Senior Telemaster

Airframe Todo List

(Sep 4, 2010 edit: I’ve moved way beyond this todo list, we are on our 2nd and 3rd gen autopilot hardware now. The software has been totally revamped. Still, kind of fun to look back to see where we were a few years ago. Never did clean the gummy residue off the load struts though …)

  • Setup a heading hold mode and implement waypoint following.
  • Setup auto-throttle control module.
  • Install secondary “tapped PPM signal” receiver.
  • Install MaxStream radio modem.
  • Low priority: clean gummy sticker residue off the aluminum load struts.

November 16, 2007

Today I logged about 50 minutes of fully autonomous flight. Since I fixed the MNAV firmware bug, I have not seen any servo glitching or problems. The (patched) system is starting to earn my confidence.

August 14, 2007

Today I did two test flights in some non-trivial wind.


On the first flight the gps gave me a couple sequential bogus readings so the aircraft thought it was -1000 MSL (i.e. below sea level.) This caused it to climb towards what it thought was target altitude, but in reality was way too high. So we shut down the motor and brought it back. I’ve never seen the gps glitch like this so I’m not sure what to make of it.

On the second flight we had good altitude readings, but I observed that I constructed my sequence of waypoints in a way that it was impossible to reach the 3rd waypoint from the 2nd waypoint given the existing wind conditions. I “cheated” and helped tighten that turn with the rudder and with that small amount of help, it was able to reliably negotiate the course.

July 28, 2007

Today I flew twice to collect airspeed data from the pitot tube. The data is noisier than I hoped, but seems to be pretty plausible and certainly usable if I filter/smooth it a bit.


July 27, 2007

Today I installed a pitot tube on the Telemaster and routed some flexible out the center section of the wing so I can plug it into the MNAV when I install the wing. I used a length of carbon fiber tube rather than metal. These pitot tubes take a lot of abuse, even when you know they are there and are trying to be careful. 🙂



July 13, 2007

Flight report: The goal of today’s flight was to excercise the altitude hold system after making a couple small tweaks to the gains. It was a beautiful morning, temps in the mid-60’s, not a cloud in the sky, wind out of the west at about 4-5 mph, nice smooth air.


Here is the plot of altitude vs. time. The red line is the target altitude of 426.72 meters MSL (or 1400′ MSL which is about 500′ AGL.) Altitude statistics for the flight: average altitude = 426.44m, min = 419.6m, max = 434.4m, standard deviation = 2.85m.


Here is a motor/battery plot from my eagle tree data logger:


Here is a movie of a synthetic replay of a portion of the flight:

June 23, 2007

I took the telemaster out and ran out the 8000 mAh battery over the course of two flights. I didn’t quite have a full charge on it to start with and ran it up at full throttle most of the time, so I only got about 22 minutes of motor on flight. It is a bit surreal to see such a large airplane fly so quietly, smoothly, gently, and slowly — but also very pretty to watch. The air was really smooth this evening and winds were light so the big telemaster was super stable and solid.

On the first flight the MNAV performed great. I saw one momentary servo glitch, but other than that, it tracked rock solid each time I activated the autopilot. I would fly down to one end of the field, turn around, line it up, and activate the autopilot. The autopilot would hold a pitch attitude and lock the wings level, so I could then steer with the rudder and fly a pass. At the other end of the field I’d flip back to manual control, wheel around, line her back up, and reactivate the AP again. I flew a dozen “long” passes like that and it worked well.

On the second flight, the MNAV servo output was in glitch mode the whole time. Each time I activated the AP, it drove the servos to full stop which puts the controls hard over. So I just flew it manually around the pattern, and checked periodically to see if the MNAV revived itself, which it never did. I should point out that I’m 95% sure the servo output glitch is something specific to this particular MNAV unit. It’s not a more general problem as far as I know. I should also point out that the sensors and flight computer runs fine during the time the servo output glitches out. Once I get the other MNAV back from Xbow, I will probably send this one in for service.

June 9, 2007

Today I flew two more flights with the MNAV involved. I activated the pitch angle hold and that seemed to work reasonably well to. I think the control outputs jitter more than the simulator and the plane reacts to turbulence more than the simulator, so the result is not a super smooth solid lock on the target values, but at the same time the bouncing seems really random like turbulence, not regular like an oscillating, not quite tuned, PID.

The flight control computer/mnav blipped out on me once and kept trying to put the controls hard over to full stop. I’m not sure what happened there, but it eventually recovered itself. I need to do some more ground testing to see if I can reproduce the problem.

June 8, 2007

First flight with the MNAV helping!

For the past weeks I have been battling one problem after another getting all the components together and working on this new airframe. The biggest problem has been interference and noise and a corresponding dangerous drop in range for manual control.

Ok, so tonight, I finally have everything together (I think — I hope! — it’s always those pesky oversights that can kill your airplane.) I pass my range check (not by a lot, but I pass), there is a 10 mph direct cross wind which I’m not especially fond of, but not too worried about. Skies are perfectly clear, temp is 70 degrees, I’m the only one at the field tonight, I need to be especially careful not to put my finger in the prop, I can’t dial 911 with out it! I power on the flight computer and verify it is up and running, I verify the borrowed IMU/GPS sensor is up and running, I verify I have gps lock, I double check all the controls are working and in the right direction, and finally I verify I can hand over control to the autopilot and yank it back reliably with a switch on my transmitter.

Tonight I only had the aileron channel routed through the flight computer, so when I hand off control, the AP is only going to do wing leveling, I still have manual control on all the other axes.

With everything in place, I took a deep breath and taxied out to the end of the runway and lined myself up. I took another deep breath and smoothly advanced the throttle. As the airspeed came up, the aircraft weathervaned strongly into the direct cross wind so I quickly shut down the motor and aborted the take off run. This was my first try at a cross wind take off in this aircraft so I didn’t have my thumb gains set right! 🙂 On the second try I pretty much held the runway center line, and after what seemed like an eternal take off run in the grass, the tail came up and I coaxed it off the ground. It climbed out more lethargically than I thought it should, but it was climbing … temps were warmer than when I’ve flown before and I had an extra battery pack on board … I’m not too worried just yet.

I circled the field for a couple minutes to gain some altitude and catch my breath. I was paying careful attention to whether or not I had solid manual control with no glitching. Flying this airplane manually is a no-brainer, so I’m not nervous about that. But this is the first time I’ve flown all the equipment plugged together and powered on. I’m very worried about the effect of noise and reduced range. If my receiver is overwhelmed with noise once the aircraft is at altitude, and I can’t control the airplane … it goes in hard and half my life for the past couple months goes home in a garbage bag (I did bring one just in case!) Failure at this critical point in the project could mean giving up all my UAV hopes and dreams. Success here is the gateway to all sorts of interesting future projects and fun. I’m usually not a highly dramatic type of guy, but for some reason, I felt a heavy weight that this was a very critical junction in my life. So I was *very* apprehensive about possible equipment failure or some oversight that would lead to the aircraft being destroyed.

But so far so good, I’m up and flying, control inputs feel solid, no glitches that I can see in my receiver. This was the main focus of my test flight, to validate the manual override board I have recently added to the whole system and to make sure I can fly the aircraft safely and reliably no matter what goofiness the flight computer might attempt.

So now I’m at altitude, tooling around, no more excuses … I flipped the autopilot switch to activate it and immediately my wings leveled out and stuck there. I wiggled the rudder back and forth to full stop in each direction. This aircraft has a lot of dihedral and strong roll coupling, so even though a good whack (sorry Bill) of rudder causes quite a bit of roll, and even though holding full rudder deflection puts a lot of pressure on the airframe to roll, the flight computer is able to bring the wings back to level again and hold them there.

This is the bonus part of the flight. I honestly didn’t care if the flight computer put the ailerons hard over and held them there, as long as I could recover manual control any time I wanted. But the nice bonus was that the roll angle estimation and wing leveling code worked great. The gains maybe weren’t exactly perfect, but they were ok, and did the job just fine.

I landed and then repeated everything again with a second flight. I figure anyone can hit the three point shot, but if you can do it twice in a row, maybe it wasn’t entirely dumb luck.

This UAV is running an almost exact copy of the FlightGear autopilot algorithm, it even has the same xml parser, property system, and almost the same autopilot config syntax.

The Rascal 110 dynamics model we have in flightgear is not perfect, but it’s not horrible either. I developed a bunch of flightgear autopilot modes to work with the Rascal 110 model, including a wing leveler, heading hold, pitch hold, altitude hold, speed hold, etc. etc. These autopilot configs were all developed entirely in FlightGear using a JSBSim dynamics model.

Now with this UAV I am developing, I can take the *exact* autopilot config file off the flight simulator (make a couple small syntax modifications) and the actual, real-life, real world uav autopilot can run the same algorithm on the real uav in real flight with the same gains. If the simulation model used to tune the gains is good enough, the real life UAV autopilot should work out of the box (maybe not optimally) but well enough to be stable and do the job.

In this case, the wing leveler config developed in simulation with JSBSim and FlightGear worked out of the box on the real UAV running the same algorithm. (The Rascal and the Telemaster aren’t exactly the same airframe so this isn’t a perfect conceived test case, but they are close enough …)

I hope to head out tomorrow and repeat everything with the pitch axis. So tomorrow night I could be easily be sobbing, but for tonight at least I am happy. 🙂

No pictures tonight … Ruth misplaced the digital camera …

May 28, 2007

Test flight with MNAV and GumStix collecting data

Today was really windy, but right down the pipe so I put in 3 flights on the telemaster. This weekend I mounted the Xbow MNAV connected to the GumStix flight computer. I can log about 60 hours of flight data on-board with the Gumstix’s 1Gb MMC card.

I can take that flight data and replay it in FlightGear to “visualize” the flight and also to get a sense of how well the sensors and flight computer estimate the location and attitude of the aircraft.

Here are some pictures of the initial installation of the equipment.




April 18, 2007

Maiden flight!

Temps were about 60F, clear sky, winds out of the East at 10-15 (runway is N/S.) The winds were very erratic, gusty, and changing directions. Not the nicest for flying but not insane either.






March 30, 2007

Today I tightened down the motor mounts, prop, and spinner. I plugged in the battery and fired up the motor. She seems like she generates plenty of thrust to fly. The speed controller seemed like it was getting a tiny bit warm in my short run tests, so just to be safe I moved it back out on the front of the firewall.

I installed the receiver and secured the antenna. I put a bit of effort into routing and securing all the wires so there wouldn’t be stuff flopping all over inside the cabin. With the power on, I setup all the servos to move the correct direction, and adjusted the linkages so all the surfaces start out centered. I think the current control throws look fine.

The aircraft balances pretty close, but is slightly tail heavy … I’m not sure it’s enough to worry about with this design. I think everything is just about ready for a maiden flight.

March 29, 2007

Installed the battery in the nose compartment and packed foam all around it so it can’t move or dislodge (I hope!) It’s a two pound battery. I moved the speed controller installation to just inside the front of the nose hatch.

I found a long tube to route the receiver wire through and stuck that down the tail. I’m not sure I’m happy with that, but it’s a start.

I’m to the point now where I need to finalize the receiver installation, balance, and I should be ready to fly.

March 28, 2007

Mounted the speed controller on the lower outside front of the firewall with double sided velcro tape. I found a 5/16″ washer laying around the lab for mounting the prop. Also, I ran the prop through a balancer and I think it is very close.

On the way home from work I stopped by the hobby shop and picked up a Y-harness for connecting the wing servos, a short servo extender for the throttle channel, and a packet of metal 4×40 x 1″ bolts for installing wing load struts.

March 26, 2007

Today I bent the load strut mounting tabs to the correct angles and glued the plugs into the aluminum tubes with the correct insertion distance. I “pinned” the struts for added structural security, i.e. drilled a hole through everything crossways and epoxied a wire pin in to make sure everything is solid.

I then drilled out the mounting holes and installed the blind nuts in the wings and fuselage.



Here are some various angles of everything assembled as far as I can right now:






March 23, 2007

Load struts: assembled plugs and straps that I will fit into the ends of the aluminum tubing. Sanded the wood plugs down to the inner shape of the tube for a nice snug fit … well, one is a little loose, but it will be ok.

“All” that is left is to glue the plugs into the tub ends, bend the tabs to be flush with the fuselage and wing, drill the mounting hole, feed the bolt through, and we are pretty much done.

The one small thing that I’m worried about is where I’m going to remove the wing covering in order to install the blind nut on the back side of the wood mounting block.



March 21, 2007

Today I fabricated and installed the aileron linkages. I also cut my wing load strut tubes to their proper size (aluminum tubing in a tear drop/airfoil shape.) I have brass strips to mount in each end which will bolt on to the fuselage and wing. I need to cut those and fit them into the ends of the aluminum tube.

March 19, 2007

Today I installed a light weight 1.75″ tail wheel with a collar. I also installed servo extension security clips between the aileron/elevator servos and their 24″ extension wires. This is a small added measure of security to make sure these never can pull, vibrate, or wiggle free.

This weekend I purchased short extension cables to facilitate plugging the wing servos into the receiver (so I don’t need to plug and unplug cables directly into the receiver every time I go to fly.)

Finally, I fabricated and installed a piece to re-enforce the wing leading edge mount.

Detail of tail wheel and rudder linkages …



Detail of elevator linkage …


Two shots of how the overall project is shaping up …



March 16, 2007

Progress made today:

  • Installed steerable tail wheel. I forgot a small step before I glued on the vertical stab. This forced me to go to the hobby shop to buy a slightly different tail wheel mount. The 2″ tail wheel is slightly too big, so I will probably need to get a 1.5″ wheel and a 3/32″ wheel collar.
  • Installed an elevator joiner wire. I bought something pre-bent from SIG and removed the nylon control horn. I cut a groove into both elevator halves and epoxied it in. I also notched the rudder in one spot for clearance. Note: this is another one that would have been good to do before gluing on the elevators.
  • Installed the elevator and rudder servos with their extension wires.
  • Installed the elevator and rudder control horns and linkages.

This is the elevator joiner I used …

IMG_4319 Low res cell phone shots …







March 15, 2007

Epoxied the tail surfaces onto the fuselage. Installed the main gear.

March 9, 2007

Pictures with the tail surfaces dry fitted …

After these pictures were taken, I ripped off most of the remaining yellow on the fuselage and replaced it with ferrari red.








March 6, 2007

Detail of tail servo mounting modification …




March 5, 2007

Detail of shipping damage (fixed) and re-covering plan …





UAV Research at the University of Minnesota

August 3, 2007

Today we took a small field trip to do a couple systems tests.

Here is the basic setup with Paw, Troy, and Brian:




Here is a closup of the Unicorn airframe with all our instrumentation:


Here are some pictures of the flying area:





Here are some pictures of the Unicorn in flight:




Senior Telemaster Early Flights

While rummaging around my hard drive I stumbled on some early footage of our Senior Telemaster flying. This footage includes some of my earliest successful autonomous flight using the original MNAV sensor head (now discontinued.) There isn’t anything especially great about this footage, other than I love the way the Telemaster looks in the air, especially it’s “scale” take offs and landings. These videos date back to sometime in mid-summer 2007.

The following movie includes some take offs, landings, and some early autonomous flight. One thing that is purposely highlighted in this video is an MNAV firmware bug that would cause the servos to go hard over full stop and send the UAV spiraling out of the sky. Fortunately we could quickly flip back into manual mode and manually pilot the aircraft to safety. Eventually I personally tracked down the firmware bug and fixed it. (Now we no longer use the MNAV in our UAV work.)

Here is a second movie showing a segment of successful autonomous flight.  The Telemaster is flying a bowtie — or figure 8 — pattern.  This is interesting to me because it offers a visual reference for the “quality” of our early MNAV based autopilot system that we can compare our current system against.

Sig Rascal 110 #1 Instrumentation

MicroGear and the Xbow MNAV

March 5, 2007

Paw finalized the RxMux assembly after I verified everything worked.

March 3, 2007

Today I ground tested my RxMux manual override board and everything appeared to work great.

February 20, 2007

Flight testing the MNAV, Stargate, and Maxstream again.





In our test last November, we had problems with our gps antenna connection. We made an attempt to fix that issue, but have been waiting through about 40 days of brutally cold weather here in MN to find a flyable day. Finally today the weather and our schedules matched up so we put in 2 very nice test flights.

The MNAV worked well. The secured gps antenna connection held up well though all RPM ranges, taxiing, take offs, landings, and some relatively agressive in flight maneuvers. Everything seemed solid. That is great news because it lets us move on to more interesting things.

In addition to just collecting the data, I fed it into a running copy of FlightGear in real time so you could visualize a synthetic view of the real world flight on the computer. This includes the world position, attitude, and flight surface deflections. It’s very cool to see the virtual view on the computer match exactly what the real world airplane is doing as it happens.

This synthetic view will be useful for a number of things down the road. For example, we can click anywhere in the synthetic view and get back the real lon/lat/elevation of the ground point we clicked on. Now imagine that we have overlaid live video on top of the synthetic view in a way that matches the two views "conformally".

The synthetic view also gives us the ability to highlight obstacles, restricted airspace, target routes, approach paths, other uav’s in the air, etc.

Takeoff …

Roll …

Fly by …

Loop and Landing …

November 22, 2006

First flight with the MNAV, Stargate, and Maxstream on board

First, no crashes to report!

Here are some observations.

1. We powered everything on and it all worked at the field (including gps.) Then we did a test flight with the MNAV and radio modem *off*. We landed and then powered everything back on to do engine on tests. From that point on … with engine running or not running we were unable to get any gps lock. We were pushed into a corner with battery life so we couldn’t sit around and wait 30 minutes to see if the gps could find itself. So we flew without gps and collected attitude, airspeed, pressure altitude, and transmitter stick inputs; the gps never came to life in flight. Dang, no gps data. 🙁 We need to investigate this further, we don’t think it was an antenna connection issue since we *very* carefully secured that. I’ve observed in the past that the MNAV gps can be a bit persnickety, and the MNAV hides all the nmea strings so we can’t see if the gps is seeing zero satellites (antenna problem?) or seeing a few satellites but is lost and trying to find itself.

2. I was really underwhelmed with battery life for our maxstream radio modems and the MNAV/stargate. We were running standard 9v batteries for the maxstream and a 3.7v lipoly (1320mah) battery for the stargate. We got maybe 20 minutes out of everything before the batteries died. I’d really like a bit longer life for a better safety margin when we fly with the system in the critical control path.

So the two main technical issues for us to iron out are (a) figure out what happened with the gps … probably just needed to let it sit for a while longer to find itself and (b) we need longer battery life for a bit more comfortable safety margins. A 10-15 minute flight with 20 minute total battery life is pushing things a lot tighter than I’m comfortable with.

3. We did collect attitude data. I don’t have a very good sense as to the quality of the attitude data, and without location/gps to go along with it, it doesn’t really make sense to replay it in flightgear as a visual sanity check … that was what I was hoping to do. We don’t have an absolute truth reference, but when you replay the flight you can visually catch a lot of abnormal things that wouldn’t show up in the plots.

4. The pitot tube dynamic pressure system seemed to work and yield plausible results. Top speed for the day with our Rascal 110 was about 90mph. Throttled back cruise was maybe 35-45 mph. Approach speed was about 35 (makes sense since I didn’t change the elevator trim between cruise and approach so speed shouldn’t change too much if I’m flying hands off.) Touch down looked like in the neighborhood of 25 mph. (Units converted from m/s to mph indicated, not knots.) Sense since, not knots … how do I stumble into this clumsy english?

5. static pressure/altitude data seemed plausible. The minimum/ground value for the day was 738, max altitidue was 1230, so we kept it under 500′ AGL. Actual field elevation is close to 900′ so these values weren’t calibrated for local weather conditions and that is expected.

6. We updated the demo code to log transmitter stick inputs (and onboard battery voltage) as well. Those seemed pretty good overall, but there are several places where the servo channel data glitched out. The system didn’t register this as a complete transmitter off situation, but the data would get really noisy: 1-3 values at the high end, 1-3 values at the low end, 1-3 correct values, repeat. There is one particular stretch of about 7-8 seconds where the data got *really* noisy and bad. We never flew very far out, so I really didn’t want to see this kind of noise in my data. Makes me very nervous, makes me think that I could easily fly out of range and not be able to recover control in time to save the airplane, if the MNAV is in the critical control path. I didn’t try to fly especially far away, but at the same time we didn’t have gps working so I can’t know what the cutoff range is. We might be able to filter the transmitter data to help things out since there are clearly some good values in the midst of the noise, but still, not what I wanted to see. We’ll need to think very hard about a safe way to proceed. Aside from the one ugly 7-8 second stretch, the other drop outs seemed much more momentary and probably wouldn’t hurt us too bad.

So that’s where we are at. We need to figure out why our gps refused to work. We need to get some bigger capacity batteries. And we need to solve our MNAV interference/range issues before we can proceed on to autonomous flight.

November 18, 2006

Pictures of the MNAV and StarGate mounting detail






MaxStream Xtend Radio Modem




GPS Antenna and Ground Plane



External Power Switches


Adding and Routing the data and power cables




Power and data for the MaxStream radio modem


Packaging everything together







Installation in the fuselage







Sig Rascal 110 #2 Results

Sig Rascal 110 #2 – UMN UAV Project

IMG_4073 Purchased August 24, 2005.

This is the second Rascal 110 purchased by this project. It comes with a Zenoah G26 gas 2-stroke ignition engine. The engine seems to be a very nice match for the aircraft. With an 18×8 prop, it will idle at about 1800-2000 rpm and tops out at about 7000 rpm. This isn’t quite the 10k rpm that our OS 1.60FX would turn (with the same prop) on our first Rascal, but it is more than enough power for this aircraft.

October 25, 2006.

Pictures from our morning at Jensen field. We were flight testing to collect data from our 3-gps carrier phase differential attitude determination system.

Warmups on the ground …



Fly by’s …





Landing approaches …






Some of the ground crew helping out …


Taxiing …



Closeups of the instrumentation …




Detail of one of the wing gps antennas mounted internally …


Overview of the airframe …



July 11, 2006.

Here is an article written about our project: sensor-0517.pdf.

Here are the raw (color) photos.




May 31, 2006.

Today was a beautiful day. Temps in the upper 70’s. Wind about 5mph out of the N and NE. Our mission for today was to collect good MIDG data. We have had problems with our aerocomm radio modem link so Greg rigged up a little gumstix unit to record the data directly on board. We lose our real time link for this, but we get solid data with no drop outs. We want solid data for another little adventure.

This was our first outing of the season. We put in two flights today and the Rascal flew beautifully. The engine ran great and was rock solid. I even spotted my landings on the runway, although with the extra weight, she rolled out a little long onto the grassy area which is no big deal. The Rascal handles mowed grass with no problems.

Here are a couple new pictures of the Rascal with Greg and I.




December 12, 2005.

Movie time.

Here’s a word of explanation. On 10/26/05 we did a test flight and captured the live video stream from the onboard camera as well as the data stream from the onboard IMU/GPS/INS unit. (See the entry for 10/26 below.) We replayed the data stream in FlightGear with overlayed instruments (this can also be done in real time as the data is captured) and saved that out as a movie. Then we edited the two streams together in two different ways: side by side, and blended overlay. The result is interesting. You can see at the start of the flight where the IMU was pretty far out of whack … as much as 5-20 degrees off in yaw. But as the flight progresses you can see the error diminish and later in the flight the match can be quite good.

Both videos are about 50Mb to download. You will need DivX6 to play them. Linux or open source people might find they can play these with xine or mplayer. Windows people might need to go download the DivX6 runtime codec from

Blended Overlay

Side by Side

November 30, 2005.

Minnesota! This here is what we live for! We put up with above freezing temps, oppressively humid summers, terrible mosquitos for a few months of the year, and high taxes, but days like today are the payoff: Standing outside in +20F temps (and often colder) freezing our butts off or playing hockey, or both. 🙂

Today we flew Rascal #2 at Jenson field. Temps were about +20F. Winds were some where between calm and very light. We probably had about 2″ of snow on the ground … right on the borderline for operating with wheels. The first takeoff run was a little “S” shaped and we had one nose over on a landing (zero damage) but other than that everything worked fine as I adapted to the conditions. We also tested our new video capture hardware.

Just look at these picts and see what you folks in sunny warm climates are missing out on. Oh yeah, and these pictures look dark for good reason. The sun barely cracks the horizon up here this time of year. And when the sun is up (or not up) it’s usually cloudy.





November 11, 2005.

img_3142 Here is a diagram of some of the Rascal 110 dimensions.

October 26, 2005.

Today we flew Rascal #2 at Jensen Field. Weather was perfect, winds were light and out of the east. Our primary goal was to test the live video and live synthetic view right next to each other to verify they track each other as they should since both reflect live data. We also practiced flying over specific objects and buildings to see if we could recognize things from the live video in flight.

Here are two pictures taken nearly at the same time. The first is from the onboard video camera, the second is the FlightGear based synthetic view:



October 17, 2005.

Paw and Greg fixed our radio modem / RC system interference problem so now we get excellent range checks. We setup down at “Jensen field” in Rosemount, MN.

Here is a picture of the field and the row of hangars:


Our ground station is basically setting up our equipment on a picnic table in the shade:



Here are some pictures of our aircraft posing:







Our wireless video camera refused to work so we flew without that. Our primary objective though was to flight test our “synthetic visual system.” This consists of a MIDG-II IMU/GPS/INS transmitting data to the ground via an aerocom radio modem. On the ground we have some custom software that reads the MIDG-II binary data and sends the results over to FlightGear for real time rendering. Ted built us a “photo-real” model of the area that was compatible with FlightGear.

We were disappointed with the data rate we were getting (maybe 5hz.) This prevented us from doing any serious flying under the hood. But I think we validated our approach and when we track down our data rate issues we will have a very powerful remote piloting tool.

I made a short movie of an approach and touch down. I replayed the data on my laptop so it is smoother than in real time because the replay code can easily interpolate between data points. I pointed my little digital camera at the laptop’s screen and here is the result, low quality, but you can see things in action. Notice the live working instruments at the bottom of the display. We are landing to the west and have a pretty stiff turbulent wind coming from the NW and rolling over the hangers. I’m not showing off my piloting skills here, but the touch down was nice ad gentle. 🙂

Click on the following “text” link to view the movie:


October 3, 2005.

All our instrumentation was moved over to Rascal #2 so today we flew a series of test flights to make sure the plane flew well with it’s payload. We couldn’t pass the range check test with our radio modem on so we scratched our heads for a while, but couldn’t come up with any clever ideas so we unplugged the power and grudgingly continued with no telemetry data.

We were very happy with the aircraft’s performance and handling with the load. We seemed to be getting solid video with our patch antenna. We also wanted to put more time on our buddy box/safety pilot system so we did one flight where we switched off control to each other.

Finally, happy with how everything was working, we walked over to the ground station and took turns flying “under the hood” with video only. Our camera points 45 degrees down which gives a good view of the ground, but not the horizon. We were able to hold a reasonably straight line via video alone, but turns were very difficult because you couldn’t see the horizon to get a good judge of bank angle and pitch angle. This is where the synthetic view would have been really nice … but no telemetry data.

Next up is to figure out our interference problems and get back to where we can fly with telemetry turned on.

September 26, 2005.

Friday and today we have been working on extracting our instrumentation from Rascal #1 and testing it to verify everything still works after our crash. We need to do a bit more testing, but we are reasonably confident at this point that everything survived intact (minus our radio modem which had it’s antenna sheared off.)

Today I continued to modify my MIDG-II replay software so I can read the live incoming data directly from the serial port and pass it to FlightGear. This gives us a live, real time, synthetic view that should match the real aircraft very closely. This should allow us to insert things like restricted airspace, important objects, mission goals, flight route, etc. into the synthetic view. We even have the capability of setting up a virtual ILS for our landing approach. Our sensors are not good enough to allow us to do a full stop landing, but they could assist us in lining up and flying most of the approach.

In addition, the synthetic display can be viewed from inside a live virtual cockpit with working instruments, from a chase plane, or from a tower view. This should be really cool!

September 21, 2005.

After crashing Rascal #1 we decided to take Rascal #2 up for it’s maiden flight. Everything worked great and the plane flies just as beautifully as Rascal #1. I think we get a lower idle and more braking action from the prop on approach, so Rascal #2 is easier to spot in on the landings. Also, there is much more fined grained precision on the low end of the throttle range with the G26 which means a notch up or down of power actually gives you a notch up or down of thrust. This makes controlling the rate of decent on approach work just like the text books say it should. (With smaller glow engines you typically have poor precision on the low end and seem to have not enough or too much power on approach with no fine grained control.)

One small bit of strangeness did occur in our 2nd flight. I think we had the idle set a little too low so when I pulled power for a slow fly by, the engine quit . I was downwind, but way to high to make a good approach to the runway. I kept the wings level and touched down on the extreme far end of our mowed grassy area. I bounced back up and sailed into the farm field (alfalfa???) bounced one more time in the field and the nosed over and went tail high. Zero damage, not a scratch, but it ended up *inches* from where Rascal

Sig Rascal 110 #1 Rebuild

Sig Rascal 110 #1 – UMN UAV Project

Next steps …

  • Come up with a better solution for securing the tank.
  • Redo top nose cover job?
  • Cut and fit cowl.
  • Purchase replacement windshield (?)

October 20, 2006.

This morning I re-maidened Rascal #1. I had to work through some minorhardware issues, but the structure and rebuild all held together well,and the airplane flew straight and true and as good as it ever has.I’ve very happy with the outcome, and very glad to have this airplaneback on active flying status.

October 19, 2006.

This evening I ran the engine for the first time after the crash. Everythingseemed to perform well. I think I may try to re-maiden tomorrow if the weather is ok.

October 2, 2006.

The battery and reciever and crystal arrived today. I tested to makesure they all work. I brought the aircraft home this evening andmounted the receiver, battery, and volt-watch unit. I still need to secureantenna. I’m running out of things I can think of todo before test flying!

September 29, 2006.

Begin putting the damaged right wing back together. I epoxied thewing joiner box back together so it is secure again and then a rebuiltand sheeted the first inboard section of leading edge back to the mainspare. Finally I sanded and covered it and (tada!) the wing is done!


We are *really* close to being ready to fly! Just waiting on thereceiver and battery now. I will cut and fit the cowl after wesuccessfully test fly and after I’ve regained my confidence in thisengine.



September 28, 2006.

Mixed up some epoxy and sawdust and used that to fill in the shatteredend of the right wing strut. I sanded this down to shape, redrilledthe hole, and threw a quick coat of white paint on it. Good asnew. 🙂 Here are before (damaged) and after (fixed) shots.



September 27, 2006.

Installed the engine and muffler.Reassembled and installed the main gear. Cleanedup the wings in advance of inspecting and repairing them. Inspected thedamaged wing strut and determined it is fixable.


September 26, 2006.

I recovered the front of the fuselage in white monocote. The top of thenose didn’t turn out nearly as well as I had hoped so I may cut that offand redo. We ordered a receiver and battery. I reinstalled the tailwheel assembly and the canopy.



September 25, 2006.

Test mounted the engine and rigged the throttle linkage. Next step is tocover the front of the fuselage.

September 22, 2006.

Worked on installing the fuel tank. Routed the throttle linkage housing.

September 19, 2006.

Secured the front wing support (where the wing dowls plug into.) Gluedin forward cabin support rods.

September 18, 2006.

Epoxy seal/paint the outside firewall.Filled in some of the gaps/cracks with balsa filler.

September 15, 2006.

Today I glued in the top nose support stringers. Then starting with thecracked up nose sheeting from the original, I drew a rough template of theshape the sheeting needed to be. I transfered that to 1/16th sheeting anddid some test fitting and trimming. Finally I squirted it up with windex which was what I had on hand and the sheeting pretty much melted around the curve …cool. 🙂 The final results looks better on the left side than on the rightbut I guess that just means I need to do a little filling and sanding.



Posing with the cowl and the cabin support rods.


September 14, 2006.

Reinforce left front nose side splice internally with some hardwoodsquare stock.

Finished sheeting both sides of the front fuselage (i.e. sides of the nose.)



September 13, 2006.

Reinforce wing leading edge bulkhead reinforcements. I’m compensating herefor an earlier mistake where the bottom portion of these reinforcments didn’tget clamped in where they should have been and thus there is some ugly gaps.It will be non-visible when the fuselage is all sheeted, but I just wanted tomake sure it’s solid structurally.

I installed (and tack epoxied) the engine mount blind nuts onto the backside of the firewall. These are a horrible pain to deal with once everythingis sheeted in … probably the hardest part of assembling the stock Rascal.So they are installed now and I don’t have to worry about them later.


Finally, I cross sheeted the bottom of fuselage forward of the main gearblock. This adds a surprising amount of rigidity to the nose section whichis what I was hoping/planning. I also attached the original sheeting to therear of the main gear block. (Little details …)


September 12, 2006.

Today I spent a few moments fabricating and installing the firewall sidereinforcements. I also fabricated the nose top stringers.



September 11, 2006.

Today I glued in the main landing gear block and supporting structure.I also dug around the shop and found the hardwood stringers I’ll needto support the balsa sheeting on the top of the nose section. When Itest fit the cowl, she’s actually starting to look a bit like her oldself again! Maybe there is hope after all. 🙂


September 8, 2006.

I spent a few more minutes fiddling with the fuselage. I foundseveral cracks and splits in the cabin roof where the wing dowlsinsert, specifically the sheeting forward of that.

I also secured several more cracks and splits in the bulkheads andvarious places I found them. I epoxied the split off pieces back ontothe landing gear mounting block.

I glued the cracks and splits in the right side forward nose section(shown in test fit configuration in IMG_3993 in the Sep. 7 entry.) Ifinalized my scheme to splice in the left side nose piece to theoriginal. It will involve a number of doublers, some beefy squarestock, and a big mess of epoxy.

Finally, I began to reassemble the nose section.


September 7, 2006.

A week ago I scored a new replacement cowl from a very kind fellow modellerwho was willing to donate his spare to the cause.

Today I did some work gluing the split clamshell that was the fuselage backtogether. From the wing trailing edge foward split out like a big clam shell.From the wing leading edge forward is just splinters. After today I shouldmostly have the wing leading edge back to the tail all fixed up and solidagain.

Yes, it still looks pretty ugly, but an amazing amount of rebuild progresshas actually been made:





May 30, 2006.

Today I made a replica of the fuselage bulk head that is between thecabin and the firewall. I also started tracing out the piece for the leftside of the nose (which exploded in the crash.) It is important to get thesize and shape just right so the firewall has the proper amount of right anddown thrust. I think I’ve got it, but it’s something I have to be carefulabout. I’m not 100% sure yet how I will fit/splice the new piece into theold one, but once I figure out how to get the side pieces glued on in a structurally sound way, I’m home free for this rebuild I think.

The replacement firewall:


The replacement rear fuel compartment bulkhead:


The replacement "instrument panel":


The replacement left side of the nose area:


May 26, 2006.

I pieced together the shattered bits of the original firewall and usedthat as a pattern to trace out a new firewall. The firewall is two layersof 1/8" light ply, so I made two copies, and sandwiched them together withepoxy.

My goal is to use the original bits as patterns to build new pieces when theoriginal is just too shattered. There is still some thought that needs togo into how best to proceed in some areas, but I’m making progress.

May 19, 2006.

Happy birthday to me. 🙂 I spent the evening cleaning up the engine.The dirt was surface only, nothing even made it into the carb. The carb wasslightly shielded by the cowl and the engine wasn’t running when it hit.There was a small amount of dirt/grit inside the veturi, so I popped offthe carb and blew everything out from the backside with carb cleaner. Hopethat doesn’t attack the rubber gaskets (not to mention my fingers.) Thealuminum spinner appeared to have no damage, the engine seemed to turn well,the motor mount was 100% intact. The only thing that shattered was thefirewall. I’m going to have to build me another one of those.

May 17, 2006.

Some of the parts fit together, some don’t. Some pieces are just not thereanymore.








May 16, 2006.

Today I started laying out the parts to try to piece them together. I madea lot of progress figuring out what goes where, but haven’t glued anythingback together yet. Some pieces I can probably glue together and use, butsome of the more load bearing structures I’ll glue together to make a formand then reproduce the part.

For what it’s worth, the left side of the front fuselage got compressedand exploded into bits. The right side of the forward fuselage is"reasonably" intact. This would indicate that the aircraft had some rightlateral motion when it impacted the ground. That is the direction the tailswung around after impact and it was the right wing that ripped off.

After staring at the pieces for a few minutes today, I believe this Rascalcan be rebuilt and will fly again, but it will take some effort. It won’tbe a completely trivial rebuild.

September 26, 2005.

Until further notice, Rascal #1 is offline. We are transfering all ourinstrumentation and cameras over toRascal #2.Follow that link to the most current interesting info.

September 21, 2005.

Today we flew 3 very nice flights testing out a new patch antennafor our video system. The new antenna seemed to yield much better results thanwe were getting before. Plus we had determined that the ground transceiverfor our radio modem link had also been interfering, so we put some goodseperation between the radio modem transceiver and the wireless videoreceiver and that all worked much better.

The video was working well so the next thing we wanted to try was havingme fly by video only (using a buddy box system and a safety pilot.)Take offs and landings would be done visually as per standard R/C procedures. The fly by video would only happen during a short segmentof the flight.

Shortly after take off (with maybe 75′ altitude) the engine sputteredand died. I thought I had plenty of altitude to turn back to the fieldso I initiated a turn. By my recollection I stayed off the elevator toavoid any chance of stalling, so the nose dropped substantially duringthe turn. However, once I got pointed down wind and tried to roll out ofthe turn (still with 20-30′ of altitude) the plane was unresponsive anddove straight in at a pretty sharp angle. As you can see there wassubstantial structural damage.





img_2936Case of dumb thumbs? Did Iride full elevator all the way into the ground? I didn’t think so at thetime, but the consensus of the audience was that I stalled it in. Buthere is my thinking: 1. TheRascal is nearly impossible to stall, 2. I am aware of this issue and I*thought* I was intentionally staying off the elevator specifically toavoid this mistake, and 3. the plane seemed completely unresponsive inthe final one or two seconds.

But I also don’t trust my recollection and I know my mind can play trickswith me. I am hoping we have good MIDG data from this fateful flight.I am hoping that I can compare the planes directional vector with it’sorientation to get an estimate of velocity and alpha. I don’t think Istalled it in, but I’m hoping the data can shed some light on what reallyhappened. We don’t have a way to record control inputs or indicatedairspeed, so we may never know for sure what happened. Just so I don’tforget, wind estimate for the time of the crash was 5 mph out of the south.The MIDG will give me speed relative to the earth, not speed relative tothe local air mass.

Update (Sep 23, 2005):

  • For some unexplained reason (and this has never happened to us before) our MIDG didn’t have a gps solution for the final flight. This means we had no position or velocity data, only attitude data.
  • The attitude data clearly shows the take off, climb out, and turn back to the field.
  • Our engine died during the climb out before the turn. But during the 180 turn back to the field, the nose immediately drops to between 10-15 degrees pitch down. This supports my intention to lay off the elevator during the entire turn so that gravity would keep the aircraft above a safe airspeed and eliminate the risk of stalling.
  • Because of data buffering and the fact that when the main gear departed, it sheared off our radio modem antenna, I believe we lost the last second or two of data. Our wonderful video capture software automatically deleted the video for us because it detected too much snow (after the crash.)


Note that I am speaking unofficially here, and from the perspective ofthe pilot in command with an ego to defend. Whatever the evidence,the conclusion will be that it was not my fault. 😛

I believe I properly executed my plan to turn back to the field withzero elevator input. The resulting natural dive during the turnshould have kept the airplane at safe flying speed since it naturallyseeks an equilibrium. This aircraft is *very* difficult to stall andin all previous stall tests, stalls were slow, required a tremendousamount of forced up elevator, they were gentle not sharp, and somelimited control authority was always preserved even during the stall.This makes it hard for me to believe that I could have been in a stallregime, and even if I was, I would have expected different behaviorfrom the aircraft. I believe I had sufficient and safe airspeed.However, when I tried to roll out of the turn and pull out of the diveI had nothing. The plane gave no response and continued to divestraight into the ground.

My conclusion then is that given my recollection of the control inputsand my intentions (supported by the attitude data) combined with myunderstanding of aerodynamics and my specific knowledge of theparticular flight characteristics of this aircraft, I believe theaircraft maintained safe airspeed throughout the 180 turn backmanuever, and very likely I over compensated and gained more airspeedthan needed through the turn/dive. Based on my understanding of thespecific characteristics of this plane, I find it highly unlikely thatI was any where close to the stall regime. The more likely scenariois the relative orientation of the plane’s R/C receiver antenna to theground transmitter, combined with the interference patterns of the twoon-board transmitters (1 for video and 1 for data) put us in atemporary "dead" zone. Unfortunately our close proximity to theground when this occured meant that we were unable to fly through thedead zone and recover … we hit the ground first.

I think I can rule out pilot error in the direct operation of theaircraft, however there are still higher level issues we have controlover that likely contributed to the crash. Specifically engine tuningprior to the flight. We did run up the engine on the ground beforetake off and it sounded perfect, but perhaps we missed something.Also we were using a buddy box system for the first time on thisflight. Did that contribute in any way? Crashes seem to always be along sequence of events where the initial problem leads to, but is notthe source of the final crash. What else could we have done prior tothe flight, with the setup of the airplane, the setup of the buddy boxsystem, the setup of our instrumentation, our flight plan, etc. tohave prevented this crash? Are there things we can do to ensure morereliable engine operation?

This big Rascal can be rebuilt and will fly again.

Sig Rascal 110 #1 Flying

Sig Rascal 110 #1 – UMN UAV Project


Project started March, 2005.

I am involved with the University of Minnesota Aero Dept. on a UAV project. My part of the project involves assembling the airframe as well as being the chief test pilot.

June 7, 2005.

Today we flew the Rascal 110 on it’s maiden flight! Winds were about 10-ish out of the SE, gusting to 15+. It was a bit on the windy side and the gusty cross wind was tricky, but we managed.

Posing at the start of the day …


img_2547 Curt at the controls …



img_2551 Turning final …

img_2550 Landing …

img_2552 Ready for another flight …

img_2553 Greg at the controls …






img_2560 Still in one piece at the end of the day … 🙂

img_2561 The Rascal is a big beautiful flying airplane. It’s a tremendous floater. Even at 1/4 throttle, the tail comes up quickly on the take off roll, and it’s airborn soon after. We powered the Rascal with an OS 1.60 FX 2-stroke. That is plenty of power and she can sure fly with a lot less engine, but the airplane is big enough to handle all that extra power just fine. It will go unlimited vertical, but just barely. It was a bit tricky to handle in the gusty cross-wind, really wanting to weathervane into the wind, but I’m sure with some practice and more flight time I will get a better feel for how it handles on approach. It’s a great flying aircraft and should be able to carry quite a load.

June 16, 2005.

Today I made two very short flights with the Rascal. Both ended early with the engine quiting. I safely dead sticked both times, but I wasn’t in the mood to practice dead sticking today. I couldn’t get the engine to run reliably, even after a fresh glow plug so I gave up for the day. I’m going to rip the cowl off and play with it at home to see if I can figure out what’s going on. No clouds, barely any wind, temp in the mid-70’s, other than engine proplems, it was a great day for flying. 🙁

The prop got damaged on the trailing edge midway between the center and the tip during the maiden flight, so today I was flying with a new prop. I went with an 18×8 (instead of the original 18×10) to try to get more “braking” action on landings to counteract it’s tendency to float forever. I’ll have to wait to get the engine running reliably again before I know how much this will help.

June 18, 2005.

Today I put in 3 really good flights. I yanked the cowl at home and ran a tank through in my front yard which I think cleared out the cobwebs. I think it was still a little tight from being so new and I was a bit off on my needle valve adjustment last Thursday. But today I had the engine running great. It pulls the Rascal through the air authoritatively and allows you to do *big* beautiful maneuvers. I probably ran 60+ oz of fuel through the engine today.

July 2, 2005.

Today we tested a simple telemetry system consisting of a Garmin Etrex GPS and an AeroComm radio modem. The aircraft flew great. The telemetry worked great. The GPS worked great and we got WAAS correction. Everything went pretty much as good or better than expected. From the data collected the max speed we hit with the Rascal 110 was 88.9 kts and I suspect that was with the wind which was running 5-10 kts at the surface. The Rascal can cruise comfortably at 25 kts and can lollygag and putter around at maybe 15-ish or even a little slower. But fire up the throttle and she get’s up and goes.

Last minute tweaks and engine tuning before a flight.


img_2658 Slow fly by and final approach.


img_2660 I also hacked together a system to load in the gps track, interpolate/smooth the 0.5hz data to 60hz and fake roll/pitch, then blast the result to FlightGear via UDP packets. The result is a virtual replay of our flight and it turned out pretty reasonable for a linear interpolation. FlightGear provides a working virtual instrument panel (AI, ASI, Alt, DG, and VSI) as well as a working HUD and a 3D sythetic world view.

These are virtual views from the flight playback. Notice the working instruments in the virtual cockpit view. Oh, and it’s flight gear so I tuned in a nearby VOR station. 🙂 If we recorded control inputs we could animate yoke, pedals, and control surfaces in FlightGear as well.

Virtual-UAV-01 We can also do external chase views and include a HUD if we like.



Virtual-UAV-04 Here is a plot of a portion of one of our flights.


August 5, 2005.

We had to scrub today’s tests due to radio interference problems (eventually traced back to the receiver.) Our intention was to test the newly installed wireless video system with two cameras. One pointed straight down and one pointed forward 45 degrees.




August 24, 2005.

Today we did 3 really nice flights to test our MIDG II IMU and our 2 camera, 2 channel wireless video system. We think we got excellent looking data back from the MIDG and it appeared to work *very* well. Unfortunately, our wireless video was really bad. We are going to have to do a lot of work on the video system to get it up and running satisfactorily.

We bought a second Rascal (RTF). This one is powered by a Zenoah G26 2-stroke gas engine.

Update: I worked over the weekend on parsing the MIDG binary data and feeding it into FlightGear. The result is a really nice animation of the 3 flights. The 50hz data rate on the MIDG captures a lot of the subtle nuances of the flight, dutch roles, wind gusts, twitchy thumbs, and even does a good job capturing aerobatic maneuvers–loops, rolls, wing overs, etc.

Sig Rascal 110 #1 Construction

Sig Rascal 110 #1 – UMN UAV Project


Project started March, 2005.

I am involved with the University of Minnesota Aero Dept. on a UAV project. My part of the project involves assembling the airframe as well as being the chief test pilot.

November 3, 2004.

It appears that our small UAV project got funded here at the U of MN. Hooray! I get to be paid (for a short time) to build and fly R/C airplanes. We plan to purchase our first hardware in early February ’05 and immediately work on assembling and test flying the airframe .

February 24, 2005.

U of MN UAV project update: We have done the initial airframe, engine, and R/C gear order. A couple items were backordered so we don’t have any fun toys to play with quite yet. The airframe will be a Sig Rascal 110 running an OS 1.6 2-stroke engine. Initially we will hand fly it (perhaps using an onboard camera rather than direct line of sight?) but eventually we will develop autonomous capabilities as well.

April 4, 2005.

Installed ailerons, aileron servos, linkages, and routed servo leads. The wings are essentially complete. Here are some pictures of the different pieces:





April 6, 2005.

Here are a few pictures of some of the toys hanging around the Aero work shop:




May 2, 2005.

Installed OS 1.60 engine into the Rascal with special ordered beefier engine mount.

May 4, 2005.

U of MN UAV project update: I cut, fit, and installed the cowl today. I found an 18×10 prop (in the recommended range) floating around the lab and slapped it on temporarily. Yikes … it is big! I will have to make one more opening for the mixture adjustment. It looks like we will need to special order a spinner for this beast. I also took a heat gun to the wings and fuselage and shrunk out most of the wrinkles.







May 6, 2005.

Today I bought a larger tank (24oz) than stock and fit it. I haven’t locked it in place yet, but that’s [hopefully] a quick thing. I also glued in the fairings which provide a bit of extra support for the horizontal stabalizer where it attaches to the fuselage. Next up is installing the elevator and rudder servos in the tail. Here are a couple pictures from inside the cabin.




May 9, 2005.

Today I installed the rudder and elevator servos in the tail of the Rascal. This minimizes the length of the linkage used (thus reducing slop and the risk of flutter.) I then attached the horizontal and vertical stabalizers and the additional two fairings on the top side of the horizontal stab. Then I attached the elevator, tail wheel, and finally attached the rudder. Next up is the rudder and elevator linkages.





May 10, 2005.

Today I fabricated and installed the rudder and elevator linkages. I also installed the springs that attach the rudder to the steerable tail wheel. After that I turned my attention to the inside of the aircraft and installed the onboard radio on/off switch and the throttle servo and linkage.



img_2529 Then I assembled and installed the main gear. The Rascal wheel pants come completely finished, even with blind nuts already installed. It’s about a 10 minute job to assemble and attach the main gear, wheels, and wheel pants, including taking them out of the baggies.



img_2530 With the main gear installed, the Rascal can now stand on her own, so it was time to pose for some pictures.


img_2519 Now I add the wing.




img_2523 I almost forgot about the cowl.


img_2525 And I might as well put on a prop while I’m at it.






May 11, 2005.

Items completed today:
Installed the side windows.
Installed a new prop (appropriately drilled out for our shaft diameter.)
Secured fuel tank.
Installed an extension to the receiver on/off switch.
Initial balance tests indicate that we might come out pretty close with no added weight.

May 12, 2005.

Items completed today:
Pad and secure battery and receiver.
Route the receiver antenna.
Touch up and shrink covering in a few areas.

May 13, 2005.

Today I checked the control surface throws to verify they matched the manufacturer’s recomendations. I also moved the battery as far forward as possible to put the aircraft in balance. I think we are now balanced with no need to add additional dead weight. With the exception of final checks, this plane is ready to fly.

May 19, 2005.

Today we fired up the brand new OS 1.60 FX 2-stroke engine and ran 24oz of fuel through it. The engine behaved well and pulls *very* strong. The next big step is the maiden flight. I will be gone most of next week so we will likely shoot for a day the week after next weather permitting.

May 30, 2005.

Fit and installed new spinner. We still need to get all the right tools so I can properly tighten everything up. Right now a couple pieces are only finger tight, but they look good.

June 6, 2005.

New pictures … all ready for her maiden flight tomorrow (weather permitting.)







Roll/Pitch Stabilization System

Pictures of the FMA Direct Co-Pilot installation …



The FMA Co-Pilot has some complexity to it, so it’s important to read through the manual carefully to get it setup properly, and to get it calibrated properly at the field before flight. The manual explains quite a bit about how it works, and what it can and can’t do. It’s important to understand these things so you know the devices limits and what you can expect. I mounted my sensor on the centerline of the fuselage immediately behind the trailing edge of the wing. This gives me good visibility left/right and makes for a very clean installation. Exhaust comes out the bottom-left of cowl so it’s fairly protected from residue. I pick up more bugs there than anything else.

I discovered that you have to be very careful and precise with the field calibration procedure, otherwise it will try to drive the aircraft into a bank and the plane will be constantly turning. Conveniently, my transmitter trim still works with the Co-Pilot device activated, but one caution, if your aircraft is trimmed for normal manual flight and you activate the co-pilot and it is not perfectly calibrated, you need to retrim. That can lead to (possibly substantial) trim changes with the device on versus off … it’s workable, but you have to be aware of it.

I setup the Co-Pilot so I could turn it on/off and adjust the gain with my “flap” channel. That worked well, and I found that with my big, slow Kadet Sr. I could fly with the gains dialed to full max just fine. Initially I only activated it altitude, but eventually I tried flying lower, and even landing with it activated. It worked so well that I eventually did takeoffs and touch and goes as well as landings. I observed no ill tendencies and it seemed to help make my landing smoother because it can compensate for gusts more quickly than I can (and I was able to practice this because the winds were getting gustier as my test progressed.) With the co-pilot activated, it wants to drive the wings level and sort of tries to hold pitch. But it still passes through your manual inputs “additively” so you are able to fly fairly normally and override the stabalization controls.

Ok, so the big question after playing with the co-pilot for a few flights is “how will it work as a UAV stabalization system?” My answer at this point is, yes very well for many applications and airframes. However, it’s not perfect and it’s not magic. The Kadet is big and slow, so even with the gains dialed up to max, it can’t keep the plane perfectly level all the time. It is highly sensitive to the field calibration procedure, so you need to perform that carefully, then ensure that you are well trimmed before cutting it loose to do anything on it’s own. It does do the job though and keeps the plane reasaonably stable. With the system activated, it is very safe. You can input full rudder deflections and while I do observe some banking, the system holds it’s own and limits the bank to 10-20 degrees and keeps everything under control. Note that this is a “simple” proportional controller so it can’t cancel out all errors or biases, but it produces a “stable” system. That’s why it can’t hold the wings level against rudder input, but with neutral rudder it does just fine.

I think I’m happy enough with the co-pilot to move forward and start looking at getting my flight computer running.



EGN Project Overview

My first goal is to have fun and use this as an outlet for a few ideas that have been bouncing around in my head. I have a life long love of aviation, airplane models, and computers so I would like to mix these together a bit.  I have an FMA Direct CoPilot IR stabilization system that I plan to use to keep the aircraft self stable. Notice that this is instead of any type of gyro/accelerometer/IMU unit. An IMU typically reports orientation as input to a flight computer which then does mathemagic to calculate servo positions to keep the plane level. This IR unit does all that itself in a$100 unit. The downside is it can only hold the wings level, the upside is that is usually exactly what we want to do.

I also have a small flight computer, a First Robotics Minicontroller. This has a CPU and can drive servos directly. I hope to attach a GPS, do a small amount of crunching and then drive the rudder servo to steer the aircraft to a waypoint (such as home.)

I also hope to use this to do some sort of aerial photography … either wireless video, or digital stills, or both.

Beyond that it would be fun to add a radio modem to pass telemtery information to a ground station and perhaps pass commands back to the onboard computer system. I could envision some integration with FlightGear to use that as synthetic vision or overlay the live camera view on top of the synthetic view.

I am funding this project on a hobby budget. So I plan to scrimp and save, reuse existing equipment, buy off of ebay, etc. and only move as fast as my spare funds will allow. My goal is to build a self stable, self navigating R/C aircraft for under $1000 total cost.


Let me just say a brief word about safety and politics. I understand the AMA is concerned about UAV projects. They don’t want irresponsible behavior ruining the hobby for everyone else. I plan to always operate this aircraft under the constraints of the AMA safety code. (i.e. always in visual and radio range, always within the R/C model altitude limits, always with a human pilot able to assume manual control at any instant.) I also understand the FAA doesn’t really know what to do with UAV’s yet and has no immediate plan for fitting them into the USA airspace. Again, I plan to operate this aircraft entirely as a R/C model aircraft which the FAA is not interested in regulating.

Throughout this project I plan to pay keen attention to safety issues, fault tolerance for the onboard systems, safety for the pilot, safety for the aircraft, and safety for everyone else. I know I can’t control every aspect of every circumstance, but I wish to be very thoughtful, and very considerate of a wide range of safety issue so that (1) the aircraft itself is fault tolerant as much as possible, (2) the on board intelligence will be able to detect some problems before the pilot on the ground and take steps to minimize or avoid damage to the aircraft, and (3) I will operate the aircraft so that if something does happen, it will happen as far away from any person or property as possible.

I believe this approach is essentially restating the spirit of the AMA safety code with fewer specifics.