Series / Parallel + Order Switching?

[Lewham5115]

I have a project where I will have two guitar pedal effects circuits in the same project enclosure and I would like to be able to turn each one on independently of eachother, but also to change their order in the signal chain, in addition to being able to have them both in series or parallel.

So something like

In - - Effect A - - Effect B - - Out
In - - Effect B - - Effect A - - Out
In - - Effect A - - Out
      - Effect B -


With four switches, one to turn each effect on and off, one to change between series and parallel order, and one to reverse order of the circuits when they are in series.


But I'm struggling to wrap my head around how to achieve this. Having the series/parallel switch is what's really messing with me.

Can this be achieved with mechanical switches? 4000 series logic? Do I need some sort of microprocessor to control relays?

Any help would be greatly appreciated.

[Focalized]

This is a start. Others would be better at expanding to how you want it, and with out the jacks I think.

http://beavisaudio.com/projects/looperswitcher/img/Effect-Order-Switcher-Wiring-Diagram.png

[Lewham5115]

This is a start. Others would be better at expanding to how you want it I think.

http://beavisaudio.com/projects/looperswitcher/img/Effect-Order-Switcher-Wiring-Diagram.png

Ya, the order switching isn't a huge deal, but I'm just struggling to get my head around the series/Parallel in addition to that.

[ElectricDruid]

Since you've got three options, I'm not convinced footswitches would be the way to go. They only ever have two options. You'd either have to use two switches with some redundancy (order switch does nothing when series/parallel is in the parallel position) or you use a different switch type. A 3x4 rotary seems like it might be a good option. Whether a fancy 3-position toggle would be able to do it, I'm not sure.

Otherwise, yes, relays and logic or a uP.

There are different ways to think about it, but one way is to look where the effects ins and outs come from/go to.
Effect A gets its input either from the Input jack, or from the output of B. Its output goes to either the Output jack or the input of B.
Effect B does something similar. So there are three options, and four places where that's switched, so it's a 4P3T switch.

HTH

[R.G.]

The 1999 "Juggler" article has an easier to follow diagram of how to use a 4PDT switch to swap pedal order. Near as I can tell without tracing every wire, the beavisaudio pic uses the same wiring.

The switching to do the series/parallel/swap either way can be done in two general ways. One is to draw up all the desired combinations and then make a long series of drawings to dope out what switches are needed. The other way is to put in an any-to-any crosspoint switch IC and pick the paths that you need. The only disadvantage of the crosspoint variant is that it needs a uC to interpret the stomp switch inputs, then translate that into which path is needed in the crosspoint switch.

I did a design/pcb for an 8x8 non-blocking crosspoint switch made up from CMOS logic and CD4051 analog switches; search for geofex crosspoint to turn up the ASMOP 1d Matrix Controller. The design was the guts of an any-order effects switching controller with memory for patch orders that was run entirely from stomp switches. The CMOS IC design would appeal more to people not used to fancier ICs, but a Mitel MT8800 series crosspoint IC would be easier to use.

A crosspoint (also called crossbar) switch is a very old idea, originating (I think) in early days of the wired telephone. Imagine a group of wires stretched left to right, say, four of them. Call these H1 through H4. Now imagine four wires stretched vertically, but spaced above the horizontal wires by about a half-inch. Call these V1 through V4. Finish the switch assembly by installing an SPST switch between a horizontal wire and a vertical wire each place where they cross over one another.

It happens that if you decide that the horizontal wires are all inputs to the switch and the vertical wires are outputs to the switch, you can switch any input wire to any output wire by just flipping the SPST switch where the wires cross. If you are careful about how you control your 4x4 switches, you can make this non-blocking; picking any switch path does not block any other path.

Now make the switches be electronically controlled; either relays or analog switches. Hook an input jack to H1, Hook an output jack to V1. Hook one effect input to V2, the other to V3 (as the effect inputs have to be connected to the switch array outputs). Hook one effect output to H2, the other to H3. Now you have any-to-any order switching. Getting parallel variants requires repurposing one vertical to be the mix of two outputs and switching inputs into it. This requires non-blocking switches and a mixer outboard of the switch matrix. I'd have to think a bit more about whether another horizontal/vertical is needed for all possible effect combinations including parallel. I didn't do that in the ASMOP design.

Mitel, now owned by Microchip, used to make the MT88nn series of analog crosspoint audio switches in at least 8x4 and 8x8. I messed with these for a while for a possible crosspoint switching design, but went with the ASMOP 1d instead because of most DIYers not wanting to mess with higher integration chips. Mouser still has stock on the MT8816, which is an 8 by 16 crosspoint switch in a dual in line package, or $10.00. You could probably find an MT8808 8x8 chip somewhere.

Edit: I found an article on this kind of thing. It's here:https://cplcrapper.wordpress.com/2017/01/17/stompbox-switch/

[Lewham5115]

The 1999 "Juggler" article has an easier to follow diagram of how to use a 4PDT switch to swap pedal order. Near as I can tell without tracing every wire, the beavisaudio pic uses the same wiring.

The switching to do the series/parallel/swap either way can be done in two general ways. One is to draw up all the desired combinations and then make a long series of drawings to dope out what switches are needed. The other way is to put in an any-to-any crosspoint switch IC and pick the paths that you need. The only disadvantage of the crosspoint variant is that it needs a uC to interpret the stomp switch inputs, then translate that into which path is needed in the crosspoint switch.

I did a design/pcb for an 8x8 non-blocking crosspoint switch made up from CMOS logic and CD4051 analog switches; search for geofex crosspoint to turn up the ASMOP 1d Matrix Controller. The design was the guts of an any-order effects switching controller with memory for patch orders that was run entirely from stomp switches. The CMOS IC design would appeal more to people not used to fancier ICs, but a Mitel MT8800 series crosspoint IC would be easier to use.

A crosspoint (also called crossbar) switch is a very old idea, originating (I think) in early days of the wired telephone. Imagine a group of wires stretched left to right, say, four of them. Call these H1 through H4. Now imagine four wires stretched vertically, but spaced above the horizontal wires by about a half-inch. Call these V1 through V4. Finish the switch assembly by installing an SPST switch between a horizontal wire and a vertical wire each place where they cross over one another.

It happens that if you decide that the horizontal wires are all inputs to the switch and the vertical wires are outputs to the switch, you can switch any input wire to any output wire by just flipping the SPST switch where the wires cross. If you are careful about how you control your 4x4 switches, you can make this non-blocking; picking any switch path does not block any other path.

Now make the switches be electronically controlled; either relays or analog switches. Hook an input jack to H1, Hook an output jack to V1. Hook one effect input to V2, the other to V3 (as the effect inputs have to be connected to the switch array outputs). Hook one effect output to H2, the other to H3. Now you have any-to-any order switching. Getting parallel variants requires repurposing one vertical to be the mix of two outputs and switching inputs into it. This requires non-blocking switches and a mixer outboard of the switch matrix. I'd have to think a bit more about whether another horizontal/vertical is needed for all possible effect combinations including parallel. I didn't do that in the ASMOP design.

Mitel, now owned by Microchip, used to make the MT88nn series of analog crosspoint audio switches in at least 8x4 and 8x8. I messed with these for a while for a possible crosspoint switching design, but went with the ASMOP 1d instead because of most DIYers not wanting to mess with higher integration chips. Mouser still has stock on the MT8816, which is an 8 by 16 crosspoint switch in a dual in line package, or $10.00. You could probably find an MT8808 8x8 chip somewhere.

Edit: I found an article on this kind of thing. It's here:https://cplcrapper.wordpress.com/2017/01/17/stompbox-switch/

Really appreciate the input and point in the right general direction.

I don't know why more-than-standard physical switching trips me up so much.

I'm thinking using a microcontroller to control soke sort of 40xx is looking like the way to go.

Now to go do some more reading on that.

[R.G.]

More-than-standard switching trips me up too. I wind up in most cases drawing out how the wires have to run in all switch cases and then looking at the differences. It gets complicated very quickly.

If you're looking at using a CMOS switch chip setup, do some study on the ASMOP 1d case to start. It's a partial crosspoint, with a few limitations, but it covers a lot of cases. The CD4051 is a one-to-8 analog switch. That is, one common pin can be set to any one of eight different "throw" pins.

If you label the common pin as an "input" then it can be switched to any of eight "outputs" by setting the A-B-C address pins to a binary code of 000, 001, 010, 011, and so on. Paralleling more CD4051s with their eight "outputs" to the first CD4051 gives eight inputs, any of which can be switched to any of the eight output lines.

It's not a true crosspoint, as each "input" line only gets to contact one output line at a time, not zero to all of them. But it's probably enough for the use you have in mind.

Here are a couple of things that took me a while to tumble to. Each CD4051 also has an inhibit pin. When this pin is active, the common doesn't connect to any of the throws; in effect, it's a mute on all of the chip's connection. In counting up how many inputs and outputs a crossbar switch will need for guitar effect routing, the switch matrix itself needs one input and one output; as an example, if you are routing six effects in any order, you need seven inputs and seven outputs. Each effect has an input and output, and if you want to bypass them all, the matrix needs an input and an output, too.
And you can parallel inputs or outputs by switching two inputs to the same output line, or two outputs to the same input line.

Hmmm. Now that I think of it, you will need extra lines in the matrix for this paralleling. I just exceeded what I can do in my head without drawing it out. You may really need all eight of the possibilities on the 4051 to have the paralleling work out right.

You're in for a fun project!

[Matthew Sanford]

Here are a couple of things that took me a while to tumble to. Each CD4051 also has an inhibit pin. When this pin is active, the common doesn't connect to any of the throws; in effect, it's a mute on all of the chip's connection.

This made me smack myself. Though I want the OP's idea (who doesn't want delay in front for a bit of ambient before the end for some psych?), I've been trying to rid myself of switches and pots, one thing being turning off the signal or... inhibit it! Duh! Thanks RG

[R.G.]

It made me whack my forehead when I figure it out too.  The ability to turn off the whole switch section is not normal for people used to thinking of mechanical switches.

Guys, remember - for non-popping switching with CD405x chips, bias every input and output pin to half the power supply for the chip. Put a 100K (ish) resistor from each in/out pins to the bias supply, and a coupling cap to keep this DC off the external world. Biasing these things in the middle makes the internal leakages be equal and opposite, so they tend to internally cancel any feedthrough pretty well.

[Mark Hammer]

An article by Dennis Bohn (formerly of National Semiconductor, IIRC) in the Nov/Dec 1981 issue of Polyphony (available for download on the "Muzines" site) presents a circuit for doing series/parallel and sum/difference.

Here's the circuit described/detailed in the article.  You are recommended to consult the article, rather than just work from the schematic.

[PRR]

Dennis Bohn (formerly of National Semiconductor,

The Nat Semi gig was just a steppingstone. He was a force at Rane and in the AES. Dennis left us in 2022 so there are good summaries. (Boogle "dennis bohn audio" cuz there's other guy(s) of like name.)
https://obituaries.seattletimes.com/obituary/dennis-bohn-1086968301
https://www.aes.org/par/
https://www.prosoundweb.com/author/dennis-bohn/
https://www.amazon.com/Audio-Handbook-Dennis-Ed-Bohn/dp/B000GYKY9W

[Mark Hammer]

Thanks, Paul.  Wasn't aware of anything past the NatSem gig.  A very productive and esteemed life.  We should all live that long and get that much done.