ROB SEWARD   ::   WORKS   BIO   CONTACT   BLOG

Make your own True Random Number Generator 2

Intro

This generator uses avalanche noise, and is based on a designs by Will Ware and Aaron Logue. Herein are instructions as to how you can use Arduino to analyze a noise source and output random data serially. 

There are two types of random numbers: true and pseudo. Pseudo random numbers are created by an algorithm. The problem with this is that if someone knows what algorithm you use, it is theoretically possible predict what numbers you will create. True random number generators create sequences that are impossible to predict. They use random physical phenomenon as their source or randomness. They are used for encryption and Princeton's micro psychokinesis research.

Notes on the New Design

The previous RNG had a problem: after about 3 months of use the noise signal would drift and the ratio of 0s and 1s would go out of balance. 

While trying to debug this problem I redesigned the whole board.  My redesign was based off of Aaron Logue's design.  It functions on the same principle as the first RNG, except it uses fewer components and the there is no signal interference from the RS232 chip.  

Even with the new design, however, the new RNG started to behave just like the old one. After 3 months the output became not random. I was just about to build a new board from scratch, when it dawned on me that the signal (even though it looked OK on the oscilloscope) may have changed and caused the ratio of 0s and 1s to drift. I suspect component (or board) aging was the culprit.

So I wrote a program that calibrates itself.  Thus if the signal changes, the micro-controller adapts. It fixed the problem.

I also rewrote everything for Arduino. The Arduino API is better designed than the PIC libraries I've used (I've tried MPLAB ICD 2, PIC C, and PicBasic), way cheaper, and as far as I can tell, every bit as capable. Also it's open source. Also, if you have a breadboard and the parts you can assemble the circuit very quickly. I've also designed a breakout board. I haven't built it yet, but if you are feeling adventurous contact me and I'll email you the CAD files.

The two transistors with their bases touching create "avalanche noise in a reverse-biased PN junction." This noise is amplified by the third transistor, and sent across a voltage divider to Arduino. On the oscilloscope, the signal looks very fuzzy. In software, that randomness is converted into a stream of 1s and 0s.

The code works like this: for the first 10 seconds, record the signal from the circuit on the analog input pin. Find the median. Then for each subsequent reading of the input pin, see if it is above or below the median. If it is above, output a 1, if it is below, output a 0. Filtering can be applied to the output stream of 1s and 0s to reduce bias.