State variable bandpass filter opamp selection

I have to design a band-pass active filter which will be a part of receiver of acousto-magnetic EAS system. Actually I'm doing a simulations using LTSpice and try to select a few opamps for further prototyping. It works fine, but I have a few questions. The key parameters are the following:

- f0 = 58kHz
- Q: ~10-50 tunable
- gain: ~20-40dB tunable
- input signal: from receiver antenna (LC network etc.) 1...50mV with a lot of mess, especially interferences from neighbouring SMPS etc.
- load: antialiasing filter and ADC in microcontroller

Goal: Detection of pure 58kHz signal from tag. It will be finally done using DSP in a microcontroller. A weaker signal

My questions:
1) How to select the GBW of opamp for state variable topology? Is the rule 100*f0*gain is correct? So 100*58kHz*100V/V = 580MHz? :| Should I reduce gain?
2) How to bring the rule above together with the following advice:

"unlike the single op amp architectures, the open loop gain (3Q) need only be slightly higher than the filter's output gain (Q), and the low-pass gain is Q, which reduces the requirements on the op amps GBW"
https://www.maximintegrated.com/en/app-notes/index.mvp/id/1762

Click to expand...

3) How about noise in opamp, should I care about it in this application? On the one hand I have high gain but on the other this is not a precision dc or audio application.
4) Does the type of power supply (single or symmetrical) do matter?
5) Can I use a digital potentiometer for adjusting filter parameters? Do you have any advice about that? Small capacitance?

Hi,

GBW: I don´t see why a factor of 100 shoulld be necessary. at least 5 maybe 20.

Noise: I´d use an Opamap feedback to create a 58kHz bandpass. This reduces unwanted OPAMP noise. Then you don´t have to worry about Opamp noise.

Power supply: Make it simple.

Potentiometer, filter adjustment: Why not a fixed bandpass and the rest inside the microcontroller?

Klaus

Coolrob - which of the variuous state-variable topologies are you using?

Filter adjustment: my experience shows that analog tuning brings higher sensitivity in case of low external noise. That's why I'am asking about digital potentiometers and they impact on this circuit.

I use this topology:


The chart below shows the filter response for 1ms 58kHz 20mV signal:


I have tried with DABP filter as per KlausST suggestion - I had to put another one opamp to achieve the same gain and the response is very similar.

I forgot about one important thing: duration of the input signal is about 1ms. So the settling time of the overall analog path is very important. Unfortunately the higher Q the longer settling time. How to play with it?

You could use any of the modern opamps offered by the major semiconductor companies.
They will be low voltage, single supply capable, and some have really impressive AC performance.

Very important, use good quality poly film capacitors. They are critical for good performance, specially with high Q circuits.

With respect to Q, how stable and accurate is the 58khz signal? Otherwise you may end up attenuating the desired signal.

My opinion would be to employ the lowest Q consistent with a S/N ratio which you may later process with DSP techniques.

I have no clear data about accuracy of the frequency of the signal. It results from accuracy of the tag, which rings with its own resonant frequency. I will check it. All in all I'm going to make Q and gain tunable.

Regarding to the poly capacitors - whether they are really necessary if they are located in the feedback loop only (in state variable topology) - so not dc biased like capacitors between signal path and ground? NP0 are not enough?

NP0 should work well.