That's sweet Brian.

I did a lot of surfing for autopilots but didn't find that one. It looks very similar to my own. That's the first one I've seen that discusses the hydraulics. There's another similar one that uses a tiller pilot electric ram. Mine could be of interest and educational to others. I guess since it's open source I'm compelled to get the details up on the web but at this point I'm just too busy working and sailing to fuss with that kind of stuff.

I wanted to mimic the hydraulics on my model boat with tiny little cylinders they make for model excavators but they were super expensive and didn't have the same kind of pump.

Most hydraulics have a permanent pressure created by a pump that runs continuously and they open and close valves to actuate the cylinder but these autopilot cylinders use a different kind of pump that runs on demand. I think it's just a reversible gear pump and I looked for tiny ones of those as well. But in the end the model hydraulics would cost a pretty good fraction of the full sized hydraulics so it didn't make sense to model that.

I ended up actuating the model rudder using robot stuff like pulleys, belts, and a gear motor. I used a hall effect rotary sensor for the rudder position indicator. That's got to be way better than the $1k rudder references you buy from the big boys.

Long term though if I were to get serious about the development I'd like to have the rudder position sensor built into the hydraulic ram so there's no possibility for error. I'm pretty sure you can also get cylinders with hard limits on them so they are physically incapable of going beyond the desired swing. The fluid pressure thing is also something I'd like to look into.

In addition to that mine would have some error handling built in so it didn't do crazy things like gybe the boat when the compass data fails. I mean really.... if one second ago everything was fine what makes it Ok to the next second swing the helm hard over? That failure mode would be so easy to code out.

My rant these days about boating is that instrument makers create cheap consumer grade cr@p but they want to charge top of the line prices for it. They want to stick with proprietary interfaces and software even though it's clear the public wants something else. At this point the only instrumentation I would consider would have to be compatible with OpenCPN. Companies who are hostile to OpenCPN will not get anything from me.

I'm just now getting around to putting the model boat in the water so I've got some development to do. I'll try to make a movie and post that. It's the only idea I have so obviously I've got to stick with my day job, but there's a lot of people designing their own autopilots so clearly there's a desire to get away from the Simrads, and B$G's, and other assorted a$$wipes. If I had the money to really prototype a system it would sure be fun. My system wouldn't be cheap but I would give you part numbers and supply sources as well as source code a theory of operations etc. It would be the last autopilot you'd ever have to buy.