Subject: Re: insufficient entropy for rnd
To: None <>
From: Philip Stortz <>
List: tech-crypto
Date: 08/24/2003 03:10:45
interesting, though most of those sources will work at a relatively low
frequency (i.e. require a relatively long sample time per random #).  i've
considered making a really nice noise source for some time and using it for
random numbers.  does any one know of any links or papers?  i've looked at the
stuff at who make noise sources and it looks good.  i've also
considered making my own as follows:

i have a number of old crt's from o'scopes, they have some loose electrodes etc.
and aren't usable in scopes.  i've considered carefully breaking the nipple
where they were evacuated, and attaching a metal tube. then adding a small
amount of air/argon/helium etc. (maybe methane etc. since they are nice big fat
organic molecules with lots of ionization states).  the idea being to run the
tube "normally", i.e. run the electron gun at normal voltages and normal second
anode voltages and monitor the anode current with a resistor (using capacitive
coupling to eliminate the hv dc).  with the random collisions of the beam with
atoms/molecules/ions in the beam path, and their random arrival at the anode or
internal coating that's at the anode voltage.  this should produce nice high
frequency noise and be somewhat adjustable by varying the amount and type of
gasses added (i.e., make it a "dirty" vacuum).  the current signal would be
amplified if necessary and applied to a comparator to generate a random pulse
sequence at a high rate so "entropy" can be extracted rapidly (i.e. lots of
randoms in a reasonable time period).  the nice thing about scope tubes is that
they are designed to use a low second anode voltage and hence generate very  low
energy x-rays that are easy to block, so low energy in fact that people don't
normally worry about it (typically 5 or 6kv compared to 25kv+ for a color tv, or
100+kv for a dental x-ray machine).  this does occasionally happen with tv sets
and causes a very odd blurring which varies over time and over the screen from
very small amounts of air getting in a tube with a tiny, tiny crack.  another
nice thing about the scope tubes is that they are relatively long and skinny
providing a long beam path in a smaller space than most tubes.

is this a "good" way to generate noise?  is it "insane"?  is it no better than a
noise diode or traditional noise tube (i'm not too familiar with those, other
than having used a "radio shack" signal generator that used a noise diode to
generate a broad band signal suitable for injection into any stage of an analog
system).  is any one interested in this?  at best i probably wouldn't have time
to do this till next spring, but i have spent a lot of time thinking about noise
sources for randoms and other uses.  i suspect that one advantage over a noise
diode would be a higher power noise signal.  i tend not to trust randoms
generated by state machines etc., particularly because so much information
relevant to those systems has undoubtedly been classified and the large amount
of disinformation put out by those who know what's important better than we do.....

Daniel Carosone wrote:
> On Thu, Aug 21, 2003 at 11:19:35AM -0400, Michael Richardson wrote:
> >
> >     Daniel> pointed at lava lamps, fish tanks, busy streets, etc.  - various
> >     Daniel> hobby or commercial rng devices (diode noise, etc)
> >
> >   Check out:
> >

Philip Stortz--"Only love can kill the Demon", from Natural Born Killers, the
American people and Government would be wise to heed that simple Wisdom.  Blood
Shed feeds the demon and makes it stronger, Love and Forgiveness starve the
demon.  hatred and violence are the demon.  It's a good movie, if you can
realize it's a farce and know what a farce is....