PN Shift Register Generators
In my first or second year on the job, around 1976, I was tasked to come up with a noise generator to be used in some test equipment. I don't remember the requirements, but I do remember considering whether to try a digital approach, using shift registers, or an analog approach, using an avalanche diode at the input of an operational amplifier.
The digital approach is very much along the lines of a Digi-Key 2018 article by Art Pini, Use Readily Available Components to Generate Pseudo-Random Sequences and White Noise, available here
I seemed to think 16 bits would be a useful size for the shift register, but the only literature I could find listed the critical feedback taps for smaller length registers.
Had I seen at the time the just published book, Spread Spectrum Systems, by Robert C. Dixon, I would have been able to select from the tap configurations listed for shift registers in lengths from 2 to 33 bits.
The literature seemed to indicate the only way to discover the feedback taps for the required maximal length sequences was by trial and error. I had access to a 16-bit mini-computer, the TI 960B, so I came up with an assembly language program that would run a simulated generator until it paused when a maximal sequence was detected. While I laboriously entered the program using the 16 front panel toggle switches, a programmer had mercy on me and fixed up a keyboard that allowed me to enter hex values for the instructions.
It was exciting when the program paused when finding taps corresponding to those I had seen in the literature. But when the register length was about 11 or 12 bits, the waiting time became uncomfortably long. I started to estimate how long it might take to search all the tap combinations for a 16 bit register. I sensed my career might not last long enough, so I proceeded with the avalanche diode solution.
At least twice in the years to follow, I had the opportunity to make and use a Shift Register Generator. They are very beautiful and fascinating circuits.