Storing presets

[mikeb]

I'm building a microsynth for myself, and have been thinking that the complexity of the thing makes it awkward for live use; also, it would be neat to be able to switch sounds in an instant. How was storage of presets done in the early days with digital pedals? Was it using digital pots or some sort of analog interface (voltage control of FETS or of an OTA or?)?

I'm trying to get enough of an idea of how this might be approached. Obviously a PIC or similar can be used to store and retrieve the 'presets', but does this necessitate a nifty way of sampling the current position of each control needing 'memorising'? I'd like to approach this from the point of view of retrofitting an existing circuit WITHOUT redesigning too much of it from scratch.

TIA

Mike

[R.G.]

There's only two ways: you can sample the present position and memorize it or do all the pot-twiddling digitally so the PIC knows what you're saving.

The simplest thing to do is to use all digital pot ICs, and only up/down buttons for the controls.

The next-simplest is a dual pot (one control, one sampler) with a stepper motor arranged to drive the shaft through a rubber tube clamped on the motor shaft and pot shaft. When off, the stepper lets you turn the pot. The sampler pot lets the PIC sample where the control is, and the stepper when on lets the PIC turn the pot mechanically to the preset.

I love the PIC pot turner myself, but it's considerably more complicated than the others.  

[niftydog]

perhaps there's another way... this is off the top of my head, and it depends on exactly how the circuit operates;

For simplicity, you have one control you want to store presets for. Say you want THREE presets. You retrofit three trim pots (or even fixed resistors) and a CMOS switch. Now, you "connect" each switch in turn, set the trim pots to the desired setting and they can now be recalled by simply closing the appropriate switch.

You might still use a PIC or a PICAXE to control the CMOS switch.

[mikeb]

Yes, I'd thought of that way Steve - that's the 'Yamaha CS80' way of storing presets. It would take a bunch of CMOS switches, but for a 'step through a bunch of presets' approach, it would be relatively easy and wouldn't require much brain power (always in short supply).

RG, the price of digital pots means I'd unlikely to go that way. Hmmm ... what about if each pot controlled a LED/LDR combo, and the pot was also connected to an analog sampling input. Then in 'live' mode the pot would control the brightness of the LED, and hence the LDR value - additionally the micro could sample the voltage. In 'preset' mode the pot would be switched out and the analog would instead become an analog out (I know some PICS are capable of this - I think), and then would control the LED value itself. I need to go re-read the capabilities of PICs and the Atmel micros again, I think .... but the trimpot presets sounds like #1 choice for now!

Mike

[puretube]

free information block:

[niftydog]

PICs are indeed capable of this, but, short of some fancy coding, you're limited to one or two "analogue" outputs that can easily produce a PWM waveform.

[Mark Hammer]

I posted a scan of an old (>20yrs) Polyphony article on my site sometime back that shows one way.  This is the Quad Sequential Switch article at http://hammer.ampage.org (scroll down a few pages).

This uses CMOS switches driven by a 4017 counter driven by a 555 clock to step through settings in round-robin fashion.  There ARE limits to what you can do with this, and the inability to have direct random access to settings without first stepping through the others is a serious operational constraint, but it CAN work and is still pretty much in the analog domain....which makes it feasible for us mere mortals.

The CMOS switches can be ganged, such that each step turns on, say, 4 switch sections, which in turn might select between different pot settings.

As a simple example, suppose I have 2 inverting op-amp sections.  Feedback resistance in stage 1 determines the amount of drive for a set of clipping diodes immediately after it, Distortion+ style.  Section 2 sets the volume level, which can be greater or less than 1, given the manner in which feedback resistors in inverting op-amps can alter gain.  You could effectively "flip" between a few distortion/level settings by lifting/enabling the connections between one end of each pot and the corresponding op-amp pin using a switch section per pot.

Here, every preset requires its own corresponding set of pots.  So, while it can be relatively simple to implement, you may need a lot of chassis-panel space, not to mention pots, knobs, and connecting wires.

[R.G.]

This uses CMOS switches driven by a 4017 counter driven by a 555 clock to step through settings in round-robin fashion. There ARE limits to what you can do with this, and the inability to have direct random access to settings without first stepping through the others is a serious operational constraint, but it CAN work and is still pretty much in the analog domain....which makes it feasible for us mere mortals.

Yep - that does work. I can see random access by using the one-of-N CMOS transparent latch trick to select the set of things you want.

... I think I think of this approach as the "Sea of Pots".  

[Paul Perry (Frostwave)]

The whole basis of the analog synths, was to have voltage control over all parameters. Not all did, but if you do, then the problem is just one of checking voltage levels & then regenerating them.
"flying fader" technology is what you want.. motorised dual pots, one element active, the other used as a rotation sendor. Won't be cheap..!

[austin]

i would get a bunch of 8 pin micros with ADC  and PWM built in.  each control would have it's own micro.  then have another micro that is the master.

this is assuming your pieces are voltage controlable.

[niftydog]

ATtiny13 AVR micro would by my PIC (ha ha ha... ha... ehhh.  :? )

ehem...

2 PWM outputs in a 8 pin DIP!

PIC don't seem to have a small micro with PWM. I'm prepared to be proved wrong though!

[mikeb]

The CMOS-switched trimpots look like the best option for me currently. Still need to check if they will work ok with the voltages in a micro-synth though.....

Mike

[niftydog]

the inputs accept up to Vcc +0.5V, from memory.

[Paul Perry (Frostwave)]

PIC don't seem to have a small micro with PWM. I'm prepared to be proved wrong though!

PIC16F87X ?

[toneman]

"memorizing" pot settings is 'easy" if all U r doing is
digitizing and storing a voltage value.
the Paia Proteus monophonic synth had 16 recallable presets.
it used a scanning technique with a 4bit flash digitizer.
it used NO computer!!!
flash converter was descrete comparators and resistors.
(see "dual 4-bit digitizer")
one recently sold on eBay.......
all readings were voltages converted to binary by flash converter,
stored in battery backed RAM.  U had to step thru the presets
to get to the one U wanted.
there was even an LED on (below)each of the knobs to indicate
when the memory value and pot value had changed/were different.
VERY clever.  VERY effecient.  VERY long time ago.
Just got my blue LEDs in for my Proteus Restoration.
afn
tone

[niftydog]

PIC16F87X ?

well, yeah, but 28 pins for just two PWM outputs.

I was thinking along the ATMEL lines of the 8 pin micros.