As you may know, the oscillating frequency of ring oscillator changes significantly with supply voltage variation.
Are there methods to reduce this effect?

 

Fixed the supply voltage?

This effect allows them to be used as vco.

 

Probably, the key would be to make a stage with a delay that has only a small dependency on power supply voltage, and that would be highly dependent on the topology of the inverters in your oscillator.

If you were to change the inverter to latches clocked by a stable (example: quartz) clock source, taking the outputs from NOT-Q there would be little dependency on the power supply voltage.

 

By any chance would you have a schematic of the circuit you have in mind so some specific comments can be made?

 

Thanks for the help.

Fixed the supply voltage?
This effect allows them to be used as vco.

No, supply voltage varies a lot. So I need to fix frequency.

Probably, the key would be to make a stage with a delay that has only a small dependency on power supply voltage, and that would be highly dependent on the topology of the inverters in your oscillator.

I am using the inverter topology as below. Are these some ways to reduce the dependency?

If you were to change the inverter to latches clocked by a stable (example: quartz) clock source, taking the outputs fromNOT-Q there would be little dependency on the power supply voltage.

I can't use quartz here so it would be difficult.