MMR-40 Active Audio Filters – U8a
MMR-40 Active Audio Filters – U8a
The next step was the Active Audio Filters.
- U8a - Audio Amp
- U8b - Filter Amp
- C42 - Coupling Capacitor
- C47 and R16 - Low Pass Filter
- R14, R40, R41, R42
- C72, C73
Block Diagram from http://www.web.mac.com/petermarxy/MMR-40BlockDiagram (no longer active)
There is some rolloff at the high end from R16 and C47 and U8a. This forms a low pass filter. It is bounded by the Equation:
Where Fc = Cutoff Frequency, Rf = Feedback Resistance, Cf = Feedback Capacitance
Cf = 1/2(PI)(fc)(Rf) and fc = 1/(2(PI)(Cf)(Rf))
or
Cutoff Frequency = 1/(2(PI)(Cf)(Rf))
Cutoff Frequency = 1/2(PI)(0.00033)
Cutoff Frequency = 1/0.0020734
Cutoff Frequency = 482 hz
That seems kind of a strange place to roll off the audio to me, so I did some investigation to see if I was getting this right
Looking on the web, I find a Texas Instruments Application Report (SLA058 - Nov 2000) that confirms the formula I was using.
Still not totally convinced, I built the circuit in ltSpice. I had to find and install a model for the LM358, but it turns out that using the generic opamp2 in ltSpice gave pretty much the same results. I kind of expected that.
The resulting plot from running the simulation looks like this.
As you can see, the response starts dropping off at about 400 - 500 hz.
Finally, I used my noise injector to input about 0.025vpp wide-band noise into the input of u8a.
And the resulting output from u8a shows the expected rolloff. Notice that the rolloff is down almost 10 db at 1600 hz. So everything agrees that the low pass filter rolloff starts at about 480 hz.
Migrated Comments:
June 13th, 2015 at 11:31 pm
It means, that the R_C combination proposed to be soldered have the wrong value? Other schematics shows that there shall be 470 pF and 470 k. Which is the correct value, than?
June 14th, 2015 at 7:35 am
I think it’s probably a subjective choice. Notice that the rolloff is very gradual. If it had a steeper rolloff, then I certainly would choose a higher cutoff frequency. But with it being so gradual, if we moved the cutoff up by 1000 hz, I think we would be inviting more high frequency hiss to our bandpass.
In the end, the rig sounds good as designed, and I would not change the circuit.