Re: S-Tec 50 Autopilot

["James F. Agnew" <RRT0162@IBM.NET>]

Brian K. Michalk wrote:
> 
> > I have been reading off & on about problems with the S-Tec 50 two-axis
> > autopilot and opinions from different guys on how to fix such problems.
> > My personal experience was that both, the aileron and the elevator servos,
> > were over controlling and producing oscillating movement which could go from a
> 
> I've got a beef with using only altitude as an input for an autopilot.
> As most of you may know, I'm rolling my own integrated computer
> into the "Goose".  One of it's functions will be autopilot.
> Actually the autopilot will be handled by an embedded PC from Zworld,
> and it will be instructed from the main computer via RS485 for the altitude
> to hold.
> 
> We know that when one is sufficiently "in front of the power curve"
> (I hate that saying) that attitude will at least temporarily
> affect altitude.  So a pitch up results in an increase in altitude.
> The AP should take altitude as an input and change attitude
> by a prescribed amount.
> 
> Correct me if I'm wrong, but if we have a perfect non-precessing
> gyro, and we know that we are X amount of feet away from our
> altitude and we are currently straight and level, then a Y degree
> change in pitch would be called for.
> 
> In turbulence this would be great.  If you are at altitude, and the aircraft
> is upset, the AP would instantly correct for pitch rather than altitude.
> 
> In reality, I think a cheap piezo electric gyro would
> work.  Damp out the really long wavelengths on the order of more than 5 minutes
> to fix precession.  Also one of the sensitive ADXL01 chips
> 
> that measure acceleration might work.  Mount it with it's sensitive
> axis longitudinal to the aircraft.
> 
> Oh well, just my two cents.  This is one of the reasons I'm
> building my own airplane.
> I think it's gonna be a hoot to try out
> all these ideas I have knocking around
> inside my head.
> 
> --
> Brian Michalk  <http://www.awpi.com/michalk>
> Life is what you make of it ... never wish you had done something.
> Aviator, experimental aircraft builder, motorcyclist, SCUBA diver
> musician, home-brewer, entrepenuer and SINGLE!

Brian, something to think about, on my sail boat I had an autopilot that
used an analog computer with a flux gate sensor for heading
information.  This system used a feedback loop from the "rudder" so it
could tell how much it was using to turn.  It used the rate of change
from the sensor to determine how much and how fast to to move the rudder
and used a variable speed (pulse width modulation) electric motor to
move the rudder.  Because of this setup it could react quickly to large
movements and back off as it came to center so it wouldn't overshoot. 
It also looked at the number of corrections in the same direction and
after several would find a new centering heading to set a crab to offset
drift.  Something to think about in your design.  You can now buy Flux
Gate sensor black boxes with several types of output, I don't have the
address but I could find it.

Jim

PS - this unit was for a 9 ton boat and had a 1250 pound thrust
hydraulic ram so it is probably is a litle heavy duty for aircraft
use!!  (:>)
-- 
//------10--------20--------30--------40--------50--------60--------70--------80
// James F. Agnew
// Tampa, FL
// Velocity 173 FG Elite ( http://www.VelocityAircraft.com/ ) under
construction