so ive been messing around with this for a few weeks now and i think im overcomplicating it but here goes. i have an old prop pitch rotator that i picked up at a hamfest and the controller that came with it is pretty much toast, the pot inside reads all over the place and i dont trust it for anything more than rough positioning. anyway i got to thinking that an arduino mega with a couple of relays and a good analog read on the feedback pot might do the job better than buying a replacement controller which honestly the aftermarket stuff for these is not cheap.

the basic idea is just read the pot voltage, do some math to convert that to degrees, compare it to a target heading sent over serial from ham radio deluxe or whatever, and then drive the relay board to run the motor CW or CCW until its close enough. ive got the serial comms working and the relay switching is fine but im having trouble with the pot readings being noisy — like plus or minus 8 degrees of bounce even when the antenna isnt moving. tried a 100uf cap across the wiper but it helped maybe a little. running 5v reference off the arduino which is probably part of the problem i guess.

has anyone gone down this road with a rotator project or something similar where youre reading a noisy analog signal and trying to get clean position data out of it. wondering if i should just switch to an external aref or maybe do some software averaging. the raspberry pi version of this might be cleaner but i really prefer keeping the control logic on something that just boots instantly and doesnt need an OS

yeah the onboard 5v aref on those arduino megas is pretty dirty, its riding on whatever noise is on your power rail which in a ham shack is basically a disaster waiting to happen. try wiring the AREF pin to a 3.3v regulator or better yet one of those little LM4040 precision voltage references, makes a huge difference with position feedback pots. also are you using shielded cable between the rotator and the controller box? that whole run is acting like an antenna and picking up everything especially if youre near your feedlines.

the software averaging thing works too, i usually do a rolling average of like 16 or 32 samples and only update the position variable if the new reading is more than a degree or two away from the current stored value. kills the jitter pretty effectively. i did almost exactly this project with a CD-45 about two years ago and ended up with maybe plus or minus 1.5 degrees accuracy which is honestly better than the original controller was doing anyway

cant speak to the prop pitch specifically but i built a raspberry pi based logger and rig control box a while back and ran into the same noise issues trying to read an analog input — ended up just bit-banging to an ADS1115 over i2c which is a 16 bit ADC and that thing is rock solid even in the shack. obviously thats an i2c device so on the arduino side you can use the same chip no problem, its like 2 bucks on aliexpress. might be overkill for a rotator but once you stop fighting the noise problem you can actually focus on the fun part of the project