I think these are alternatives to solid-state relays for audio switching but do not provide galvanic isolation, which I don’t need for an audio preamp application.
I am interested in what to look for in datasheets (or what to avoid!), part recommendations, etc. THD is on my mind, but this is not an instrumentation application, so a typical amount is tolerable.
TI’s John Caldwell has written the best doc on this I’ve seen (link below). It explains how to think about the problem when distortion cannot be compromised. And that takes you down the path of on-resistance flatness. The TMUX4827 you mention has 1 mohm of flatness which is exceptional.
If your aim is to provide selectable gain via analog switches, then the most straightforward solution here is run parallel gain paths using precision resistors AND THEN you use the switch to pick off the midpoint and route that to the opamp inverting terminal. So, let’s say you wanted an adjustable gain of 5 or 10 in a non-inverting amp. You’d have two feedback paths around the op-amp: One using 1k and 4k resistors (for a gain of 5)and one using 1k and 9k resistors (for a gain of 9). And then use a switch to the opamp negative terminal to pick the midpoint of the 1k + 4k combo OR the 1k and 9k combo. That way, the current through the switch is just the amp negative terminal bias current, which is so small the voltage change that comes from the switch Rdson+Ronflatness is inconsequential.
If your aim is to provide routing (eg either input X or input Y) then you can do something similar reconfiguring the opamp as an inverting amp and pick one of two branches) OR pick your signal source via a switch and run straight into the non-inverting input of the opamp.
The key from a distortion perspective is to get the current through the switch to very nearly zero. The higher the current through the switch, the more the Ronflatness matters.
This is so helpful — I need an internet service that can send you a nice bottle of wine!
I need to digest this a bit and read the Caldwell note. I’m currently on holiday (Tuscany) pondering new hobby things to do when back (hence the comment about wine)!
Yes, variable gain is what I was thinking. Something like a guitar amp pedal so quite forgiving in the THD department. But it is very good to know how the professionals approach this.