"Skip Gundlach" wrote:
snip

Meanwhile, one of the guys working with me in the rudder restoration
suggests using a "ducktail" on the rudder to improve the autopilot
stability (it hunts busily, now, and has since we have owned the
boat). I'm assuming that if that were efficient, today's high-priced
yachts would have them. To me it seems counterintuitive - even the
lowest-slowest airplanes don't have such a configuration to the rudder
or elevators, and anything which creates drag (as I assume a blunt
end, particularly one increased in size from the exit, to perhaps the
width of a foot into the chord from the aft, would do) isn't a good
thing, either.


Focusing only on the autopilot issue presented, I would suggest that the
hunting issue is not likely related to rudder shape. The autopilot is just
a servocontrol loop, with PID configuration (Proportional, Integral, and
Derivative). Feedback comes from the fluxgate compass or wind direction
transducer. Your percieved hunting behavior is typical of servoloops
adjusted with too little forward, or proportional gain. They cannot react
fast enough to changing conditions, and are forever correcting for the
situation of a few seconds ago. Under certain circumstances, a servoloop
with too much forward gain can exhibit instability in a way that resembles
hunting, even though the mechanism is different.

Some servocontrollers also allow adjusment of the integral and derivative
parameters. The integral feature (sometimes called "reset") improves the
accuracy of response when conditions don't change much. This is often
automatic and not adjustable. The derivative adjusment stabilizes the
control loop, allowing more forward gain and hence quicker response by
introducing the rate-of-change of the feedback

The short version is; experiment with the adjustments available in the
autopilot and on its control panel, and note whether increasing/decreasing
each parameter improves the response of the system or not. If it makes
things worse, go the other way.

I know this is a little obtuse, but I hope it might lead you to another
possible cure for your control problem.

John Hart

--
-------------------- http://NewsReader.Com/ --------------------
Usenet Newsgroup Service $9.95/Month 30GB

wrote in news:20070314163424.069
:

-------------------- http://NewsReader.Com/ --------------------
Usenet Newsgroup Service $9.95/Month 30GB



John, for less than $4 more, on a 3-month payment at a time, you can
download 24/7/365 with unlimited GB and 100 days retention at 99.9%
completeness on www.usenetserver.com, which is FAR greater than any other
news server I ever encountered.....works great!

I download about 15GB/day...(c;

Larry
--
How much price inflation is caused by illegal
aliens gobbling up goods and services, creating
shortages for the natives? I heard 40%!

inscribed in red ink for all to know:
"Skip Gundlach" wrote:

snip


Meanwhile, one of the guys working with me in the rudder restoration
suggests using a "ducktail" on the rudder to improve the autopilot
stability (it hunts busily, now, and has since we have owned the
boat). I'm assuming that if that were efficient, today's high-priced
yachts would have them. To me it seems counterintuitive - even the
lowest-slowest airplanes don't have such a configuration to the rudder
or elevators, and anything which creates drag (as I assume a blunt
end, particularly one increased in size from the exit, to perhaps the
width of a foot into the chord from the aft, would do) isn't a good
thing, either.



Focusing only on the autopilot issue presented, I would suggest that the
hunting issue is not likely related to rudder shape. The autopilot is just
a servocontrol loop, with PID configuration (Proportional, Integral, and
Derivative). Feedback comes from the fluxgate compass or wind direction
transducer. Your percieved hunting behavior is typical of servoloops
adjusted with too little forward, or proportional gain. They cannot react
fast enough to changing conditions, and are forever correcting for the
situation of a few seconds ago. Under certain circumstances, a servoloop
with too much forward gain can exhibit instability in a way that resembles
hunting, even though the mechanism is different.

Some servocontrollers also allow adjusment of the integral and derivative
parameters. The integral feature (sometimes called "reset") improves the
accuracy of response when conditions don't change much. This is often
automatic and not adjustable. The derivative adjusment stabilizes the
control loop, allowing more forward gain and hence quicker response by
introducing the rate-of-change of the feedback

The short version is; experiment with the adjustments available in the
autopilot and on its control panel, and note whether increasing/decreasing
each parameter improves the response of the system or not. If it makes
things worse, go the other way.

I know this is a little obtuse, but I hope it might lead you to another
possible cure for your control problem.

John Hart


Very well said, John.

Another cause of failing to settle at the setpoint is too much lag in
the system. So, too much gain, or too much lag, or both will give an
unstable system. Adding rudder area will have the effect of increasing
the gain. It is unlikely that this will improve the system.

Another big cause of hunting: deadband. If there is slop either
between the actuator and the rudder panel, or between the rudder sensor
and the rudder panel, this is a big no no.

In all, I'd remove any slop, and then follow John's excellent advice above.


bob
s/v Eolian
Seattle