Klondike - 16 chip ASIC Open Source Board - Preliminary

[gyverlb]

You really should set up a P2Pool

P2pool does not work with Avalons, so I wouldn't expect Avalon-related projects to work with p2pool either.

It doesn't work optimally (you don't get 100% of your hashrate). There's not much information going around but it probably is due to Avalons taking their time for returning shares to the miner software. The chips themselves don't have any limitation it seems to be a problem with buffers between the chips and the software.

BkkCoins, can you comment on that? Do you expect any problem between Klondike and p2pool with its short interval (10s)?

[marto74]

[bassclef]

^^I'd like to see a single version of that heatsink.

Something very low-profile is what I'm interested it, so I can stack them close together using spacers and point fans with good static pressure at them. If I remember right, Bkk said something about the heatsink requirements not being much, so something with tall fins might be overkill.

[KS]

Old design or you're reverting to full cutout for the PCI-E connector? Or I'm missing sth?

[Bicknellski]

Got an update on my board prices for fab in Jakarta and I wanted to share.

Prices in IDR - FOB Jakarta
    
Klondike 1 - Nano ASIC Miner  

Quantity   25

PCB   80,000
Components   100,000  
Assembly   55,000  
Per Board   250,000      

Total Order   6,000,000      


Klondike 16 – ASIC Miner Board          

Quantity   25

PCB    100,000  
Components   650,000  
Assembly   120,000  
Per Board   850,000  

Total Order   21,500,000


http://www.xe.com/currencyconverter/

[btceic]

Got an update on my board prices for fab in Jakarta and I wanted to share.

Prices in IDR - FOB Jakarta
    
Klondike 1 - Nano ASIC Miner  

Quantity   25

PCB   80,000
Components   100,000  
Assembly   55,000  
Per Board   250,000      

Total Order   6,000,000      


Klondike 16 – ASIC Miner Board          

Quantity   25

PCB    100,000  
Components   650,000  
Assembly   120,000  
Per Board   850,000  

Total Order   21,500,000


http://www.xe.com/currencyconverter/

Did I do this right?

http://www.xe.com/currencyconverter/convert/?Amount=6000000&From=IDR&To=USD
6,000,000.00 IDR    = 611.091 USD

~$600 for 25 USB for ~8 GH/s
~$2,189.74 for 25 16 boards for 133 GH/s

What is the time to market on these? Lead time and does these prices include the chips?

[Bicknellski]

Got an update on my board prices for fab in Jakarta and I wanted to share.

Prices in IDR - FOB Jakarta
    
Klondike 1 - Nano ASIC Miner  

Quantity   25

PCB   80,000
Components   100,000  
Assembly   55,000  
Per Board   250,000      

Total Order   6,000,000      


Klondike 16 – ASIC Miner Board          

Quantity   25

PCB    100,000  
Components   650,000  
Assembly   120,000  
Per Board   850,000  

Total Order   21,500,000


http://www.xe.com/currencyconverter/

Did I do this right?

http://www.xe.com/currencyconverter/convert/?Amount=6000000&From=IDR&To=USD
6,000,000.00 IDR    = 611.091 USD

~$600 for 25 USB for ~8 GH/s
~$2,189.74 for 25 16 boards for 133 GH/s

What is the time to market on these? Lead time and does these prices include the chips?

COST only not including chips etc... just boards. Same as everyone else when we get chips to boards. I am not selling outside Indonesia and doing very very small run in September / October.  Let's keep the thread to BKKCoins design etc... PM me or check my sig for my thread. I just wanted to post so people get an idea of board costs here in Indonesia. As others posted for Europe and US and China.

[marto74]

Old design or you're reverting to full cutout for the PCI-E connector? Or I'm missing sth?

I mentioned before .
When we are doing a heatsink with threads only then it's with pocket 4 mm
When there are through holes this is going to be with full cut out.
Here is the drawing of a heatsink for 2xK16.

[Beastlymac]

Got an update on my board prices for fab in Jakarta and I wanted to share.

Prices in IDR - FOB Jakarta
    
Klondike 1 - Nano ASIC Miner  

Quantity   25

PCB   80,000
Components   100,000  
Assembly   55,000  
Per Board   250,000      

Total Order   6,000,000      


Klondike 16 – ASIC Miner Board          

Quantity   25

PCB    100,000  
Components   650,000  
Assembly   120,000  
Per Board   850,000  

Total Order   21,500,000


http://www.xe.com/currencyconverter/

Does that price include heatsinks?

[Bicknellski]

PM me... for anything related to that post sorry people. Need to keep thread clear for BKKCoins.

[marto74]

One of participants in my group buy for chips did sent early extra 0.04 BTC for developer , he will order 1 K16 board assembled by me later.
I'm sending them to BKKcoins donation address
Payment sent
https://blockchain.info/tx/06e955aa7a2dbcbda8e785d0141582b0bd3aadc848fa8835d7bbc4ffa57ca481

[evilscoop]

FYI
4UCONN parts 10987 & 11182 have an MOQ of 1k parts

Im trying to source alternatives with a much smaller MOQ...
Other board assemblers, you need to be aware of this

[marto74]

FYI
4UCONN parts 10987 & 11182 have an MOQ of 1k parts

Im trying to source alternatives with a much smaller MOQ...
Other board assemblers, you need to be aware of this

Thank you for the info

[BkkCoins]

FYI
4UCONN parts 10987 & 11182 have an MOQ of 1k parts

Im trying to source alternatives with a much smaller MOQ...
Other board assemblers, you need to be aware of this

This is only about $65 total, and I think amongst us we'll use them all (4 2x2 pin hdrs/K16). The real problem with 4UCONN is the lead time is 4-6 weeks currently. I've been thinking of going back to thru hole for these, or at least having holes there just in case, since they're out near the edge a smaller heat sink won't contact the underside. I was going to order 1000 pcs of both those, to spread out amongst forums members or kits, until they told me it could be 6 weeks. Now, not so sure. AFAIK there isn't another source I've seen for SMD pin headers - though, having said that I expect everyone to dig up a half dozen in China...

This pioneering effort truly puts you up there with Stallman/Wozniak/Linus Thorvalds types of heroes for sure (imo)

<Blush> I think you've gone too far, but if it weren't for the lack of communication I would have put Yifu in that category for the ASIC design.

One of participants in my group buy for chips did sent early extra 0.04 BTC for developer , he will order 1 K16 board assembled by me later.
I'm sending them to BKKcoins donation address
Payment sent
https://blockchain.info/tx/06e955aa7a2dbcbda8e785d0141582b0bd3aadc848fa8835d7bbc4ffa57ca481

Thank You. If people do want to mine to me then they can use 50btc.com and set user name to be 1E1gPLHVo5qoeB5BZqHZhUvLC7yy3VoEPZ, or any pool or method where you can set the target address - I've created that one separately so people can monitor how much is mined to me - so let's keep it for mined coin only. My regular donation address is: 1BvpJvejSmtWjuDfytZL5HTBUjoXokazp9 and even little bits help with travel and shipping costs.

Marto - I received your heat sink. DHL! Looks good and is waiting here for testing.

BkkCoins, can you comment on that? Do you expect any problem between Klondike and p2pool with its short interval (10s)?

I believe that found nonces should be relayed as quickly as any other method, like GPU. I guess it will depend on the driver end and how often it polls. But in the PIC once a share is spit out by the ASIC, it gets sent immediately (< 1uS) to the host, unless it's "down the I2C chain", in which case it could take a few mS due to bus polling. Maybe later I can work on an I2C interrupt hack. Hopefully I can get the host to support a callback so that it immediately gets relayed. That would be ideal. I don't know what the issue with Avalon is and why it takes longer - unless, maybe the ASIC doesn't report it immediately and in that case I'd be in the same boat.

Thanks again for the statements of support. I'm doing my best to keep the things moving forward.

In case it's not obvious - I'm back from Bkk again. I have the scope and tested it with built in test signal and it appears to be A-ok. Looks awesome so far. The K1 boards look good too and they sent me 12 instead of 10. I'm going to probe them randomly a bit with my meter and see if the traces are as expected before I put parts on. Total, I got 9 new packages this trip so lots of stuff to work on now. I'm off for a shower and maybe rest a bit before digging in again.

Next up - make a parts inventory to see if I really have everything I need. Then I think I'll try cutting some SMD stencils on the Silhouette SD. Apparently it should work and there's a nice open source app for doing it directly from gerbers (see here). So if that works as expected I'll stencil up a few K1 boards and place parts and cook. Then I'll test out my firmware chaining functionality.

Anyone, who has ASIC samples in bound to them - please PM me with a rough estimate when you expect arrival and re-shipment to me so I can plan my next trip for their arrival. I'm expecting k16 boards any day now as well, and by then should have fully working firmware, and all needed parts for a K16 test. Fingers crossed.

[simon66]

I would love to see some pics of the K1

[emanymton]

FYI
4UCONN parts 10987 & 11182 have an MOQ of 1k parts

Im trying to source alternatives with a much smaller MOQ...
Other board assemblers, you need to be aware of this

This is only about $65 total, and I think amongst us we'll use them all (4 2x2 pin hdrs/K16). The real problem with 4UCONN is the lead time is 4-6 weeks currently. I've been thinking of going back to thru hole for these, or at least having holes there just in case, since they're out near the edge a smaller heat sink won't contact the underside. I was going to order 1000 pcs of both those, to spread out amongst forums members or kits, until they told me it could be 6 weeks. Now, not so sure. AFAIK there isn't another source I've seen for SMD pin headers - though, having said that I expect everyone to dig up a half dozen in China...

Found these during a search:    

http://www.globalconnectortechnology.com/connector/?series=BG050 - 2 row
http://www.globalconnectortechnology.com/connector/?series=BG055 - 1 row

Not sure if this is exactly what you're looking for but there seems to be no shortage of suppliers on ali-express anyway.

[BkkCoins]

I would love to see some pics of the K1

Here's a couple snaps with the new microscope. Amazingly it worked first time with Cheese in Linux by just choosing it as video source (usually it uses my webcam). Right off, I can tell the internal LEDs are going to give a bit too much glare so having external illumination will be better. But it works, and yes, that is an ant on the board! Don't know how that happened as it was gone when I looked. One of the things I have to put up with working out here in the jungle. They get into everything, everywhere.

The whole 2-up board.


Close up near the 1.2V buck reg showing part of thermal pad.

[Sophokles]

Cool! Great pictures. Keep up the good work! 

[BkkCoins]

Found these during a search:    

http://www.globalconnectortechnology.com/connector/?series=BG050 - 2 row
http://www.globalconnectortechnology.com/connector/?series=BG055 - 1 row

Not sure if this is exactly what you're looking for but there seems to be no shortage of suppliers on ali-express anyway.

Great. Yes, the 2 row 4 contact one is the one. Hopefully they have stock and not 6 weeks lead time. I've saved that for when I have some time.

The other one is a 3 pin horizontal smd pin header. Drawing here,

http://www.4uconnector.com/online/object/4udrawing/11182.pdf

[LaserHorse]

[…] and yes, that is an ant on the board! Don't know how that happened as it was gone when I looked.

Obviously you have a natural talent for debugging.

thnx for giving us a peek of the boards - lookin' good.