Junction Noise

[Magnus Danielson]

From: Brockr0 at cs.com
Subject: Re: Junction Noise
Date: Wed, 6 Sep 2000 19:54:20 EDT

> Magnus, cfmd at swipnet.se writes:
> 
> > So similar but oh so diffrent.
> >  
> >  Are you sure you meant NOR? XOR or XNOR is more like it.
> >  
> >  
> 
> I'll get this right yet.
> 
> NAND and XOR actually, for CRC generation. Input data NANDed with last stage 
> output fed back to the first stage and XORed at the polynomial taps,except 
> they are inputs so they are not really taps... 
> 
> If you want a pseudo-random pattern generator, you don't have an input data 
> stream, so change the output NAND to NOT and feed it back to the first stage 
> and the polynomial XOR inputs. I don't know if every prime polynomial will 
> give a maximal length non-repeating string with no invalid states but I 
> believe the common communications CRC polynomials such as CRC-32 and CRC-16 
> work this way as they are chosen for this convolution  characteristic.

They have to be irreduceable also, and you need to test it specifically for
this.

Cheers,
Magnus