REFLECTOR:Navaid Autopilot

[Alexander Balic]

Brian,

About the PID control, I have used them regularly in industrial
applications- usually for temperature controllers, but for other things as
well, they are great when they have the chance to tune themselves to the
oscillations of the system that they are connected to, to reduce undershoot
and overshoot. What they are designed to do as you said is to reduce the
amount of oscillations that occur within a corrective action- as in a curing
oven for example, the unit tunes to the mass of the oven, and the lag in the
heating system, so if the oven is under temp, it will turn the elements on
with the proper amount of amperage and the proper amount of time so that as
the temp approaches the set temp, the elements will turn off just at the
correct moment to prevent the temp from overshooting ( thermostats
overshoot) we used them for viscosity control with great success also.  The
problems that all of the PID controllers that I have used (Omron and AB) is
that they are not that good at correcting rapid oscillation, nor are they
good if you change the system in which they are controlling on a regular
basis. I suspect that this is what might be happening if indeed the AP uses
this technology to determine servo throw.  If the turbulence was the same
every time the aircraft was upset, and the aircraft CG and trim situation
was the same, then the PID system would have the required number of
oscillations to tune. but since none of those elements are constants, I
would think that the PID system would remain in a tuning state, and have
great difficulty determining the correct outputs to send to the servos to
dampen the pitch oscillations properly. But if the AP can not even keep the
aircraft level in a non-turbulence state, then it seems that there is
something else wrong, since the PID system should have no trouble keeping
things in order.

-----Original Message-----
From: reflector-admin@tvbf.org [mailto:reflector-admin@tvbf.org]On
Behalf Of Brian Michalk
Sent: Friday, July 11, 2003 7:15 AM
To: reflector@tvbf.org
Subject: RE: REFLECTOR:Navaid Autopilot


Where does the Navaid get it's inputs from?

This is a classic coupled oscillation problem.

If you gyro lags the actual orientation of the aircraft, then the Navaid is
always going to correct (overcorrect) for a condition that existed in the
past.

If turbulence upsets our airplane, we want the autopilot to immediately
correct, just like we do.  However, unless there's a good PID algorithm in
the autopilot, it's going to correct nice and easy.  I doubt they have a CPU
controlling this, as it's got to meet a whole other set of rules for
certified installation.  I could be wrong, I don't know if the S-tec is
approved for certified aircraft.  Anyway, it could still be a hardware PID,
and PID's can be difficult to tune.  Back to turbulence.  If he have it be
nice and snappy in turbulence, then when we are in calm air, it will tend to
oscillate, because it is trying to quickly bring the plane back to level.
However, if the gyro is lagging the actual state, then the AP will
overcorrect, causing the oscillation.

It's actually a lot more complicated, but that's the gyst.  Other culprits
are slack in the control surfaces.  Even though some people have added
damping to the mechanical output, slack could have been the original
problem.  All you need to do is move the oscillating frequencies of the
aircraft and control system away from each other such that they tend to damp
each other out.

PID algorithms are pretty cool.
P---- proportional: move a proportional distance towards the target
	The main parameter that makes the PID work.  If I have 10PSI of pressure,
and an air cylinder moved a weight 50% to my target, set the pressure to
15PSI.
I---- integral: the longer I've been off target, move more forcefully
towards the target
	This one keeps one from being "almost" at the target.  If there's friction
in the system, you need a little bit of push to move small distances to get
right on top of the target.
D---- derivative:  The quicker I'm getting to the target (actually
increasing rate of error), move less forcefully.
	This one keeps us from overshooting the target.  If I am a mile away from a
stopsign, going 60miles per hour, my error rate is fairly low, it's the
error from second to second.  At a mile away, it's 1/60.  However as I
approach the stopsign without reducing speed, then the rate of error becomes
1.  I'll overshoot unless I pay attention to the error rate.

 Brian Michalk  <http://www.michalk.com>
Life is what you make of it ... never wish you had done something.
Aviator, experimental aircraft builder, motorcyclist, SCUBA diver
musician, home-brewer, entrepreneur and mostly single


> -----Original Message-----
> From: reflector-admin@tvbf.org [mailto:reflector-admin@tvbf.org]On
> Behalf Of Chuck Jensen
> Sent: Friday, July 11, 2003 7:48 AM
> To: 'reflector@tvbf.org'
> Subject: REFLECTOR:Navaid Autopilot
>
>
> I found the comments about the S-TEC and its oscillation phenomena
> particularly interesting.  I'm running a Naviad AP and Wing
> Leveler and see
> the same type of behavior, especially above approx. 8,000'.  When
> the WL is
> turned on, it rapidly oscillates L/R.
>
> I really didn't have a clue what was wrong and hadn't had it
> looked at yet,
> but now I wonder if this is, in fact, the same thing that occurs in the
> S-TEC.
>
> Has anyone had a similar behavior with the Navaid?   Solution?   Is this
> problem likely similar/same as with the S-TEC?
>
> Chuck
> N27GV
> _______________________________________________
> To change your email address, visit
> http://www.tvbf.org/mailman/listinfo/reflector
>
> Visit the gallery!  tvbf:jamaicangoose
>

_______________________________________________
To change your email address, visit
http://www.tvbf.org/mailman/listinfo/reflector

Visit the gallery!  tvbf:jamaicangoose