so ive been going back and forth on this for a few weeks now. I have an old Yaesu G-450A rotator that works fine mechanically but the controller unit is pretty beat up, the meter is off and one of the pots is flaky. instead of hunting down a replacement controller or sending it out i started thinking about just building something around an arduino and a couple of motor driver boards.

the basic idea would be to use an arduino mega, read the azimuth pot through the ADC, drive the motor through an H-bridge (was thinking the IBT-2 module, it handles enough current), and then either run a small LCD for local display or just do everything over serial and use a python script on a raspberry pi zero to handle the actual interface and maybe tie it into hamlib or rotctld so i can control it from WSJT-X or whatever.

has anyone actually done this end to end? i see a lot of half finished projects online where guys build the arduino part and then sort of trail off before integrating it with anything useful. the rotctld piece is what im least sure about, not totally clear on how to set up a virtual rotator interface that software will actually talk to properly. any gotchas i should know about before i sink a weekend into this

done almost exactly this with a cheap chinese rotator that came with a basically useless controller. used a nano instead of a mega since i didnt need the extra pins but same concept. the IBT-2 is a solid choice, i used a BTS7960 based one same thing really, just make sure you put a decent flyback diode setup in there or you'll get weird resets when the motor kicks off, learned that the hard way after about two weeks of head scratching.

the rotctld integration is actually not that bad once you understand what its expecting. you write a simple sketch that listens on serial for the hamlib ROT commands, basically just P for position set and p for position get, and return the azimuth in the format it wants. theres a github repo by someone called ok2cqr or something close to that which has a pretty complete implementation you can adapt. the pi zero is overkill for just the serial bridge but if youre already planning to have it there for other stuff then sure, i run mine headless and just ssh in when something breaks.

one thing i wish id done differently is add some software end stop limits early on rather than relying on the mechanical ones, just a few lines of code but saves you from a bad day if something in the position feedback drifts.

yeah rotctld isnt too scary, i had it talking to my az-el setup within an afternoon once i found the right model string. honestly the hardest part for me was getting the ADC readings stable enough — rotator pots are noisy and if you dont do some averaging or filtering in firmware you get this jittery azimuth display that drives you nuts. i ended up just doing a rolling average over like 16 samples and that cleaned it right up.