Oscilloscope digital filter effect, adding ringing on sharp edge. why is that?

[David_]

Hello.

I have a budget Rigol DS5062C 20MHz dual channel scope, I am playing around with the digital filter options while observing the scopes square wave reference used to compensate probes.

While I add a lowpass filter it appears ringing on the edges of the square wave, why is that?
I can't think of why.

Regards

 

[D.A.(Tony)Stewart]

A perfectly balanced probe is flat response. Any imbalance in R ratio or C ratio results in a 1st error response. Any additional filtering will be 2nd or higher and likely have ringing unless overdamped.

-even if the probe is perfectly balanced, any N>1 order LPF will have ringing unless critically damped. (butterworth) which is unlikely if it is a Nyquist filter...

 

[vGoodtimes]

@SunnySkyguy -- critical damping would be a Q of 0.5. Butterworth has a Q of 0.707. Even Bessel has ringing (Q = 0.577).

@David_ -- Not sure what the filtering method is. Many lowpass implementations can do this. Do you have any information of the lowpass filter implementation?

 

[D.A.(Tony)Stewart]

@SunnySkyguy -- critical damping would be a Q of 0.5. Butterworth has a Q of 0.707. Even Bessel has ringing (Q = 0.577).

@David_ -- Not sure what the filtering method is. Many lowpass implementations can do this. Do you have any information of the lowpass filter implementation?

Critical damping (ζ = 1)
Over-damping (ζ > 1)
Under-damping (0 ≤ ζ < 1


There are many ways to design a LPF.

 

[tggzzz]

Hello.
I have a budget Rigol DS5062C 20MHz dual channel scope, I am playing around with the digital filter options while observing the scopes square wave reference used to compensate probes.
While I add a lowpass filter it appears ringing on the edges of the square wave, why is that?
I can't think of why.
Regards

You will get more relevant responses from people if you characterise your problem more accurately. In this case the key parameters you omit to mention are the LPF bandwidth, the frequency of the ringing, and the relative amplitude of the ringing.

You may care to understand the information in the Tektronix references given here: https://entertaininghacks.wordpress.com/library-2/scope-probe-reference-material/

 

[FvM]

Unfortunately I have add another correction to the quoted "filter theory". Q is a parameter of the second order building block ("bi-quad"), not the complete filter characteristicly unless it's a second order filter. The bi-quad sections of a higher order filter have typically different Q factor.

It's in fact unusual that a signal filter provided by an oscilloscope shows ringing because oscilloscope signal processing is focussed on correct time domain representation. There's no information about filter characteristics in the Rigol manual.

A bessel characteristic is often appropriate for time domain related filtering because it has very little (less than 1 %) overshoot. To get absolutely no overshoot, I would go for a gaussian characteristic.