Stretch Aegea (Supra Wannabee)

I put together an Aegea from the Phil Barnes wings in the Spring of 2004 to try out the latest Drela airfoils for TD.  I now have far too many Drela design based RC sailplanes including 3 Supergees, a Bubble Dancer, a 3m Skydancer and my own design SpinCycles.  The Aegea I have came out at 67 ounces ready to fly, so it is pre-ballasted to the mid range of the recommended weight.  I have flown it in winds up to 25 mph with only 7 ounces of extra ballast so the speed range is very good at this light weight. To see just how slow a TD plane can fly I decided to build a stretched Aegea to 136" and build it far lighter than the commercial production wings. I also had the chance to observe the performance of the Supra at the 2004 F3J WC flying against some of the world's best pilots and planes and came away with the desire to build one. The real Supra construction files are here.

By the time anyone reaches this level of construction experience they will develop routine methods for building common items like joiner boxes, booms, and tailfeathers. Many other sources of widely varying information on vacuum bagging and molding are available on the net so there's no need to regurgitate that info on this particular plane. The major difference in the home build planes versus vacuum bagged production wings is the spar system. 

Whenever I build a plane, I take shortcuts to ease building and to use the materials/tools I have on hand, using techniques that have worked for me in the past. Some little changes are incorporated into each new plane that is built. Having designed, built and competed with my own composite sailplanes, these choices are simply my own decisions without criticism of the original design choices. On the other hand, I also realize that the changes will mean that the plane will not make the specified weight targets.

These are the changes I chose to make:

1. Planform is a a stretch Aegea, same chords and panel breaks, 136" span (135" projected) but with the forward sweep of the Supra. This means that the aspect ratio is basically retained, but the panel breaks are slightly different.
2. Spar and skin design is my own.
3. Tail design is my own, but moments are retained.
4. Pod and boom fuselage is not a pylon mount, I have a shoulder mount open class pod mold.
5. No RDS.

The center panel has a 1/2" wide spar with full balsa/plywood core and the lightweight 1k carbon sleeve. The caps are laminated precured CF strips tapering from 0.124" to 0.062" at the dihedral break. The tip panels have a partial spar 12" long to hold the joiner tubes. The remainder of the tip panel is stressed skin with a unicarbon leading edge and kevlar trailing edge and hinge. I put two 1/8" wide weblets under the top skin only. Joiner rods from 3/8" diameter CF are straight and the joiner boxes give the dihedral at the break of 6 degrees. The center panel is flat. The main skin is all bias Kevlar, with doublers and CF reinforcements at the stress points. Root ribs are 12# balsa. 12# balsa is easy to get, just ask for the Midwest balsa in the rack of your LHS. 

This shows the panel skins (bottom). The spars are full depth and grouted into the core. On the tip panel, there is a 5" overlap from the partial spar to the uni-CF skin. Top CF skin is 4.9 oz. uni and the bottom skin is 2.9 oz. uni. There is no finish layer of fiberglass on the wing to save weight. All Kevlar is 1.7 oz. 

I tried the integral skin hinge method for the first time and like the results but dislike the extra work. Only the center panel flaps have these, the ailerons use the Barnes skin hinge method and I have faced the aileron gap. This tool was made from a block of oak with a snap off blade at an angle to cut the flap free. You only get one shot at this cut so make it early in the core prep.

With too much time on my hands, a proper stab V-mount mold is needed (read about the issues I had on the first day). The platform mold is made from 6mm Lexan with 3/4" oak backing. To cut the groove for the pivot tube, the poor man's milling machine is used. Just chuck up a Dremel ball end cutter in the drill press and use a fence to cut the groove. To make the stab platform, three layers of 5 oz. CF are used and the brass pivot tube is buried in the groove. The V mold is made from 3/4" oak cut with a thin bandsaw blade. When making parts, the molds are clamped into a vice to squeeze out excess resin. So this design does not have two hinge pins, the hinge pin runs from one side of the V-mount, through a brass tube in the pivoting platform, then into the other side of the V-mount arm. The same block mold is used to make the CF control arm with the molded in pushrod hole. The control arm is L shaped and glued to the stab platform. Finally a tiny plywood block is tapped to engage the hold down bolt.  Making the V-strut mold and the platform mold, then a couple of these stab mounts is virtually as much work as the typical ARF. The design load on the mount is 20 pounds, which seems unlikely to be achieved in any flight condition. Using a bathroom scale and nerves of steel, I put a compression load of 30 pounds on the mount without breaking it.

To prevent dents and rash, all my planes have fitted boxes made out of 1" extruded foam. Little pockets and holders make sure that you do not leave joiners, ballast and hardware at home. 

Equipment is as follows:

• Aileron servos - Volz Wingmaxx HP
• Flap servos - JR 368
• Tail servos - Multiplex Digi Micro Speed, these servos are CRAP. Read this if you have them or are considering buying them. Replaced with JR368.
• Receiver - Futaba 8 slim.
• Battery - 5 cell Kan 1050 NiMH

The tail uses the HT series airfoils. The rudder is Kevlar with CF skin spar and the stab is fiberglass because I ran out of Kevlar. I mold my own V-mounts. Pushrods are 0.050 CF rods with bondable Teflon sleeves, run on the outside of the boom. The boom is a tailboom.com 5 layer BD boom that I have stiffened with 1/16" balsa shear webs at the pod end. Rick Walba says that a heavy weight open class boom will be available early in 2005.

Final weight is 54.2 ounces (1537 grams), with nose tooth and heavy nosecone. A 11 oz copper/lead slug can be bolted under the wing.

With winter closing in, there's nothing to do but build another one. This plane is really a tremendous floater with some range. Even though I did not make the target weight, it is still 20% lighter than any other plane of this size, stiffness and strength.

In summary this is why you build composite planes, to build a design that is cannot be purchased (good luck finding someone to build one for you), that in some ways will outperform anything in the market. Thanks to Dr. Drela for putting out another interesting plane.

Flying (Sept. 2004)

The following is intended for your benefit, learn from my poor decision. On my (old)  V-mount the stab is removed by sliding out the pivot pin and detaching the pushrod clevis. This gets to be a pain since you have little room to work between the struts. The specified design is better but I did not want to make that complex a mold or purchase the part. To make the stab easier to remove, this plane switched to a ball link on a stud. Flexibility came into the linkage from the plastic ball cup and more importantly, the pushrod is offset on the control horn allowing it to twist under load. The centering of the stab was solid but there was some springy flex. After two moderate launches, I decided to give the plane a max tension launch into a 10 mph headwind on my competition winch. Right at the top of the launch before the zoom, the stab fluttered and broke the pivot tube from the stab.

Slight digression - you can fly an open class sailplane without a stab by rolling inverted and using crow to modulate the pitch attitude. Rudder and ailerons work as normal. I've now done this four times, on a Thermal Eagle where an aileron broke off and took out the stab, a Bird of Prey where the stab fluttered in a dive, a Maestro where the chute hit the tail and this time.

The plane was landed and the tip panels creased slightly from striking the ground first. Other than that, no damage and the plane flew the next day. Addendum: I now fabricate a proper stab V-mount with the pivoting platform so that the pushrod does not have to be detached.

After about 5 hours flight time on the model, what else can you say about a 1105 sq.in. sailplane with radically thin airfoils that weighs 55 ounces? You would expect it to launch hard with that big wing. You would expect to easily out climb traditional competition designs in thermals. You would think that it would not go upwind at all. Well the last one is not true. After reading earlier posts about the camber setup on the AG4X airfoils and many hours flying my Aegea, I have the 4 flight modes on the wing. Launch is normal. Speed mode is reflex -2 degrees and most of the flying around is in this mode. Thermal mode is 4 degrees of droop and plane slows to jogging speed. I added a reflex speed mode where the flaps are reflexed 4 degrees and the plane moves out much better when flying fast upwind.  

I have the plane set up quite stable at 32% of MAC, so all of these trailing edge settings require significant elevator compensation, but the plane flies very well in each configuration. The tail volume is at minimum so the CG is not as rearward as any of my other planes, far too much correction is required with the rearward CG when trying to do a minimum sink thermal turn. The plane will pull out in any reasonable dive, so fast cruise configuration requires a down elevator preset. For flying back upwind to the field, the plane is set up for a slight tuck, then when the plane disappears coming at you, slowly pull back on the elevator until the plane is visible again. This way the plane will continue to penetrate while you are trying to regain sight of it.

November 20, 2004

The weather is dead calm, 100% overcast and 4 ^oC, a perfect day for trimming out planes for launch stability and hang time. First up is the recently acquired Escape, fiddling with the tow hook, camber and a minor adjustment to the CG. There is some gentle lift where the plane will circle and lose 4 or 5 feet per circle at the slowest thermal turn.

The Supra is next, looking to maximize the zoom off the top. The lift is now so weak that you cannot see the gain or loss in height unless the plane is low and close by. Whereas the Escape could not climb, the Supra will gain a tiny amount of altitude in the mild lift, and I manage to circle for 10 minutes while gaining less than 50 feet and drifting 200 yards. Any adjustment to the camber is demonstrated by the sink rate, so I play with turn radius and camber setting to find the most comfortable minimum sink configuration.

After molding the new stab platform, and the repairs to the old stab, I bagged a new stab with at higher aspect ratio and 96 sq. in. of area. The old stab at 85 sq. in. was a little undersize for my preference. I could not get smooth thermal turns unless the CG was forward so a larger stab will allow the CG to be further back, plus some weight can be saved. After 4 hours of work building and refitting a grand total of 16 grams was taken off the plane. The time would have been better spent watching Endless Lift thrice, but we'll see how it flies.

December 2004

Another identical model on the building table, a little cleaner construction job since I can putter away at the tasks over the winter. Finally got a reasonable paint job on the wing. This one has a much stronger home made boom, 79 g with 46 g of CF as compared to the 37 g (total) Bubble Dancer boom.

The stiffer boom resulted in the total weight of the plane increasing to 55.6 ounces. Still a very light plane at this stiffness and 1105 sq. in.

Differences in weight are primarily due to the heavier boom, heavier radio components, dropping the 1.0 oz/yd Kevlar and a balancing problem with the heavy boom. Supra #1 has less than 2 oz. of nose weight. Relative to the Supra the wing chords are 3% larger, there's 5% more area and 2" more span. The spars are stronger with more carbon, full balsa/ply cores and all Kevlar is 1.7 oz/yd including the tails. 

January 18, 2005

Stretch Aegea # 2 was finished the first week of January but it did not fly until today. This one still came out at under 56 ounces and is built much stiffer than the last one, and probably 3 times stronger than the Barnes Aegea that I flew last year.

The earlier stab design appeared to be too small, so this plane has a higher aspect ratio stab and 7% more stab area. This guess proved to be correct since the CG of #2 is at 39%, a full 3/8" behind the optimum CG of #1.

The stiffer wing and boom gives more control during the launch, so the zooms can be even more violent. I did a load test on the stab and mount so it won't break off in the bucket. Float of the plane is as to be expected, plus more droop programmed for landing mode has the plane slower on final than any open class plane I've seen.

Here is a minimum sink test flight on Model 2. Temperature 11 C, Pressure 89 kPa, Wind 1.4 to 3 m/s (est.), sunny.

You can see that there was very little vertical air movement. The launch was from my winch with 400 feet of line to the turnaround. Launch height is 548 feet with a 228 foot zoom, and I leveled off about 10 feet too soon.  The plane was doing at least 70 mph at release. There's a crosswind so the initial drop is getting turned into the wind. Total flight time was 5:45. Average sink rate, 1.38 ft/s excluding the landing setup. Assuming a Cl of 0.8 results in a forward speed of 22.7 ft/s (6.92 m/s) and a L/D of 16.4. Polar for the AG41 has a Cl/Cd of 50 at Re of 90k. Camber of +2 degrees was on for the entire flight after the zoom. At the end of the flight, flaps were used to land, hence the sudden drop. These altimeters are interesting tools. Unfortunately there's no other data out there to compare this performance against so I will continue to test my own planes relative to each other.

So the plane will make 6 minutes in dead air off a very short winch line. With a normal 600 to 700 foot winch line,  7 minute tasks will be a piece of cake and we only need an 830 ft. (253m) launch to make 10 minutes.

Lots of chatter on Allegro-lite and RC Groups, but nobody has stepped up and built a Supra yet. Kennedy (Ava guy) says he will sell a molded version, but the low weight is 90% of the advantage of this design and a molded model isn't going to come within sniffing distance of the target unless there's a huge leap in technology. Look out, I have two of these ultra-light Aegeas for the upcoming contest season.

Return to Main Page

copyright 2004 W. Man-Son-Hing