The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> Mixed-Signal Design >> Cont. Time Vs Disc. Time Sigma Delta Modulators
https://designers-guide.org/forum/YaBB.pl?num=1291297050

Message started by jefkat on Dec 2^nd, 2010, 5:37am

Title: Cont. Time Vs Disc. Time Sigma Delta Modulators
Post by jefkat on Dec 2^nd, 2010, 5:37am

Hi guys,
Could anyone say what are (if any) advantages to using continuous time sigma delta over discrete time (Switched Cap) Sigma delta modulator?

Thanks for your time
shaf

Title: Re: Cont. Time Vs Disc. Time Sigma Delta Modulators
Post by raja.cedt on Dec 2^nd, 2010, 7:32pm

hi,
you can find this in many text book or else you can have look at this below website. The main advantages i could see is implicit anti -aliasing filtering and variations in the the input impedance of the modulator is less so that poles and zero locations of the loop won't change much.

http://www.ee.iitm.ac.in/~nagendra/videolectures/doku.php?id=200801vlsiconf:start.

Title: Re: Cont. Time Vs Disc. Time Sigma Delta Modulators
Post by vivkr on Dec 3^rd, 2010, 6:51am

raja.cedt wrote on Dec 2^nd, 2010, 7:32pm:

hi,
The main advantages i could see is implicit anti -aliasing filtering

and variations in the the input impedance of the modulator is less so that poles and zero locations of the loop won't change much.

The first point is clear of course. You have implicit antialiasing which helps a lot in some cases.

The second point is a bit obscure to me. Why should the loop pole/zero locations have anything to do with the input impedance. At any rate, the fixed absolute location of poles/zeros in the NTF is one of the major DISADVANTAGES of using a continuous-time modulator, especially since you are likely to want to use a modulator for different frequency ranges, e.g. in case you want to make it for a wireless application, then for different standards.

A discrete-time solution in this case is more elegant as the NTF remains independent of the operating clock frequency, allowing the same modulator to be scaled and used for different applications, provided that the settling requirements are always fulfilled.

And there lies the catch. Since discrete-time modulators require very good settling and are typically realized in form of switched-capacitor filters, they consume much more power for the same operating frequency, and are also not useful once you start clocking at really high rates.....

There are other differences too of course, but these seem to be the most important ones.

Vivek

Title: Re: Cont. Time Vs Disc. Time Sigma Delta Modulators
Post by jefkat on Dec 3^rd, 2010, 8:06am

Thanks for the responses...
The amplfier in Cont Time SDM has to settle as well since the DAC feeds in a step ( or I missed something ?)
In disc. time case the amplifier has to settle to enough accuracy so the question is are the requirements on amplifier (bandwidth wise) in case of Cont. time less stringent because may be they have more time?

Title: Re: Cont. Time Vs Disc. Time Sigma Delta Modulators
Post by vivkr on Dec 9^th, 2010, 6:05am

indeed, it is true that the amp in a CT DSM has to settle too. However, let's put it this way - comparing a DT DSM to a CT DSM is like comparing apples to oranges as the two have very different features. let me try to answer your first question though:

A DT modulator is usually restricted by its step response, and more specifically by the slewing in its amplifier. In a CT modulator, slewing is not allowed, which means that the amplifier technically is always operating in a "small-signal" mode, and the step response is not the limiting feature by design, provided that the design was done correctly. A couple other points help too, like the fact that you will likely have (will need) a multi-bit DAC in a CT DSM to get away from the problem of jitter. So you will intrinsically have a smaller DAC step in most cases.

The other side concerns linearity. DT DSMs typically are made for relatively high linearity => the settling of amps becomes more critical. Since the CT DSMs are usually not running in a slew-limited mode, the settling is not going to determine the linearity. Of course, the comparison is again flawed since you will anyway not be designing (and will also not achieve) high THD with a CT DSM. These aim for more modest THD.

Hope that makes sense.

Vivek