[SOLVED] VCVS Analog BPF Filter

[yyang13]

Dear All,

Im trying to build a VSVS BPF with fc = 32.8 Khz and gain = 10. Attached below is my design. Vin = ultrasound receiver(32.8Khz).
The problem with my circuit is that in its idle state, the whole circuit is oscillating around 15 Khz.
Are there any ways to improve the circuit to make it stable even when the input is idle?

 

[FvM]

I don't see a reason for instability with TL074 and the component values in the schematic.

 

[yyang13]

I don't see a reason for instability with TL074 and the component values in the schematic.

Hi, Fvm and everybody!

This is my circuit:


This is what i observed when i feed in my input


This is what i observed when there is no input


Is there any stability criterion that i will need to satisfied??
I have done a search in wiki and found that on this link
https://en.wikipedia.org/wiki/Audio_filter
under Self oscillation, something about my Q factor too high circuit will be come an oscillator
not sure if this is my case?

 

[LvW]

yyang13,
I see only one possible reason for oscillation: Unwanted feedback coupling effects due to the mechanical construction of the whole circuit.
Try to move some wires and check if there is any change.

 

[BradtheRad]

I find your circuit works well in simulation. 12x gain at 33 kHz.
As long as there is some incoming signal, there are no oscillations at any frequency from 1k to 60k.

However when I disconnect the input resistor, it oscillates at 16 kHz. It is not too different from the relaxation oscillator configuration.
The output is a square wave rail-to-rail, like your scope trace.

It might help to install a resistor somewhere, but it needs to be at a spot that will not result in Q being reduced.

 

[godfreyl]

When the input is open circuit there is positive feedback overall, and thus oscillation. To fix it, add a unity gain buffer before the filter.

 

[LvW]

yyang13,
I see only one possible reason for oscillation: Unwanted feedback coupling effects due to the mechanical construction of the whole circuit.
Try to move some wires and check if there is any change.

Ahh - I see. Of course, if you modify the circuit by disconnecting an element the circuit does not behave as desired. But is this a problem?
During operation the first resistor will always be connected to the input source, will it not?

 

[yyang13]

Pls to hear from all,

1) yes the circuit is indeed similar to a relaxation circuit,

as shown in

2) I have calculated and verified that the filter is working when im feeding in input, close to my desired fc
the only issue is how to make it stable such that when there is no input, there is no output as well

3) The oscillation stops when i removed the positive feedback, however then the circuit is no long a filter

4) I have tried with a diode place at the positive feedback, it helps to stop the oscillation, at the same time my input as well

- - - Updated - - -

When the input is open circuit there is positive feedback overall, and thus oscillation. To fix it, add a unity gain buffer before the filter.

Hi godfreyl, great advice!

I have configured my circuit as such:


the simulation test shows promising results, will update again once I am able to !

 

[FvM]

To be honest, I never suspected the filter might be operated with unconnected input. I simply can't see a reasonable purpose of this operation.

A unity gain buffer can be still useful if the source impedance is unsure.

 

[yyang13]

To be honest, I never suspected the filter might be operated with unconnected input. I simply can't see a reasonable purpose of this operation.

A unity gain buffer can be still useful if the source impedance is unsure.

Thanks Fvm! My input is connected to an ultrasound receiver.
The oscillation generated induced a pitch audible on the ultrasound receiver
Hence, the need to remove the oscillation.

hmm, do you think that the oscillation occur because the input impedance is not high enough?
I did a measurement at before connecting the unity buffer at and after connecting the buffer at Node A

 

[LvW]

yyang13,

in case you cannot afford a second opamp as a decoupling unity gain amplifier (in order to avoid the observed oscillations) - what about using another filter topology?
For example, you could replace all 3 resistors by 3 capacitors and vice versa (new values, of course). Another alternative is the multi-feedback (MFB) topology.