Potentiometer position switching

[selfctrl]

Hey People

Apologies for this slightly esoteric request, but recently I had the idea of building a RAT circuit inside an old WAH enclosure.
I wanted to fit it with two 100K pots and a 100K pot on the treadle.
Since the pots are all of the same value I want some way of being able to switch the position of the potentiometers.

e.g.

Position One
Pot :1 Volume
Pot 2: Gain
Treadle Pot 3: Filter

Position Two
Pot :1 Filter
Pot 2: Volume
Treadle Pot 3: Gain

Position Three
Pot :1 Gain
Pot 2: Filter
Treadle Pot 3: Volume

Anyone got any ideas of the best way to accomplish this or am I on a fools errand?
Any nudges in the right direction would be greatly appreciated.

Stef

[frodog]

I don't have the faculties right now to figure out the xact signal flow or draw up a schematic, but yes, with the wires going from board to potentiometers first passing through a dipswitch arrangement and/or a 3T3P switch before connecting to each of the pots/treadle, that should be possible.

[crochambeau]

You have the each pot hitting each function round robin style. So: seven circuit nodes, each interfacing with three of the nine pot lugs.

While I am sure a mechanical switch solution can be derived, I suspect user enjoyment would prefer one physical switch - meaning this looks like a solid state/CMOS paired with logic solution.

A problem exists. Pop free switching demands similar DC states (or effective pull down, but that might be detrimental inside the feedback loop of your gain stage). You might have to live with some switching noise.

While I know it can be done, I am not in the position to unravel this puzzle. I'd be looking at CD4053 (three SPDT switches per package), CD4066 (four pass/no pass paths, for when you need to connect pins on the pot or shunt to ground at output volume), and whatever logic needed to parse the path. For a switch I would select a three position ON-OFF-ON switch, and tie center to ground and each side up to V+ through something like a 10K resistor (these need to be separate lines, so 10K on each side - the resulting data line is the "lower"/switch side of the 10K resistor). This gives you a two bit word that looks like 01, 11, 10 depending on switch position.

From there you just need to "wire" three paths based on those two bit words. It may be easier to achieve your aims using CD4052, which requires two bits of information to make it through all of the states. Luckily, you don't need access to the fourth state (00) because you're only working with three paths.

I would not suggest this as a beginners project, but it could be a lot of fun.

Ultimately, each "switch" introduces a little resistance equivalence, so compensations may be required elsewhere in circuit if you start deadheading into lossy performance (this is the part where we get to relish in the fact that while it may be logically possible it could also be unusable - but don't let that stop you!)

[selfctrl]

Thanks so much for your input guys!
I am by no means a beginner in terms of building and modifying pedals.
I don't come from an electrical engineering background and am very much self taught.
This is something a little different than I am used to, so this advice is invaluable.

Thanks so much.

[crochambeau]

Since my earlier post did not scare you off, I will go a little deeper..

Playing with CMOS switching was a mind opening experience for me. I'm happy to throw in possible solutions/angle to specific problems, but you'll need a roadmap first.

My method (and please bear in mind that there are probably more efficient ways to go about this that I do not know about) would be to label the circuit nodes and pot legs with separate letters or numbers, and then write out all of the intended connections. So say you have a pot and designate that A, B, and C, and then each circuit node has a number running from 1 to 7. You'd wind up with a (partial) map looking like this:

A1, B1, C2
A3, B3, C4
A5, B6, C7

Where each column is a pot lug, and each row is a circuit node.

This way you can map out everything that needs to happen from any vantage point. Example: lug A needs to connect to 1, 3, and 5. I'd probably steer toward a 4052 for this, but it can be done with simpler switches (mental horsepower will usually need to increase in the design phase to use simpler parts).

Repeat until all needed switch modes are addressed, and then we look at the logic needs of the switches to achieve that.

So say we determine that a 4052 needs a 01 in order to meet the needs of a given state, and in this state the switch position is in 11. You can strap a NAND gate across your two bit data line, that will output a low state when both inputs are high, and connect that to the instruction pin (so long as that pin needs a high signal during the other two configurations).

Dig into the truth tables for AND, OR, NAND, and NOR for other ideas. Sometimes you need to combine a couple layers of "instruction" (even though this is hardware, it's essentially low level programming) to get your needs met.

[eatyourguitar]

I would use the CD4066 for switching. I started drawing it but it is 27 switches driven from 3 control lines. this is a 9P3T switch. I can help you with the momentary switch debounce, the counter and the LED indicators. the switch debounce resistors and caps can be increased. there is a wide range of values that will work. you could do 100K pull up 1M pull down and 470nF. compare to 1K, 10K, 100nF.