The back of a pot used as a signal ground, in a grounded chassis design; how reliable?

Question

This question relates to those large pots with solder terminals that are still used in electric guitars; they used to be common in metal-chassis radios, amps, TV, as volume and tone controls.

It was common practice that the metal chassis was a signal ground, and often signal-ground connections would be made to the back of the pot - in the simplest case, one of the pot terminals was bent back and soldered there; but also, caps or resistors were often connected between a terminal on the pot and the back of the pot.

However, in many pots, the back cover is connected to to the front of the pot assembly by crimped tabs. There is often a lockwasher between the front of the pot and the chassis, which provides a good connection there - but it seems to me that the connect made by the crimped tabs could be unreliable in the long term.

So I'm asking if people have seen issues related to this - are there some types of pots which are better than others (in particular, it seems possible that OEMs might have selected pots for this issue back in the day, whereas in newer pots the reliability of the connection may not be a big concern for the component manufacturers). Would it be reasonable when servicing any such equipment, to add soldered jumpers to ground the backs of any pots which are used as a signal ground? Or would that depend on the types of pots?

The problem is it can be very tricky to simply check, since these issues can often be intermittent, and the act of placing a probe on the back of the pot can cause a flaky ground to suddenly be OK. I ran into this issue recently with an electric guitar that started acting is if there was an open ground within; when I opened it up, the issue had (of course) gone away, but I found that it had been wired so the ground path actually passed through both pot cases in sequence (via the metal plate in which they both were mounted); I added a connection between the back of them to remedy this, and never had a problem after that (which doesn't prove that that was the problem, of course, but...)

UPDATE

@ThreePhaseEel has suggested a fix which also covers a different issue. I appreciate the remarks; I'm working on this Traynor bass amp, and had not considered isolating the ground, as suggested. It's a lot of extra trouble since there are a number of connection points - including the output tube cathodes, and also the negative ends (and outer shields) of two 'chassis mount' electrolytic caps, and the signal input jacks.
As a nearly continuous metal plane, the chassis should actually be a pretty solid ground, whatever issues are there with power supply current coupling into the signal path, are likely quite minor compared to the ripple on the unregulated supply, (and likewise this is probably not the highest-gain point for RF pickup). But it may be practical to leave the power stage connections (and associated power caps) on the chassis and just make sure everything else goes to a common point, thanks for that suggestion.

The method of adding reliable grounding wires to the backs of the pots (as rework/maintenance, not in original design) solves the separate issue of intermittently open signal-ground connections, and is not much trouble to do; I was wondering in particular

• If anyone had experience with seeing problems solved by this; and
• Is anyone aware if 'old' pots were made in a way which gave them better electrical integrity, to support this construction method.

Answer 1

What they're doing is using the chassis ground as a signal reference, instead of using a signal reference plane or bus that is connected to the chassis ground in an explicit manner. This seems to work OK at first, but creates not only the problem you're describing where you have signal grounding wires making unreliable connections to the chassis, but a much nastier gremlin: common-impedance coupling.

Since the signal return currents are now flowing through the same piece of metal as all the 60Hz and RF circulating currents the chassis has picked up, these changing currents can cause voltage drops between different "signal ground" points, which manifests in the circuit as funny noises.

The fix for this is to lift all the signal ground points from the chassis and connect them to a signal ground reference point, bus, or plane, which is then connected to the chassis at a single point near where the signal inputs and outputs are found. This keeps circulating currents in the chassis from inducing voltages into the signal ground system, and also means that the integrity of the signal reference system isn't dependent on weird mechanical connections, such as those you saw in potentiometer cases.

Answer 2

I've worked on a lot of old equipment built using those techniques - I used to build some of my old Intercom belt packs this way as well.

For the most part, this technique has been extremely reliable. In the very few cases where there was a bad connection, it was a simple fix to just add an extra ground jumper as needed.

However, I don't build equipment that way anymore. Not because it's not reliable, but instead because this technique makes it much more difficult to replace the pots.

Most of the equipment built using these techniques is several decades old and the pots now need to be replaced. Not only is it a pain in the *ss to remove all those (really short) ground wires, but the new pots that I'm installing have a plastic body.