Pseudorandom Noise Shift registers....

[Chris Crosskey]

Hi Folks, 
I've been following the thread about this and would like to add my own 
comments to teh pile. As far as  I can make out, with a pair of 4006 
CMOS shift register arrays plus an XOR gate plus an oscillator capable 
of running at approx 50kHz, well within the scope of a simple 555 
circuit, it is possible to have a pseudo-random generator with a 
repeat time in the order of five hours. The 4006's need to be 
connected as follows...

4006/A

Pin 1: Output of XOR
Pin 2: Pin  4 of 4006/A
Pin 3: Clock circuit
Pin 4: Pin 2 of 4006/A
Pin 5: Pin 11 of 4006/A
Pin 6: Pin 10 of 4006/A
Pin 7: Gnd
Pin 8: Pin 1 of 4006/B
Pin 9: 100K resistor to Ground
Pin 10: Pin 6 of 4006/A
Pin 11: Pin 5 of 4006/A
Pin 12: 100K resistor to ground
Pin 13: 100K resistor to ground
Pin 14: +15V


4006/B

Pin 1: Pin 8 of 4006/A
Pin 2: Pin 4 of 4006/B and input A of XOR
Pin 3: Clock circuit
Pin 4: Pin 2 of 4006/B
Pin 5: Pin 11 of 4006/B
Pin 6: Pin 10 of 4006/B
Pin 7: Gnd
Pin 8: 100K resistor to ground
Pin 9: To output buffer and Input B of XOR
Pin 10: Pin 6 of 4006/B
Pin 11: Pin 5 of 4006/B
Pin 12: 100K resistor to ground
Pin 13: 100K resistor to ground
Pin 14: +15V

The XOR can be a 4070, and it would be a good idea to have something 
that forces a 1 onto pin1 of 4006/A at start-up. I think if you use a 
4077 XNOR then you don't need the start-force circuit. I leave you lot to 
design your own drivers and buffer/filter arrays, suffice it to say 
that according to my maths at 50kHz clock speed this circuit will be 
giving you slightly better than audio white noise with a repeat time 
of nearly five hours. This is taken from the table in Horowitz and 
Hill for a 33-stage shift register with the XOR loop running from bit 
20 and bit 33, giving a repeat time of 8589934591 clock cycles, at 
50kHz that is 17,178 seconds, which is 286.3 minutes, approximately 
four and three-quarter hours....

chrisc