[Cryptech Tech] Avalanche noise test boards

[Benedikt Stockebrand]

Hi folks,

I've finished soldering up the boards, gave them a very quick test (one
doesn't yet work) and just started a 24h test run on the first five of
them.

Before I can ship them I need the snail mail addresses of everyone
interested.  And I leave you to deal with customs:-)

@Bernd, Basil: Since you were interested in the topic but didn't (yet)
               say if you'd like a board: Would you like one, too?

If things work out I'll ship them some time on tuesday or wednesday, as
my current money making project permits and I can get hold of some
suitable packaging material.


As expected, a few non-critical things went wrong:

- The FT232 footprint didn't get any solder mask between the pads; fixed
  by using breakout boards on the test connectors for the FT232.

- I didn't ground the test pin on the FT232; fixed with a solder bridge.

- The Eagle footprint for inductors had too small diameter holes;
  kludged up with some component butchery.  *Please* be careful not to
  break them off; these aren't very robust.

Additionally, there's a list of things I'd like to change next time, but
none of them are as significant.  And in any case, the boards should do
fine for testing.  Just don't rely on board layout and pin numbering.


Usage:

- You can get GND both from the little pad on the underside of the board
  (kind of suitable for the GND clip on a scope probe) and TP13.

- I put BC337-16's into the socket for the avalanche diode/BE transistor
  junction.  If you want to try any other, either connect them there or
  between TP2 (cathode/emitter) and TP3 (anode/base) on the bottom test
  connector row.  @Warren: I'll try to remember packing the last set of
  Zener diodes and transistors in your parcel.

- When powering on, the bottom yellow and green LEDs should give 10
  short flashes with a 1 second delay between them; this is the power-on
  self test in the driver code.  Afterwards, the bottom yellow LED
  indicates if the device is idle, i.e. stopped by hardware flow
  control, and the green one if the device is actually
  generating/sending data.  (The other LEDs are as yet unused.)

- On any Un*x the devices should show up as a TTY device (ttyUSB[0-9]*
  on Linux).  They send a raw binary stream, input is ignored.  Line
  parameters are 8N1, 500kbit/s.

- If you use Linux, consider running something like

    % ldattach -8 -n -1 -s 500000 tty /dev/$DEV
    % rngtest -c 1 </dev/$DEV

- (Especially) Joachim: You can feed the circuit 5V from the USB
  interface and get the raw noise signal at 5V from pin TP8.  Flow
  control doesn't interfere with that, so you can simply ignore the
  MCU/USB part.  Either rely on the input protection of your device or
  use a voltage divider to bring the signal down to 3.3V.


Cheers,

    Benedikt

-- 
Benedikt Stockebrand,                   Stepladder IT Training+Consulting
Dipl.-Inform.                           http://www.stepladder-it.com/

          Business Grade IPv6 --- Consulting, Training, Projects

BIVBlog---Benedikt's IT Video Blog: http://www.stepladder-it.com/bivblog/