The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> Analog Design >> Oscillator + AGC
https://designers-guide.org/forum/YaBB.pl?num=1160155101

Message started by pink_panther on Oct 6^th, 2006, 10:18am

Title: Oscillator + AGC
Post by pink_panther on Oct 6^th, 2006, 10:18am

Hi,

I am trying to design a crystal oscillator with AGC circuit built in, in order to decrease the total power consumption of the circuit. The difficulty we are facing right now is the stability analysis of the whole loop.

The topology is a classical Pierce oscillator, where a Pmos current source feeds current into a Nmos device. This Nmos device acts as a gain element. We are AC coupled to the crystal signal. After AC coupling we detect the peak of this oscillation and compare the peak value to a known reference voltage. Depending on the difference between the peak value and the reference voltage, we adjust the value of the Pmos current source. Since the whole system is pretty non linear, an AC analyis in Spice is unsufficient.

Does anybody have any suggestion how to analyse this kind of circuit for any potential unstabilities? Any help is welcome, even a title of a paper.

Thanks.

Title: Re: Oscillator + AGC
Post by ACWWong on Oct 9^th, 2006, 4:39am

I guess you have to ensure your control loop is slow enough to ensure you don;t reduce the current too much to stop oscillation or make the oscillator have start up difficulties (especially cold and dozy crystal).
I can't suggest anything other than some transient simulations (artifical de Q to speed the sims up probably not appropriate in this case).... have you searched through the papers by Vittoz ?

Title: Re: Oscillator + AGC
Post by pink_panther on Oct 10^th, 2006, 1:05am

Thanks ACWWong.

I went through the papers by Vittoz. There is not much about the stability analysis as far as I can tell.

What would be the concern, if we artifically de Q crystals to speed up the simulations?

Title: Re: Oscillator + AGC
Post by ACWWong on Oct 10^th, 2006, 2:30am

I guess the concern with artifical deQ would be that the speed at which your control loop works is artifically increased, and you might not catch if your oscillator continued to oscillate as current is decreased accuarately.

Title: Re: Oscillator + AGC
Post by pink_panther on Oct 10^th, 2006, 3:41am

That is a good point.

Thanks for your inupts.

Title: Re: Oscillator + AGC
Post by ACWWong on Oct 11^th, 2006, 8:27am

This paper is quite good for what i guess you are trying to do:

W.Thommen,"An improved low power crystal oscillator", Proc. ESSCIRC’99, pp.146-149

it presents the use of the bias PMOS gm to contribute to the overall negative R using a coupling C variant of pierce oscillator... thus lower power...