1
\$\begingroup\$

I want to create a device for my guitar pedals which can rearrange my pedals in any given order. I have the following: At the beginning of the device I have an input for my guitar. For each pedal I have an ouput from which the signal goes to the pedal, and an input, which receives the output of the pedal. Usually these pedals are linked in a predefined order, which can't be changed without rewiring. I want to be able to change to do that without rewiring. Also, the device has to be able to work with ~20 pedals max.

The first solution I came up with is to use (a lot of) rotary switches. However, this requires to change all the switches when I want to change the order in which the pedals are linked, so this isn't really useful on stage. As far as I could find there is also no way to control a rotary switch from an electric circuit.

Another solution I came up with is to use (ss) relays. I can control these from a microprocessor, so I can predefine the order of the pedals. However, I can't come up with a solution which doesn't use many dozens of relays.

Can someone help me with a solution?

\$\endgroup\$
4
  • \$\begingroup\$ What kind of signal comes out of each pedal? Is a pedal basically a pot putting out a voltage as a function of position? \$\endgroup\$ Jul 3, 2011 at 16:47
  • \$\begingroup\$ @Olin - I was also thinking of a potmeter, but if it outputs an analog voltage level it wouldn't have an input, would it? (except power supply, that is) \$\endgroup\$
    – stevenvh
    Jul 3, 2011 at 17:09
  • \$\begingroup\$ I didn't understand the OP to mean that the pedals had inputs. I thought he had 20 pedals and 20 pedal inputs on some equipment, but wanted to easily rearrange which pedals drove which of the input ports of the equipment. I don't play guitar, so clarification from the OP would help. \$\endgroup\$ Jul 3, 2011 at 17:19
  • 2
    \$\begingroup\$ @Olin, stevenh: guitar "pedals" are generally signal processors (Effect units) between the guitar and the amp. For example, OP can have a guitar output connected into a simple volume pedal (which is basically a potmeter), and then the output of the volume pedal connected into a distortion pedal, from which the output signal goes into the amp. Since the distortion pedal works by clipping the input signal, it will matter if the volume pedal is placed before or after the distortion unit. OP is looking for a way to rearrange a chain of processors. \$\endgroup\$
    – vgru
    Jul 3, 2011 at 17:43

2 Answers 2

2
\$\begingroup\$

Well, if you have that many pedals, you'll need a lot of switches, whatever type. I presume that both the pedal's input and output are low-power audio signals, correct me if I'm wrong.
What are your options?

  1. A lot of relays. Which you don't want, probably because they are bulky and expensive. But power relays are not made to switch signals in the first place.
  2. The reed relays I talked about in my answer to your previous question. These are cheaper, less bulky and fit for switching signals.
  3. Analog integrated switches. These were also mentioned in answers to this recent question. Solid-state.
  4. Analog multiplexers. Also mentioned in the answers to that question. This would considerably reduce the number of parts. (Olin rightly mentions 400 relays...)
  5. Crosspoint switches. Fair warning: these are expensive! But then you'll only need a few of them. They have a number of inputs (often 8 or 16), a number of outputs (also often 8 or 16) and a matrix which allows connections from one input to an output. They're expensive because most are spec'd for HF operation (video signals). I have yet to see one for audio. A solution with the analog multiplexers would come closest to that. (The most expensive part in this Analog Devices list is USD 300! A 16x16 switch is still more than USD 75. What I said: expensive)
\$\endgroup\$
10
  • 1
    \$\begingroup\$ @Tiddo - you're too quick accepting answers! Have you read Olin's answer? And I didn't even have the chance to edit mine! :-) And once an answer is accepted other users may not be very motivated anymore to post other alternatives. Give it a day or so, maybe there are better solutions than mine. The question had only 8 views so far. \$\endgroup\$
    – stevenvh
    Jul 3, 2011 at 17:11
  • \$\begingroup\$ You're right, I should have waited a little more. However, as I already stated in my comment to Olin's answer, I don't want to convert my signal to digital form and back. IMO your 5th solution is the best solution, but it's too expensive for me. So I still think I'll go for the multiplexers solution \$\endgroup\$
    – Tiddo
    Jul 3, 2011 at 20:58
  • \$\begingroup\$ Do you mean "But power relays are not made to switch signals in the first place."? \$\endgroup\$
    – endolith
    Jul 3, 2011 at 23:44
  • \$\begingroup\$ @stevenvh: by the way, do you think it is better to have multiple small microcontrollers to control each mux individually, or one big microcontroller to control them all at once? \$\endgroup\$
    – Tiddo
    Jul 4, 2011 at 12:29
  • \$\begingroup\$ @Tiddo - Depends on what the inputs/outputs of the different pedals depend on. If pedals are mutually related I would do everything in 1 microcontroller, if they aren't your software may be simpler if you have 1 microcontroller per pedal. \$\endgroup\$
    – stevenvh
    Jul 4, 2011 at 12:42
1
\$\begingroup\$

I'm guessing the outputs of each of these pedal things is a analog voltage? Are the pedals pots that put out a voltage as a function of pedal position?

If so, then the question is basically asking for a crossbar switch of 20 analog outputs to 20 analog inputs. I'll assume this is what you want.

This could be done with relays, but you would need a great many of them. If each input must be arbitrarily switchable to any pedal, then you need 400 separate switches. This would be physically large and fairly expensive. Solid state relays are not appropriate since they will have voltage drops accross them. They are meant for switching power where dropping a Volt or so doesn't matter.

There are such things as analog switches. These are ICs with several analog paths that are either open or closed to a few 10s of Ωs and are digitally controlled. You could put 20 channels of these switches in front of every input and have firmware in a microcontroller turn on one of them per input. This is basically the same as the relay concept, except that the analog switches will be much smaller and cheaper. Ultimately there are still 400 individual switches. You might want to buffer the output of each individual mux (collection of 20 switches with their outputs tied together) with a opamp. This avoids issues due to the on-state impedance of the individual switches.

However, there might be a better way altogether than a brute force crossbar with 400 independent switches. You could read each of the 20 pedal output voltages into A/Ds in a microcontroller, do the selecting of which inputs are used for each of the outputs, then write the selected values back out to D/As, once for each input. Instead of 400 switches you have 20 A/D input channels and 20 D/A output channels. Since the pedals change their outputs in human time, you've got lots and lots of time to scan the A/D inputs in a microcontroller. There are micros that have 20 or more A/D inputs, so all you would need is little R-C filtering and input protection, then right into 20 separate pins of the micro.

In the digital approach, the muxed pedal signals don't need to be real D/As. They could be low pass filtered PWM outputs from the micro. No micro has 20 PWM output channels that I know of, but again your signals are so slow that the PWM should be doable in firmware with a interrupt per PWM slice. For example, a 100 kHz interrupt inside the micrco that updates each PWM output would give you 1 part in 1000 resolution (much better than you can hold your foot in a given position) updated at 100 Hz rate. That would require some low pass filtering, but the output should be able to keep up with how fast a human can move a foot pedal. Each low pass filtered result should be buffered with a opamp to make the impedance low.

I'd seriously look into this last approach. The electronics will be simpler, smaller, and cheaper. The foreground code in the micro can be interpreting commands received via a UART to indicate which pedal should drive each output. It can also store the last configuration in non-volatile memory.

\$\endgroup\$
3
  • 1
    \$\begingroup\$ Thanks for your answer! However, the pedals are NOT pots which put out voltage as a function of the pedal position. These pedals modify the signal in many ways, depending on the signal of the guitar, and some settings on the pedals itself. Although your solution is really nice, I don't like the idea of converting my analog signal to digital, even though I know I'll probably won't really notice the difference (guitarists are quite picky on their equipment). So I think I'll still go for the multiplex solution, even though it requires a little more components. \$\endgroup\$
    – Tiddo
    Jul 3, 2011 at 20:45
  • \$\begingroup\$ By the way, if I'm correct I only need one multiplexer for each input, which redirects the signal to the next output. 20 multiplexers will probably fit easily in my device. \$\endgroup\$
    – Tiddo
    Jul 3, 2011 at 20:51
  • \$\begingroup\$ @Olin why did you specify 'analog' paths in your description? Is that just like a normal 74HC multiplexer but it will operate in the analog domain rather than just 0 or 5V? \$\endgroup\$
    – NickHalden
    Jul 5, 2011 at 17:59

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.