Elegant way of dealing with fader range not being full 0-10k?

[LukeThomasDawson]

Howdy again,

Some suggested guidance would be super appreciated on how to tackle the max/min extents of the ALPS 10k faders I am using...

It appears that the lowest and highest values are not actually GND and 10k, but rather a few hundred ohms off, so that reading in the ADC doesn't give a clean 0 - 1023 but more like 19 - 1006... Happy to care of this in software (so that the motor isn't constantly trying to drive down to 0 and heat up for instance) but was wondering what the sophisticated tttapa approach would be in to elegantly tackle the issue?

Happy to contribute if this feature is useful for others too.

Best,
Luke

[tttapa]

I don't think you can easily fix this in hardware.

In software, the obvious choice would be to just map() the analog input value. However, this poses an interesting problem: since you're mapping a smaller range (19-1006) to a larger one (0-1023), some target positions in the larger range become unattainable. (As a side note, the overhead of the integer division required for this mapping might not be insignificant if you're doing it at ~1 kHz.)
For the controller, it might therefore be better to map the reference from (0-1023) to (19-1006), rather than mapping the measurement from (19-1006) to (0-1023). You'll still have to map from (19-1006) to (0-1023) before sending the potentiometer value over MIDI. Essentially, you can move all range conversions to the MIDI part of the code, without modifying the controller.

[LukeThomasDawson]

Yep, love it.

The immediate thought was to jump straight in to a map() on the analog read, but really what is needed here is the setpoint to be the one where realigning happens so that the pot actually does get set to a real value that it can achieve - and then the MIDI setpoint + value return can do the adjustments.

I'll add in some pot max/min variables on the MIDI side and test out some different ways of "learning" what the true max and min values are for a given fader.

Thanks for hot tip on int division too... I'll keep an eye on cycles : )

Best,
Luke