Mind you, I'm not prosthelytizing, but..

[Larry Cunningham]

..Kim mentioned the idea of using a common RC servo to control flight surfaces on a CL stunt ship. I know the subject has appeared before, but I want to say what an incredible good idea I think it would be.

Yes, there may be questions about lag and response sensitivity of an RC servo, even a heavy duty one, as compared to a conventional CL bellcrank driven one. I can't help but think these are not real problems.

Now, let me describe Stage I of the general idea here. That is, instead of a direct mechanical driven system wherein the control lines connect to a bellcrank, which deflects a control rod which deflects flaps and elevators.. I would like to see an equivalent "fly by wire" system in place. That is, there is a transducer for the bellcrank (which might be anything from a potentiometer to a digital shaft encoder) which is continously read by a controller, whose job it is to simply "look up" output drive values for flap and elevator positions, and generate appropriate servo control signals.

RC servos use a very simple drive system, which consists of a 5V pulse width modulated signal. As the width of the drive pulse increases, the servo drives a larger mechanical deflection. It's simple, which is why an "RC Servo Tester" can be fashioned with a '555 timer, a couple of resistors, capacitors and a potentiometer.

The cool idea here is the programmability of such a system. Obviously since the controller reads the position of the bellcrank and uses that value to index a "look up" table, we can easily fool with values in the look up table to "map" virtually any function we want. Want to run flap deflections at half of elevators? Put that data into the table. Want to see how your model would fly "elevators only"? Table data can be changed for that. Exponential response? Ditto. Inverse exponential response? Same-o. Model turns easier outside or inside? Easily corrected.. Etc.

Now, the idea of Stage I for me is a trimming capability. If you're truly serious about getting things right for a particular model, you need to develop a LUT (Look Up Table) specific to it. So, do Brett/Ted/Rush all prefer something slightly different in terms of trim? Add a feature to select 1 of N LUTs, with a knob having Brett, Ted, and Rush positions on it.

Better yet, suppose you are such a fanatical trimmer that you even make changes depending on altitude, density altitude, temperature, etc., well, "we have an app for that". (Doesn't that make you want to slap someone when you hear "we have an app for that"?) ;->

OK, which brings me to Stage II. Anything worth doing is worth overdoing, and I generally prefer extending a design concept to its limits, at least the ones I can think of, then paring it down to something reasonable to implement. "Brrup! Up the wall!" and then "Blap! Back Down Again.."

So, next thing you know I have a microprocessor instead of some simple state machine controller running such a system. And I want some more inputs, more sensors. If it's electric I'd consider looking at battery voltage, motor current, and temperature. And I'd install a tach to measure RPM for an IC power source. I'd also consider measuring forces in all three axes, and/or simply looking at changes in them to detect maneuvering, using a cheapie heli gyro.

Now, since I'd have all this data, not only would I use it to influence the servos such a system would be controlling, I'd also log it, say 10 times
or maybe even 100 times a second. And I'd store this data in an NVRAM, which my controller would happily dump to a laptop on landing.

Then I'd have all sorts of need "telemetry" data, and I could actually SHOW what was going on during the pattern. (Yes, I acknowledge the published posts wherein folks have been logging and displaying such data from modular commercial telemetry pieces. It's great, and I applaud it.)

OK, there's Stage I and Stage II as I see them, I know several of the clever types here are already playing seriously with PIC controllers and such. It's so clever already it is almost cheatin'..

Stage III? Well, since you ask. This is the part where we take the pilot completely out of the loop and have the PIC fly it for us! He/she stands there stupidly holding a dummy handle, and waits for the engine or motor to quit. Yes, I knew that would raise the hair on some backs! That is precisely what we seek! Complete programming of CL Stunt. Hahahahhhhaaah oh hahahahhh (Dr. Evil laugh).

OK, blast I away. I enjoy pain - both giving and receiving it.

L.

"What you don't know will always hurt you." -First Law of Blissful Ignorance

 

[Tim Wescott]

Larry, the PRM can be much smaller these days, and doesn't need the FPGA at all.  You could probably get the weight below 1oz -- including a six axis IMU -- on rev 2 or 3.

My vote is for whatever electronics gizmos you want, as long as there's still a direct and unmodified mechanical connection between the control lines and any surface that controls vertical motion.

[Peter Nevai]

Larry, get outta my head! I was thinking the exact thing. With todays electronics you could program the entire sequence and just initiate the pattern sequence when the plane is relative to the wind where you want it. Heck my cell phone can tell precisely at what angle it is being held and has an internal accelerometer. Stick that in a plane that  only has one axis of travel and the rest is just numbers.

[Steve Helmick]

Just because it's possible, doesn't mean you should do it.  The best part (selling point) with CL, is that you really get to feel the plane.  Seems like this would reduce that tactile advantage. Call me a stick in the mud, but having electrons controlling the plane seems counter-productive to me.    Steve

[Tim Wescott]

Just because it's possible, doesn't mean you should do it.  The best part (selling point) with CL, is that you really get to feel the plane.  Seems like this would reduce that tactile advantage. Call me a stick in the mud, but having electrons controlling the plane seems counter-productive to me.    Steve

Ah, but for that you just need to hold on to the restraint handle (it wouldn't be a control handle any more, of course).  And it'd only need one line, and that could be .008" for even the biggest plane (because, of course, you could have the thing sense line tension and head in).

[Tim Wescott]

Larry, get outta my head! I was thinking the exact thing. With todays electronics you could program the entire sequence and just initiate the pattern sequence when the plane is relative to the wind where you want it. Heck my cell phone can tell precisely at what angle it is being held and has an internal accelerometer. Stick that in a plane that  only has one axis of travel and the rest is just numbers.

It'd need to be more than one accelerometer -- the centripetal acceleration makes the calculations more complex.  But newer cell phones come with 6-axis inertial measurement units, which, given a bit of work can cough up all the information you need.  Add in GPS, and it'll fly the pattern autonomously.

But -- why?

[Peter Nevai]

We are already moving down this path, with the pre-programming of motor run characteristics of electric power plants. Adjustability and control not afforded to IC engine implementations. The ability to dynamically control motor output to meet power demands at any stage.

[Wayne Collier]

In the back of my head for quite some time is the notion that if I ever become unable to (or just tired of) controling the plane from the center of the circle, I could install RC equipment, replace the control lines with a tether line, and control the plane from a lawn chair at the side of the circle.  Now you're saying I could program it to not need control input at all.  I could send it to the field with a friend and let it do it's thing while I sat home and watched a movie. Or I guess I could just watch the movie and save playing with the plane at all?

[Tim Wescott]

We are already moving down this path, with the pre-programming of motor run characteristics of electric power plants. Adjustability and control not afforded to IC engine implementations. The ability to dynamically control motor output to meet power demands at any stage.

That could be done with an IC motor, too -- in fact, it's exactly what a lot of the 'stunt tune' motors seem to be trying to do.  There's no reason you couldn't combine a Hubin timer with a helicopter engine governor to get exactly the same effect -- or just build a custom gizmo that does it all.  Servo, battery, timer, governor -- poof, you have control over your engine RPM.  Who knows?  It may even work as well as the current crop of piped stunt engine, or an FP-20 with the BBTU.

[john e. holliday]

Hey if this thing can fly the perfect pattern time after time with the winds we get here in KANSAS, we can do away with judges.  No more score sheets or tabulators.   Everybody gets a trophy for getting it to work.   

[Fred Cronenwett]

Look at this post in the CL scale section, kinda talking about the same topic

http://stunthanger.com/smf/index.php?topic=19525.0

Land Softly,
Fred Cronenwett

[Tim Wescott]

Hey if this thing can fly the perfect pattern time after time with the winds we get here in KANSAS, we can do away with judges.  No more score sheets or tabulators.   Everybody gets a trophy for getting it to work.   

Just build a module that monitors the flight.  Judging to be done by download.

[Jim Kraft]

I think I will just stick with two lines, a bellcrank, and a 3/32" pushrod on my McCoy powered Ringmaster. I flew R/C Pattern for 25 years.

[Dennis Moritz]

Slippery slope. Never imagine past the mid 20th ce. 1950s. Period. That's control line.

[Larry Cunningham]

With all the effort expended designing, building, testing, coding, debugging, programming, analyzing data, etc. of such a system, would there be time left to practice the pattern? 

L.

"I'll give you a definite maybe." -Samuel Goldwyn

[Tim Wescott]

With all the effort expended designing, building, testing, coding, debugging, programming, analyzing data, etc. of such a system, would there be time left to practice the pattern?

There will be if I do it -- 'cause I'll only do it if I'm getting paid my usual for it, which means it'll be my day job.  On evenings and weekends I'll be practicing the pattern!

[proparc]

Sounds too close to F3A to me where, damn near everything is either mixed out or programmed in.

[Douglas Ames]

In the back of my head for quite some time is the notion that if I ever become unable to (or just tired of) controling the plane from the center of the circle, I could install RC equipment, replace the control lines with a tether line, and control the plane from a lawn chair at the side of the circle.  Now you're saying I could program it to not need control input at all.  I could send it to the field with a friend and let it do it's thing while I sat home and watched a movie. Or I guess I could just watch the movie and save playing with the plane at all?

That's after you paid someone to build and trim it for you...

[Larry Cunningham]

As nearly everyone is aware, flying a full sized helicopter is quite a difficult skill for almost everyone. (I'm not sure the RC helicopters are much easier, but surely they must be). Suppose you were free to specify a control system for a helicopter which would generally be intuitive, perhaps using a single joystick, throttle, perhaps an altitude control wheel. In other words, something that most anyone could operate with little more training than what would be required for say, a motorcycle.

Would you do it? Perhaps if you could press a single button and toggle between that control and conventional controls?

This is the sort of thing that is quite feasible. One example which comes to mind is NASA's Shuttle Training Aircraft, a Gulfstream GII which is programmed to handle very much like the Shuttle, at the touch of a button. Same feel, same behavior, good for checking out landing conditions. I saw an STA flying when the Shuttle landed at White Sands (Northrup Strip) way back in 1982, on Columbia's third orbital mission. They were practicing the cross-wind landings, and we in the "press corps" were listening intently to their intercom.

While I see the value of such emulations, I'm certainly not advocating it for CL Stunt. My interest in the type of programmable system I've been talking about is as a trimming (and data acquisition) tool. I don't equate being able to set up flap and elevator deflection maps to cheating, it's gathering and using information, in a scientific manner.

I know well that there are plenty of modelers capable of designing mechanical control components to get about any sort of control deflections desired. Adjustable radius elevator control horns and control rod lengths are common, if cruder examples. I have no doubt that a mechanical system using contour cams could be devised, and likely made with interchangeable cams for different maps.

A programmable electronic control system driving servos is simply another method of accomplishing such a task. Its relatively easy programmability (vs say a cam-driven mechanical system) makes it more useful in my opinion for the work of experimentally trimming and tuning. But it's an equivalent system.

As for taking the CL pilot out of the loop, I never seriously would consider such a notion. My photo of the "Pilot Replacement Module" above is tongue-firmly-in-cheek. In reality, that is a module I designed when I was working on an Army contract at White Sands, a ruggedized programmable controller for a radar jammer. I shudder to think what it eventually cost in the actual "flight" version, as I recall, the military spec single Actel programmable gate array device pictured was something over $2500 - per CHIP!!

An inside joke is that a single chip digital controller like the one pictured is incredibly crude in terms of capabilities, it's not even a weak microprocessor.. And if you provided me with the very finest, most powerful microprocessor with all sorts of resources, including some fantastic software development environment (and requisition a WIZARD to write the code while we are at it), the sort of autonomous control, a true PRM, would not be achievable. I don't care if you have some "expert" system that learns interactively and self-adjusts, using neural nets, fuzzy logic, and whatever combinations of hardware, software, firmware, fuzzyware, wareware (!!) you might devise. It's not workable. It's a great subject for science fiction, but it's not going to happen.

(If you doubt my proclamations, I invite you to check into the military's contests to try to develop an autonomous land vehicle!) This is a very different problem than say, having a cruise missile use GPS and drive itself into a target..

So, quit worrying. Unless you are seriously worried that some or the other geek(s) create this stunter "mule" device and proceed to trim and learn to fly it so well that it makes competition nearly impossible. That's not going to happen.

But what could happen is that the geeks will gather some valuable information which may lead to insight into a practical improvement (likely a minor sort) which could end up applied to the very tip-top competition stunt machines. Then, after years of use and clear evidence that it is a worthy item, it might propagate to general use.

Don't worry, let us geeks have our fun. Much of this stuff is only thought experiments anyway. 

L.

"People seldom see the halting and painful steps by which the most insignificant success is achieved." -Anne Sullivan

[Tim Wescott]

As nearly everyone is aware, flying a full sized helicopter is quite a difficult skill for almost everyone. (I'm not sure the RC helicopters are much easier, but surely they must be).

RC helicopters these days have electronic stability augmentation -- that's what folks mean when they talk about their gyros.

The little electric toy RC helicopters are nearly full-autopilot devices -- you can practically park one in the air, with the exception of the vertical axis.  Many of them suffer from too-tight tuning -- you can watch them gently oscillate in the air, from a combination of friction and excessive loop gain.

Suppose you were free to specify a control system for a helicopter which would generally be intuitive, perhaps using a single joystick, throttle, perhaps an altitude control wheel. In other words, something that most anyone could operate with little more training than what would be required for say, a motorcycle.

Would you do it? Perhaps if you could press a single button and toggle between that control and conventional controls?

Personally -- no.  But I'm a luddite.

This is the sort of thing that is quite feasible. One example which comes to mind is NASA's Shuttle Training Aircraft, a Gulfstream GII which is programmed to handle very much like the Shuttle, at the touch of a button. Same feel, same behavior, good for checking out landing conditions. I saw an STA flying when the Shuttle landed at White Sands (Northrup Strip) way back in 1982, on Columbia's third orbital mission. They were practicing the cross-wind landings, and we in the "press corps" were listening intently to their intercom.

While I see the value of such emulations, I'm certainly not advocating it for CL Stunt. My interest in the type of programmable system I've been talking about is as a trimming (and data acquisition) tool. I don't equate being able to set up flap and elevator deflection maps to cheating, it's gathering and using information, in a scientific manner.

If you're talking about something that could be used for development, not competition, then sure -- if you're flying sport, do what you want.  If you're talking competition -- I don't like the idea.  Even if you couldn't build a PRM (and I think one could) you could do some significant stability augmentation.  It'd certainly be an event changer.

As for taking the CL pilot out of the loop, I never seriously would consider such a notion. My photo of the "Pilot Replacement Module" above is tongue-firmly-in-cheek. In reality, that is a module I designed when I was working on an Army contract at White Sands, a ruggedized programmable controller for a radar jammer. I shudder to think what it eventually cost in the actual "flight" version, as I recall, the military spec single Actel programmable gate array device pictured was something over $2500 - per CHIP!!

An inside joke is that a single chip digital controller like the one pictured is incredibly crude in terms of capabilities, it's not even a weak microprocessor.. And if you provided me with the very finest, most powerful microprocessor with all sorts of resources, including some fantastic software development environment (and requisition a WIZARD to write the code while we are at it), the sort of autonomous control, a true PRM, would not be achievable. I don't care if you have some "expert" system that learns interactively and self-adjusts, using neural nets, fuzzy logic, and whatever combinations of hardware, software, firmware, fuzzyware, wareware (!!) you might devise. It's not workable. It's a great subject for science fiction, but it's not going to happen.

(If you doubt my proclamations, I invite you to check into the military's contests to try to develop an autonomous land vehicle!) This is a very different problem than say, having a cruise missile use GPS and drive itself into a target..

So, quit worrying. Unless you are seriously worried that some or the other geek(s) create this stunter "mule" device and proceed to trim and learn to fly it so well that it makes competition nearly impossible. That's not going to happen.

I think it's feasible, although impractical (and ultimately not much fun) for a hobbyist machine.  It would take a man year or two (or five) of effort, but it's orders of magnitudes less than a DARPA-style autonomous land vehicle.  The amount of machine intelligence required for an autonomous land vehicle is huge compared to what folks have achieved with computers, and mostly has to do with figuring out navigable paths in terrain.  Flying a tethered airplane, in a controlled environment, over flat pavement, doesn't require any real intelligence as long as you're doing it from inside the airplane.  I really do think you could do it with a six axis IMU and an embedded computer that didn't exceed four ounces or so.  I've got code for a Kalman filter that I think could easily be adapted to tell the app just exactly where in the circle the airplane was, and what the windspeed was, too boot.  After that it's just a matter of controlling the airplane, and thats easy if you know your automatic control theory.  About the only thing that would be really difficult would be dealing with the line going slack -- and I think you could ignore that problem initially, and overcome it later.

[Howard Rush]

Slippery slope. Never imagine past the mid 20th ce. 1950s. Period. That's control line.

I agree, and I'm probably the only actual autopilot engineer among you. 

[John Witt]

Suppose you did just a couple of things to the model.  A pair of buttons, giving a perfect full rate (for the model) 90 degree turn, one for inside and one for outside.  RC guys have had a snap roll switch on the transmitter for years doing a similar job. It would be an easy extrapolation of that to program a horizontal square eight with the proper turn angles for all the corners.

How about a level flight routine, and perhaps a pull out at 5.5 feet routine that used an utrasound sensor to detect the altitude.

All depends on the time and money you want to invest.  F1 racing went through these spasms some years ago and a good bit of it has been dropped.  I personally think they were headed for autopiloted cars, and the star value of the drivers for the promotion of the sport was going to disappear. It was also horrendously expensive.

John

[Larry Cunningham]

Don't equate this sort of control to "autopilot" - it's routine for modern commercial aircraft to have fairly sophisticated functions, even to land them. And the classic autopilot isn't that sophisticated, it amounts to a bit of self trimming to maintain a course and altitude. Nor is a Lexus capable of parking itself that great an achievement.

I think what is being imagined by some as some sort of truly autonomous control capability, say to fly a pattern isn't viable on several levels, particularly when economics are considered.

I'm only pushing for a device that allows me to adjust and play with trim features and acquire some data from a CL stunter. Is that a slippery slope? 

Tim, I'm fascinated and very interested in your comments about using a 6-axis IMU and Kalman filter to "compute" flight position and infer wind speed. Strictly from a technical interest viewpoint. I'm convinced such a development would squeeze out some interesting information about our models and optimal trim for their intended purpose. The common example of a Kalman filter in my world was the phase locked loop, and a particular familiar application was a bit synchronizer for a noisy, encoded telemetry stream. You'd be surprised the fun there is to be had with some digital processing using specialized correlator functions on high speed serial streams.

As for F1 cars, I see quite a bit of real time telemetry data from them, even visible to the Speed Channel. There seem to be a number of engineers and flatscreens watching every aspect. Without it, I doubt these cars could be as competitive as they are. Again, data acquisition. I do agree that it's not desirable to close tuning/trimming loops interactively, particularly with anyone but the driver. And perhaps the driver oughtn't to be able to play with things during the race.

Nowadays, the cost of telemetry at a very valuable level needn't be horrendously expensive, the technology is so well defined, sensors are relatively cheap, special electronics don't have to be developed much, only repackaged. Things are modular; virtually all the "intelligence" is external, with the telemetry stream sync, decommutation, and output all straightforward. Although the environment is harsh, it is nothing compared to flight let alone space conditions. The required RF links are also trivial for such applications.

By comparison, our environment on a CL stunter is quite benign and friendly, and I think some pretty interesting stuff is feasible, whether or not ultimately desirable to most people.

The notion that somehow such geeky endeavors somehow cheapen and tarnish our beloved CL stunt seems silly and unrealistic to me. It's not that much different from what a guy named Rabe was doing playing with airfoil sections on the fender of his family car in the 1950s. That activity and inquisitive attitude yielded a lot of valuable data and several innovations which have done nothing to diminish CL stunt in my view. Quite the opposite, actually.

By now you I hope the gist of my interest in this sort of thing is evident.

L.

"I have the heart of a little boy:  in a jar of formaldehyde on my desk." -Robert Bloch

[Tim Wescott]

Tim, I'm fascinated and very interested in your comments about using a 6-axis IMU and Kalman filter to "compute" flight position and infer wind speed. Strictly from a technical interest viewpoint. I'm convinced such a development would squeeze out some interesting information about our models and optimal trim for their intended purpose.

I wonder what's the lightest Android phone with such an IMU (they're out there) and GPS?  You could shove one into the fuselage of a plane at the CG and start collecting data right now.  I've got Kalman filter code that would tell you where the plane is at each update of the IMU -- there's a slim chance that it'd even work right with the data as collected by the phone.

[Howard Rush]

Don't equate this sort of control to "autopilot" - it's routine for modern commercial aircraft to have fairly sophisticated functions, even to land them. And the classic autopilot isn't that sophisticated, it amounts to a bit of self trimming to maintain a course and altitude.

It's routine and quite sophisticated. Guys like Brett and Igor, who have studied airplane flight dynamics and automatic control will have a leg up on other folks-- at least until you can buy an off-the-shelf autonomous stunter. 

[Bob Furr]

No doubt it can be done... but as for me I love control line because I am in my heart cheap (even have a combat/sport model design named Miss Parsomony that has about $10 in materials in it before you bolt on an engine) and the whole idea of a model that flys itself just doesnt make sense to me except for FF rubber....
Bob Furr

[Dennis Moritz]

"I'm only pushing for a device that allows me to adjust and play with trim features and acquire some data from a CL stunter."
A hand on the handle.
 

[Tim Wescott]

My vote is for whatever electronics gizmos you want, as long as there's still a direct and unmodified mechanical connection between the control lines and any surface that controls vertical motion.

Dang.  I'm finding myself coming around, sort of.

While my knee-jerk reaction is for the connection between surface and pilot to be entirely mechanical, I'm thinking that maybe the Cunningham System should be tolerated, at least.  Or at least if there's no hanky-panky going on -- just an essentially memoryless following of some (possibly nonlinear) relationship between bellcrank position and servo output.

[Larry Cunningham]

Dang.  I'm finding myself coming around, sort of.

While my knee-jerk reaction is for the connection between surface and pilot to be entirely mechanical, I'm thinking that maybe the Cunningham System should be tolerated, at least.  Or at least if there's no hanky-panky going on -- just an essentially memoryless following of some (possibly nonlinear) relationship between bellcrank position and servo output.

That's mighty big of you - tolerating me and all!  

How would one tell if "hanky panky" was going on? Short of me suddenly showing up and winning the Nats, who would know? If I were using some device in a completely innocent manner, wouldn't that make my whole setup suspect, regardless of how well I competed?  

While I was being accused of whatever wrongdoing I'd ask to point out precisely where in the rules what I was doing was prohibited. Nothing requires me to have a direct mechanical linkage between flight surfaces and the lines..

If I were crafty enough to manage something that was truly revolutionary, were I competition minded, I'd likely sit on it, keep it secret, steadfastly insisting that no one gets to examine my control system unless we can cut their ship open and examine theirs as well.  

I know, I know. You're joking... And so am I. But even if regs were changed to clarify and specifically outlaw such systems in competition (which would NEVER happen, we can't even define modern BOM parameters), I could learn what I seek in non-competition.

As for thorough understandings of flight dynamics, etc that's all well and fine. I continue to believe advantages folks actually have in competition is largely a function of their physical prowess and practice (and their quality equipment, of course). I'll happily "tolerate" whatever advantages anyone can wield through their awesome raw technical knowledge!

L.

"Inventing is a combination of brains and materials. The more brains you use, the less material you need." -Charles Kettering

[john e. holliday]

I thought you had the advantage with your exponential bellcrank?   

[Peter Nevai]

23 grams 20 X 40 MM package

3DM-GX3® -35
Miniature Attitude Heading Reference System (AHRS) with GPS
 - Product Overview

Overview
The 3DM-GX3® -35 high-performance, miniature AHRS (attitude heading reference system) with GPS, combines MEMS sensor technology and a highly sensitive embedded GPS receiver. It incorporates a triaxial accelerometer, triaxial gyro, triaxial magnetometer, temperature sensors, and a dedicated 32 bit processor running a sophisticated fusion algorithm to provide orientation, inertial, and GPS measurements. Data from the GPS receiver is time synchronized with the inertial sensors and all inertial and GPS data are available as custom user packets (either by polling or continuous stream).

The 3DM-GX3® -35 is a member of the 3DM-GX3® family of inertial sensors. The system offers a range of output data quantities, including fully calibrated inertial measurements: acceleration, angular rate, and magnetic field; or deltaTheta & deltaVelocity vectors. It can also output computed orientation estimates: Euler angles (pitch, roll, and yaw (heading)); orientation matrix; or quaternion. GPS data quantities include LLH position, NED velocity, ECEF position and velocity, DOP data, UTC time, GPS time, clock info, GPS fix, and SVI. The flexibility of the system is made possible by the powerful new MicroStrain Inertial Packet Protocol (MIP).

The 3DM-GX3® -35 has a dual communication interface, which supports USB and RS-232. Starter kits include choice of USB or RS-232 interface cable

[Phil Coopy]

I like the technology approach as a great project, as an interesting project.  But for competition purposes, the game is all about learning how to eliminate mistakes while manually controlling the aircraft during the sequence of maneuvers.  Using electronics to do that job would change the competition to a competition in programming rather than piloting an aircraft.  If the tech approach became popular I would expect that it would have to be religated to a separate class of competition, more in the robotics area.

Phil

[Peter Nevai]

Correcting mistakes manually is what makes this type of technology so alluring. Imagine that a system as described above collects data from several top stunt pilots best flights. The data is averaged and then compared to real time data coming from the airplane. The system can then notify you if you are in the sweet spot and when you deviate from the parameters of the optimum path.

[Tim Wescott]

Correcting mistakes manually is what makes this type of technology so alluring. Imagine that a system as described above collects data from several top stunt pilots best flights. The data is averaged and then compared to real time data coming from the airplane. The system can then notify you if you are in the sweet spot and when you deviate from the parameters of the optimum path.

I was thinking that just a system that lets you download your exact flight path, for self coaching.  Do you bobble on the exits to the squares?  It'll be there.  Are your loops egg-shaped?  Your wingovers not directly overhead?  Are you hitting 45 degrees?  It'll all be there.

[Larry Cunningham]

I was thinking that just a system that lets you download your exact flight path, for self coaching.  Do you bobble on the exits to the squares?  It'll be there.  Are your loops egg-shaped?  Your wingovers not directly overhead?  Are you hitting 45 degrees?  It'll all be there.

When pigs fly.. to the moon.. JMHO. 

L.

Watson's Law: "The reliability of machinery is inversely proportional to the number
and significance of any persons watching it."

[Peter Nevai]

Can you say Oink

[Larry Cunningham]

Peter, that is a pretty impressive demo all right. However, that's a long way from what was mentioned: if such a system ever does exist, we can dispense with human judges entirely. I recall Bob Hunt mentioning such ideas several years ago, and I poo pooed the notion at that time as well.

It is possible that my 37 years of designing and debugging hardware/software control and data acquisition systems has permanently damaged my "optimism" gene.. I still cringe when I see a unix programmer "make" a piece of software and have it magically work. (I remain suspicious - I know there are serious bugs dwelling somewhere within. ) ;->

I remind you that nothing is impossible to the fellow who doesn't have to actually do it.

Quick! Someone prove me wrong. I was dead wrong about the viability of electric power for CL stunters (I admitted it, even decided it was coming as a superior power technology..) I also fervently insisted that we'd never see a miniature hard drive with over 100 MB capacity, and was dead wrong on that by multiple orders of magnitude. Also, in 1957 I broke a hacksaw blade, and two years later I thought I had made a (different) mistake, but it later turned out I was mistaken. So, I *can* be wrong. ;->

I do applaud the optimism and open thinking I see people demonstrating here. Grizzled old hardware design engineer farts poo poo everything.

L.

PS - BTW, did anyone else look at that flying pig sculpture and think: control line model? ;->

"For a successful technology, reality must take precedence over public relations, for Nature cannot be fooled." -Richard P. Feynman

[Peter Nevai]

Hi Lar,
The newer version of that unit also has built in gps, so yo can now corrolate spacial location and altitude with the axis data, synced to the gps sat clock. Considering that such technology has been on use for some time now in military drones the tech has been vetted. But such excersises always are born out of need. And for modeling there is really no real need. In the end it just becomes a interesting mental excersise, or perhaps a college undergraduate experiment for a mandatory paper.

We are looking into the device for simplifing the mechanics of a auto tracking ku band satellite dish antenna. Will replace a half dozen sensors and other components. Can get all the necessary data to control the dish antenna from a single tiny module.

[Tim Wescott]

Peter, that is a pretty impressive demo all right. However, that's a long way from what was mentioned: if such a system ever does exist, we can dispense with human judges entirely. I recall Bob Hunt mentioning such ideas several years ago, and I poo pooed the notion at that time as well.

It is possible that my 37 years of designing and debugging hardware/software control and data acquisition systems has permanently damaged my "optimism" gene.. I still cringe when I see a unix programmer "make" a piece of software and have it magically work. (I remain suspicious - I know there are serious bugs dwelling somewhere within. ) ;->

I remind you that nothing is impossible to the fellow who doesn't have to actually do it.

Quick! Someone prove me wrong. I was dead wrong about the viability of electric power for CL stunters (I admitted it, even decided it was coming as a superior power technology..) I also fervently insisted that we'd never see a miniature hard drive with over 100 MB capacity, and was dead wrong on that by multiple orders of magnitude. Also, in 1957 I broke a hacksaw blade, and two years later I thought I had made a (different) mistake, but it later turned out I was mistaken. So, I *can* be wrong. ;->

I do applaud the optimism and open thinking I see people demonstrating here. Grizzled old hardware design engineer farts poo poo everything.

L.

I'm currently doing a Kalman filter that takes the the output of a GPS and an IMU, and merges them into a vehicle trace that's accurate down into the tens of centimeters absolute, and probably centimeters relative.  The big barrier to applications like that is that to really work well they need the vehicle to be experiencing ever-changing acceleration -- and the rotating acceleration vector of a CL plane in level flight is close enough.

So there's no question in my mind that you could put something together that would give you a trace of where the plane was every 10 milliseconds.  Whether it would be an effective coaching tool is another question, but I think that if you could present the data in an accessible (read: graphical) format, I think it would 'fly'.  Combine that with readouts on bellcrank position and leadout tension, and I think you'd have one heck of a coaching tool.

PS - BTW, did anyone else look at that flying pig sculpture and think: control line model? ;->

I'm gonna build the flying sidecar from Wallace and Gromit's A Close Shave first.

[Larry Cunningham]

Tim, you lucky dog.

It makes me wish I wasn't retired sometimes. I can see where a sort of "hybrid" system using both GPS and IMU inputs can likely work. As for the software to present the data from such a system, I believe that could be quite straightforward (not having to actually DO the software myself!).   The hard part is the HARDware (software is easy; hardware is HARD - quote from Harry Pyle, co-inventor of the 8008 architecture when he was 18 years old) to deliver that data..

It's a short trip to display such data and compare it with a template of the ideal maneuver shapes to generate some meaningful error information. Actually USING such information might be something fairly difficult for a flier. And I predict an extremely strong resistance to any such sort of "machine" non-human "judge" for CL Stunt.. EVER. (Witness the endless controversy over the BOM, rules clarifications and updates in general.)

It's a long trip to doing anything interactive with such data to autonomously "correct" a flight path. And face it, if such a system existed and got applied, you could hardly refer to that activity as "control line" any more.

What seems evident to me here is the possibility that "error" information, while interesting technically, may be of little practical use. That wouldn't prevent me and several others from being interested in it. If nothing else, I'd like to have some meaningful line tension data, and G data for stunters during various maneuvers. It may be pursuit of information for information's sake, which some people would see as folly. That's fine, there's a lot of other stuff I see as folly.

Tim, keep us posted on your progress on this effort. I'd be interested in the details at the nuts and bolts level (in fact, when you get it working, how about just handing over your design to me? Hahahah). Keep at it.

L.

"Nature gives you the face you have at twenty. Life shapes the face you have at thirty. But at fifty you get the face you deserve." -Coco Chanel

[Peter Nevai]

Hi Tim,
Nice to find another involved in similar endeavors. Our effort is improving the performance of keeping a mobile (terestrial and marine) sat dish on the bird while moving. We are using the drop in signal strength as the indicator of target loss and using the module to report the direction and angle of drift so the software can make corrections to the heading and azimuth of the dish. The gps gives us the position of the antenna relative to the sat orbit so the software know what part of the sky to find the sat. Before now such systems had to be tied into the ships gyro compass and independent differential gps for this data.

[Tim Wescott]

Tim, keep us posted on your progress on this effort. I'd be interested in the details at the nuts and bolts level (in fact, when you get it working, how about just handing over your design to me? Hahahah). Keep at it.

What I haven't done is anything to either collect data on-board a plane, or to adapt my existing software to work with whatever comes out of whatever is on the plane.

If someone wants to figure out how to collect data, though, feel free to send me a PM -- given a flight's worth of data it'll be hard for me to say "no" to at least giving the filter a whirl on airplane data.

[Peter Nevai]

Another demo video, it is highly technical