Hello,

I am investigating an oscillator-like design in ADS.

I have a feedback oscillator using an OscPort. When I run the HB simulation (with Oscillator checked), I receive a warning message that says that the circuit can potentially oscillate at more than one frequency. ADS then chooses one of the frequencies (so far, always the lowest in frequency) and the data display shows an oscillation at the frequency ADS chooses.

My question is the following: How does ADS choose which frequency to oscillate when it has more than one potential oscillation frequency? Is it as simple as "Always use the lowest potential oscillation frequency" or is it something more complicated than that? If it is the latter, how is it chosen?

The logic behind ADS's choice will be of high importance for me.

Thank You,

Lou

I am investigating an oscillator-like design in ADS.

I have a feedback oscillator using an OscPort. When I run the HB simulation (with Oscillator checked), I receive a warning message that says that the circuit can potentially oscillate at more than one frequency. ADS then chooses one of the frequencies (so far, always the lowest in frequency) and the data display shows an oscillation at the frequency ADS chooses.

My question is the following: How does ADS choose which frequency to oscillate when it has more than one potential oscillation frequency? Is it as simple as "Always use the lowest potential oscillation frequency" or is it something more complicated than that? If it is the latter, how is it chosen?

The logic behind ADS's choice will be of high importance for me.

Thank You,

Lou

Harmonic balance (large signal) oscillator analysis starts initially with a small signal freq search. Â

Frequencies are searched between

freq1*2^(-NumOctaves/2) to freq1*2^(+NumOctaves/2)

where "freq1" is the initial frequency guess provided on the harmonic balance controller. Â As you have seen, the simulator will report when more than one potential small signal oscillation freq point is found.

The large signal oscillator solution then starts from the first small signal point it found. Â But, if there are more than one small signal points, then the smallest point is used. Â

However, just because the small signal analysis reports more than one possible oscillation freq point, it does not mean that all of them are capable of leading to unique (different) large signal oscillation frequencies. Â

In order to make the simulator use something other than the first small signal point, you can change the initial frequency guess, and limit NumOctaves such that the lower small signal point(s) will not be found (per the freq search range expression above.) Â 8)