so ive been messing around with this for a few weeks now and im finally getting somewhere. basically i have an old prop pitch motor setup for my yagi and the controller that came with it is kind of garbage - analog meter, no feedback, the pointer drifts all the time. so i figured why not just replace the whole control unit with an arduino and a couple of relays and actually get real position data out of it.

im using a cheap rotary encoder bolted onto the motor shaft with a little bracket i fabbed up, feeding that into an arduino uno, and then im talking to the arduino over serial from a raspberry pi 4 that runs hamlib and feeds into my logging software. the idea is that when i log a contact or pull up a dx cluster spot it'll just... point the antenna at the right azimuth automatically. in theory anyway.

the part im stuck on is the PID tuning for the motor control. the motor overshoots like crazy and then hunts back and forth for like 5-10 seconds before it settles. i bumped the D term way up and it helped a little but now it feels sluggish on big moves. probably a tuning issue but im wondering if anyone else has done something similar and has a rough starting point for the gains. the motor is pretty beefy so inertia is definitely a factor.

also running the whole thing off 12v from the shack supply, relays handling the motor direction, pwm on a beefy mosfet for speed control. seems solid electrically but open to suggestions if somethings obviously dumb there.

yeah i did almost exactly this about two years ago, different motor but same basic idea. the hunting thing is almost always a D term that's too high relative to your P, ironically. what worked for me was backing the D way off, getting P tuned so it reaches the target reliably even if it overshoots a bit, then very slowly creeping D back up until the overshoot just barely goes away. if you go past that point it starts to feel mushy and slow like you said.

one thing that helped me a lot was adding a deadband in software - basically telling the controller if im within like 2 or 3 degrees of target just cut the motor entirely. rotating antennas dont need to be perfectly on azimuth to work fine and fighting that last little bit of inertia is where most of the hunting comes from. i have mine set to 3 degrees and honestly for casual dx work on HF you'd never notice.

also, are you filtering the encoder signal? prop pitch motors are noisy as heck electrically and if youre getting spurious pulses on the encoder lines it'll mess up your position count and make the whole thing go haywire. i put a small RC filter on the encoder lines and it made a huge difference. just a 100 ohm resistor and a 100nf cap to ground on each signal line before the arduino input pins.

cant help much with the pid stuff but the hamlib + pi integration sounds really slick. im curious what youre using on the software side to push the azimuth commands from the logging software to hamlib and then down to the arduino. is that just rotctld talking over a serial port or did you write something custom? ive been wanting to do something similar with my setup and the software side is honestly what's been holding me back more than the hardware. hardware i can figure out, python scripts that actually work reliably are another story lol