Topic: Klondike - 16 chip ASIC Open Source Board - Preliminary - page 47. (Read 435379 times)

Has anyone been able to fully populate the K16 v.0.3.0 board yet and verify functionality ?

Sure. Just send a paypal or btc w/PM, with qty for which of the 3 you'd like. Once I add the item to the store I'll update your order with them.

Perfectly understandable. I don't want to be a pig, but I would like just one. Can I add it to my existing order? Thanks

Yes, I'll add them soon but there is a limited qty. of 6 for sale. I want to keep some for my own use. If there is demand I can add more to a future board run but that will take a while. I did three versions initially and I'll cut them out and sell as separate boards for $2 each.

I've included an image link below as I don't want this to take over discussion here too much. PM me with questions. I've added a new repo at http://github.com/bkkcoins/picngo

Three versions - from left to right, picngo-sm, picngo-sb, picngo-qm. First two use the SS PIC, last one the QFN. Middle one has PCB USB connector. Board is 1.6mm so it needs a thickener to fit snug in a USB slot, but advantage no cable. The other two have pads for a SMD mini-B connector.

All have pads for:

2 decoupling caps on VDD.
Pullup resistors on I2C pins.
LED and current limit resistor.
Ferrite beads on USB lines (can be bridged to bypass).
MCP1700 3.3V reg with 2 caps.
2x 10 pin headers (each side).
1x 5 pin header (bottom ICSP).
2x 3 pin jumper headers.

Left Jumper selects LED connection - VDD or RC0 or N/C.
Right Jumper selects VDD connection - USB or USB via LDO or N/C (external)

This is kind of a work in progress - test boards for a possible educational kit with project guide book later on (like in 6 months maybe).

PCB Layout, 2 layer, 16mm x 50mm board.

BkkCoins,

Could you provide us similar information on where to drill the holes for K16. Ima nice diagram like Bick posted for the K1 would be immensely helpful in getting the mounting holes drilled.

Thanks

Thanks for the great update I'll want one or two of the breadboard adapters. Will you be adding them to the store?

That looks right to me.
Take note on the board there are a few parts on the back side. The LEDs at the left edge and the ferrite beads at the right edge. So a heat sink would need to fit between, or have cut outs. I believe I made sure the 32mm length would fit between but you'll want to check the gerber positions quite carefully to make sure the heat sink size and holes do not sit on top those parts.

What's the hole dia? 2mm?

cheers,
kev

Many thanks BKKCoins... that puts our mind at ease. Now all we need are Avalon chips to magically appear out of the ether. Thanks for the quick KiCad lesson as well.

EDIT: So is this correct?

and theoretical max for a k16 is: 450MHz @ 1 hash/Hz: 450*16=7200Mhash/second.

I believe the last stable point that was tested with the 1st revision of the board with some additional off-board fixes was 350mh per chip, but this was not with a fully populated board.  If that speed holds up, then one board would be 5.6gh/s.

In response to the Thailand Bitcoin fiasco discussion...
It was only the bank making a recommendation that Bitcoin be illegal - they have NO legal authority to make laws.
Here is a good article that explains it:
https://btcglobal.net/blog/post/bitcoin-is-not-illegal-in-thailand

Sorry i've lost some details.

How much hashpower do you manage to record with a k16 (or k1)?
I heard about burnin reaching more than 400 Mh/s with a single chip.

I don't have a drawing but the best info is from the nano.drl file in the gerbers.
The way I got exact numeric values is to load the gerbers in gerbv and then use the arrow tool to select objects and right clock "display object details". Doing that I get the following:

board edge start pos: 3.937, -3.937  (for determining relative co-ords)

top right hole: 5.1811, -3.0512
bottom left: 4.1142, -3.8386

subtract and convert to mm, and round off:

top right: 31.6, 22.5
bottom left: 4.5, 2.5
(0,0 would be bottom left edge of board)

If I use the gerbv measure tool to draw a line and measure it corresponds as well.
this info can also be extracted numerically from the nano.drl file, taking into account the mil units conversion, except it's not coded there where the origin is located.

T3
X41142Y-38386
X51811Y-30512

The holes are 2mm diameter, notated as T3C0.079 in the nano.drl using inches just to be confusing.

Was trying to read the gerbers to get a drill hole template for the K1s and just don't trust my limited skills with the Kicad.

Are there any pdf or svg already created that have the drill hole dimensions in text on the file for the K1 V0.3.0?

Just want to double check and that our heat sink fabrication gets it right as we don't want our pretty red anodized heat sinks to be useless.

I doubt I'll actually produce any boards or design layout for K64 - though technically I kind of have one in my hand right now. A 4up panel of K16 boards. A good layup of a K64 would join the klego pins and remove extra connector pads to save parts. However, joining together all the 12V power is difficult and would require a connector assembly that could take 4x as much current. It is probably is feasible to join each pair of K16s into one PCIe connector and save on 2 anyway. I don't think K64s should be built until working K16s are tested in chain configuration.

BKK UPDATE:

I picked up first revised K16 samples from Bkk yesterday and will be working to build a final test unit now. Also, my first bulk board order arrived in Bkk today but  I won't be able to get them until Monday (after I build one here).

Other notable updates:

Have solder paste and tweezers for web orders but need to pick up SD cards still. I ran out of time yesterday - but no worries yet as the final firmware and driver are not ready anyway. I did a test of solder paste before going to Bkk to confirm between two batches. Turns out the cheaper batch was garbage. I mean literally - hardly no solder in the paste. So scratched that place off my list and went with the stuff that balls up and makes nice silver joints as it should. Unfortunately I'll be selling for about the same price I buy at now - which isn't a problem as I wasn't really expecting to make anything off these accessory add ons. Just so people know the solder paste I got is XG-50 not MCN-300. Apparently this is the new version and the jars I have are all June 2013 instead of Sept. 2012 (which is what MCN-300 would be). Checking again once home I saw that XG-50 is what I used on my boards and MCN-300 is the jar I took a photo of. Doh! And getting finicky it looks like the jars have 42g now not 50g - but that's better than getting older paste.

I have Klego connectors in hand now and Klego PCBs arrived today in Bkk along with my K16 boards. K1 boards still 5-7 days out probably.

And for anyone who wants I'll have a PIC16LF1459 breadboard adapters in the K1 batch too. Allows mounting the PIC with USB and power reg. in a breadboard. Size is 16mm (0.6") x 50mm (2"). I just made these up to fill a void in the order and allow me to have a platform for doing the PIC programming without messing up K16 boards, and they'll be handy for future prototyping projects.

All the PICs, both QFN and SS, arrived as expected. So now it's mostly a matter of finishing firmware changes so I can program them ready to ship with boards next week.

Oh ya, this one's important. I couldn't get K1 Heat Sinks this time. All the stock had been bought up and they didn't know what I wanted (my Thai is not so good, actually it's pretty terrible). I'm going back with my sample pieces again but I may be looking for an alternate supply now. which is a real drag as I thought I had a good supply worked out.

Word I got from Microchip is that they are making PIC16LF1459 to order. So they won't appear on store shelves if you're waiting. You need to place an order with Microchip Direct to trigger the production allocation needed. If people had placed their orders back in June they would be on hand now.

Bkk answered it.
[/quote]
Thanks! i guess i missed it somewhere 

Ok, that's what I figured.  Thanks.