pulses of high frequency on a signal of lower frequency in a relaxation oscillator

[bondadoso]

i made the relaxation oscillator of the picture with a tank circuit L=100uh C=1uf an it is oscillating but with pulses of higher frequency betwen transition of the oscillating frequency the picture shows, i don't know if it is for bad power source or bad gain or what this came from , because i need to clean up the signal and get a clean frequency oscillation, please help

 

[FvM]

The reason is simple. The oscillator circuit which is wellknown from LC-meters doesn't work at this high resonance frequency. With higher L and C values you get continuous oscillation. A considerably faster comparator would be needed to work with the given dimensioning.

 

[dick_freebird]

I believe this is just comparator chatter. You can add enough
hysteresis to cover up the various node-bounces that are
kicked by the first comparator switching event, of you can
move to a two-comparators, SRFF control that "ratchets"
and won't reswitch.

 

[FvM]

I believe this is just comparator chatter. You can add enough
hysteresis to cover up the various node-bounces that are
kicked by the first comparator switching event, of you can
move to a two-comparators, SRFF control that "ratchets"
and won't reswitch.

I fear, adding hysteresis will stop the oscillator from intended operation. In reasonable operation range, the circuit is working as harmonic rather than relaxation oscillator, by the way.

There are previous threads discussing the oscillator topology:
https://www.edaboard.com/threads/126119/
https://www.edaboard.com/threads/231205/

 

[crutschow]

What period oscillation do you want?

With C1 = 0.1µF and L1 = 2mH my LTspice simulation gave a pulse period of about 90µs.

 

[FvM]

The circuit operates continuously over a wide frequency range. But above a certain freqquency, LM311 gain isn't sufficient to maintain continuous oscillations. The circuit falls into relaxation mode with each edge triggering a decaying oscillation at the LC resonance. It appears as a pulse train at the output, as shown in the initial post.