The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> Mixed-Signal Design >> Sigma-Delta MOD1: Quantizer gain
https://designers-guide.org/forum/YaBB.pl?num=1195293912

Message started by neoflash on Nov 17^th, 2007, 2:05am

Title: Sigma-Delta MOD1: Quantizer gain
Post by neoflash on Nov 17^th, 2007, 2:05am

In text books, the MOD1 z-domain model assumed quantizer gain as 1 for analysis.
Is this correct?

What if the gain is different from 1, any analysis done there?

Title: Re: Sigma-Delta MOD1: Quantizer gain
Post by vivkr on Nov 19^th, 2007, 2:32am

Hi neoflash,

The quantizer gain can only be assumed to be 1 for multibit quantizers and for reasonably large signal swing at quantizer inputs.

As far as I know, the quantizer gain needs to be calculated from the signal statistics for a 1-bit quantizer.
And the textbooks also do it this very way. I would recommend that you look at the textbook by Schreier &
Temes or the TCAS paper by Baird & Fiez.

Regards
Vivek

Title: Re: Sigma-Delta MOD1: Quantizer gain
Post by neoflash on Nov 19^th, 2007, 6:09am

thanks for your reply.

Then why in z-model, the gain is assumed to be "1" to get desired NTF and STF?

Title: Re: Sigma-Delta MOD1: Quantizer gain
Post by vivkr on Nov 21^st, 2007, 7:34am

Which z-model do you mean???

If it assumes a gain of 1, then there is the general assumption that the quantizer
is not being overloaded. For a 1-bit quantizer, I would suggest that one still does
not set the gain blindly to 1. However, it is probably used because it is easy to set
it to 1. The references I mentioned cover the topic in greater detail.

Regards
Vivek

Title: Re: Sigma-Delta MOD1: Quantizer gain
Post by Berti on Nov 21^st, 2007, 10:27pm

Hi Neonflash,

In the classical model, the quantizer gain has to be 1 when you assume a
unity closed-loop gain for the modulator.

However, Vivek is right. That gain will change depending on the input signal.
Signal-statistics are usually used to calculate the quantizer gain (and to
determine stability).

Regards