focus.
|
|
|
For any board you are interested in you need to try to route a design for the chip before buying. This will give you an idea of what you will get out of it. It won't tell you what clock rate you can achieve though, this depends too much on the power supply. PS: Where to get cheap and used 32 chip boards? Not that i expect an answer... ![Grin](https://bitcointalk.org/Smileys/default/grin.gif) That's a really sad story. In 2009 the co i work for was bought out and the v5 project we were working on was abandoned. We junked most of the boards (like 20 or so, each with 4 chips). Imagine mining with this system since 2010, man you could retire on that. I've only been going since March, nothing remarkable, a coin a week or so.
|
|
|
I have a system made of 32 v5lx220 devices using the open source HDL. These are nice because they're big enough to fit two cores with no hand optimisation. I get about 250MH/s per fpga before the 1.2/1.5v power supply cuts out (these are reclaimed boards).
I also have a board with 4 V5LX155 on it - I haven't bothered with this board because I can only fit a single core - ie. 125MH/s without hand optimisation (which is beyond me).
To do something good with these board I think you'll need good FPGA skills.
*edit* also without a schematic it might be hard to get serial ports to the chips (or even jtag).
Hey how do you make 2 cores communicate with the UART and the mining software? Any link to HDL code? Thanks a lot in advance. I am trying to fit two cores into a Kintex-7 LX 325 device. Take a look at the OrphanGland stratix implementation project in the open source miner code. they just split the problem space into n and have each core working part of the same share. This way you don't complicate the s/w and h/w multiplexing multiple jobs on the same serial port.
|
|
|
I have a system made of 32 v5lx220 devices using the open source HDL. These are nice because they're big enough to fit two cores with no hand optimisation. I get about 250MH/s per fpga before the 1.2/1.5v power supply cuts out (these are reclaimed boards).
I also have a board with 4 V5LX155 on it - I haven't bothered with this board because I can only fit a single core - ie. 125MH/s without hand optimisation (which is beyond me).
To do something good with these board I think you'll need good FPGA skills.
*edit* also without a schematic it might be hard to get serial ports to the chips (or even jtag).
|
|
|
It's good you have samples already, trouble is that you will get the board finished and tested, then just sit there for months waiting on a real supply of chips. I guess you can mine on the prototype in the meantime to help pay overheads.
I, for one, am grateful there are people putting time and money into this design, thanks Terrahash! A lot of us are depending on this design, a board re-roll will easily eat a week or two, the earlier the better!
|
|
|
I have a US bank account but not Citibank. I don't want to put my email address here but anyone who gets an inkling to send PP can always PM me for that.
Perhaps take a look at coinbase, works well for me.
|
|
|
Looking at the pricing it seems the only testing optional will be with the fully assembled miner. I find this odd because most manufacturers will test a PCB assembly before fitting all the heatsinks etc to reduce the cost of rework. They'll also use a jig to test. I guess the issue here is that this board is hard to test without a heatsink? Perhaps two levels of testing can be offered, a sanity test (power, trace continuity etc but no mining) and a bake in test? Perhaps we can write some code to just mine one share and shutdown the clocks, this could be ok without a heatsink?
|
|
|
Is there a possibility of buying just the components un-assembled? Meaning, just the chips, the PCB, and all the associated capacitors/connectors/resistors, so we do all the soldering ourselves? Even better, is there an option of just the asic's and the PCB un-assembled so we purchase all the passives and whatnot ourselves and solder it together ourselves??
This may be a possibility. Creating full DIY kits requires a considerable amount of time, and depending on the option chosen for parts sourcing, may not be an option. All requests are being taken into consideration, but ultimately the most cost effective, quickest solution for the majority of the orders will take precedence. The best way to do this would be to roll the whole thing yourself. The circuit will be stable long before the chips arrive. Get a set of boards made for $25 at iteadstudio, order all the parts from Mouser or Digikey and you have a kit. Someone might even create a BOM at one of these two so you can order in just a few clicks. It really isn't worth the hassle of assembling kits, the large online component sellers do a much better job for much less. I don't think anyone is going to write instructions for assembly but perhaps I'm wrong. If you can mount the ASICs yourself, the rest is a piece of cake. *edit* I would do this except I don't want to screw up the mounting of the ASICs. I'm planning on making some K1s by hand with my extra chips if they aren't used up.
|
|
|
Oh! And just got an email that the first K16 boards have shipped today.
Wow super cool update, your timing is impeccable. I was just getting cold feet at the meteoric rise in difficulty. I should know better.
|
|
|
I saw your verification and decided that was enough for me (you you could have been working together, I only bought 32 chips in batch 1, I can still feed my family if they go missing). Since then I've been totally convinced steamboat is legit, I think he is thinking way too much about the tiny details to not go through with this. I also agree that he has a lot more to gain by going through with this than scamming. This is just the start, BFL seem to be releasing chips. These Avalon chips are fairly primitive ASICs, there will be Avalon II perhaps in something 25nm, perhaps offering 200MH/s on a AA battery, who knows. I think this is the start of an industry, these boards will be obsolete in a year but the experience and reputations gained producing them will be very valuable.
|
|
|
I saw this https://bitcointalk.org/index.php?topic=223032.msg2345354#msg2345354a few minutes ago. I'm thinking to speed things out and to help BKK coins with testing There is a member willing to sell 10 chip Avalonmodule for 14 BTC (1.4 BTC per chip) I'm thinking if we all collect this ammount in an escrow ( let say John K.) if the seller is agree and module be shipped to BKKcoins? I believe this way we may spare some time and to be ready for july/august delivery of the firs group orders from zefir. What do you think? Are we sure these chips will come off on in one piece, they're going to have to get very hot to melt the solder with all that ground plane etc attached.
|
|
|
Me too and my difficulty issues are still with me.
The proxy is reporting difficulty 2 and slush is configured as 1 so my reward got halved again. There is probably a way to make them double, this is a nasty bug.
18310 2013-06-01 02:41:35 0:52:58 7941437 3183 0.00970591 238975 25.10570000 91 confirmations left 18309 2013-06-01 01:48:37 5:15:09 49659498 32145 0.00320978 238964 25.05830000 80 confirmations left 18308 2013-05-31 20:33:28 1:42:52 16633401 13142 0.02089623 238930 25.00000000 46 confirmations left 18307 2013-05-31 18:50:36 0:18:57 3010041 2271 0.01868908 238919 25.26863000 35 confirmations left 18306 2013-05-31 18:31:39 2:17:49 21231848 17015 0.01871870 238916 25.31229000 32 confirmations left 18305 2013-05-31 16:13:50 0:45:01 7410765 5670 0.01910239 238901 25.26605000 17 confirmations left 18304 2013-05-31 15:28:49 3:24:52 34544624 24506 0.01934538 238894 25.04760000 10 confirmations left
back to BTCguild for a few days.
|
|
|
Nice! That one is really beautiful and I like the stand as well. That seems essential for steadiness. Thank You and I'm sending you a PM.
I use one of these visors for all my work now, I have a stereo microscope but I really just prefer these. http://www.amazon.com/dp/B0015IN8J6/ref=biss_dp_t_asn![](https://ip.bitcointalk.org/?u=http%3A%2F%2Fecx.images-amazon.com%2Fimages%2FI%2F41yy8%252B%252B1i7L._SL500_AA300_.jpg&t=663&c=ZH2zDCkY0rHWRA)
|
|
|
|