so ive been messing around with this for a few weeks now and finally got something that half works. basically i have an old prop pitch motor setup on a home brew rotator and the original controller died ages ago so ive just been doing it manually which is obviously painful when youre trying to track anything.

the idea was to use a pi zero w with a little motor driver board — im using one of those L298N modules i had sitting in the junk box — and just write some python to handle the positioning based on feedback from a cheap hall effect sensor on the shaft. its kind of working but the positioning is drifting like 10-15 degrees over a few rotations and i cant figure out if its the sensor or my code thats losing count somewhere.

anyone done something like this or something similar? i know there are proper rotator controller kits out there but im cheap and i already had all this stuff laying around. also curious if people have used hamlib with a pi setup like this, seems like it should be possible to feed azimuth data back to the radio software but havent dug into that part yet.

yeah i did something almost identical about two years ago with a regular pi 3 and an old TV rotator. the drift problem you're describing sounds like it might be debounce issues on the hall sensor — those cheap sensors can give you phantom pulses especially if your motor is generating any electrical noise on the supply lines, which prop pitch motors definitely can. i ended up putting a small cap across the sensor output and it cleaned things up a lot, maybe 100nf or so just to smooth the transitions.

hamlib integration is actually not bad once you get the rotctld daemon running, you can set up a custom rotator model and it shows up just like any other in wsjtx or whatever. took me an afternoon to figure out the config file syntax but once its going it just works. the python side i actually scrapped and rewrote in C because the timing was too inconsistent with python for the pulse counting but that might just be my code.

dont overthink the sensor thing, i went through the same rabbit hole. ended up just using a 10-turn pot on the shaft geared down a bit and feeding it into an mcp3008 adc on the pi. way more reliable than counting pulses and you always know your absolute position even after a power cut. only issue is calibrating the pot range to actual degrees but thats just a one-time thing.