Topic: Klondike - 16 chip ASIC Open Source Board - Preliminary - page 48. (Read 435369 times)

I would imagine that they'll send the appropriate gerber files off to their choice of board manufacturer to have them built.  BKKCoins open sourced this, he's not the exclusive supplier of boards - anyone can have the PCB's made.  The K-64 is just three additional cut-and-pastes of the K-16 with the proper spacing between the copies.  Theoretically you could saw the K-64 board up into four separate K-16s (although if I were doing it, I'd tie the Vcc and GNDs together across the boards and eliminate the redundant power connectors at the very least.)

So Bkkcoins: will you be producing K64 boards (4 K16 all on one PCB board) or no? Wondering how Terrahash will be getting the K64 boards for the DX large boxes...

Didnt a bank block a mtgox account and claim its illegal? MTGox sued them and the bank had to unblock it. So i wouldnt trust a banks decision that wants to control money.

Bkk answered it.

I'm not sure whether this will affect any thing/process here. But I believe thai government has issued a ban on bitcoins...
http://www.wired.co.uk/news/archive/2013-07/29/bitcoin-illegal-in-thailand
http://qz.com/110164/thailands-infamous-bitcoin-crackdown-is-not-quite-what-it-seems/

I believe TS is from thailand?

K16 samples likely arriving tomorrow. K16 final boards just shipped, will be a few days.
I expect the first K1 boards I'll see will be next week.
All of these are the latest revisions I've done.
K16 samples are QFN.
K16 finals are half/half QFN and SS. K1 is QFN only.

Sorry I meant are the sample/test boards for the final revisions both k1 and k16?

K1 is still production. Only K16 shipping now.
I didn't order as many K1 this time since it is less sure to be final.
But enough for my current orders, and they may not take as long as the K16 for production.
However, I am a bit overloaded with work now and I'm not sure how quickly I can get everything turned around - only one of me here, scrambling on many different jobs at once.

K1 and k16?

UPDATE:

I just received shipment notification of first Final Beta boards from China. I don't have a tracking# yet but they should arriving fairly soon - possibly by Monday. Good to know they are done.

The boards are in production now. I have shipped samples due any time now, but the main batch hasn't finished and shipped yet. They could finish any time now, and once I have shipping details I'll be able to plan for re-packing and shipping out.

I'm leaving for Bkk tonight and hope to get the samples tomorrow. Then I'll return and assemble one complete board for final testing. If it goes as expected it should be coming together by this weekend. The boot loader code is partly done but needs more work and testing. It'll be close but if I can finish it up I should be programming PICs in time for final boards delivery.

The boards will be coming DHL so once I have a shipping notice they should be not long to get here, turn around and ship out. Most of my other new parts orders are waiting for me already in Bkk, including more ASICs for testing (but none of my personal supply yet). Tonight's trip to Bkk will also see me picking up all the sundry items like solder pastes, tweezers, etc for orders.

Thanks, i'm still willing to give it a go and see what happens. Worst case scenario is I remove and replace the chips; I have the skills so i'm not too worried. Maybe I can use those ones as slaves as suggested above..

@BkkCoins  When are you expecting to receive the new PCB boards and do you have enough ASIC's to populate it fully?
To would be nice to see what a full board will do with the changes that were made.

I didn't notice that earlier post about using the F version, and sorry I didn't comment on it.

You cannot currently use the F series on Klondike boards. Even though they are very similar and do everything the same they cannot run USB on 3.3V power reliably due to having an internal LDO regulator for USB. I have been checking with Microchip on ways to bypass that but so far they have not given me reason to believe it can work.

This means it will work ok if being chained but when using USB it may not play nice. The info I have from Microchip is that the LDO will operate in drop-out mode and provide lower than 3.3V to the USB circuit. The circuit can work that way but it's signalling levels will be below USB standards and may not work reliably in some cases. You cannot run it at 5V on the K boards because the other chips and ASICs require 3.3V.

I am still in discussion with a Microchip engineer about this subject but my last info is not promising.

I have a bad feeling there's going to be a lot of failed DIY adventures.  But, some could learn a ton of soldering and debugging skills in the process.

Hi, sorry to dig up such an old comment, but i decided to take a risk based on the above advice and ordered 7 qfn PIC16F1459-I/ML (before the web store was up, Doh!)
This will be for 4 k1's and the remainder k16's. I'm a little concerend now that maybe the k1's wont have enough juice to power these chips?

The K16 has a 4 pin I2C connector but 1 pin is clipped off and used as a orientation pin for a plugged hole on the female end. That pin has VCC (3.3V) on it and was intended for later to supply power to a repeater module allowing extending the bus. GND, SDA, SCL are pins used.

The I2C provided uses serial# arbitration as an addressing negotiation to assign addresses. The addressing limit is 112 devices but that does not determine the real bus limit as physical effects such as line capacitance have greater influence. I think we can trust that 8 will work but it could be higher.

Addressing doesn't depend on pins or jumpers as in some I2C implementations. Here's how it works:

1. Master sends a query for new devices on the bus.
2. All new devices respond with their serial #.
3. Since I2C is wired-or (has pull up resistors on lines), only zero bits actively actually influence the data captured by the master. This ensures the serial# received is the "lowest one", which acts as a priority filter.
4. Master increments the device count and sends back the new address with serial# as assignment.
5. The device which recognizes it's own serial# stores and uses that address.
6. The process repeats until no response is detected by the master, in which case assignment is complete. The last address value indicates the total slave device count.
7. Periodically a new scan is done just in case devices are added to the the chain.

Each PIC device will be programmed with a unique serial# to enable this method using an updated version of the ktest app via the firmware updater facility, or via manual setting if using a PICKit3. The ktest app is automated so that each device it sees without serial# will cause it to increment and assign a serial#, and record the serial# in a file. That functionality is being coded by me right now and will be pushed to Github in time for PIC/board arrival.

If anyone still needs the Murata 38L361C inductor, I just received a couple reels of its approved replacement, the Sumida CDEP105NP-0R3NC-88, stamp 0R3NL.

AFAIK these are sold out everywhere, and I had pretty poor luck sourcing the original Murata part from Chinese vendors at any reasonable price, or at all. I definitely won't be using them all, so let me know and we'll work something out.

I2C is only 2 pins -- clock and data.  (+1 for ground, so really 3).  The 3 vs 4 is probably referring to addressing -- on I2C boards you can use jumpers to select an address. I'm not sure how the boards are picking their addresses.

4 pins.