[sdiy] Shift register sequence period

[Dave Manley]

Eric Brombaugh wrote:

> Interestingly enough, removing the inversion shortens the sequence to 
> 2^17-2. Using different start/end criteria can lead to degenerate cases 
> with shorter lengths as well. Odd...

Not so odd really.  Using this:

	http://www.ece.cmu.edu/~koopman/lfsr/index.html

as a reference, there are only 7776 taps arrangements that are maximal 
length.  All the rest are going to have a number of degenerate sequences.

To get a feel for this simulate a shorter polynomial, like a 4-bit or 
5-bit, you can exhaustively try all the feedback tap combinations and 
see all the possible sequences.

And besides, it's good enough for synth-diy!

And...at least it doesn't go chuff, chuff, chuff.

:-)

-Dave