wing levelers

On Fri, 25 Feb 2005 00:12:06 GMT, "Dick" wrote:

Sitting around the hanger and discussing the possibility of somehow using
off-the-shelf electronic/computer components or gadgits for semi-automatic
wing leveling on lightly loaded, short wingspans.

To paraphrase the manual for my auto-pilot. When encountering more
than moderate turbulence, turn off the autopilot. I believe Pete
touched on this as well.

The same is true for altitude excursions.

Mine can do a much better job of holding altitude and attitude than I
when the going gets really rough.

That's when I turn it off and manually hold the attitude (more or
less).

Yes it can be done and I'd think the simplest would be to get a solid
state gyro, or turn coordinator... and build around that.


Not being computer guys, the current thought is that something should be out
there that is available cheaply and modifiable to use a simple wing
leveler.

All it takes is money. The less ingenuity, the more money.


Having heard someplace that automotive computers, as an example, sense data
millions of times leads us to believe that sensing the wing tip initial
movement and counteracting very very quickly would be a good thing.

Counteracting very quickly with a lightly loaded wing could possibly
break something very quickly as well.


As an example: One fellow pictured a rolling ball bearing inside a tube
somehow activating a magnetic switch which in turn pulses a dc linear motor
to operate the aileron minutely.

Rule number one...no... sorry that one is already taken. An airplane
and any thing in it does not know up from down in anything except
straight and level flying. However a gyro tends to stay in the same
position you put it when starting out, so it makes a good reference
platform. (as long as you don't do anything drastic to confuse it such
as causing it to tumble)
I haven't seen any prices on the components for a solid state gyro.


We are not sure how to detect the very earliest initial wing tip movement or
drive the small trim type motor and would appreciate some thoughts or site
recommendations to investigate.

Find an old DG, or TC and experiment from there.
Build in sufficient disconnects so if it goes TU you can put the
greasy side on the bottom again.

You can over-ride a wing leveler or AP, but it always gives me a
strange feeling to be telling the airplane to do one thing when it's
fighting me to do something else and its attempt is readily apparent.

Roger Halstead (K8RI & ARRL life member)
(N833R, S# CD-2 Worlds oldest Debonair)
www.rogerhalstead.com

Thanks, Dick

"Roger" wrote in message
...
To paraphrase the manual for my auto-pilot. When encountering more
than moderate turbulence, turn off the autopilot. I believe Pete
touched on this as well.

Driving an auto-pilot too hard can put your servos on the rate limit. On a
rate limit, a servo develops serious amplitude-dependent lag, which can
destabilize your loop closures. THe way arount this problem? Big, huge,
powerful, fast servos. This solution opens up a whole other can of worms.

All it takes is money. The less ingenuity, the more money.

Being intimate with the hazards associated with such a project, I would
never recommend this as a casual development project. In this case, less
ingenuity means great exposure to serious hazards. It all seems so simple in
concept, but the devil.......

Counteracting very quickly with a lightly loaded wing could possibly
break something very quickly as well.

And that's just one of the hazards.

I haven't seen any prices on the components for a solid state gyro.

The ones the R/C guys use, which are, in my opinion, quite rugged and
accurate enough for this kind of an application, pretty innexpensive. I
think I saw one model that sold for under $200. Systron-Donner makes
single-axis chips for (working from memory) under $500. Full 6-dof Motion
Packs go for around $20k. I think Crossbow has a 6-dof package for under
$12k.

Build in sufficient disconnects so if it goes TU you can put the
greasy side on the bottom again.

Yup. On the stick. Fly it with a gun to it's head.

Pete Schaefer wrote:

snip
I think Crossbow has a 6-dof package for under
$12k.

Crossbows prices have dropped over the past 2 years. The AHRS they had
at $10K is now under $8K per UI of 1.

Now if the made and interface to drive synchro instruments....

Dan, U.S. Air Force, retired

You can get really nice R/C servos for way under $100. Ball
bearingsand the works. The quarter scale size servos would probably be
about right to fly a control surface.

Piezo gyros are also under $100 for R/C applications.

Regards

wrote:

You can get really nice R/C servos for way under $100. Ball
bearingsand the works. The quarter scale size servos would probably be
about right to fly a control surface.

Piezo gyros are also under $100 for R/C applications.

Regards


The hard part is the electronics package between the two. I know the
systems I worked on, but I would be reluctant to attempt builing a
system. Not my bowl of rice, but I'd like see what others come up with.

Dan, U.S. Air Force, retired

"Dan, U.S. Air Force, retired" wrote in message
news:rhLVd.21112$Sn6.10965@lakeread03...
The hard part is the electronics package between the two.

It wouldn't be that bad, really. At least not for just a rate damper. The
R/C gyros put out pulse-width modulation signals, which are pretty easy to
work with. Also, there are a lot of decent servos out there that can read
it.

I think the hard part of the rate damper project is making sure you don't
overtorque something important in your airplane, and setting things up so
that, when the damper servo goes haywire, that you can turn it off and
revert to a normal airplane quickly and without hurting anything/anybody,
and that while you're busy figuring out that something is going wrong, that
you can overpower anything it's doing that you don't like.

More than a few people and quite a few airplanes have gotten killed in the
process of engineers trying to figure out how to do all this automatic
stabilization and fly-by-wire stuff. It's not something approached casually,
unless it's for a toy that's OK to crash (R/C model). In my somewhat short
career as a flight controls engineer (12 years), I've seen 3 airplanes lost
and more than a couple of close calls due to control design issues. Even
when everything works as designed, there are man-machine interaction
problems that can kill you. Even apparently innocuous things. Think back to
that Airbus that busted up a couple of years ago in New York. I think
there's an article in Flying this month about it. Nothing there that jumps
out at you as an obvious hazard, but it got a couple hundred people real
dead.

I'd really hate to see someone on this newsgroup go out and get himself
killed trying to invent something without full cognizance of the hazards.

Pete

Dan, U.S. Air Force, retired wrote:
wrote:

You can get really nice R/C servos for way under $100. Ball
bearingsand the works. The quarter scale size servos would probably be
about right to fly a control surface.

Piezo gyros are also under $100 for R/C applications.

Regards


The hard part is the electronics package between the two. I know the
systems I worked on, but I would be reluctant to attempt builing a
system. Not my bowl of rice, but I'd like see what others come up with.

Dan, U.S. Air Force, retired

Dan, did you ever get a chance to work with fuzzy logic?

It would require a stepper motor to control the control surface, but it
could easily maintain wing level or altitdude with grace and smoothness.
Have you ever seen the balancing trick with the mortorized car.

A weight on the end of a stick is hinged on top of a programmed electric
car. The car acclerates quickly to flip the weight vertical, and then
jostles back and forth to balance it there. All with not input except
an indicator of the angle of the stick holding the wieght. Impressive.

I have the information for the EZTrim altitude hold system. After I
have everything else on the plane working, I have a goal of reworking
the software to use a fuzzy algorithm.

"Ernest Christley" wrote in message
.. .

Ernest:

This reply is going to sound like a flame, and I really honestly don't mean
it to be, but I gotta step in and say something to ease my conscience in
case you go out and get yourself killed. If I were around and you were about
to go hop in your airplane to test something developed with this attitude,
I'd feel obligated to wrestle you to the ground, take away your keys, then
send you back to the lab to do a very thorough and formal system design
before allowing you to procede.

If I were ever to say the kinds of things you said in a design strategy
meeting, my coworkers would laugh their asses off, then beat the hell out of
me for suggesting such a thing. Then I'd probably end up in the tech pubs
department or fired or something like that.

Dan, did you ever get a chance to work with fuzzy logic?

I have. It's been a while. I'd never, ever use it on an airplane. Maybe a
washing machine controller or something like that. I'm not saying that it
can't be done, but just that it's a risky design approach. There are other
methodologies much better suited for aircraft.

Rule #1 of Flight Controls Design: KNOW YOUR PHYSICS! At the end of the day,
F still equals ma, and you ain't getting past that doing any fuzzy stuff.

It would require a stepper motor to control the control surface

Another big no-no. Steppers are fine for inkjet printers and stuff, but
initialization of position (need to be able to do a power-on reset in
flight), hazards of getting the windings out of sync (immagine you hit a
bump, and your underpowered servo gets knocked off a few ticks....now it's
running backwards.....yes, I've seen this happen), complexity of the power
electronics to drive it...... all these problems disappear with a decent
servo.

could easily maintain wing level or altitdude with grace and smoothness.

The way you say this, I can tell that you've never tackled a problem like
this before. There are tons of things to consider.

Have you ever seen the balancing trick with the mortorized car.

This is the undergraduate "intro to controls" lab experiment. It's meant to
illustrate the basic concepts of closed-loop control. Mastering this problem
only gives you a very small taste of what it takes to design even a simple
autopilot. If you've gotten that far, then next step is to either take a
flight controls class (grad level), or maybe start building some R/C models
if you don't want to go for more school (I'd suggest doing this anyway). The
school of hard knocks is fine with R/C, since the knocks aren't really all
that hard on you.

PLease, get yourself more experience with aircraft control before putting
yourself at risk. I'm guessing you already have a start in learning this
stuff, and don't think you should abandon your goals. However, take baby
steps.

Pete

Pete Schaefer wrote:

"Ernest Christley" wrote in message
.. .

Ernest:

This reply is going to sound like a flame, and I really honestly don't mean
it to be, but I gotta step in and say something to ease my conscience in
case you go out and get yourself killed. If I were around and you were about
to go hop in your airplane to test something developed with this attitude,
I'd feel obligated to wrestle you to the ground, take away your keys, then
send you back to the lab to do a very thorough and formal system design
before allowing you to procede.

If I were ever to say the kinds of things you said in a design strategy
meeting, my coworkers would laugh their asses off, then beat the hell out of
me for suggesting such a thing. Then I'd probably end up in the tech pubs
department or fired or something like that.


Dan, did you ever get a chance to work with fuzzy logic?


I have. It's been a while. I'd never, ever use it on an airplane. Maybe a
washing machine controller or something like that. I'm not saying that it
can't be done, but just that it's a risky design approach. There are other
methodologies much better suited for aircraft.

Rule #1 of Flight Controls Design: KNOW YOUR PHYSICS! At the end of the day,
F still equals ma, and you ain't getting past that doing any fuzzy stuff.


It would require a stepper motor to control the control surface


Another big no-no. Steppers are fine for inkjet printers and stuff, but
initialization of position (need to be able to do a power-on reset in
flight), hazards of getting the windings out of sync (immagine you hit a
bump, and your underpowered servo gets knocked off a few ticks....now it's
running backwards.....yes, I've seen this happen), complexity of the power
electronics to drive it...... all these problems disappear with a decent
servo.


could easily maintain wing level or altitdude with grace and smoothness.


The way you say this, I can tell that you've never tackled a problem like
this before. There are tons of things to consider.


Have you ever seen the balancing trick with the mortorized car.


This is the undergraduate "intro to controls" lab experiment. It's meant to
illustrate the basic concepts of closed-loop control. Mastering this problem
only gives you a very small taste of what it takes to design even a simple
autopilot. If you've gotten that far, then next step is to either take a
flight controls class (grad level), or maybe start building some R/C models
if you don't want to go for more school (I'd suggest doing this anyway). The
school of hard knocks is fine with R/C, since the knocks aren't really all
that hard on you.

PLease, get yourself more experience with aircraft control before putting
yourself at risk. I'm guessing you already have a start in learning this
stuff, and don't think you should abandon your goals. However, take baby
steps.

Pete


Agreed.

Dan, U.S. Air Force, retired

Ernest Christley wrote:
"Dan, did you ever get a chance to work with fuzzy logic?"

Pete Schaefer responded:
"I have. It's been a while. I'd never, ever use it on an airplane.
Maybe a washing machine controller or something like that. I'm not
saying that it can't be done, but just that it's a risky design
approach."


Why is fuzzy logic "risky"?



Pete Schaefer continues:
"There are other methodologies much better suited for aircraft."


Why are other methodologies "much better suited for aircraft"?



-Will Dwinnell
http://will.dwinnell.com