Speaker and microphone shenanigans

[Mr. Scram]

Looking at various designs, I came across the following circuit a number of times. On one occasion, it is on the microphone lines (top left). On another occasion, it is on the speaker lines (top right). Somewhat surprisingly, in the last case, another circuit is on the microphone lines (top right). In both cases, the values for the parts are exactly the same. One of the designs (bottom) also has another section with speaker and microphone lines, but uses simple inline 68 nH inductors in those cases.

I assume all are for filtering, though I can't find more details on the design on the top. I'm probably using the wrong term. The second design (bottom) with the inductors is probably meant to mitigate high frequency noise. However, why each one is applied in its specific case eludes me. Is there anyone that can shine a light on this?

[Paul Rose]

The upper right looks funny, because the capacitor values are in microfarads ( uF ) instead of picofarads ( pF ) like the rest, but the value and configuration is otherwise identical.   I suspect this is a misprint, and that all values are supposed to be picofarad ( pF ).

It looks like RF suppression to me.  The capacitors bypass RF to ground, but not the desired audio, the inductors block RF but pass the audio.

[Richard Crowley]

Yes, I agree, rather common forms of RF suppression.  To keep RF (from input and output wiring) from entering the active electronic circuits.

Clearly the capacitors in the upper right circuit are mislabeled as uF when they should be pF.  Those uF values make no sense at all.

[Mr. Scram]

Looking for more specifics than the answers above, I decided to try and simulate the circuit to see what its response is. I'm running into a problem though. One simulator throws a fit because capacitors are connected in a loop, and after fighting the clunky LTspice UI, it gives me a single voltage throughout, presumably because one side is connected directly to positive and the other to ground.

What would be a good way of examining the response of this filter, other than physically building it?

[Richard Crowley]

Which of the three different circuits are you asking about?  What exactly is LTspice calling a "loop".

The circuits with capacitors aren't really a "filter" without a series resistance (assumed in the external and/or internal wiring).  What resistance are you using for your simulation? 

And similar situation with the filter using inductors.  A capacitor or an inductor by itself doesn't form any kind of "filter" in the perfect, theoretical world of simulation.

[Mr. Scram]

Which of the three different circuits are you asking about?  What exactly is LTspice calling a "loop".

The circuits with capacitors aren't really a "filter" without a series resistance (assumed in the external and/or internal wiring).  What resistance are you using for your simulation? 

And similar situation with the filter using inductors.  A capacitor or an inductor by itself doesn't form any kind of "filter" in the perfect, theoretical world of simulation.

Sorry, I had to rewrite my post and apparently left some bits out doing so. I'm talking about the circuit that's twice in the image. LTspice isn't calling anything a loop, that's another simulator I tried.

I've used an 8 ohm resistor as a stand-in for the speaker, as a lot of speakers seem to have that resistance, but that leaves the problem of having both sides of that resistor directly attached to the voltage source. Just to be sure I understand how to simulate something, I made a bode plot for another simple RC filter. That seems to work all right. Though when I try to do the same with this circuit, I get something that doesn't quite resemble the bode plot I expected. The graph posted is measured at the positive voltage side.

It's obviously better than what I initially got, yet not the result I expected. I'd like to understand what I'm doing wrong, or maybe I'm having the wrong expectations.

[Phoenix]

As far as your simulation is concerned your source is connected directly to your load.

In the real world the source has some impedance (e.g. the cable itself has resistance, inductance and capacitance). The function of the filter will be dependent on this impedance.

[glarsson]

I've used an 8 ohm resistor as a stand-in for the speaker, as a lot of speakers seem to have that resistance, but that leaves the problem of having both sides of that resistor directly attached to the voltage source.

The speaker is not connected between left and right. You should have one speaker connected between left and ground and one between right and ground. You should also have two signal sources, left and right. Alternatively you can simplify and only simulate one channel.

[Mr. Scram]

As far as your simulation is concerned your source is connected directly to your load.

In the real world the source has some impedance (e.g. the cable itself has resistance, inductance and capacitance). The function of the filter will be dependent on this impedance.

Yes, that seems to be the problem. Introducing artificial parasitics isn't that hard, getting them somewhat relevant and accurate is a little harder, but not impossible. You think I would see a more approprate plot that way? I've been using that "trick" in real circuits before, by adding something like a diode between ground and the rest of my circuit, so I can see what kind of signal comes from the actual circuit.

[Mr. Scram]

The speaker is not connected between left and right. You should have one speaker connected between left and ground and one between right and ground. You should also have two signal sources, left and right. Alternatively you can simplify and only simulate one channel.

It's not a stereo setup. It's one speaker with two leads. I've tried replacing the negative side of the source with ground and doing the same at the resistor, but that didn't seem to make a difference. I'm also not sure that circuit is equivalent to the schematic, as it eleminates one side of the capactor array.

[StillTrying]

Drag the LT probe across the bit to measured, the 8R load.
You'll have to provide a guess at the input's and output's impedances to get any sort of result, but even then all you'll see is that the filters are well above audio frequencies, and in practice the exact RF frequencies will depend on the actual layout, lead lengths.

[ciccio]

In the right-top circuit the component connected to the microphone seems to supply power to it (an electret one?). The MIC+ label seems to indicate a positive power supply.
I don't see where the mic signal is going... I believe that without a known input point is really difficult to simulate anything.
If the schematic designer worked for me, I'll surely try to fire him: Ground labels/symbols everywhere pointing to anywhere. 

[Mr. Scram]

In the right-top circuit the component connected to the microphone seems to supply power to it (an electret one?). The MIC+ label seems to indicate a positive power supply.
I don't see where the mic signal is going... I believe that without a known input point is really difficult to simulate anything.
If the schematic designer worked for me, I'll surely try to fire him: Ground labels/symbols everywhere pointing to anywhere. 

How would you improve on the schematic? I'm trying to learn what good practices are.

[Richard Crowley]

How would you improve on the schematic? I'm trying to learn what good practices are.

At least ground symbols (of whichever style) should point "down".  Having them point in other directions could lead to confusion with other power buses. (Not to mention aesthetics.)

And while components like resistors and capacitors (and even transistors) can be shown horizontal or vertical, it is preferred to have the associated text (reference number, value, etc.) read horizontal.