Goofball_star_dot_etal wrote:

I can see two options.
The first is to attach a foil to the 'small' outboard
eg. http://www.freepatentsonline.com/3991700.html
This would be steered through your existing arrangement with or
without an autopilot. I assume that this small outboard can be raised
when not in use, removing the rudder from the water, but if it is
robust enough could be left in the water at speed.. The disadvantage
of this scheme is the large load on the autopilot of having to control
both motors though the steering mechanism.

The steering is hydraulic. I haven't done any testing, but it doesn't
seem like much of a load. I can disconnect the arm which connects the
two motors - the force required to turn the wheel doesn't change very
much.

In the case above, if I read the link correctly, the "rudder" is nothing
but a flat plate behind the propellor. And, yes, the motor can be
raised. I don't understand how the "rudder" adds anything as steering
is still accomplished by moving the motor. The only thing added here is
direct feedback from motor pointing (via steering). That must be
important for the system to function as it was suggested by the
Sportpilot (if that's what it was). I don't see what it adds as all it
would do is provide an indication of potential slack or "play" in
steerage.


The second option would be an auxiliary rudder similar to a typical
sailing dinghy rudder with a lifting blade to be used only when
trolling. This could be large enough to do the required job but quite
light and weak. If the rudder/boat response is good enough the
cheapest autopilot could be attached and used just for low speeds,
with a remote control if required. This arrangement would be a
nuisance and likely be damaged at speed so needs to be retracted then.
I would assume that any autopilot would net out yaw, but I don't know
and don't want to pay all outdoors to find out. Are you saying it's a
large factor?

No, I mean that most boats at normal speed turn more or less like a
car, with a turning circle radius that reduces with helm movement. The
assumption is that autopilots are designed for typical boats behaving
in this way. Observing autopilot behaviour, they seem to move the helm
in proportion to the error, although there is also a dead-band, some
integration of error and some damping involved. It seems they would
assume a rate of turn proportional to the error signal and actuator
movement whereas a jet or outboard at low speed produces a very small
transverse thrust in response to the angular error signal, causing
angular acceleration that increases the rate of turn with time.
It seems a better plan to try to fix the boat's low speed steering
than to expect a control system to cope with this. You need the gain
of the factors that relate angular heading error to the boat's rate of
turn (yaw rate) to be more or less constant and within the bounds of
the autopilot design, then all should be well. You don't want the
error integating up because nothing has happened yet!

My understanding of yaw differs from yours. You're probably correct.
My understanding of yaw is coupled with pitch and nets out but does
provide considerable movement. I _thought_ a fluxgate compass enclosed
a gimballed affair with sufficient time constants to do that (net out
pitch and yaw).

With the Sportpilot, it seemed as though if the time constant of the
response curve could have been lengthened, the system would have worked.
What I perhaps should have done was to measure the amount of time the
boat took to go from one compass point to another (say 45 degrees apart)
and compare that with what the current boat does, both at the same
speed. Actually, that sounds like a plan and I can still do that as I
will be seeing the owner of the Utopia in early July and will have my
boat with me. At trolling speeds, if the boat turns substantially
faster, then the old system might work - without a rudder!

Your explanation makes sense to me, but unless the control electronics
are very "smart" that gain has to be variable and it wasn't. Well,
there were, as I recall, three steps, but the longest response time
wasn't long enough and horrible overcorrection occurred at low speeds.
In fact, the longest response was what I had to use at higher speeds.

"Bigfoote" seems to indicate that a Sportpilot would work and, as I
recall, that was nowhere near as expensive as other units. Of course,
you get what you pay for, yada, yada, yada...

Thanks!
--

Larry R
rapp at lmr dot com