No, the glider (mk 2) isn’t flying yet. It has been making progress though, in pulses, and it’s almost there. One day…
In the mean time, it’s sailing season. With no access to a keelboat this year, I’ve been getting into Dinghy Cruising, which is just about what it sounds like. You upgrade a daysailing dinghy (a small sailboat launched from a beach, that has no lead keel) for rougher conditions than normal, and maybe add oars or a trolling motor, then go camping. Similar idea to sea kayaking. It’s very popular in the UK, somewhat on the east coast, not so much in BC.
In other words, most people around here think I’m a bit nuts to do this, excepting the rare physics geek who’s been known to go along for overnight trips.

(under spinnaker off Bowen Island. No, that’s NOT me, that’s an actual physicist)
So I’ll mostly end up going singlehanded, which is a fun challenge in itself. When singlehanding, you wind up being a “slave to the tiller” for many hours on end. Grabbing your water bottle or adjusting something at the mast can end in a minor emergency, not least because the skipper is also the ballast. And dinghy’s don’t hove-too (stop with the sails up) all that well, so taking a leak is an even bigger issue.
The traditional answer is to lash down the tiller with a line, hopefully easily adjustable, like this, but the heading tends to drift.
So, an autopilot! Keelboat size tiller-autopilots use a linear servo and magnetic compass, burning a fair amount of 12v power, and cost around $550. I figured I could make my own.

The idea is you replace the usual single line bracing the tiller with a clothesline arrangement, with the autopilot on one gunnel and a pulley on the other. You then have a jam cleat on the tiller to grab the top line, and just attach or detach it from the line at will.
It has about 12 lbs of maximum pull, which is plenty, and gets NMEA data from a Garmin GPSMAP 76 for compass heading, track, and waypoints. So far it’s seen one day of testing, and worked well in light to moderate wind conditions or under oars, but not so well in waves. About what I expected; the major limit to performance is the only 1/2 hz updates from the garmin handheld.
It’s also watertight, including the servo, and has a small solar panel that should keep it full charged. An buzzer sounds to warn about error states. The microcontroller code wound up being 700 lines or so, 90% of the ROM available (4kb); close.
Not bad for a < 40 hour project. Hopefully it’ll perform well on the next longer trip.
Comment [208]
* * *