Noise is something we normally avoid at all costs, in electronics as well as day-to-day life. But sometimes you need it to test the noise handling ability of some piece of equipment. Although I'm intending to use this for testing the NBTV/SSTV-625 up-converter it can be used in many applications by simply rearranging the output filter.
As shown the filter approximates a typical HF receiver's response of 300-3000Hz but the digital output has noise components up to 50kHz. It's based on a pseudo-random number generator which has 33 stages in the shift register, the random pattern goes though 2^33-1 patterns before repeating, about 8.6 billion. At the execution rate of the micro that's about once every 6.5 minutes - close enough to random for me!
You could do this in CMOS/TTL hardware but it would take six 14/16-pin chips plus an oscillator and the filter. I have seen a design for a hardware-based version that takes 913 YEARS to repeat using nine TTL chips plus oscillator and filter. It was used for gambling...which I abhor.
The maximum output voltage is +/-2.5V p-p. To measure the RMS value you need a true RMS reading measuring device with adequate bandwidth - not your average multimeter!
Steve A.
'The Brat', think of a noisy kid!