anelok: an ugly case and a bit of randomness

Werner Almesberger werner at almesberger.net
Mon Oct 14 12:38:37 EDT 2013


I made a first try at the top half of the case. I was hoping I
could make a rough model for the form factor from MDF, but the MDF
basically fell apart already during machining and I had to cut more
things than planned manually. The not so pretty result is here:

http://downloads.qi-hardware.com/people/werner/anelok/tmp/mdf-start.jpg

The "sunken" wheel is probably caused by untrimmed MDF fibers
pushing the PCB away from the border on which it should rest. The
holes for OLED and LED wouldn't exist in the real case - I'd just
use transparent acrylic for the whole piece.

Alas, I don't have acrylic of the right thickness at hand, so
this will have to wait until later this week.

This is what it should look like:

http://downloads.qi-hardware.com/people/werner/anelok/tmp/top-mesh.png

If anyone wants to play with the mesh:

http://downloads.qi-hardware.com/people/werner/anelok/tmp/top-mesh.stl.bz2


Last but not least, I wrote a first test for the hardware random
number generator in the transceiver. This will be the system's main
if not only source of entropy, so it's important that the RNG works
flawlessly. Getting the RNG in a crypto system wrong is what earns
you this kind of press coverage:

http://arstechnica.com/security/2013/09/fatal-crypto-flaw-in-some-government-certified-smartcards-makes-forgery-a-snap/

My test wasn't quite so exacting. I just pulled seven bits from the
RNG, used them as index for a "bin", and then displayed the bins and
the number of times they've been chosen on the OLED, repeating the
operation until one bin was about to reach 64 and thus leave the
128 x 64 display.

One result is here:

http://downloads.qi-hardware.com/people/werner/anelok/tmp/mdf-rng.jpg

At a first glance, the result looks reasonable: all the bins fill
up to some extent, some faster, some slower, and there are no
extreme outliers. The pattern also changes from run to run.

Of course, this needs much more analysis, but it's encouraging that
the random numbers do indeed appear to be random.

- Werner



More information about the discussion mailing list


interactive