Current News

Report from Dan (yours truely)

I think we would all admit this first-annual event was challenging, to say the least! The April weather played out with winds aloft in the 20kt range, from the West on Saturday and from the South on Sunday. What difficult conditions!

Saturday started out with a pretty good breeze on the ground, showing little surprises for the conditions aloft. Team Old School managed to cheat the wind and gather some altitude to start the course first, around 1pm. Those of us back at the field watched them slowly make progress North, meanwhile fighting the wind and fast-drifting thermals. Team ALOFT finally found some altitude after over a dozen launches and attempts chasing lift downwind. ALOFT was commanded on the course and crossed the start line at 1:42pm.

Conditions on course were rather ridiculous for ALOFT. Despite tracking straight down the road on the Birdie to Gas Station leg, the plane was crabbed a solid 45 degrees into the stiff West wind. Duane and Doug did diligent duty managing driving but weren't ready for the first challenge, as ALOFT rounded Gas Station and picked up speed. With a 30kt airspeed and 20kt winds, ALOFT cruised from Gas Station to Stop Sign with a 50kt ground speed, prompting Duane to safely go the speed limit of 55mph (ahem) to slowly catch up. Adam's eagle-eyes strained hard and kept a safe lock on the wings, but man the plane sure moved out.

On the back leg, ALOFT's ground progress slowed and the Jeep caught up with no problems. We have to maintain about 1/4 mi behind the plane to keep a good data link (which gets better with distance actually because of the antenna arrangement), so this was now perfect. The ALOFT soaring code did its best to fight the wind, but on a due South course and winds blowing from the West, any altitude gains meant drifting far off course and burning most of it to get back on track. The leg to Saratoga cost a lot of altitude. Now tracking back North West into the wind yet again, ALOFT did not manage to find any good lift and Adam took over piloting duties at 750ft and within sight of the Birdie start/finish line. Unfortunately, the lift just wasn't there and Adam landed softly in a grassy green field. For the first lap, we made 8.74mi around the 9.52mi course.

After heading back to the field to change batteries and get some food in us, we heard Team Old School made 7.65mi and re-lit to leave just left 10 minutes earlier. The competition was on! Team ALOFT spun back up to altitude and started gingerly down the course again. The day had opened up a bit with a few more hours to get the boundary layer height higher. Winds were still up, but at least we were able to work for more altitude. After some more fighting, ALOFT came around the course with 2300ft to spare and decided one lap at 9.52mi was just plenty. Team Old School attempted two more times, but the winds were just too much and their first distance of 7.65mi stood.

Overnight, the winds dropped for Sunday's speed task. The forecasts were for less vertical motion and lower boundary layer height, but the winds would be smoother early and build throughout the day. Sunday was definitely a task for getting on course early and working the weak lift that was to be found.

Dave Beach and Gordy Stalls had driven in specifically to attempt their LSF Level 4 task of a 10km goal-and-return. After driving around Sunday morning, they found a straighter run than the DESS Lap course could provide, so they drove out and decided to gamble with a three person crew that included two pilots, one driver, and no spotters. From what I heard of the trip, it was difficult for George Oakey to call back to both pilots of upcoming trees, so Gordy ended up losing sight of his plane and fortunately landing in a field a mile away with only slight damage to his Supra. Dave managed to push all the way to the turn-around at 10k, but said the lift just ran out of gas and he had to put it down at the turn around. Despite the struggles, Dave was in good spirits. Good shot at it guys!

Meanwhile, over at the DESS course, Teams Old School and ALOFT were having good luck catching lift off the winch. Team Old School switched to the 24 year-old (and well seasoned!) Constellation that has seen much XC action before. With the trusty Constellation getting tiny in the sky, the three-man crew hopped in a pickup and headed out. Team ALOFT took a lot longer to get started. Launching at 11am, the soaring code just couldn't break through 2000ft. The day developed a lot and the thermals topped out a bit higher, so at 2400ft and 1:42pm (after already a long flight), Team ALOFT set out down the course.

The evening before, I programmed in some additional wind-compensation, hoping to account better for the drift of the lift. Even with the winds aloft at only 7kts to start the day, the drift compensation was helping. Also, Matt Hazard brought a new Unscented Kalman Filter method for Thermal Identification and we were able to try it to great success. On this run of the course, we zipped around the first couple turnpoints and thought we were doing super when we caught up to Team Old School. We played the best game of leap-frog for 45 minutes before Team Old School finally pulled ahead and disappeared down course. The soaring code was doing better at making wind-decisions and I tried extremely hard not to click any buttons such that the performance was all due to the code. At approximately 3000ft and with only 3 miles to go, we realized we had the course made and I turned soaring off to let speed-to-fly dolphin soar the plane back to the field. Hitting amazing lift, we carried 2300ft into the start/finish line and decided to go again!

Team Old School was only slight ahead and we caught up to them after a few minutes on course. Another great game of leap-frog commenced and we all were really starting to struggle with the now 20kt South wind. Both teams rounded the Stop Sign turn and got stuck on the due South leg toward Saratoga. The wind turned out too much for the venerable Constellation and Brady set it skillfully down in a field. Team ALOFT rode several thermals up and then made zero altitude gain coming back down course. We turned soaring mode off to see if we could just make the next set of fields and give the code another chance to find lift. This turned out to burn all the rest of the altitude and Adam tried valiantly from 750ft to find some lift at a small field just before the 264 crossing. We dabbled a little here and there, but the wind got to us too and we landed at 3:05pm after a 4hr flight.

Results drum roll please......
• Day 1: Team ALOFT managed a complete lap, 9.52mi
• Day 1: Team Old School managed just shy of a lap, 7.52mi
  
• Day 2: Team ALOFT managed a complete lap at an average of 6.14mph
• Day 2: Team Old School managed a complete lap at an average of 6.24mph
• Overall: Team ALOFT:1984 points
• Overall: Team Old School: 1804 points

Team ALOFT squeaked it out .... however, Brady posted a 25.5km continuous flight on Sunday and earned his 25k XC pin. He also posted a personal best duration of 3hr 16min.

I must say, this was an extremely fun event and everyone who participated was asking for another one next year. Thanks to everyone for coming out and coming together to fight against mother nature. I learned a whole lot about flying in wind and was able to make some code updates to begin compensating better for the drift. Brady was super stoked that his 24 year old rudder-elevator-flap ship beat a $2000 composite full-house moldie with LOTS of electronic goodies installed. Goes to show, it really is the pilot making good decisions that wins a contest.

Report from Adam (the safety pilot):

It was horrible! We had to sit out in the sun for hours and hours while riding around in the back of some pickup truck, and Dan forced me to stare up at this little dot in the sky all day long ;o)

On Saturday, we had eight people at the field with five XC planes (two SBXCs, a Comet, a Constellation, and an Escape). Conditions were challenging to say the least. We got everything set up and were waiting for the day to develop. The air was buoyant on the surface, so the hand launch was doing quite well. Once the bigger planes made it into the air, we found it difficult to get up to starting altitude. The winds aloft ranged between 15-20kts sustained, which pushed any thermal that we managed to find quickly past the trees at the edge of the field, forcing us to lose most of the altitude gains limping back upwind. There were several instances where the Piccolo was holding the plane at near zero ground speed back behind the trees! We had at least a dozen or so good into-the-wind launches, but just weren't able to connect with the boomer needed to get above the low level stuff and massive sink holes. Brady Baggs was able to get his SBXC up and on the course by lunchtime. After an hour of hard-fought progress, he returned with plane in truck, landing out about two miles short of a full lap.

As the day opened up, we were able to slowly work our way up to altitude. Once comfortable with conditions and altitude to make an attempt, Team ALOFT was on the course. After a few thermal stops on the north leg, we made it to the first turn. The plane made the turn about half a mile ahead of the Jeep and immediately rocketed downwind. We had plenty of altitude, but found a sink area and were plowing ahead at 35+kts with a 22kt tailwind. We were happily making 55kts down-course progress.

On the South leg, things changed. Once again, we managed to get behind the plane. Dan announced that it had captured the turn point and was heading South while we raced ahead to the turn. I was in disbelief since the plane looked like it was headed back towards us and now off at a 90 degree angle. As we made the turn, it was apparent that the SBXC was being forced to fly at a 45 degree crab in order to maintain the ground track due to the wind! Ground speed was apathetic, and any thermal stops were challenging since we had to return to the course after being blown downwind. We made the last turnpoint, but were still struggling with the wind and a lack of thermals on the last leg. After dropping below 750ft, I retook control and tried to squeak out a save. I managed to find some positive air, but was not able to follow it downwind due to some power lines by the road. At less than half a mile from the field, we probably could have made it with the altitude that we had, but there would have been no margin for landing and we safety would have been out the window with such a low crossing over the highway.

After returning to the field, we found that Brady had begun another run. He shortly returned after only making it to the first turnpoint. We returned to starting altitude and jumped back on course ourselves. This attempt would prove much easier than the last as the thermals were a bit stronger, allowing us to break 3500ft at points. We completed a single lap with copious altitude to spare, finishing all 9.6 miles in 69 minutes. Brady made a third attempt and could be seen from the field working hard on the backside of the course. Everyone was rooting for him to finish strong, but he was once again forced to land short a couple of miles.

On Sunday, the winds were much calmer, with nice smooth air and gentle thermals. Several guys decided to attempt their LSF 10k out and back runs, but wanted to find a course that was better aligned with the roads, making the task easier. The left the field and eventually returned to the field later in the day with only half of the 10k completed by two pilots.

After playing handlaunch to evaluate the air, we decided to put the SBXC back up at around 11:00. The inversion layer slowly lifted through the morning, allowing us to slowly work higher and higher, from one thermal to the next. Unfortunately, we were getting stuck at 1600ft, but the light winds allowed Dan and Matt to switch out thermal ID codes and evaluate their performance. I'm happy to report that Matt's unscented Kalman filter code worked quite well on its first outing, performing on par with Dan's code! We also tried out the Gedeon model and found similar performance. Dan was able to implement a new feed forward wind adjustment that seemed to help the drift estimate.

Brady managed to work his RES Constellation up to about 1800ft and decided to go on course. After moving to the start point to allow us to climb out in position, ALOFT managed to finally break the 2000ft mark that we had decided would be good enough to attempt a run. We all took bets on what the max altitude of the day would be, with guesses of 2400, 2800, 3200, and 4000ft (Dan was being a bit optimistic there ;o). Progress was slow, due to the lethargic updraft velocity, but we eventually made out way all the way around, maxing out at around 3100ft. As we approached the start point, we caught up with Brady, who had a 20min head start. The first lap took around 81 minutes I believe. For the next 30 minutes or so, we leapfrogged Brady as we progressed around the course. The winds picked up on the back leg, with a 12-15kt headwind. The thermals were getting blown down course fast enough that all our gains were lost by the time we returned to where we started. I don't know if Brady was just trying to stay competitive and pushed too hard, but he made progress upwind, but at a dangerously low altitude. Eventually, he was forced to land out about halfway around the course. We concentrated on staying above the curve, but just weren't able to pick up altitude and make progress. Over a 30 minute period, we only managed to go from 1600ft to 1900ft with zero course progress. Deciding that we simply had to push ahead, we made some progress, but just weren't able to find a thermal when we needed it. Once again I took control at 700ft and managed to work up to 900ft or so, but at 1/3 mile downwind and at the edge of the safe landing field. Unable to follow further downwind, I returned to overhead, but couldn't find anything else. We landed approximately 6.3 miles into the second lap, after 2:20 of course time. For the day, we had one flight, so I guess that was compensation for all the consecutive launch and landing cycles on Saturday! Brady also broke his single flight TD endurance record at 3:16 and 25.5km distance! Woo hoo!!

All in all, it was a successful event!

Report from Brady (the other competitor):

Ok I am just happy riding around in the back of a truck (former dog in a past life). But put a transmitter in my hands and a plane in the air and it doesn’t get any better than this!! I want to thank Dan for having this event. We need to do it again before next year!! I would like to especially thank Dave who drove for us on Saturday! Dave If you need a driver sometime let me know I will do my best to be there. I would also like to thank Mihai and Dick for there support in driving and spotting/co-pilot, moral support etc. Now we are team Old School!

Going up against team ALOFT with all of the gadgets is very intimidating and I tried to play my own game and not consider this as a competition. Well maybe at least until Dick became my co-pilot on Sunday! I think that some day soon, Dan and Adam will break all XC records. They surely are on the right track.

A couple of things that I will never forget. Saturday, I had several low saves with the SBXC and Mihai coaching me on telling me "YOU CAN’T LAND YET." We rode the same thermal almost half way down the back leg drifting with it all the way. Somewhere close to the Saratoga turn as we were flyng down wind, it was if the elevator came off!! All at once, the SBXC was flying very fast. I had little control as the trees blocked our sight. Then as we cleared the trees Mihi calmly mentioned: "I think it's upside down." I said: "you know I think you are right." I regained control as there seemed to be nothing wrong. We didn’t see it loop. It just came over the trees inverted??

I soon realized that all the turbulence was caused by a strong down-wind thermal. I worked the bottom edge over a house and trees intersection power lines etc to try and save the flight. But I was a day late and a thermal short. It was a tight landing, having to clear the house and trees and getting it on the ground without running out of room.

Sunday started out nice as could be: no wind, warm, early signs of thermals. I really had only planed to put the old Constellation up just for kicks. I never figured that I was going to go on the road with it. I really had planned to wait for Dick to get there before making an attempt.(Unless I got into some good lift). Well, I caught a nice thermal right off the end of the winch line. So what the heck, let's fly some. So after about 45min of playing. I couldn’t stand it any more and I asked Mihai if he wanted to drive and we would go. About then Dick came driving up and we yelled out "want to go?" I don’t think his feet hit the ground from his car to the back of the truck. We realized that we were not prepared to go. So Dick ran back to his car and got his stop watches, a pen, and paper.

I really never figured we would make it past the first turn. I didn’t leave as high as I would have liked to, but it was nice to get the old XC back into the air one more time. We made the first turn easy as pie, caught the next thermal right were it should be at the corner, and quickly got back on course. We were just talking and flying just nice and easy. I realized about the same time Dick said it, that it sure is flying good and I agreed. We caught another very nice thermal along the way. And another just were it should have been right after the last leg stop sign. Just then Dick caught a glimpse of Team ALOFT's plane. It was like JAWS lurking behind us!!! DD… DD…. DD… (don’t know how to do the sound track) eeK!!

At that point, Dick nudged me on to get a little more serious by telling me they were getting closer. I’m not sure were they passed us at, but the Constellation leaped into a another good thermal soon after ALOFT blew by. I think it was a wave that we caught, but when we left that thermal, we rode it all the way to the Saratoga turn, passing ALOFT along the way. We still had good altitude at the turn so we made the turn and I had enough to get to the next big field on the right that I had picked out. Sure enough, the next thermal was there waiting for us right where it should be. I really was only thinking about going one lap and only want to gain enough altitude to make it back to the field so that Dick could put his plane together and give it a go. Well, we over-estimated and hit the start stop line with plenty of juice. So we made a rolling decision to go. Quickly back to the first turn and the thermal was waiting on us right at the corner. Quick climb and back on the road. Caught another thermal around the Broken House and then all the way down the back leg. Again right in front of the noisy Dirt bike house. There was good lift there but I just couldn’t think straight with the noise of the dirt bike buzzing around. So I told Mihai to go a little more down the road.

At this point again somewhere in here, we got a glimpse of Team Jaws-ALOFT and soon we were overtaken. It went from calm to very windy. I caught several thermals only to make little headway. So I decided to not thermal as high and next attempted to try a push forward as much as possible. We also noticed Team ALOFT having the same trouble and I could have piggy backed the thermal that they were in but only to end up back at the stop sign again. So I knew that this was my last ditch effort when I left this thermal to go upwind. My goal was to go to the tree line to the next field in hopes that a thermal would have gotten stuck in the corner. I lost a lot of altitude all the way passing ALOFT again and not getting to the next tree line. It just wasn’t going to happen. I tried worked a small thermal drifting back downwind only to lose it all coming back. I did every thing that I could think of surfing up wind and catching some small bumps. I have to say I pushed the OLD XCC to its limits it just didn’t have anything left to give. I hit one last low bump and decided not to risk it .

On the ride back we were thinking that I would be very close to getting the 25K distance. After adding up the mileage including the distance from the winch to the road, it came out at 25.5km.

I like this time of year for our little 9.7 mile course. The trees are not filed out. It’s not so darn hot and thermals are smaller/harder to find. It also makes it more equal for smaller planes and guys without gadgets.

It was a heck of a lot of working your tail off flying all weekend. We all were very tired but very happy.

I learned yesterday that the new plane needs a slightly more forward (~1/4") CG than is recommended. This is definitely noticeable in signaling thermals and upwind flight. Before moving the CG forward, the plane pitched quite well to signal thermals, but a little bit too aggressively flying into strong thermals. Upwind legs looked frustrating and penetration was difficult. Removing two quarters from the tail made all the difference. It could stand to move perhaps slightly more forward, but Adam was happy thermaling it after the CG shift.

I also learned yesterday that the power system I spec'd is a bit under-powered for climb-out. We winch-launched to ensure plenty of initial altitude to play with if the power system was lacking. I started with the lowest power combo I picked (Neu 1105/3Y, 6.7:1, CC 35A ESC, CAM 11x10) thinking I might get away with it. For the second flight, we bumped the prop to a 12x10 and it still seems like it needs a bit more static thrust for a safe hand-throw takeoff. It is sacrificing some efficiency to tailor for takeoff rather than cruise, but I really don't want to need a winch. Both power systems were indeed enough to sustain cruise though, so I'm in the right ballpark. We didn't even move to any of the spare battery packs!

I suppose next is going to be more analysis to update the propulsion system, update some code I/O snafus I found, measure the inertia of the new setup, and a laundry list of other to-do items. However, it is quite nice and a satisfying feeling to have a new plane in the air successfully, especially to go from maiden flight to full auto-soaring in the span of one flight test day.

Thanks for tuning back in and hopefully you enjoyed the news and the fun pictures I get air back under my wings :-)

Dan

I am in an absolutely wonderful mood despite it being past my bed time. Thought I'd post an update.

I finalized the eSBXC tail controls today. The pushrod for the elevator had long since been trimmed to length and ends bonded together, but now the servo arm has been trimmed to match the travel and clearance for elevator mix with flap has been confirmed. The rudder received pull-pull cables made I believe from Kevlar. I was more than slightly wary, instead thinking of manly-sounding steel fishing lead, but I put together the stock Kevlar setup just to see and I am pleased. I still need to tighten it down all the way and lock-nut the clevises, but as far as my mindset goes, this I'll pick up on the pre-final inspection.

I also set up the program on the RC transmitter. It was a decision point whether to copy the setup from the glider version and try to figure out where to put the throttle control versus assuming the eSBXC is a powered plane that just happens to also be a good glider. I settled on the first and copied the SBXC programming. After zeroing out the sub-trims and any travel adjustments, I was left with the proper mixing and correct switches/knobs setup to our expectations. After examining the options, I settled on having the throttle on the left slider (index finger actuated) of the JR 9303. Since the plane is going to relatively underpowered (especially compared to the Windex), the throttle will get essentially set to full power for takeoff and climbout anyway, so not much modulation will be needed. Likewise for flying, throttle will get set to a high efficiency setting and the pilot will be lenient on altitude changes due to environmental effects (which is why this project exists anyhow!). Landings will be conventional glider landings, with the motor off and having full authority with flaps on the left thumb stick for glide path control. I feel this is a good comprimise and makes trivial switching between the two SBXC.

I also have been doing some wiring. The safety switch now conforms to the above discussed channel setup. I have two more wire runs from the safety switch to the wing connector and the fuselage wiring is more or less complete. Add the motor in when it arrives and clean up with several more zip ties (I love zip ties), and it should look pretty good. I still have the power switching and associated wiring to do still. The old switch area was much larger due to the more aft location of the hatch. I'm not totally sure physically how to make the part I want for this switching setup. However, I made the last one twice and I figure this one I'll know what I want better when everything else is installed. Worst case, I simply directly connect power in with no switches for the first flight.

I started playing more with the PC104 board too. I sort of bounce between my project efforts until I get them all done, if you haven't figured out my style yet. The PC104 hasn't done much since I set it up in its new enclosure. I still need to make an image of the compact flash card so I can get back to this known state quickly if I have any memory storage kinds of hardware issues. So far (knock on wood), the CF card seems to be quite content running XP.

I coded pretty hard for the last couple weeks on some updraft model work. I'm hoping to test my algorithms more in a 6dof sim. This is perhaps backwards from most research; typically a lot of ground testing results in a shorter flight testing segment. I know I'm not typical, I actually like to fly and was perfectly content learning my lessons in the air. It was nice to know early on that the actual flight vehicle responded positively rather than a super simplified simulation model. Anyhow, off topic. The updraft model was heavily based on Michael Allen's paper and I hope to publish the new stuff I added as well as perhaps a bit easier to use model. I'm slowly learning that academia/research thrives on people writing papers with half their time, so perhaps it's time I write another. The updraft model is as good a topic as any. I sure have captured enough and spent enough time staring at the sun watching the plane work them...

Related to the updraft modelling, if anyone happens to have a paper describing simply the structure of a mushroom cloud or a hurricane, I think these may be good stepping stones to adding more improvements to the updraft model.

What else? Well I think that's it for this update. Sleep is finally catching up with me...

I have been steadily cranking away at the eSBXC conversion. I still have a few parts on order that have yet to arrive, but progress has been pretty steady. Major to-do items left are:

It snowed today!

I like snow. We got 3-4 inches, which is more than enough to completely shut Raleigh down. The funny part is, the roads are simply slushy and folks are driving 5 miles an hour. Not much different than rain at this point folks! Anyhow, maybe I should get out and drive around myself to see if things have changed since my walk earlier.

Anyhow, back to airplane stuff on a snowy day. If you haven't been keeping tabs, I have been uploading new pictures of my electric SBXC construction in the gallery. The current progress is focused on electronics hardware installation brackets. I finished up the main battery tray underneath the wing earlier. The vario tray now is complete too, matching almost identically the one in the old SBXC. The receiver tray is under work, just needing to get some 4-40 blind nuts and nylon hardware. I have a whole bunch of wire work still to do, but it's a whole lot easier when the components are mounted already and simply need to be routed together.

As for balance, I still don't have a good feeling if I'm nose heavy already. The 3s3p 7.8Ah main battery and the 4s 8Ah propulsion battery are as aft as physically possible. The Piccolo is in the same location as in the glider. The vario and receiver trays ended up having to scoot about 3" forward because of the hatch location being farther forward. The only other items that goes in the nose are the motor/gearbox/prop, speed control, power switches, and potentially the PC104 I've been working on.

Oh you haven't heard about the PC104? Well I have made substantial progress on this over the last two weeks. I have had the hardware for a while and even was successful getting it running mid fall, but a file corrupted and I ended up putting it down. Now, I'm back with force and reinstalled everything from scratch. The compact flash card is doing just fine so far running XP, despite being not the best option to replace a hard drive. I have been told to expect as little as 2 weeks of time before the CF card poo's out, so only time will tell.

What more about the PC104? I now can remote desktop into it so I can have a nice headless computer to install in the airplane. Simply use the ethernet connection and no monitor/mouse/keyboard needed. Much lighter that way ;-) The only other connections are serial to speak Piccolo and a power input for the obvious reason. I haven't measured the power draw yet, but I have run successfully off of a propulsion battery. Next step is to make a pretty shielded box; I like boxing components when I can afford the weight because it accents the modular approach and sure makes the components easier to handle! I'm measuring I need a box 5.75" x 4" x 1.5" in size. I'm certain the PC104 is overkill for the algorithm, but it is very nice for algorithm development rather than spending all my time figuring out how to squish down the code that is still in periodic flux.

I tell you, I am quite excited to get this PC104 in the air and really watch the plane making its own decisions. A buddy and I were talking after UAV Safety class that this is likely one of the few, if not the only, UAV that could potentially fly-away despite having no motor onboard. Surely this means more safety measures will be in place before attempting any flying with it, but it was a neat thought-experiment to figure out the safety issues.

Enjoy the pictures as they come in over the next few weeks!

At the turn of the year, everyone always seems so retrospective, so I figured I'd give a short look back at the news-makers of 2008.

Probably the biggest news was breaking the FAI goal-and-return record, raising the bar to 30.2mi (60.4mi round trip!). While this is unofficial, I think it conclusively shows that ALOFT is really on to something. This project in roughly two and a half years has gone literally from an idea over a pool table and pizza to breaking records for distance and not even pushing the bounds of soaring potential.

The big "graduation exercise" was competing at the 2008 Montague Cross Country Challenge. Being the first ever fully autonomous glider entry into a soaring competition, and understanding some of the deeply rooted limitations (like having no terrain knowledge), I find the 3rd place overall standing to be quite a testimate to how well the soaring algorithms work in real-life. Not just an academic simulator exercise, the ALOFT project has gotten out into the field and proven its meddle against the best in the world.

Early in 2008 the SBXC received a well-deserved upgrade from the loaned NASA Piccolo Plus. The Naval Research Laboratory funding kicked in and brought a new Piccolo II to the table, giving updated telemetry rates and more accurate air data sensing. The newly released Piccolo version 2.0.4 was also flown and tweaked in early 2008. This opened doors for even higher telemetry rates and a full GPS/INS integrated solution for cleaner state information, resulting ultimately in less necessary filtering in the ALOFT code.

The SBXC also received some general hardware updates after a scary 100mph dive. New servos, new wiring, new flight batteries, a new 2.4GHz spread spectrum manual control system, and general attention to some of the finer details rounded out the hardware upgrades. With all the flight hours and several landings starting really to add up, it was very good to perform some routine maintenance before anything really important gave up the ghost at the wrong moment.

Possibly one of the only downers from 2008 is the flutter incident with the Windex powered glider. Despite being designed for aerobatics, a poor control horn installation resulted in the Windex exciting aileron and flap flutter at 70kts during a relatively routine maneuver. Suffering only what must really be called minor damage, the aircraft (under excellent piloting skills) limped home and protected its expensive contents. Jokingly, it would have been nice to have a binary crash: either no crash or beyond repairable. With the repairable damage to the wing, it put the project in a pickle, whether to repair or take an intermediate step backwards.

As "team conservative," after the Windex incident, a new SBXC is under construction, this one with a motor and a meager battery pack. Despite the supposed high endurance potential of the Windex, it seemed more fitting to learn integrating electric propulsion into the soaring equations on an airframe more similar to what was working so well over the past years.

How could I forget Hawaii!? My autonomous soaring paper at the 2008 AIAA Guidance, Navigation, and Controls conference in Honolulu went over well and has allowed me to put in the public domain my first academic work. Documenting several updates to Allen's widely cited GNC paper, my hope is that others can pick up this document and begin their own autonomous soaring projects.

Currently also under progress is my PhD dissertation. Some 130 pages so far, it has been terrific to put into writing the lessons learned and solidify many of the nitty gritty details that aren't common discussion. In the process, several new bugs and issues have been routed out, so this has been especially insightful at times.

Looking ahead to the new year, several items on the agenda are either already under development or at least beyond the drawing board.

The good news is that it looks like the expensive parts are all okay. The bad news is that the Windex fluttered at 64kts.

Adam and I took out the Windex Sunday 11/9/08 with the intention getting a better idea of how well it can catch thermals and evaluate the way to soar with it. I had a folding prop ready, but thought to stick with the fixed prop for the first flight. Adam flew for powered for roughly 4 minutes and then soared around a bit before finding a really nice thermal. He was able to ride this thermal to about 1700ft from about 500ft entry. He was remarking how he was having to hold the wings banked and what speed felt comfortable and etc. Just learning the new plane. It is much, much harder to see by the way.

So after the 1700ft thermal topped out, we saw from the telemetry that the Windex thermaled around 35kts and typically cruised around at 40kts. Since we had some altitude, I suggested we do a CG dive check or two and evaluate how it handled and watch the higher speed performance. Starting from 40kt, Adam pushed the nose and had a 55kt max speed the first dive. It was showing slightly nose-heavy tendencies, wanting to stably pull out in a reasonable arc. We did one more (now into the wind) and held it slightly longer and reached 65kts. Now at this point, the airplane still didn’t look overly fast. It was still fairly reasonably fast, but not to the point where we should have been concerned. This wasn’t up to good high-speed pass speed yet.

We heard a buzzing/humming. About that same time we saw a piece of white fall off the airplane and start fluttering down. Almost immediately with the sound I called out flutter and tried to get someone to watch the piece, thinking we’d lost a full aileron. Adam quite rightly tried to get rid of the speed while we watched both ailerons wagging very quickly. The drag must have helped slow it down too. Adam found his rudder worked to turn the plane and keep wings level, so he made a full 360 turn, said he was putting down flaps to try slowing down more, and unceremoniously landed short of the field in the beans. The beans probably saved the plane from more damage. Adam certainly earned his wings saving the plane.

We both half ran through the beans to unplug the batteries and forgot the camera SD card to document the site. Getting up to the plane though, we didn’t see much damage. There were beans everywhere of course and some slid under the canopy so there were beans inside even. The lipos were all in-tact and still attached to their mount which of course popped its glue joint from the fuselage. We pulled them out first. The TEK was slightly bent but the pitot looks untouched. The fuselage is intact and looks untouched, save the beans rolling around inside. The wings… all four control surfaces were completely free, not attached anymore to servos. The lower inner half of the right aileron is missing, which must have been the part we saw falling (too far into the woods to find). There is some damage where the wipers impacted the overhanging trailing edge skin, but mostly a cosmetic fix. All four servo control arms were shattered or broken. I took pictures of each as soon as we opened the hatches. The wings structurally were still attached and nothing looks really broken with the main structure. The hinges look okay.

Looks like we just hit the flutter point and the internal linkage setup couldn’t cope. I knew the servo arms were short because of the internal linkage setup, but I figured this was recommended (as in they gave the parts for it) by the manufacturer. In retrospect I could have used metal servo arms; since all four servo arms are broken, I’d say that the plastic ones must have been the weak link. I recognized that the old linkage setup wasn’t as slop-free as say the SBXC elevator, but first two flights didn’t show flutter problems during normal flight speeds. Making the linkage slop-free is a top-priority on this model.

Not sure what the plan is for the Windex at this point...

More pictures of the aftermath in the Gallery

I've been doing a little reading into what would be necessary for an honest to goodness world record attempt. The FAI is the officiating body, so I found their official rules for F3H: Cross Country Soaring

The F3H section is for Cross Country Soaring competitions. In particular regards to the aircraft, it must comply with section 5.3.1.3 which I have copied from the January 1st, 2008 edition:

In particular, item d. above limits the use of telecommunication devices from the aircraft to the pilot-in-command and was added in the 2003 edition of these rules. This prohibits use of a variometer system (i.e. Skymelody, Piccolario, Wstech, Ace Thermal Sniffler, etc) in a cross country event. Ironically, Joe Wurts' record in 1989 was well before this rule was enacted and his use of a thermal sniffler was grandfathered in.

Also in particular, section 5.H.1.g is relevant: "There is no restriction on the number of controls or sensors." In this statement, a vario is not expressly prohibited, but the transmission equipment of said variometer would not be allowed.

One thing I don't understand fully is if a variometer by 5.H.1.g is potentially the intended "sensor" target, then why would the telecommunications device be prohibited in 5.3.1.3? Backing up, why are telecommunications devices in 5.3.1.3 prohibited at all?

As I currently read the rules presented above, ALOFT does not qualify for an FAI XC competition event on the grounds that it transmits back data to the ground in the form of aircraft telemetry and in the form of a variometer signal. Also as I currently read the rules, if I were to put ALL of the soaring knowledge onboard and eliminate the transmission of the telemetry and variometer signal, then this part of the rules would be satisfied. (I will note for safety purposes this would be a bad idea; thought experiment only).

Rule 5.H.1.g about no restriction on the number of controls or sensors does not seem to prohibit a "very smart servo system." Say when I'm flying, I have a mode switch that is "super thermal camber mode" that also is savvy enough to find the lift, then I consider that within "no restrictions on the number of controls or sensors."

Am I off base here? Was the 2003 change perhaps NOT intended to prohibit variometers from XC flying? Do folks still fly with varios at FAI sanctioned events?? Does a world record attempt necessarily FIT within the F3H rules or does the vehicle merely need to qualify as an "aeromodell" to set a record?

Dan

PS: I know ALOFT is over the 5kg weight limit, but that is relatively an easy thing to deal with.

Update 11/3/08:
After hearing back from several folks including the VP of LSF, I have realized my mistake. The wording in rule F.5.H.f states to follow rule 5.3.1.3 "with regards to vehicle size and weight," which excludes the part about telecommunications. So indeed variometers are allowed in F3H rules and the world moves on.

Now if I could just shave that pesky weight to get down to 5kg...

ALOFT traveled 30.2mi goal-and-return (60.4mi round trip) on October 5th, 2008 in approximately 3hr 42min. This beats the FAI record of 24.3mi set in 2005 by Gary Fogel by over 5miles!

Also, ALOFT had to travel more road miles, so it also simultaneously qualified for a 100km distance pin from xcsoaring.com by covering 114.14km (70.9mi) worth of turn points!

Also, ALOFT set a new personal best flight time of 4.5hrs, over 98% of which was autonomously controlled by Dan's custom soaring software and the Cloud Cap Technologies Piccolo II autopilot.

The flight report...

We arrived at the field around 9:30am and kept our eyes skyward to look for the normal buzzards that are our first signs of thermal activity, but none were to be found. We feared the day was going to be a bust. Pressing on, we set up the winch, assembled the aircraft, and put together the ground control station in the bed of Brady's pickup truck.

We figured we needed to get moving around 11am to have a good shot at the record, so all the waiting around was starting to get unnerving. Finally, right about at 11am, the birds started cruising around, though mostly low on the horizon when we spotted them. Wanting to start feeling out the air, we launched at 11:20 into soft air which was not really letting go of the ground. After 6 minutes, Adam and I agreed we should just re-launch and try a different part of the field. Once again, no success.

Before launching the third time, we were looking around the field for signs of where the next thermal would be and Adam shouted "there it is, right there, I see it!" Now normally I would laugh at someone claiming this, but I was wearing Adam's scratched pair of identical sunglasses, so he showed me the same wisp of a cloud he was seeing. "Sure, why not try," I thought. Off the launch and Adam steered directly to the wisp and the vario started chirping away! I leaped into the pickup truck bed and turned on soaring mode. Like a happy bird, the plane wheeled around and started going up.

It took some effort to really get to our minimum course-starting altitude. This initial thermal only took us up to 1500ft before it petered out and the plane started into its standard box search pattern. Coming back down gently while searching, we figured it was going to be re-launch time again. But, ranging out to an area of the field we hadn't been able to reach earlier, the vario started creeping up again and soaring mode latched and started making gains. Working a pair of thermals up to just over 3000ft and getting fuzzy just beneath a cloud, we decided to start the run.

Jumping into the truck and moving to the first turn point is always an adrenaline rush. Especially with the air so spotty. But it was 12:10 and we were really eating into our day's flight window.

Crossing Turn 1 and Turn 2 was a breeze, without stopping even. Then just past Turn 3, we had really eaten into our altitude buffer and were back to the 1000ft range. I started poking at the computer, trying to get the autopilot to realize it had bumped through some reasonable thermals. Finally, we knew we had a 3/4mi stretch of trees coming up and we knew we needed plenty of altitude to make any more progress, so I sent the plane back to the previous waypoint, hoping it would find a thermal. At 850ft with the plane getting awful large, I played conservative and reluctantly handed Adam the manual controls. In the hands of a skilled pilot, he flew back to the bumps he heard on the vario and snagged a great feeder thermal. I pecked away quickly to make the soaring software a bit more conservative and we quickly reengaged into soaring mode. This thermal took us back up to cloud level just shy of 3000ft and we all breathed a huge sigh of relief.

Pretty soon after, we rounded Turn 3 and found a massive thermal under a cloud. This one took us up to 4500ft getting near the edge of visibility and we knew the day had developed. It was about 1:00pm and we were only 15 miles into the 70 road miles we'd have to travel. The pace was only so-so.

We had several areas we knew had overhanging trees that would obstruct our view if the plane wanted to soar on either side. These concerns turned out to be unwarranted because we kept the plane out in front of us a good ways. Most of the obstructed areas the plane was soaring well beyond before we got to them, so it was a simple matter of us driving down the road watching the plane cruising straight ahead. Flying ahead of us, the antenna in the glider's vertical tail gave our mobile ground station better link anyway, so this was a great position to be in.

I suppose I should say the hours crept by, but time really flew. The plane would catch a nice thermal, ride up on the edge of a cloud sometimes, the thermal would get mediocre, and the plane would set off down the road toward the next turn-point. It was a hurry-up and wait kind of day. Adam and I kept each other company in the truck, feeding flight altitudes to each other intermittently or calling out soaring mode latches as we saw them happen. It was breezy in the truck bed going down the road, so I spend a lot of time yelling at Adam or asking Brady for truck speed changes. When we stopped, little flies would find us and threaten to get in our eyes and we would start to get hot.

Coming into the mid-point turn-around, we were fairly happy with our altitude but dissatisfied with our pace. It was 2:10pm and had taken us a bit over 2hrs to cover 30mi. From the soaring forecast, we knew the soaring day was going to be over at 5pm, so we were getting concerned about having enough time to get back. Brady made a meager suggestion to fly the plane slower when at altitude, to cover more distance without burning as much altitude. We had been really hammering the speed ring to keep the plane very fast (~40kt) between thermals, which really ate into our altitude quickly. I settled on shooting for 30kt inter-thermal speeds. Oh how this helped! We cleared the next 19 course miles in just under 50 minutes.

The last dozen miles to get home were a breeze. The day had fully developed and the thermals the plane found would take ALOFT for a ride back to cloud height. One super sucker even hauled us to a tad over 5000ft, a new personal best and actually still quite within visual range when directly overhead. We easily cruised the last miles and had almost 4000ft still coming into the last 6miles. Adam and I were holding back celebration in the back of the truck but we all knew we had it made. Cruising the last 2 miles still holding over 3000ft, all the excitement finally came out and we were honking the car and hi-fiving and just enjoying the moment. Landing was no-issue and we let the autopilot spiral the plane down the last 2500ft slowly so we could make the 4.5hr flight time mark. Landing occurred at 4:02pm less than 100ft from the winch used for takeoff.

I want to thank Adam Propst for staring into the sky and for keeping a watchful eye on the airplane the entire flight, despite bad sun angles sometimes (I owe an apology to his eyes). I also want to thank Brady Baggs for his driving and moral support; a driver really can make or break a run, so a huge pat on the back goes to him!

Some quick facts about the flight:

I have posted pictures from the flight: Mayo's XC Record Flight
I have posted altitude and flight path telemetry logs (Google Earth).

Thanks all who have supported this effort, not limited to the Naval Research Laboratory and the Down East Soaring Society. Thanks everybody! The ball-game isn't over yet ;-)