Re: TRNG with 18 Mbit/s using Thermal Noise

Rob Warnock wrote:
Jo Schueth <jo@xxxxxxxxxxxxx> wrote:
+---------------
| in a home-brew project, I recently built a TRNG based on thermal noise
| that achieves a net data rate of 18 MBit/s after post-processing.
...
| http://www.schlaupelz.de/TRNG/Highspeed_TRNG.html
| | The point is not that you can build a TRNG with good quality output
| from readily available parts at home, but rather the high bandwith
| achieved with this design. ...
+---------------

Congratulations! It's always rewarding to build something and get
it working.

I would just note, however, that since you already have a comparator
and at Atmel micro board, I don't see why you really need the
additional complexity of the LNB & antenna. A standard microwave
noise diode (of the sort used to calibrate pwoer meters) connected
to your comparator (possibly with a cheap op-amp pre-amp stage)
should give you thermal noise every bit as fast & good as your
LNB+antenna, but mountable right on your Atmel board.

Or if you're willing to settle for Zener noise a.k.a. avalanche noise
(*almost* as good as thermal, for RNGs at least), then you might want
to consider one of these approaches:

http://world.std.com/~reinhold/waynesrngcomp.gif
http://www.maxim-ic.com/appnotes.cfm/appnote_number/3469
http://willware.net/hw-rng.html
http://www.cryogenius.com/hardware/rng/
http://www.cryogenius.com/hardware/isarng/
http://www.ciphersbyritter.com/NOISE/NOISRC.HTM
http://electronicdesign.com/Files/29/6356/Figure_01.gif


Thanks for these references. I had experimented earlier with Zener diodes
and also with transistors used like Zener diodes, but never got high
bandwidth, and even observed some long-time correlation effects like
flicker noise. So I really have more trust in thermal noise.

You would, of course, need to do de-biasing (Von Neuman or eq.)
and whitening (SHA-1 or eq., perhaps), but you need that anyway
for your LNB+antenna design...

I don't do de-biasing, as the differential tuner output and the comparator
have reasonably low offsets. Instead of SHA, I use AES for entropy extraction
(16 bytes in, only 12 bytes output used).



-Rob

-----
Rob Warnock <rpw3@xxxxxxxx>
627 26th Avenue <URL:http://rpw3.org/>
San Mateo, CA 94403 (650)572-2607

.

Relevant Pages

  • Re: Capacitor-feedback for low noise
    ... then you need a model for the noise behavior. ... Equivalently, its thermal noise can be modeled by a Norton equivalent circuit, where the current source in parallel is random with a spectral density of 4kTRe) A2/Hz. ... The usual way to derive the Johnson noise formula for a resistor is to use classical equipartition of energy, which predicts that any single degree of freedom, e.g. the charge on a capacitor, has an RMS energy of kT/2. ...
    (sci.electronics.design)
  • Re: TRNG with 18 Mbit/s using Thermal Noise
    ... I recently built a TRNG based on thermal noise ... and at Atmel micro board, I don't see why you really need the ... noise diode connected ... should give you thermal noise every bit as fast & good as your ...
    (sci.crypt)
  • Re: TRNG with 18 Mbit/s using Thermal Noise
    ... I recently built a TRNG based on thermal noise ... and at Atmel micro board, I don't see why you really need the ... noise diode connected ... should give you thermal noise every bit as fast & good as your ...
    (sci.crypt)
  • Re: Why not make the sensor larger?
    ... Thinking common f/ratios deliver the same photons ... Thermal noise is negligible in most situations. ... Where N = total noise in electrons, P = number of photons, ...
    (rec.photo.digital)
  • Re: looking for article in EDN: Test Whether a Noise Source is Gaussian
    ... I prefer Gaussian or non-Gaussian) or not, of various noise ... Next is "Analyzing Noise Generators". ... this is the normal output of a zener noise generator. ... I hope using 2 to 5 ns comparators, ...
    (sci.electronics.design)
Loading