Digital presets for analog circuits?

[varialbender]

I'm working on a knobby phaser, and it would be so much easier to use if it had presets. Is it possible to switch between a couple presets of 8 knobs and 3 or 4 switches? Will learning with the Atmel AVR STK500 lead me in that direction? Anything else I should start learning? I'd love to be able to build that. Any help is much appreciated.
Thanks

[R.G.]

Using a uC makes it simple. Use external pots that only go into the uC. The real controls are digital pots which the uC manipulates to be like the external real pots. All you need from there is a "remember this" switch for the uC and you can have as many presets as your uC has flash/eeprom for, which could be a lot.

Oh, yeah - you have to work out some way to tell the uC to remember setting #7 instead of #4.

See ASMOP - user interfaces for some of the problems. To store presets, you have to be able to give the presets a unique identifier. A mechanical switch works. A simple toggle gives you two - live pots or remembered preset. A rotary switch can give you up to about 12. The uC just has to know which switch position you want to recall, and bingo, you're there.

[varialbender]

I'm hoping to have this interface:
StompSwitch1 changes to the next preset or saves the current settings over the current preset if held down for 2 seconds.
StompSwitch2 is just an ordinary bypass switch.
The current bank (of 4 maybe?) will be displayed by an LED.

So the STK500 is a step in the right direction, I take it. Ya?

[Peter Snowberg]

If you want to select between a couple of sets of knobs, you can use a single switch to operate some simple CMOS MUX chips like the 4053. You get three switches per 4053 so figure out which connections need to get switched on each of the controls and wire all the 4053 select lines together. Bring that select line to a SPST switch going to ground and a 10K resistor going to +V. To get more complex, you could do the same thing with a 1-of-4 mux and have up to three pre-sets along with your normal controls.   Removing the complexity of the microcontroller is usually desireable. I love 'em but nothing tops discrete logic in some places.

For the control you just mentioned, you're going to want a microcontroller and much more complex control mechanisms like ePots.

The STK500 is an excellent way to head in in the direction you're looking at. 

[varialbender]

If I understand you correctly, that method would involve switching from one bank of real pots to another bank of real pots, right? So if I had 8 controls, I'd need 32 pots for 4 presets? I can see this being useful for smaller sets of knobs, but as I thinnk you pointed out, that wouldn't be right for my design. Am I understanding you correctly?
Multiplexers are fun though, eh? When I first learned about them, I immediately thought sequencers were just a clever use of them. A couple days ago was my first viewing of a sequencer design with the ZVex thread. Pretty cool.
Thanks for the advice, I'll definitely start working towards that with the STK500.

[Peter Snowberg]

Correct.

I would use the MUX method with the front panel controls and from one to three sets of trimpots for the presets on a pedal that uses two to four knobs. If you have 8 knobs and you want more than one preset, a microcontroller is probably the way to go.

[R.G.]

So the STK500 is a step in the right direction, I take it. Ya?

I'm not familiar with the STK500. From where I am with the PIC16F819, it would take me
1 - PIC 16F819
2 - MC42050 dual digital pots
1 - plugboard/breadboard
misc LEDs, switches, logic chips etc
1 - victim effect to use the pots on

and about three hours start to finish programming, simulation and debug.

Of course, I have already programmed up my subroutines for writing to the digipots, reading external pots and switches, etc. so I can stand on my own shoulders.

The program pseudocode flow is simple enough just to write down.

Hey - here's a thought.

Write down for us the pseudocode for how to make this work. That doesn't require ANY specifics of how the uC operates, the language, the digipots, etc.  I think it will be an enlightening exercise.  Do the design in a uC and language independent way, post it here.

[The Tone God]

Digital pots have a number of restrictions on their use. They cannot just be dropped into any application without proper consideration of both the circuit and the desired result.

As a somewhat more general answer I would sooner support the idea of having a set of pots for each parameter and selecting them with some kind of bidirection multiplexer controlled with a uC.

Andrew

[Peter Snowberg]

Mesa used a really cool trick for emulating ground referenced rheostats. They used a dual-element Vactrol with the common tied to ground and the used one of the LDR elements to provide feedback for driving the LED.  Quite cleaver!

The main problem I've seen with ePots is that they don't want to work with signals higher in voltage than the switching logic supply voltage.  It's just not a perfect world (yet).

With a microcontroller, you may be able to move some functionality like the LFO into digital. That makes control much easier.

[R.G.]

Digital pots have a number of restrictions on their use. They cannot just be dropped into any application without proper consideration of both the circuit and the desired result.

Right.

You may have to:
(a) restrict the signal range to never get outside their power supply voltages
(b) restrict the currents to their safe range, particularly paying attention to the wiper current
(c) buffer before or after them to prevent unfortunate loading effects on the circuits before or after
(d) diddle one or more time constants to keep the same range in the circuit; in a phaser, particularly, you might have to change cap values on the phase networks

If you ...really... want analog presets, do this:
(1) get several small stepper motors with shafts coming out both sides; I cried over the demise of the old 5.25" diskette drive because the head positioner motors were ... perfect, and for a few years there, the whole thing was free if you asked.
(2) mount your stepper to the panel, put a knob on the motor shaft (they were dang near always 1/4") and slip some 1/4" id rubber tubing over the other end of the motor shaft and the pot shaft, perhaps with some glue or a hose clamp
(3) Ideally, use a dual pot. This lets you have one section to read into the uC.

Now you can still turn the pots by hand, and the uC can read the pot positions and remember them, as well as turning them to remembered positions via the stepper. As a bonus, the cogging of the stepper imparts a neato-feeling stepping to the pot rotation. If the motor pushes too hard, the tubing slips on the motor shaft and you don't twist the shaft off, no limit switches needed.

I think Panasonic makes motorized pots, available through DigiKey and/or Mouser, too.

[DavidS]

When I was a kid/early teen, my dad got a Sony home stereo setup that had a motorized volume knob. When you pressed the Vol. Up button on the remote, the knob turned! I loved that, I played with it all the time...

[MetalGuy]

A Digital Pots Control Unit is about to be born very soon so stay tuned!
The DPCU will be able to store the settings of 8 to 16 digital pots /controlled via mechanical rotary encoders/ into patches and banks. The simple version / a floor pedal/ will include 5 footswitches for 5 patches. Think a simplified version of JMP-1 + compressor and noise gate. The more complex version will support LCD and other features.

[mimmotronics]

A Digital Pots Control Unit is about to be born very soon so stay tuned!
The DPCU will be able to store the settings of 8 to 16 digital pots /controlled via mechanical rotary encoders/ into patches and banks. The simple version / a floor pedal/ will include 5 footswitches for 5 patches. Think a simplified version of JMP-1 + compressor and noise gate. The more complex version will support LCD and other features.

I'm interested! Was this ever implemented?