Sallen Key distortion? - simulated vs measured

[HiFiNutNut]

Try running the measurement again at a higher output volume.

MJL,

It sounds like a great idea and I will do that. I now remembered that for the woofer measurement the sound card output from the computer was on the low side comparing to the input channel, while with the tweeter it was nearly matched. Although not level matching, I shall increase the output and measured the woofer independantly from the tweeter.

Thanks for the good tip. I know there are experts here.

Regards,
Bill

 

[MJL21193]

Thanks for the good tip. I know there are experts here.

No problem
I'm no expert, I just have some experience with both the measuring and the filters.

 

[BobEllis]

You might consider changing the order of the sections in the band pass section. Put the high pass first and the low pass filters out the noise of the high pass section. It won't have much if anything to do with the issues you are seeing, but may help keep your mid quieter.

You can use Speaker Workshop to measure the filter output directly. Just attenuate the signal to something your sound card can handle with a resistive divider and feed that into the line input. You can include the amp in the measurement if you want, just attenuate the output more.

 

[gootee]

For future reference, there are some good chapters on analog filters in this downloadable book, "Op Amp Applications Handbook", Walt Jung, Editor Emeritus:

ADI - Analog Dialogue | Op Amp Applications Handbook

And please, the name of the filter topology is Sallen-Key, named after two people whose names are Sallen and Key. In particular, there is no letter "H" in the name "Sallen".

 

[HiFiNutNut]

You might consider changing the order of the sections in the band pass section. Put the high pass first and the low pass filters out the noise of the high pass section. It won't have much if anything to do with the issues you are seeing, but may help keep your mid quieter.

This is something I thought about before. But which is ultimately quieter? I am still thinking it through.

I have considered many factors. If the circuit itself generates high frequency noise then the way you describe would have lower noise, this may include EMI induced / inducted, ground noise, resistor noise, thermal noise, etc. However, the noise level would be relatively low.

On the other hand, opamps have high PSSR at lower frequencies and low PSSR at high frequencies. In other words, unless the power supply impedance is 0, opamps would produce more distortion / noise at the output at higher frequencies.

The idea I had was to place the low pass filter at the beginning of the bandpass so that only the first opamp needs to deal with high frequencies. And this first opamp is placed very close to the power supply regulator output which has an impedance below 0.001R (in theory) up to hundreds of kilo Hz. This should produce the lowest distortion figure for all opamps.

I guess this gives an overall less noise comparing to placing the low pass filter the last. But I am not sure unless I can measure it, and with the cheap and old equipments I have I can't.

 

[HiFiNutNut]

For future reference, there are some good chapters on analog filters in this downloadable book, "Op Amp Applications Handbook", Walt Jung, Editor Emeritus:

ADI - Analog Dialogue | Op Amp Applications Handbook

And please, the name of the filter topology is Sallen-Key, named after two people whose names are Sallen and Key. In particular, there is no letter "H" in the name "Sallen".

Thanks, Gootee.