Spicing Colpitt osc?

[walker1]

I am trying to simulate a crystal Colpit oscillator in the free B2-Spice edition. (release 4.2.1)

I do have oscillation, but only a few mVpp, and only very near a perfect sine. (No harmonics it seems :-(

The crystal is composed of 20fF, 1.25mH and 30 Ohm resistor in series. All shunted by 4.5pF. (Roughly 30MHz.)
These parameters seem to be in the correct area if compared to various ex. out on the web.

I have tried other oscillator circuits, but more or less with same results.

If I instead change the series LC parts a factor of 1,000 (down in impedance) I get a nice Colpitt waveform of more than 1 Vpp

Is it my model or is it Spice, that is incorrect?

 

[MacGyverS2000]

Semantics, maybe, but SPICE is (almost) never incorrect... it's a computer simulation, which is key. There can only ever be two issues when a computer model does not match up with the real world:
1) The model does not accurately represent the real-world item.
2) The simulator does not take all necessary real-world variables into account.


Dan - Owner

http://www.Hi-TecDesigns.com

 

[IRstuff]

SPICE is also simulating an IDEAL environment. There is no noise, no mismatch of gain, etc. If you start SPICE out in the metastable state, it will most likely stay in the metastable state. You might try initializing your circuit into an unbalanced state.

TTFN

FAQ731-376

 

[Comcokid]

Spice is great for simulating many circuits. But there are some circuits that are very difficult to get an accurate simulation. Oscillators are among those circuits that don't spice well.