"Josephine ... was designed to steer herself ..." - page 89. Something to do with her hull form, I guess.

Slieve McGalliard wrote:

The whole point of the trim tab and the later servo pendulum is to amplify/ augment the weak signal from the sensor vane to help it drive the actual rudder

That's the truth, but it's not the whole truth. " In order to obtain [negative] feedback, we must have a two stage system" - page 54 in Bill Belcher's book. That is, either a pendulum servo or a trim tab. Or, if true negative feedback is not present, we must have what Bill Belcher calls "synthetic feedback".

The fact is that boats with a very well balanced spade rudder are the most difficult to steer with a vane gear. Going back to the early 1970s, I spent a lot of time with Blondie, trying to get the SP to steer his Kingfisher 20, with its spade rudder, and it wouldn't, when used in the way that most boats used it, with 30 degrees of pendulum movement = 30 degrees of tiller movement. We had to bring the lines to the tiller very much further from the rudder axis, so that there was very little rudder movement. The boat itself was unstable downwind, and the balanced spade rudder did nothing to help stabilise the steering. We invented a "gate" gear, a servo that swung around a vertical axis, and this worked, but never lead to a commercially available gear as it wasn't what the generality of boats needed (though there was a Windpilot gear made to this principle at one time). Since that time, I've always made my pendulum gears with some trail to the servo blade, ie, the power axis is not horizontal but angled. This very effectively "puts the feathers on the arrow".

The point I'm making is that power applied to a balanced rudder is only half the equation. Keeping that power under control  by adding in negative feedback is the other half. A horizontal axis vane turning a balanced rudder directly is inherently unstable, and therefore the stability must be supplied by the boat. A vertical axis vane will work (so long as the rudder is not fully balanced),  by virtue of the negative feedback supplied as the vane quickly lines up with the wind and loses power, but must be very large. 

With the Willing Griffin type of gear, the main rudder is acting as a large skeg, adding in the downwind course stability that the hull lacks. That's why I think I can get away without making a pendulum gear.  A vertical axis vane driving an auxiliary rudder that has some balance area but is not fully balanced has a chance of success. A horizontal or inclined axis vane driving Weaverbird's rudder directly has no chance whatsoever.

Some drawings can be found here.

The concept is an update of Willing Griffin’s OSTAR 1972 gear, an auxiliary rudder mounted on the main rudder, driven by vertical axis vane. David Blagden found that a Hasler trimtab gear just wouldn’t work on such a small lively boat as a Hunter 19, as it responded too slowly, being intended for larger boats. Then Blondie came to see her, and reversed the linkage so that the trimtab became an auxiliary rudder, with the main rudder fixed, and all was well. Weaverbird is a little larger than Willing Griffin, but has the same dinghy-style rudder, with no skeg, and is similarly lively to steer. A servo pendulum gear with a trailing blade would probably steer better, but would take a lot more time to build.

I have bought some SeaSure dinghy rudder pintles and gudgeons to mount a 120mm wide auxiliary rudder blade.

Moving upwards, there is a pin and slot linkage that is arranged to give quick movement through the middle of the auxiliary rudder’s arc, where little power is needed, and then more leverage at full deflection.

I have a problem that will be not commonly found - the main rudder is arranged so as to be lifted by about 125mm when the boat takes the ground. This makes linkages between the aux rudder and the vane a bit tricky, if the vane is mounted on the boat. I have decided to mount the vane assembly onto the rudder, so that everything moves together, and the only type of vane that can be fitted this way is a vertical axis vane. A horizontal or inclined axis cannot be used,as positive feedback is introduced, and steering becomes unstable. I have a length of 12mm diameter stainless steel bar, which I shall mount in a block attached to the rudder as a fixed vane support shaft. Rotating on this will be two 30mm diameter GRP tubes. The lower tube carries the pin that links to the aux rudder, the upper tube carries the vane, with the latch gear linking them. I have bought a bicycle’s 52 tooth alloy chainring as the basis of a toothed wheel and pin style of latch gear. This will give an adjustment of just under 7 degrees per tooth (the Aries ratchet ring has 60 teeth - 6 degrees per tooth). The main rudder is used to trim out weather helm, and can be set so as to bias the gear slightly, one way or the other, so 7 degree steps should be OK. A worm and wormwheel latch gear would give step-less adjustment, but would take a lot more time to build.

The vane itself will have a frame of 10mm and 6mm carbon fibre tubes and will be covered with spinnaker nylon, to keep it very light and responsive. It may not need a counterbalance weight. The triangular shape is so that, like a delta wing, it will develop very large amounts of lift at large angles of incidence without stalling, and also, the centre of area is as far away from the axis as possible.