I am looking to make the inputs of a simple inverting op amp mixer that I have, AC coupled so I can monitor a very simple mix of various keyboards and what not. I thought of just adding a series capacitor with a resistor to ground on each input (say 47nF/220k), making a highpass filter, but it doesn’t seem to work very well. I get plenty of hum, the potentiometers do not fully attenuate the input signal, there is some kind of a cross-talk low-pass filter among the inputs, and, depending on the combination of the values in the RC filter, I get some oscillation when multiple inputs are cranked up!
Obviously, there are several variables at play here as the potentiometer value needs to be factored in (among other things), so I am hoping that someone more experienced or knowledgeable would have a better suggestion for the capacitor and resistor values at the inputs.
Where did you put the cap and resistor? If before the attenuator then the attenuator is in parallel with the 220k and the 100k is in parallel with the bottom of the attenuator, so that alters the cutoff of the filter. If after the attenuator then the 100k is in parallel with the 220k, again altering the cutoff although not in an attenuator dependent way. If after the 100k resistor then the 220k essentially is useless since it connects a virtual ground to ground, and the attenuator + 100k + 47 nF becomes a low pass filter.
If you were to use just a 100 nF cap between the attenuator and the 100k resistor then they’d form a high pass filter with about the same cutoff as 220k and 47 nF, with no resistances in parallel to mess it up.
I don’t understand all the symptoms but knowing where you put the filter would help.
OK, I don’t see how that could cause all the symptoms you mentioned. The high pass cutoff would be anywhere from 1/(2π *47 nF * (220k || 100k)) to 1/(2π * 47 nF * (220 k || (100k || 100k))) = 49 to 83 Hz depending on the attenuator setting, higher than you presumably intended, but I can’t see how it would cause the hum, incorrect attenuation, cross talk and low pass, and oscillation so I would guess there’s another problem elsewhere.
I’d try a 100 nF cap between attenuator and 100k resistor. That should work fine, with a ~15 Hz cutoff. The attenuator affects the cutoff slightly but not enough to matter. (Red/blue/green lines are different attenuator settings; solid lines are amplitude, dotted lines are phase.)
The mixer was working fine as such, but since I am building a couple of different ones I thought to repurpose it by a slight modification. I was unpleasantly surprised by how bad it got just by adding a couple of capacitors, but fact is that I tried it with different signals (not the usual oscillators) so that must have played a role as well. I will try the 100nF capacitors after the attenuator and report back. Thanks!
I tried the 100nF capacitors between the attenuators and the 100k resistors but couldn’t get it to work with line level signals. The persistent problems are as follows:
There is considerable hum which only goes away when the attenuators are all the way up. The hum was not present when I used the mixer before adding the capacitors.
There is a signal bleed in all channels. The signal goes through even when turning down the attenuator all the way down.
Turning up two channels with line level signals beyond 1 o’clock results in oscillation.
The problems are the same across all inputs and persist regardless of what connects to the mixer or what the mixer is connected to.
This has to be the stupidest mistake ever on my part (at least for now, I will probably top it soon). The hub and bleed were a giveaway that a ground connection was missing. Obviously, I forgot to connect a ground wire that I disconnected to make working space for connecting the capacitors. The mixer works perfectly now! Thanks to @analogoutput for suggesting the most efficient way to connect the capacitors as well.
