to avoid oscillations

H

Thread Starter

Henri

Does anyone have a good solution to avoid oscillations in a window comparator, handling slow, but high impedance, input signals?

I prefer a solution that doesn't affect my feedback & setpoint signals.

thanks.
 
R

Robert Scott

About the only solution is to add hysteresis to each threshold, but that does affect the setpoints.

Robert Scott
Real-Time Specialties
Embedded Systems Consulting
 
1- add a low pass filter to reduce signal bandwidth to the minimum your process requires, as opposed to whatever is happening by default now. This will have the effect of reducing noise bandwidth.
2- add or increase hysteresis to your comparator until it is larger than the noise floor.
3- reduce noise coupling into your input signal
4- reduce sensor output impedance. Place signal conditioning at the sensor rather than after the long high impedance cable.

The above combined will eliminate most output instabilities, assuming they're not due to the window comparator internal noise, which is most often power supply generated (conducted) due to inadequate filtering or regulation.

best regards,
Matt Tudor, MSEE - RF Consulting
 
C

Curt Wuollet

If the design is clean otherwise (single point grounding, short leads, etc.) the choices are bandwidth control or hysteresis. The second will affect your setpoints. I seen some layouts that simply could not be persuaded to be stable. And protoboard lashups can be counted on to be among these.

Regards

cww
 
Y

Yeasir Rahul

Dear Henri,

Use a Schmitt trigger like feedback in your comparator. It will affect the setpoint a little bit, but that is necessary for avoiding the oscillation during cross-over.

If you are already using Scmitt triggering techniques and still experiencing oscillations, then consider increasing the thresh-hold gap.

Cheers.

Yeasir Rahul
VoltSmith Technology
http://www.voltsmith.com
 
Thanks Robert & Curt for your response. I have to digest this, and let you know what solution I decided on.

Henri
 
Top