Sallen-key filter + general amp.

[Pax Writer]

Hello experts

I promised a friend to try to help him build a low-pass op-amp circuit, although I have next to no experience with that kind of circuits.
Being lazy, I "constructed" my circuit from an application note called "Filter Design in 30 Seconds" (I know it looks nothing like Audio Guru's illustration, but I thought if a bunch of engineers would publish that app note, it couldn't be all wrong).
Based on the schematics and calculations (and what I happened to have in my component bin) in that app note, I made the circuit, which I attached to this thread.
It uses a TL072, and the first opamp is supposed to be a low-pass filter with a transition frequency of 300Hz. The second was just used as a sort of general amplification circuit.
The whole circuit is powered with 9V from a lab supply.
The first circuit seems to work to some extent while the second is just dead.
I'm aware that my circuit probably has many problems, but what are the main mistakes I've made? Is there any way to make this circuit work, or do I need to start over?
I have tested the circuit both by humming into the mics, but also using an old B&O tone generator, moving from 50Hz to around 10kHz.

Note: Its getting late around here, so I'll hop to bed and won't be able to answer for the next seven hours or so, just so nobody thinks I'm ignoring their answer

 

[crutschow]

Don't see any obvious problems with the circuit, although R4 and R5 are rather low resistance. C2 should be a nonpolarized cap.

A quick simulation showed a -3dB rolloff at 410Hz.

Is the second amp dead for all positions of pot R9? If the pot wiper is set at the top extreme position, the feedback resistance will be essentially zero and you'll get no output.

 

[Roff]

My sim puts the -3dB point at 291Hz. This assumes a zero impedance source.
Pax Writer, are you aware that you need to cap-couple the output if you have a grounded load? The AC output will be riding on Vcc/2.
What is the value of Vcc?

 

[audioguru]

The circuit has losses, a very small input signal and no voltage gain. The output is 1mV to 5mV which is almost nothing.

Maybe the load impedance is lower than the 1k ohms that an opamp can drive. A speaker?

 

[Pax Writer]

Hi
... And thanks for your answers. Here are some replies:

crutschow: C2 was polarized because I only had that size as a tantal. I'll see if I can find something better instead. Is it impossible to use a polarized cap here? If not, why? (I ask, hoping to learn something, not because I feel a need to have a polarized cap here).
R4 and R5 were calculated to 798 ohms by my spreadsheet (made using the equations in the app note).
EDIT: Yes, the second op amp is dead in all positions of the pot, but since you don't seem to see a problem with the schematic, I suspect it is due to a soldering error. I'll check up on that.

Roff: By cap-couple, I guess you mean that I should have a capacitor in series with the output? How do I dimension such a capacitor?
I forgot to mention that Vcc right now is 9V single supply.

Audioguru: The load of the output right now is just my scope - I didn't want to complicate the circuit too much by adding another amplifer until I had this part working. I do it this way because I have so little experience with amplifiers. In the end it is my intention that the signal will be processed digitally, but that part of the circuit has yet to be constructed. I was, however, hoping that the second stage would also amplify the voltage, but I guess I need another value pot, then?

And in general: Is it okay that the two op-amps are DC-coupled, or should I have a cap between them?

 

[audioguru]

You might as well do it right:
1) Amplify the signal from a single microphone, then filter it.
2) Decouple the battery.
3) Decouple the power to the mic and the bias voltage to the opamp.
4) Increase the value of most resistors so that reduced values can be used for the capacitors so that very low distortion plastic film capacitors can be used (except the 10uF one).

Three mics pickup three times as much background noise as a single mic.

I couldn't remove all the yellow dotted lines all over the schematic.

 

[crutschow]

crutschow: C2 was polarized because I only had that size as a tantal. I'll see if I can find something better instead. Is it impossible to use a polarized cap here? If not, why? (I ask, hoping to learn something, not because I feel a need to have a polarized cap here).

If a cap is carrying an AC (bipolar) signal without a DC bias then it should be non-polarized to avoid reverse biasing the cap (although at a few tenths of a volt AC a polarized will probably work.) Also polarized caps generally have a very loose tolerance (sometimes 50%), and you normally want tighter tolerances for a filter so that the filter characteristics stay close to your design point.

DC coupling the two amps together is fine

What is the DC output voltage of the second amp? It should be ½ the supply voltage.

P.S. My sim now shows a rolloff of 291Hz, the same as Roff. Had a cap connected to the wrong node.

 

[Pax Writer]

Thank you both from your answers
I'll try your suggestion, Audioguru. Your idea of increasing resistor values to decrease capacitor values is really good, too bad I didn't get it myself. Same goes for switching the order of the op amps to amplify first and then filter.

If I wanted the low-pass filter to also have amplification, could I add a resistor between 1OUT and 1IN-?

The DC out of the second amp was in fact ½vcc, but I think I simply made a soldering error, and I'll weed that out when I rebuild the circuit.

I got the idea of using several mics in parallel from a user here on the forum once. The point was supposedly to increase the overall sensitivity of the microphone circuit, but as you point out, Audioguru, I'll also get more background noise.

Thanks for the clarification about the capacitors, crutschow. I'll go to work with the improvements, you suggested.

EDIT: The dotted yellow lines are there, because the schematic is a screendump made from a pdf generated by Pads Logic. In the pdf, you can click the components to get additional information about them.

 

[audioguru]

If I wanted the low-pass filter to also have amplification, could I add a resistor between 1OUT and 1IN-?

Then the opamp will be an oscillator instead of a filter.

I got the idea of using several mics in parallel from a user here on the forum once. The point was supposedly to increase the overall sensitivity of the microphone circuit, but as you point out, Audioguru, I'll also get more background noise.

The active mic is loaded down by the other mics which reduces its sensitivity.
the 1k resistor powering all the mics also loads down the active mic and reduces its sensitivity. Plus the added background noise.

EDIT: The dotted yellow lines are there, because the schematic is a screendump made from a pdf generated by Pads Logic. In the pdf, you can click the components to get additional information about them.

Your faint yellow dotted lines are nothing compared to the horrible schematics posted by others:
1) A negative with a black background.
2) Covered in chicken-pox dots.
3) A very fuzzy JPG file type instead of a very clear GIF or PNG file type.