Thread Tools
This thread is privately moderated by BAF23, who may elect to delete unwanted replies.
Dec 21, 2015, 05:08 PM
The sky is the limit
BAF23's Avatar
Thread OP
Build Log

Fournier RF5b Sperber scale1:6




Although I had made up my mind not to buy any other model before I finished all the ones awaiting in the attic, my eyes popped when I saw the picture of an ultra-rare 2m80 span Fournier RF5b Sperber advertised for a reasonable price on a well-known German RC forum.



Besides having met an old friend who bought a full size RF5 and promised me a ride in 2016, I also remembered how often there had been suitable soaring weather but nobody to tow during the summer months. I had hoped to fill that void with the foam Hype DG1001M with retractable engine, but it proved too much of a toy and after selling it I was on the lookout for a a suitable traditional material powered glider with a motor in the nose, preferably in the 3meter class that I also could use it on the slopes, and representative of a real motorglider. I immediately wrote a mail to the seller and he rapidly agreed to sell it to me although he had two more offers. Two days later I drove to Germany to pick it up amidst a cellar full of other models young and old.

The seller had bought the model years ago from a German living in the alps, but had no idea if it was a kit or homebuilt because you never saw another one although the fuselage was in strong polyester/fiber. The model was more than complete and ready to fly except for a battery. It had an expensive engine in the nose, an ESC, Graupner servos, yellow, blue and clear canopies, two horizontal tailplanes and two different wingspar joiners. The gear was fixed with removable scale type outrigger pogos and besides the original fixed prop with scale nosecone, it now sported an Aeronaut folding prop. Both wings and tailplane were stored in a bubble protector, and the fuselage rested on a custom made wooden cradle. As soon as I got home I made a detailed inventory:
Span 283cm, length 135cm, weight dry 2700gr,
Graupner Ultra 1300-7 brushed engine, Wotan 20v 5 amp opto ESC, Servos: elev: Graupner C508, rudder: Graupner C507, ailerons: Vozz mini standard, fixed prop 12x8, fold prop Aeronaut 11x6,5



After searching the web I found out that the model was a 1976 kit from a German engineer who later became famous in pioneering model helicopters of all kinds. Here are the links for the RF5 pages in his 1976 catalog
Http://www.dieterschlueter.de/Katalo...0Sperber-1.pdf
Http://www.dieterschlueter.de/Katalo...0Sperber-2.pdf
The kit had Abachi covered yellow-foam wings with a one-piece center section and two outer panels as per original. I suppose for ease of transport the builder of this model at some time modified it to removable complete half-wings with obvious different covering material in the vicinity of the aileron servos (or installed different servos). The original blue tinted thick and heavy canopy was intact but had been replaced by a much thinner (thus lighter) yellow version that covered twin pilots who sat too deep in the cockpit and watched an instrument panel including a Mach indicator and upside down needle and ball. The model had been decorated to portray the prototype full-size Sperber. A much lighter horizontal tailplane had been used but it looked extremely flimsy and anyway the tail surface covering left much to be desired.



It could be flown as such, but was obviously so old that an iron would never get the unknown material really tight. That said, I still was very happy about my purchase and gave priority to my new toy to get it flight ready in the attractive elegant colors of the 70's scheme of the Belgian Air Cadets PL70 of which very few pictures could be found.




The aft fuselage job

To produce this I first had to make sure I could get the damaged clear canopy back into shape. I first painstakingly removed all the remains of the contact type glue that had been used, including over both canopy bows. I then used my hair blower to get rid of all the bulbs and creases on the port side and after some time achieved an acceptable result. I now could use the knife to pry the yellow acetate canopy off the frame and remove the pilots from their base plate. The scale of the pilots looked good but the builder just glued their torsos and half-arms to the base and painted different sweater colors on them. Both bearded pilots looked like twins and were unacceptable to man the Air Cadets airplane. These figures were made of thin plastic and my first attempt in cutting the beard off with a knife resulted in a gaping hole in a jaw. I resolved the problem by applying lightweight putty over the beards and long hair, and also used that material to make ears because Cadets were not allowed to wear their hair over their ears. Following picture of the PL70 in the 1988 more visible (but less attractive) scheme shows that the pilots sit very high under the canopy



I thus used scrap foam under the pilots to elevate them in a more scale flying position, the back seater needing 0,7mm and the front seater twice as much. Lengthening the torso was peanuts, but now the arms came into full view above the bottom plate so those had to be fully constructed, as well as the sticks and their fingers. The instrument panel received an update including anti-glare bow.



During that time I also dismantled the fuselage and tail, only to find that the old decoration was nearly baked into the white painted fuselage and couldn't be removed without traces by peeling off, even using the iron. The clear protection material came off much too easily leaving the old paint sticking to the soft wood underneath. That not only occurred to the markings, but also to the material applied over all the wooden surfaces, causing a real mess and much time to remove everything to cover on blank wood again. That wood showed so many deep nerves that I used lightweight filler over it before even attempting to cover it with new white Oracover. Needless to say, I quickly realized this quick restoration rapidly expanded into a full overhaul. An intermediate picture illustrates the works in progress.



It took another full day to just get rid of all the paint on the tail surfaces and a first attempt on an aileron using liberal amounts of acetone. Filling and sanding the delicate surfaces was alternated with painting the interior of the cockpit and pilots. It was obvious the old balsa and Abachi wood had to be treated with care because of their 40years, not that there was any real rot, but it felt soft and brittle, complicating the sanding. Metal parts such as the gear showed superficial rust indicating the model must have been sitting idle for quite a while in humid environments. After the purchase of a thin but very long screwdriver I was able to loosen the screws that held the fixed gear against a bulkhead. That was not only necessary to be able to cover the fuselage again, but also opened the door (pun intended) for modifying it into a retractable gear. I then finally had the tail feathers ready for covering and I decided to start with white Oracover on the rudder.



I tried applying a one piece flap around the back and worked the heat iron on low temperature to make it adhere. So far no problem, but when I dialed the temperature up to obtain shrinking of the material I sensed something wrong when I started applying it over the Oracover surface. The wood that I painstakingly had straightened and smoothened with filler didn't look or feel straight anymore, so I stopped the application and tore away one side of the cover. Inside was a disaster, the thin wood had delaminated from the much swollen yellow foamish core that now felt very spongy. Normal Oracover methods could not be used with that type of old material and that probably was the reason why the previous owner(s) weren't able to get the wrinkles out of the hollow flight control surfaces and made another elevator without foam. Applying heat to shrink new Oracover over the completely wooden planked foam-core wings was a gamble I wasn't going to undertake.

Two days later a combination of glue, filler and sanding produced an acceptable rudder again and a second Oracover covering attempt was made, but this time using smaller precut areas that were already affixed in a flat manner with a very low temperature iron setting. With only one side done, heat stretching was cautiously applied over the hollow surfaces allowing the heat to be dissipated through the back, and allowing me to closely monitor visually and tactually the reaction of the foam all along the progress. The other side was then undertaken but even more cautiously because the inside heat couldn't dissipate as fast anymore. After doing the white I applied the orange portions in a slight overlap for tensional strength and applied a white cheat-line as per original to finish. It wasn't the prettiest Oracover job but taking the previous disaster and present limitations into account I found the result acceptable and started printing the Belgian Air Force roundels and registration number on white transfer paper.

I allowed the setting spray to dry overnight before cutting the excess support material off with a pair of sharp scissors. When they were applied on the rudder it all looked very pretty so I gathered my courage and decided to use the same careful method for the horizontal stabilizer. Because it didn't have hinges for the elevator I started applying a single piece of Oracover over the complete bottom of the stab and elevator, following the contours of both surfaces that were mirrored with their flat hinge-surfaces aligned in a straight line at the back. After ironing the lower surfaces I was able to lay the assembly flat on the belly in order to apply a single Oracover surface over the top, thereby joining the elevator straight to the stab end, having the Oracover grab the one that ran up and down into the lower cavity. I reckon this method is the opposite of what most people do, but I wanted the overlap seams to be just below the leading edge and a slight overlap at the trailing edge, and found no other way to obtain that. Whilst the very flimsy flat horizontal stab assembly that the previous owner used only weighed 49 grams, the original thicker and more rigid one now weighs 93 grams all covered and decorated. 44 extra grams is peanuts compared to the extra solidity which might come in handy for landings on the slopes, but will require weight in the nose because of the long arm of this aft increase.



Orange Oracover was then used at the tips and a white cheat-line for the final touch. I then painted the control horns white again before screwing them in the old holes that had to be found and punctured, and a drop of CA glue into them to ensure better grip in the old wood. In between the tail works I had removed the inappropriate instruments on the panel (even a Mach-meter, altimeter indicating 25000ft, artificial horizon and upside down needle and ball!) and replaced them with a set of dials that could normally be found in a motorglider like a Winter variometer, RPM indicator to 3000RPM, cyl head temp, altimeter without drum, etc. As period cadets wore orange flying suits the front guy was painted as such, and his face painted much younger than the one of the green flight-suited instructor in the back, the latter wearing longer but graying hair and ranks on the shoulders. Flat varnish plus eye and mouth painting completed the figures which then were glued on the baseplate, after which the clear canopy could then be glued in two times to the baseplate and the plywood arcs. After thoroughly drying, silver colored Oracover strips were applied all around to simulate the aluminum frame of the real aircraft canopy.




The wings

It took me 6 hours to completely strip a single wing to bare wood, reducing the weight from 551 to 515 grams in the process. I got a headache from using almost a liter of acetone in a non-ventilated apartment during winter, and used half a roll of paper towels and 5 rags to remove all the glue/white pigment remains. Lightweight filler was then applied in major depressions or leading edge damage, and after sanding it finally was ready for covering. I later saw that I had better had used filler on the entire wooden wings, but at that time choose not to do so for time constraints, to keep the weight down, and to facilitate peeling off the vinyl if corrections had to be made during the application in such long one-piece surfaces.



The removal system that worked best consisted of first applying very moderate heat with the iron on the overlaps/fold-down areas to first pry-off the sides with a sharp knife, then use the iron and a gentle almost parallel lifting of the old (Monocote type?) material from the wing. Too low temperature and I just got the transparent plastic to lift, too much heat and the glue/pigment separated more between the clear plastic than from the wood. With just the right temperature and speed I was able to take about half of the pigment together with the plastic but still leaving some glue on the wood. The process was further complicated by a previous wing modification to incorporate different (quality) aileron servos but necessitated polyester reinforcements that had just been painted over. Combine that with filler that was used at various parts and you can imagine that whatever was left on the wing after this first removal was just a sticky mess (as illustrated in previous picture).

To reduce the amount of acetone, paper towels and rags, I then used 60 grit paper to sand away the highest buildups of glue/paint. The glue quickly clogged the rough paper and sanding the thin Abachi wood layer had to be avoided. Next came the liberal application of acetone on the paper towels that quickly softened the glue/paint, but immediately filled the paper so that after a few square centimeters another unsaturated part of the paper had to be used. Even careful use on both sides produced an incredible amount of waste before all the paint was gone. I then used 180 grit paper to rub away the still soft glue that rolled up and just made more mess. When the acetone had evaporated I used 240 grit paper to get rid of the last sticky remnants and then wiped the wooden surfaces with liberal amounts of cleaning benzene on rags. Only then was I able to feel where depressions or transition zones between materials could be filled again with lightweight filler. Luckily the fuselage was polyester because all those other surfaces took me a full week (of working at least 8 hours a day) to prepare for new covering (just one plain aileron took more than two hours to strip even without having to use a filler). The toxic fumes of the products and sawdust made my apartment, clothes, hair a mess and gave me headaches for a few days (who said winter was the time to build?). No partner ever could stand so much invasion of model parts all over the limited space of an apartment, I'm afraid this hobby dictates me to remain single for the rest of (building) times, I even hardly found a flat spot to eat in between the airplane parts!

In-between I also roughened the shiny fuselage and discovered the (more recent?) white paint had been applied without first roughening the original more yellowish polyester and was therefore quite brittle and prone to pop, necessitating more sanding and filling out. The vertical tail also showed serious bulbing at both sides and that could not be completely sanded away because it would puncture the balsa. I can assume that only the main fuselage is strong polyester, but am unable to see or feel where it merges into a probably full foam tail with balsa planking. Did the bulging originate when somebody applied too much heat on the German flag marking on the tail? Having sanded to the pure balsa on that spot meant one single paint coat was not sufficient to produce a uniform shiny white over the entire tail, so I applied a second coat of white lacquer.

Although I already had covered many large planes in vinyl before, for some reason I experienced many problems getting these simple wings done, even one side at a time. That still meant pieces of 1m40 that had to be applied in one go, and any minimal deviation in initial angle application meant a large offset at the outer wing. Only two of the five panels on the first wing fitted nicely without wrinkles (3 ended up in the bin halfway the application), but because the white vinyl is so shiny it also showed that the upper wing profile was anything but smooth convex. Just as on the vertical tail there were bulges here and there on the wood, that didn't clearly show on the bare wood. Having not used any heat I can only imagine the bulges had been created during earlier heat applications, maybe even by me when I experimented with heating some packing adhesive over the glue rests in the hope of removing it faster than with solvents. Using heavy pressure with my thumb I was able to push most bulges flat, proof that it must have been expanding foam beneath. It's still not perfect but at least not that obvious anymore and aerodynamically much smoother. Regarding weight I used the scale to measure 551gr as bought, 515gr after complete strip and 569gr after vinyling. 20gr extra weight per wing is no big deal and well worth the strong shiny finish without having to use heat for application over those old foamy wings. Using more lateral excess material I was able to apply the vinyl on the other wing much faster and more uniform without a single crease.

Next entry is about the completion and maiden flight of the model.
Last edited by BAF23; Dec 22, 2015 at 04:49 AM.
Sign up now
to remove ads between posts
Dec 21, 2015, 06:00 PM
Registered User
Thank you for a very interesting story of your project!
Dec 21, 2015, 06:14 PM
The sky is the limit
BAF23's Avatar
Thread OP

Part 2 of the RF5b restoration


Installing retract and gear doors plus final decoration and assembly

After getting the express delivery (too awful expensive) of a set of trailing suspension gear-legs from the Hobby King International warehouse (nobody else I tried in Europe had anything similar in stock), I replaced the plastic looking wheel with a 60mm Kaman wheel that fitted behind the structural firewall. The retract had to be attached somehow to the strong 5mm plywood fuselage frame that had held the fixed gear before. Fore-aft location was dictated by the total gear length behind the firewall and CG location, which in this case necessitated the tie-downs to be bolted behind the existing frame (that had to be cut to allow the gear leg to pass). Upper/lower position was dictated by the need for sufficient room for a 4S4000 battery to be positioned between the gear mechanism and the cockpit pan. Mounting it too low resulted in the relatively large wheel not to be fully enclosed in the fuselage after retraction. Luckily the gear leg on the real aircraft also was mounted with a forward rake and reproducing that allowed me to fulfill all objectives. The spring was soft enough to completely compress and stay there at flying weight with minimal help (slight push or landing absorption). On the ground the model then assumes a correct fuselage angle and the wheel is close to the real position near the CG on the full size aircraft. The strong engine pull on the model could make it tip over if the wheel was this far back. Installing the wheel at that knee angle would prevent the tire from being sufficiently high not to prevent the geardoors from closing completely, and that is why the unloaded spring extension is still essential.

With all these factors in mind, I drew pencil lines where I wanted the gear doors and used the angle-grinder on the Dremel to cut the doors away from the fuselage. Unfortunately the former gear frame also formed the separation between the kit's polyester fuselage and balsa modified wing center section. After delicate removal, the remaining door panels thus had to be strengthened over the material junction. With the doors away, I gained access to the lower part of the frame so I also could make the vertical slit for the gear leg. Previous builders had reinforced that area with liberal amounts of epoxy and the smell using the grinder was awful, alternating between burned triplex, molten polyester and thick epoxy. Neither from top nor bottom was it possible to directly make the 13mm vertical channel in the frame. A combination of tens of 45 angled slices through both top and bottom were necessary to finally allow a circular sanding device to produce the clean vertical cut. 4mm plywood was then cut to shape and trial fitted/adjusted for the retract to be mounted with strong hold-downs in an additional 4mm plywood smaller backplate. Don't forget this plane has only one gear to absorb the touchdown forces, and because of the accessibility the gear has to be mounted and screwed from the top through the cockpit room onto the custom cut angled plywood baseplate.



With the dirty work done I then could further decorate the fuselage and fin. With the latter being adjacent to the highly glossy white Oracover horizontal tail and rudder, I decided to apply white vinyl over the paint, not only to better cover discolorations from the various underground surfaces, but also to better match the other tail surfaces. Aesthetics dictated I used a single large sheet to cover both sides in one go, and I went from the top (better base for the orange) to the complete bottom of the fuselage, the forward junction line being hidden under the elevator. I then temporarily attached the rudder so the lower separation line between white and orange on the top of the fin was perfectly lined-up when ironing the orange Oracover over the vinyl.

I then cut an 8mm wide long strip of orange Oracover which I aligned aft of the cylinder fairings and parallel under the cockpit line till the rudder and elevator push-rod exits under the horizontal stab. Just as for the others I used self adhesive pinstripes, but instead of 2,1mm width as on the wings and tail, the 1,3mm width along the fuselage proportional to the pictures of the full size period Sperber. The front of the line stopped short of the back end of the cylinder aerodynamic cover, from there on I continued forward with Revell nr 30 orange enamel applied with a small brush to obtain the flowing contour lines around those complex curves. The lower part of the cowling was Oracover again that had first been cut after a paper template to follow the various curves and join halfway the hand-painted bow. After mounting the gear leg and the engine cowling plus spinner and prop, I assembled the tail and wings and for the first time admired the stunning looks of this very attractive model.

I completed all that in just two weeks time and finished only half an hour short of loading the model in the car for the annual early winter evening where club members show their winter projects. I got many complements and most couldn't understand how I had integrated vinyl, Oracover, white lacquer, orange enamel and regular pinstripes to obtain such uniform colored surfaces. I really had to point the various areas and seams to them before they believed me. There still was work to be done on the retract and the gear doors, plus the complete electronics, but it was clear that this rare 40 year old model was at the start of a new career with many admirers.



After a day of rest I resumed the works and started by aligning the gear. Having worked so fast, the plywood plate must have moved during the overnight drying and that resulted in the gear not pointing straight when retracted and even worse, to hang up behind the fuselage at the door cutout. The only possible solution was to elongate the holes on the retract flat plate so it could pivot just a few degrees on the solidly glued plywood. I used metal washers to spread the forces of those 4 screws on a larger surface of the plastic gear baseplate. The up-angle also got a slight correction by use of a single large washer on the forward side that also corrected the lateral angle when out and locked. As usual with thicker scale looking gear legs, they sometimes press so hard on the motor spot that the gear stubbornly refuses to extend, but that was easily corrected by grinding away half a millimeter of aluminum on the leg and plastic plus metal on the motor side.

The wings slid over a tube that passed through a slightly larger messing tube that had been epoxied to the fuselage when they modified the 3-piece into 2-piece wings. The lighter hollow fiberglass tube fitted well but looked too flimsy for anything but very conservative gliding. A much stronger full resin tube was provided but had never been used because it had a 10,2mm diameter that was 3mm too wide to slide through the copper tube. As nobody seemed to offer a solution to correct the problem, I had a try with grade 60 emery paper and was surprised that the fibers got shaved without peeling or so. After an hour of twisting and sliding the rod through many emery clothes (they rapidly clogged), I finally could pass the rod through the tube and in the wings. It might have been technically ok, but I doubt it was medically acceptable because during 4 days my hands itched like hell after having sanded and polished the rod with ever higher Emery numbers to get it smooth. No matter how often I thoroughly washed my hands, their tops kept itching for more than a week. At that point I believed the much stronger rod would allow me to confidently fly in more turbulent conditions and perform mild aerobatics just as the full size Sperber is capable of.

After the perfectly working retract it was time to tackle the dreaded gear doors. The flimsy remains of the curved polyester I had cut out of the bottom fuselage had to be solidified to accept the HK door-hinges and actuators. As from the start I had decided to use a 6 channel receiver, my doors thus had to be somehow mechanically operated by the gear retract system. Without having made up my mind about how it would eventually work, I started by gluing the door hinges to the doors at the places they could best take for both solidity and away from the wide parts of the gear leg and wheel. I use a grinder to deepen and prepare the surfaces (holes in the nylon and deep scratches on the flat parts) before gluing them with 30 minute epoxy, the other side of the hinge resting on loose pieces of wood piled at the ideal height so the angle of the doors would match the curved contour of the fuselage at that spot. I then decided to use parts of scrap servo arms to make attachment pivot points for an autonomous spring system that would push and keep the doors open. These were glued on the forward edge (to minimize airflow induced torsion) and just cleared the balsa plates that I had glued to the fuselage to form a mechanical upper door travel limit (the air stream wanting to push the doors further in).

After considering many possibilities I chose for Kwik-link actuators that were attached to a scrap cruciform servo arm that is held down by a spring attached to the lower part of the firewall. This system ensures that both doors are pulled open and stay open into the air stream. Each time I used epoxy I smeared the excess prepared glue on the inside of the gear doors to reinforce them so they would keep their curves even with actuators pushing or pulling at the various attachments. With the glue setting times it took two days to just complete the mechanism that couldn't be tested before the hinges were glued to the fuselage, but after that it would be too difficult to change anything and the access became very limited, a typical catch 22 situation. I then applied Oracover to the outside and orange paint to the inside of the gear doors before actually gluing them to the fuselage. To have a better idea of how I prepared those parts I made a picture before painting/covering.



To close the doors I tried installing a lever that when pushed up by the tire would also push the crucifix servo arm up and thus pull the doors closed. After spending 3 hours on that I had to abandon because the gear servo had insufficient force to raise the gear and pull the doors against the spring force, it just cut-off before reaching the full up position. I then fashioned a flexible link between the aft door brackets that got caught by the retracting leg to pull the doors closed. This worked because the arm distance was much shorter and thus more force ended up on the doors. It wasn't ideal because the doors got pushed open at the front by the spring and pulled shut at the back by trip wire, that caused torsion on the door in the closed position so back home and during transport the gear better be in the down position.. With a workable gear/door system I could move on to making a battery tray. The previous owner had no idea where the CG had been so I decided to completely assemble the model, measure the wing root width and wingtips, stretch a rope between the 25% MAC points (leading and trailing edges are not perpendicular to the fuselage), and mark the inside of the fuselage at the intersecting point which will be the temporary CG location till test-flying dictates moving it further aft. A 33% MAC (Mean Aerodynamic Chord) for most aft CG limit also was measured and marked on the inner wing rib. For my Sig balancer it means that the CG has to be between 7 and 8,5cm behind the leading edge at the root.



I then supported the fuselage at those marks and was surprised my battery needed to be pushed fully forward against the firewall for a rough balance. After disassembly I studied the nose portion with the gear up and canopy in place, and with a flashlight through the nose found a suitable and practicable spot for the battery. I then cut the too light older wing spar joiner to have two tube lengths that were glued perpendicular to the fuselage, one just above the retract motor, and one in the nose at a point where the battery could lay at an angle so the forward top rests against the top of the nose just behind the firewall, the bottom of the battery clears the retract in the up position and the aft battery top rests against the back of the angled instrument panel when the canopy is installed for flight. A plywood guide plate joined both tubes to make it easier to push the battery up during insertion, and another plywood plate was glued perpendicular to the fuselage against the former gear attachment frame to prevent the battery from moving backwards during acceleration or climbs. Small vertical cutouts were made in that plate for the battery power and balance wires. Foam blocks were then shaped and glued to the fuselage sides to prevent lateral movements. That pretty much filled the complete nose section aft of the firewall.



Forward of the firewall I took the motor out of the cone and checked the condition of the brushes. These showed very little wear so replacement is not due soon. Engine mounting bolts were replaced and shorter conical ones installed so they were flush and held firm with blue thread locker. Mounting the motor at an angle for side and down-trust didn't seem possible without making an additional wedge ring around the motor front. Prop holder got seated and secured and the complete nosecone now comes closer to the cowling without any bolts interfering. The fixed pitch APC 12x8 propeller was used installed at that time but I still was undecided if I better used less pitch for better climb under power, or more pitch for less drag during soaring. Between the Graupner Ultra 1300-7 brushed motor and the cti-aichtal Wotan 20v75Amp ESC I found a kind of permanently attached green Multiplex plug where 3 pins were used per wire. This definitely cannot take 75Amps for long, but previous owner used it and I kept it for the maiden. The battery plug was an original Deans which I cut off and replaced by an EC5 for compatibility with my other planes and batteries. With seven identical (quality) 4S4000 batteries I can switch them between my B25, Spitfire, T28 and this RF5 (PC7 will use two in series), which makes things much simpler at the field. I thus can simply swap the EC5 equipped FAS100 volts/amps sensor from one plane to the other, just plugging it in between the ESC and the battery. The signal then passes through a variometer before it enters the S-Port of the receiver so I always have full telemetry available.

Because this ESC had no internal BEC I soldered an external 5V5Amp UBEC directly to the ESC battery lead to power the receiver. So far I do not plan on using a second battery and SBEC with Schottky diode for redundancy. The ESC has no low voltage cutoff but my telemetry will keep me up to date about the consumed power and if I don't get an audio warning, whenever I'll see that engine power slowly drops, I figure that if I close the throttle I still can glide it in with more than 5V out of the 14,8V to power the receiver.



Time to assemble the completed model again and check where it balanced after all the frontal weight increases. Even after moving every gimmick as much forward as possible, it still needed 100gr just behind the prop to fall within the calculated CG bracket. I really wonder how the previous owners flew it because I saw no traces of weights having been installed in the nose. On the previous picture you can see following items from left to right: mandatory identification plate, rudder and elevator servos, small copper tube for aft wing compression and torsion, aileron servo wires that are connected to loose individual receiver extensions for ease of field assembly, perpendicular mounted receiver antennas, spring preventing wings from spreading after assembly, messing tube through which the wing join rod passes through, red FAS100 power measurement unit, black variometer and BEC, receiver top forward deep in fuselage, ESC temporarily held by foam against frame, 4S4000 battery between fuselage glued foam blocks.

At this point I also performed the initial programming of the Taranis transmitter and adjusted servo movements and throws on the model after binding and fail-safe was set. Firmly holding it in place on the kitchen floor, I unlocked the kill switch and slowly opened up the throttle. That engine/prop produced a lot of wind and at full throttle the telemetry reported a whopping 74 Amps. That meant more than 1100 Watts ! (reason for the Mach-meter on the old instrument panel?). With the additional lead on the nose the ready to fly model tipped 3500gr on the scale (400gr battery inclusive). It was definitely overpowered so I will reduce max power to fly it and even more to minimize the torque during takeoff. I then measured the wings and calculated a total surface of 51dm, resulting in a 68gr/dm wingload, just a tad more than my large Ka8 glider ,which seems quite realistic after having talked to period pilots who flew both full-size types through eachother.

Next I dismantled the model and took the nose section off so I could place a custom folded 85gr piece of lead under the motor at the extreme front. 15gr of small lead was then attached to the forward frame to be easily removed if flight testing revealed the model to be nose heavy. After reassembly of the complete model on the kitchen floor (the only space large enough in the house) I delicately turned it upside down to put on the Sig balancer. That is easier said than done with a 283cm wide model under a ceiling of only 250cm. With the gear down it balanced at 30,5% MAC which is more aft than I like for a maiden but still within limits, and with the gear retracting forward it brought it to 29% MAC. After putting the model on its wheels again I noticed that all the weight in the nose completely squashed the suspension spring after the least vertical push, and then remained kneeled. That worked as planned. With the gear up the doors conform well to the original fuselage curves and produce very clean lines for soaring.



That completed the restoration of this 40 year old kit, testflying came next with main concerns being the CG position, power requirements and the evaluation of sink rate for possible programming of reflex on the ailerons to bring the model down from strong thermals and steepen the approach because unlike the real Sperber, this model has no spoilers on the wings.

Maiden flight

Luck was on my side, two days after the restoration and just 4 days short of Christmas 2015, an exceptionally warm weekend (13C) with only force 3 winds along the runway axis, provided the opportunity for a maiden flight. I wasn't alone on the field but seeing 3 gliders being towed gave me the boost to at least assemble my model and watch what the wind would do because it was really a bit too stiff for a maiden. During assembly I discovered the MPX plug between the engine and ESC had disconnected and I just blamed myself not having substituted a connection better suited for the high power draw of the motor. I took the risk for the maiden but it was a bad omen. The rest of the assembly went well and after mounting the battery I performed a satisfactory range check and had a buddy double check the setup and controls. Whilst the towship gassed up, the experienced RC glider pilots came over and listened to the strange sound that probably was produced by the ESC. None of them could help me if that sound was normal or not for such old brushed combo. A 30 second power check initially drew 80amps but then stabilized around 73amps. All agreed this was way too much power for that model and recommended using a 3cell battery in the future, but as I need weight in the front...



With everything powered up and ready to go I didn't hesitate and took my model to the runway before I could change my mind. Staying behind the model I slowly opened up the throttle until acceleration seemed appropriate and kept it there (short of half throttle). With that limited power and the programmed throttle to rudder mix, I had no problem countering the lack of side-trust or torque, and just a tad of rudder was needed to keep it tracking straight. It flew off by itself after a ground roll of about 40 meters and even at that reduced power had a positive climb that I checked with a few clicks of down trim. I felt the effects of the turbulent air in downwind but was quite surprised from the flexing of the wings that I observed. I reduced the throttle to 1/3rd and turned upwind again with the model passing over me for trimming, but to my surprise it tracked perfectly straight and didn't need any aileron nor rudder trim. I opened up the throttle to half and climbed during the following pattern (without touching the cruise elevator trim for the rest of the flight anymore). As soon as I rolled out upwind, I slowly throttled back and the model flew really very slowly (against the stiff winds aloft) before I twice times let it slow to a stall (from which it immediately recovered without wingdrop). Assured about that, I throttled completely back to check how it glided, but knew it wouldn't be terrific because as planned I flew the complete maiden with the gear down (old-skool testpilot habit). Aileron and elevator response felt correct so the throws and expo had been a good guess. Time to open up the throttle and gain altitude for a CG dive test. With the gear still down, the nose came up too slowly but I expect this to be more pronounced with the gear up and the CG shifts slightly forward. Due to the turbulence and all too visible wing flex, I postponed the steep turns and turn reversals for a later flight.

With the main objectives of the maiden completed I came down to check the sink rate in the pattern. I flew a power-off normal glider pattern but even without spoilers and the throttle fully idle, the prop kept windmilling till I initiated an overshoot a meter over the end the runway. Coming in a bit lower on the next approach I still was too high for a comfortable landing. Third attempt was better but the touch must have been at too high a speed and when I bounced about down the runway I opened up for a go-around. Although I had opened up the throttle swiftly to half, I could not see any tendency of the model to swerve left of raise the nose too quickly (which I had expected due to lack of down and side thrust on the motor). Next approach had the glideslope pointing at the threshold and after a gentle flare and slightly holding the RF5b off, it touched down very smoothly on its single wheel before the tail settled and the model came to a stop 10 meters further down without a hint of breaking off to one side or another. I taxied back and was very satisfied about that maiden flight in less than ideal weather conditions. I immediately opened the canopy and felt at the ESC but that wasn't warm at all. Initial field inspection revealed no visual post-flight shortcomings. Puzzled about the wing flex, I asked a club member to lift one wing tip while I took the other up.



We both were surprised the wing flexed more than on a Boeing Dreamliner! I called another member to make pictures of it and we then started looking for the cause of the dihedral increase compared to the normal position on the ground.



First we saw that both wings physically separated at the bottom and somebody suggested to use a tight strap to keep the wings together instead of the spring. That indeed diminished the flex and the bottom of both wings didn't separate so much anymore. After removing one wing I felt how much I could bend the glass reinforced fiber rod and was surprised that it flexed so much as I thought those thighs were rigid, certainly if they only penetrated the wings for a mere 8cm each side. After taking it out completely it became obvious that this was not the stiff strong rod I had imagined. The wings themselves showed minimal flex, it was the rod that was the main culprit. The previous owner flew with a much lighter hollow (but rigid) aluminum rod that normally couldn't be bent but which I didn't trust for overall strength. This full fiber rod will not break, but flexes too much from an aerodynamical standpoint. That had been the main reason for not exploring the flight envelope further during the maiden. Next flight I will use a 9,8mm Graupner solid carbon rod that I found in my stock and already cut to length. I might end up using a full steel rod but that probably would add another 100gr in total weight. On the other hand it will add a ton of confidence to fly in turbulence, at high speed, and allow me to perform mild scale-like aerobatics. Having somebody to handle my camera I took the opportunity to have some pictures taken of the elegant nose when the gear is retracted and the doors closed in the soaring attitude.



I'm really looking forward to see it flyby like that during the next flight. Before that I will first strongly glue the lead in the nose and even add 7 more grams. With the nose off I will also replace the unsuitable MPX plug by 4mm bullets but only after investigating more about that ESC. If I cannot find more info, I'm afraid I will have to install a recent model ESC that has features such as low-voltage cutoff and a possibility to program the prop to stop completely in the air for less air resistance when gliding. A finer prop will be installed to reduce top speed. The throttle range will anyhow be curtailed to reduce the possible max amps but still allow for 30 nose up climbs between soaring sessions. Reflex on the ailerons will be programmed without using an extra channel but I want the model to be able to loose altitude faster in case of strong updrafts, and to provide more sink rate in the pattern during no wind conditions. I still am very much satisfied about the purchase of that rare timeless model and my ability to calculate and (gu)estimate essential but unavailable data for CG, control throws, mixes and expos to conduct a safe maiden flight without too many surprises.

Stay tuned for further reports and hopefully also in-flight movies around spring 2016.
Last edited by BAF23; Dec 22, 2015 at 05:03 AM.
Dec 21, 2015, 09:20 PM
rip
ripacheco's Avatar
Any videos yet?
This is an awesome airplane.
Dec 22, 2015, 07:59 AM
guamflyer - Avid Sloper
guamflyer's Avatar
Dude.. what an awesome restoration..love that plane alot..what a great find..
Dec 22, 2015, 04:57 PM
Registered User
Hi all,

Im the previous owner and first of all I have to say that Im really sorry how much work this plane caused to Laurence. I think you did a really admirable job on the Sperber and I really like the way it looks now.

I may explain a few things on this plane as well. First of all swear I did no paint job or covering to that plane at all. So all the blame should be on my pre-Owner.

I flew it once 12 years ago with 10 cells of NiCd accus. This explains that I had no problems with the 75A ESC and the MPX plugs as it draw below 60 Amps with the mounted Prop. I exchanged to the smaller cam-folding prop as I planned to use 4S lipos so therefore it was mounted already.

On the other hand, those NiCds where a lot heavier and I used the lightweight horizontal Stab which led to the fact that I did not add lead at all. Im not really sure about the weigh at that time but I think it was around 3kg. I even used the light weight wing joiner and had by far less dihydral than in the pictures above.

On my maiden the Sperber had a tendancy to tippstall if flown too slow, so perhaps my CG was too far aft. I had no info about that as well.

Once again: excellent job done and I hope to read more about this lovely plane. Hope you gain your health back soon after all these chemical emission you had to suffer!

Jan
Dec 30, 2015, 10:04 AM
guamflyer - Avid Sloper
guamflyer's Avatar
I'm Following this...YOU Sir.. Have motivated me,,..pulled out outrunner to power a 1/6 or 1/5 scale ......
Dec 30, 2015, 01:23 PM
The sky is the limit
BAF23's Avatar
Thread OP

Essential corrections after the maiden flight


Thank you all for the nice comments. Before storing the model for the winter I decided to correct the major flaws that I discovered during the maiden. I took the engine out and was glad to see (and smell) that there were no traces of over-temperature, which is a small miracle after I ran it up to 80amps (meaning 1000watts+) during ground testing. Further Internet searches revealed a nominal voltage value of 12, but acceptable 10to14volts. As I flew it with 4S, the fully loaded battery pumped out a little over 15v. The motor apparently was designed to take max 32Amps and thus producing only 314Watt. That must have been around what I used during the maiden so I think that the motor is adequate for scale flying. Culprit for the excessive load was the 12x8 prop that I used, combined with my lack of knowledge of brushed electro combos. When I inserted the recently found figures in a prop-finder webpage it turns out that a 11x6 prop seems to produce the best results for this motor and model, causing 32Amp to produce 400watt for takeoff with the same 4S Lipo. With those numbers in mind I even won't change the MPX plug that is safe for up to 45Amps.

That is how I will test-fly it when the weather gets better, but I didn't glue the lead more solidly in the nose because if the power seems marginal, I still could install a 420gr Graupner 1600-6 (compared to my 340gr 1300-7 motor) that a friend doesn't use anymore. Her engine is better suited for voltage (nominal 14 but ok from 9to24) and in that case I just will remove the lead in the nose and drill new engine mount holes (it uses 3 instead of 2 in the front plate). The stronger engine comes with an ESC that has engine-stop capability. I'll await the results of next flight before I decide.

The excessive wing bend had been my biggest concern but the use of a stiff carbon rod didn't boost my confidence regarding the fact it only penetrated the wings for 8cm. The forces created by each wing's 136cm leverage on the small carbon stubs frightened me. Prodding deeper with the carbon rod into each wing, I felt I wasn't really hitting anything solid at the maximum depth. Having no idea about the wing construction I progressed very cautiously, but progression got jammed each time by friction. I then bought a 10mm solid steel rod that I sanded 0,2mm using the Dremel so I could slide it through the fuselage and the wing tubes in messing. It took me hours to get it right but the 31cm long rod still was difficult to engage into the full wing depth because of friction. Further sanding of the steel and use of machine oil whilst pushing and rotating the rod in a wing revealed that I slowly gained ground, but it was very laborious and painful for my hands. I hardly could believe it that on one wing I was finally able to push it in for 11cm, and on the other wing for 10cm, before hitting something definitely solid (sounds like wood) at the end of the tunnels.

That meant that my rod could end up 11cm deep in one wing, but only 7,5cm in the other, totally unacceptable (the fuselage portion for the rod being 13,5cm long). I thus took the grinder and made a new rod of 34,5cm long, which then again was painstakingly reduced by 0,2mm in diameter by hours of sanding and polishing over its total length. The friction at the ends was probably caused by the messing tubes having undergone some bending due to the shorter stubs during the past flights. After hours of labor (including soft forcing) I finally got it right and now both wings assemble like a charm. When testing the bending, the wings now are kept rigid at the root, but flex in a more natural way span-wise. That boosts my confidence a bit because not only is the rod more than 10% longer, it penetrates each wing 20% deeper. This is still only a mere 8% of the total span, but at least an improvement. Negative side of this: the GFK rod weighed 47gr, the carbon rod 37gr, but the steel one a whopping 209gr. That is another 150gr extra, luckily at only one cm behind the CG.

On the last day before New-Year 2014-2015 I assembled the model again and started programming my transmitter so the lack of spoilers could be compensated by raising both ailerons if an increase of the rate of descend would be needed. That could be either high in the air if thermals become too strong, or to loose altitude quickly when men-carrying aircraft or helicopters infringe the airspace we normally use. In the pattern this will allow to steepen the glide angle and float less between finals and the touchdown. On my pure gliders I use the throttle to fully proportionally operate the spoilers but on this motorglider I already use the throttle to control the engine. I toyed with the idea of using the the upper throttle quadrant for the engine and the lower quadrant to raise the ailerons but abandoned it because of the vagueness of the mid-position (Taranis transmitters can technically be programmed that unusual way).

I reasoned that having a proportional throttle, I very well could use that to control the final glide after selecting fixed “spoiler” positions as desired. For practical and standardization purposes I decided to use the 3-position flap-lever on my transmitter. Flap-lever up means no mix and ailerons streamlined with the wings when neutral. Flap-lever neutral means a 20% mix with both ailerons moving up in sync, and flap-switch down mixes a 40% ailerons up with the 50%-differential ailerons that were reduced to 60% of their travel when using only the stick. This means that with full flaps and full aileron selected, one aileron points up at about 45 while the other one is nearly streamlined with the wing. This feature will also completely eliminate any tip-stall tendency when using ailerons at slow airspeeds.

The actual deployment/retraction of simultaneous aileron deflections happens with a 4-second slow duration to give me the chance to compensate for the associated changes in lift during the shake-down flights. To largely neutralize the estimate changes of lift, I mixed a compensation of the flap-switch with the elevator, also with a 4 second slowdown. Test flying in powered straight-and-level downwind will allow me to fine-tune the elevator mix so that the airplane continues its flight path without having to compensate with elevator stick whatever flap-position is selected. I also programmed an audio-warning every 10 seconds whenever the flap-switch is not in the full-up position.

After I purchased an assortment of props I also got a Graupner V60 ESC for brushed engines from my club president. It could handle 16,8v en had an internal 3A BEC. Last but not least, when the two black wires were connected, it had a prop brake. Although the nominal value was 60 Amps, it can handle 100 Amp peaks. After dismantling the nose of the RF5 I soldered 3,5mm bullet connectors on the motor and mounted the ESC with nylon straps on the former gas-engine mount, perfectly in the airstream for adequate cooling. On the battery side of the ESC I soldered a XT60 plug so I could eventually easily substitute a FCS40 for the FAS100 power sensor. On the other end of the latter I made a stretch of wire with XT60 into the sensor, and EC5 for the battery, the stretch allowing much easier connection than before. The ESC yellow control wire runs directly to the receiver but the red wire goes to a Shottky diode which has the other entry connected via an external BEC to a 2S900mah emergency power battery. The output of the diode powers the receiver together with the common negative for both sources.

That setup provides dual battery supply and eliminates the risks of a depleted power battery because the ESC has no minimum voltage cutoff possibilities. The internal ESC BEC delivers 5,6V with only 3Amps and in a normal mode delivers all the power for the receiver and servos. If it becomes short of power (weak battery or more than 3Amp draw) the diode will allow the 5,3v 5Amp of the external BEC and emergency battery to ensure no break power for landing. I programmed my telemetry to give me an audible battery low call whenever the voltage at the receiver drops below 5,4v, and a battery critical call when the voltage drops below 5,05v. By first connecting the emergency battery and listening to the battery low call I assure myself of the proper working, confirmed by the audio disappearing after the power battery is connected, meaning the ESC BEC is still operating correctly. The power sensor allows my telemetry also to display Amps and power in Watts, but even better alerts me by voice of the Mah consumed from the 4S4000 battery so I know when it is time to land or if I have sufficient capacity for another climb. I also installed a high precision (20cm) altimeter/variometer and programmed the audio as per my other normal gliders.

I then mounted 3 different props and checked my telemetry which gave respectively 47, 48 and 53 Amps. For the second flight the Aeronaut 11x6 prop is used, delivering 614Watt at 53Amps which is still above the normal motor rating but 60% less than the setup during the maiden. The idea is to find out how much (if any) extra power this setup delivers for scale flying (speed) and climb between glides (steep slow speed climb). A copilot will then reduce the throttle to read 32 Amps and 400 Watts on the telemetry and we'll see what that gives. I feasible, the prop will be changed for a 10x6 or 10x5 and if below 40Amps, the power sensor exchanged for the FSC40. That will bring the power close to the certified motor rating, and the smaller prop will create less drag when stopped during soaring. The wiring has been cleaned out and all accessories are forward of the CG and cooling air passes over them.

the RF5b is ready to resume further test-flying whenever suitable warm weather and wind-still conditions come back. I'll keep you informed of the progress via this channel, hopefully with in-flight pictures or video's.
Last edited by BAF23; Jan 19, 2016 at 11:40 AM.
Apr 22, 2016, 05:20 PM
The sky is the limit
BAF23's Avatar
Thread OP

Flight tests nrs 2 and 3


Flight test nrs 2 and 3

Spring 2016 had a few flying days early April, but the crosswind made it tricky to operate that motorglider. Assembly of the wings required a lot of fiddling and brute force, something was not sliding well. During the first flight in the new configuration I was pleased not seeing the wings bend anymore. They flexed a bit spanwise but the steel rod ensured the center section remained rock-solid. It still climbed very well with the 11x6 prop and power had to be seriously reduced to produce scale speeds, but I'm puzzled by the amps on my telemetry still showing 63A. Selecting flaps (actually moving both ailerons up) produced the desired increase in sink rate without too much pitch change. The mix I had (gu)estimated was too pronounced so figures have to be reduced.

Those were the things I wanted to know during that second flight, and I decided to land because longitudinal stability was insufficient to my like. I suppose that was the result of the heavy steel rod positioned just aft of the 33% MAC point. Not having fully tested those up ailerons I opted for a no flap (no spoiler) landing and on the third approach I got her down but couldn't stop her before she ran a few meters into the grass overrun, causing her to stand on her nose without any other damage.

Hoping the wind would swing more into the runway later in the day, I swapped the 11x6 prop for an aeronaut 10x6. I also displaced the backup battery as far forward as the power cables allowed. The sun was getting lower and although the wind remained fully cross, it didn't blow so hard anymore and I took her up for another shakedown flight. That prop still produced more than ample pulling and the slightly more forward CG (no measurement on the field) made her at least longitudinally neutral. I allowed her to climb to altitude for a CG dive check without any sign of recovery, just a straight dive. I then retracted the nosewheel (forward) and this time it showed at least a bit of leveling signs during another dive check. The model now behaved more stable so I knew this would be the most practicable aft CG point for this bird. My telemetry on that prop still displays weird figures at full power, 53Amps and only 280Watts, impossible for a fully loaded 4S (14,8v).

The regained longitudinal stability also allowed me to perform a couple of stalls (left wingdrop with severe nose down), and an attempt at soaring under a dark cloud. This smaller prop seemed to produce significant less drag and with the vario beeping dialed up, I got reassuring audio signals that it would thermal. Back down for steady cruising and a better assessment of the spoiler mix. I also went to the stall with half spoiler position. Ailerons could still be used and the wingdrop was much more gently. After entering the pattern I cracked half flap and noted the sink rate to be much better in finals. After one low approach with uncomplicated flap retraction, I flew a second pattern with half flap and landed out of it, using much less runway. During the taxi back I noted that my throttle inputs were not proportional to the engine response, the motor kicking in rather late and immediately producing more trust than needed for taxiing slowly back to the pits. This third flight of my RF5b showed the way to go to improve the model further. Most difficult part was pulling the wings apart from the fuselage, I needed the help of a second person to get that done!

Back home I put the assembled model on the Sig Balancer to locate and mark this most aft CG position which was 7,7cm with the gear down. With the gear up that amounts to 7,5cm behind the leading edge and is sufficiently stable longitudinally. I also reduced the spoiler elevator mixes from respectively -20/-10 to -13/-6 hoping the model will maintain its body attitude whatever spoiler setting. The takeoff weight with the iron rod now was 3710gr which results in a 73gr/dm wingload.

A close investigation to the wing attachments revealed multiple problems. With some more grinding away (maybe a 10th of a mm) on the iron rod it penetrated more freely to the end of each wing's receptacle, but that still didn't help much pushing both wings into the fuselage. The lineup of the big rod wasn't the problem, but the smaller 3mm rods onto the compression tubes were, even without the large rod in place. As the torsional forces of the wings around the large rods were mostly spread on the fuselage fairing over the wings, I dared to grind away half a mil from the short stubs. This was done on the lower side so the wings wouldn't rub as much against the fairing during assembly/disassembly. All those tricks worked well and took away most of the friction but even if the starboard wing nestled close to the fuselage rib, the port one had no gap at the trailing edge but a 3mm gap at the leading edge. With this inflexible iron rod, the gap could not be closed by force. Using a straight edge and measure tape I found out the first wing rib was perfectly aligned, but the 5mm thick port fuselage rib was not parallel to the fuselage and starboard rib. As I didn't want to damage the vinyl over the wings, I saw no other solution than to eliminate a 3mm wedge from the wing joiner to the trailing edge. Here is where I marked the part to be cut.



As the eliminated slice was under the fuselage fairing, it wouldn't be visible and probably hardly reduce overall strength. The thick plywood rib and awkward position underneath the flared polyester fairing gave me little other option than to use a grinder on the Dremel and be very careful not to slip. Too slow a rotation speed and th would jam and catch the different wood types of the plies (at this extreme shallow angle), too fast and the wood started to burn. Off course I then also had to grind away 2mm from the iron rod but in the end the result was near perfect. Before painting the fuselage rib I took the opportunity to cut away any excess glue that was in the way and add PU glue where the joint between rib and fuselage seemed flimsy or broken. Assembly and disassembly of the wings over the rod and against the fuselage is nearly effortless now and looks the par.



I then attached the external BEC slightly more forward from the bottom to the side of the fuselage to gain some extra wire length to connect to the standby battery that for CG reasons was relocated from the center of the cockpit to aside the main battery. An indentation in the foam holding the main 4S battery was sufficient to ensure that the 2S battery wouldn't move in flight. Both of them are kept in position by the cockpit/canopy bathtub, no straps nor screws are required for swapping them between flights. It all looks a bit cramped in that forward part, but CG dictates it. Also note the spring that is used to connect both wings, on this picture is rests between the servos for transport. At the field I only need a screwdriver to fix the single nylon horizontal tail screw and both outrigger wheels, all the rest is done without tools so nothing can get lost. I swapped the 10x6 for an APC 10x5 prop but had to drill it out to 8mm for the axle. A ground run test still shows 51Amps and the same 270Watts power, again very doubtful figures.



With the model now approaching its final configuration and setup, I wanted a better way to stock and transport it. I measured the available room on top of the driving seats of my camper and found out that a single box containing all the parts of the RF5b could be inserted through the wide closet door. With the fuselage being 1m36 and wings 140cm, the design of the box had to be clever to enter the available 98x24 closet door, luckily inside the compartment was 180cm wide and 56cm deep so with a bit of inventivity a clever design could be drawn. During all the flying season, this motorglider thus could have a permanent place in my camper. The advantage of that being that I always will have it with me, filling the gaps whenever there is no towship available or like in Germany only electric flying is allowed between 12h and 14h (for noise). As it uses the same batteries as my Spitfire, T28, Piper, B25 and Pilatus, I will always have a backup plane available on the field. The first attempt for a box looked good and enabled both the wings and fuselage to be lowered next to each other in an upright position.



Problem was that when I painfully inserted that square box through the slit above the driver's cabin of my camper, it rapidly bumped into the slant forward part so there was lots of space that became useless, and the intended door of the box was much wider than the chest opening so it wouldn't open. That meant back to the drawing board again. Not square one because I wanted to make use of the already cut and assembled material. I kept the largest panel and both small endplates, used half the other side as an additional cover over the baseplate to house the horizontal stab, wing joiner bar and outrigger wheels. I then copied the first rib of a wing and carved a thick wooden block to shape after which I glued a rod in it to engage the fuselage tube for the wing joiner rod. This certainly would keep the model from moving around. The vertical tail rested flat on the accessory box door and lightweight isolation panels cut into shape around the tail to block it.



A similar construction was glued to support the nose under one of the fake cylinder cowlings. The result was a slim height front which just fitted against the narrow forward part of the camper top. The endplated were then sawn off to hug the slanted walls and allow the widest part of the wings in their protective cover material to rest on it at an angle, the light tips resting on the other cylinder cowling. When everything is in place and the landing gear is retracted, a pivoting panel is raised and held by magnets at the top extremities, effectively protecting the complete model from anything else that I still can stuff around it in that closet. It all took a week to complete with many trial and error fittings, but results in a very practical way to always have the model with me without sacrificing any significant space in my camper.



Stay tuned for further reports and hopefully a video or so.
Last edited by BAF23; Jul 28, 2016 at 11:20 AM.
May 16, 2016, 04:37 PM
The sky is the limit
BAF23's Avatar
Thread OP

Flight test nr 4 - 5 -6


Early May I attended a 4day international gathering of large gliders on the Tongeren grassfield. On day two I was up early and assembled my 3 models on the field (RF5, Flair Ka8b and Blanik). The wind was still calm and soon a queue started forming but the towship had problems and the operation came to a standstill after only 2 tows. That was the moment I was waiting for to prove the use of a motorglider. Because this was the first time I would operate it from a grass field, I wasn't sure what to do with the outrigger wheels. These were so tiny but on the other hand the field had been well groomed. Would they get ripped off or prevent a tip from digging in and cause the Sperber to pirouette? Having mounted them for the good looks for static display, I decided to leave them on and have a try.

Takeoff was remarkably uneventful and the outriggers seemed to still be at the correct angles. I waited till I was overhead again to retract the gear, that way I could observe it retracted normally and both doors closed properly. I think I will maintain that procedure for a while because I'd hate a buzzing jammed gear draining power and getting hot during half-hour flights. The 10x5 prop still pulled well and allowed a positive yet scale like climb to the 200m altitude, the legal max for the event. I cut the engine and stabilized the glide with a few clicks of elevator. I was amazed how little altitude I lost during soaring.

Being alone in the air I took the opportunity to fly in downwind to check my new mix figures between ailerons partially up and elevator. All 3 settings (still controlled by the flap switch but intended to be used as spoilers) were selectable in increments or just from off to full and back, without any effect on the aircraft pitch. These test were repeated with engine out and at various power settings, demonstrating the ability to use those spoilers at will to get out of too strong thermals, to increase the rate of descend in final, and to perform a go-around at any setting, all without having to retrim in pitch.

I allowed the model to glide in (but keeping the gear up) so I got used to an engine-out pattern and found it very similar to my Ka8, both for speed as for height loss. A meter above the threshold I opened up the throttle again to climb to 200m in a single pattern. I had programmed the 120m and 200m altitude to be voice announced at each passage, but further altitudes and vario beeps only when I activate the feature on my tx (on the regular hook open switch). With the consumed battery power also being announced at 1000,2000, 2500, 3000 and 3500mah, I was able to pretty well follow up my power available situation all along the flight. Besides the trim-mix tests, I performed 4 climbs back to 200m in that half hour flight. Even fully depleting the power battery wouldn't be a disaster because I get a voltage warning and still have the 950mah 2S backup battery to glide back in full control. On this flight it wasn't necessary as I measured 19% battery power remaining after landing.

Because I knew I had little power left, I intentionally landed in pure glider mode without the aid of engine nor spoilers. I dropped the gear in final but had expected it to increase the glide angle more than it did. Although I crossed the fence at only one meter, it floated half the field before it touched down very smoothly. The model continued rolling out for quite a bit but at the end when the speed was negligible the tail lifted and the model stopped on its nose. Walking towards the model I thought that maybe the single wheel got caught in a small hole or so, but I couldn't see anything wrong and pushed the tail back on the ground. As soon as I taxied half a meter even with full up elevator, it nosed over again. I initially concluded my model was not suitable for operation on soft surfaces, but I was wrong.

Back at the pits I was unable to raise the gear completely for stowage. That is when I discovered the gear to be at a too rearward angle, caused by the 4mm pin between trunnion and leg to be bent. The knicked pseudo oleo looked scale but obviously was not sufficiently solid. It was designed for 3kg models, but those normally stand on 2 or 3 wheels, not all the weight on a single wheel. I already had noticed that with the batteries installed, the gear could be compressed and not spring back anymore, but so far haven't been able to find an adequate stronger spring. With the compression moving the wheel back, the leg was pushed even further back when rolling over the grass, and the slight opening between trunnion and leg allowed the pin to bend even further, the leg moving 1cm back into the balsa fuselage bottom. With the wheel that far back, almost under the CG, any power application just nosed the model over.

During disassembly next to the camper, I had the canopy assembly behind me whilst a German pilot helped me keep the model straight, whilst I slowly forced the leg back forward so I could retract the gear for transport. Although I needed an immense force, my solid angled construction base for the retract didn't give in and everything went well. I was so happy that my gear retracted and extended as before, that when I stood up to thank the guy, I accidentally stepped with my espadrille on my canopy in the high grass. I didn't like the cracking noise I heard, but at least the clear canopy seemed unscratched. A closer look revealed the forward interior wooden bow to have separated and broken in four pieces, what a bugger.

Back home I started with the canopy. That had been glued all around and over the bows, with canopy glue that held more solidly than first expected. Getting a slit on the side made it possible to retrieve and already glue parts together, but there was no way to reconstruct the complete bow without ungluing the complete clear canopy. Getting that done I saw the opportunity to change the position of the back seat pilot. I erroneously first had positioned him too far back (dictated by the back canopy former), besides not looking good and seated too low under the lowering back of the canopy, moving him forward would also help moving the CG forward just a bit more. I pried him free, cut slits in the bottom of his arms, added foam under his half-body, grind away plastic and foam under his arms and shoulders, and after some putty and paint glued him in the correct position. What a visual difference that made, it now looks like the instructor position on the pictures of the real RF5b motorgliders. I then scraped away as much old glue as I could from the canopy and cockpit side before gluing them back together and applying new silver Oracover all around.



I took out the retract and tested and dismantled the complete assembly. The pin between the trunnion and leg was a few mm too long because I had shortened the alu leg. Having no 4mm piano wire in stock, I used part of an old screwdriver bit which I first cut to the correct length before grinding the indentation flats for the screws to grab on. That is when I also discovered only one screw through the trunnion axle held the pin, contributing in the pin mishap. Having the wide leg base now flat against the trunnion and kept together with 4 locktite treated screws, plus inserting a 4mm plywood plate where the leg had pushed the floor back, gives me sufficient confidence to try to operate my Sperber from the grass once more. If I encounter the same phenomena, I will either restrain from operating from soft surfaces, or drill out trunnion and leg to accept a 5mm pin.

With the readings I got with this motor/ESC/prop combo and the performance I have seen, I replaced the FAS100 sensor (unavailable anymore) with an FCS40 type (still in production). Only with a full battery do I get 50 Amp at full power, but I do not even need that for takeoff. Afterwards I only see figures around 42 Amps on my telemetry. The 40 Amps sensor produces the same figures and seems to cope with this slightly over-specs power. Next flight will be performed in that configuration and climbs with 40Amps evaluated. If that works I will reduce the throttle range to stay within the specs of the Amp sensor, even with a full battery.

The model reaches its final setup stages and is a delight to fly. Its pure lines on the ground and in the air attract a lot of pilots who comment how much better it looks than the more popular single seat RF3 and RF4's seen at most fields. It took me a while to get it right, but I'm used to need around 5 setup flight before being satisfied about the performance of a new model. This one really is under way to beat all my expectations, both esthetically and performance wise.

Flight test nr 5 and 6

A few days later the weather was finally perfect for r/c flying, with fair weather cumulus and force 2 winds in the runway axis. Takeoff was swift with the 40Amp sensor so it really doesn't limit the current. I throttled back and made a low approach to check the gear was fully up and the doors closed before I climbed to 200meters. I first flew too slow for being steady in the thermals and fell through. After the second climb I really caught on to the thermals and I quickly reached 300m so I got out and lost altitude under the blue skies. At 200m I quickly got another lift so I tried to maintain between 200 and 300m (the full size gliders were also soaring around the field). Even with the spoilerons fully extended I sometimes kept climbing, a sign that I could augment their angle even more. This proved to be the good moment to try some aerobatics at altitude. Loops, cuban eight, barrel rolls and even aileron rolls were easy and looked elegant without engine.

I was surprised the roll rate was still good because due to the spoilerons and dialing in 50% aileron differential, aileron movements with spoilerons IN are not that large anymore. It was no problem staying airborne and even if I hadn't used half of my battery during those two climbs, I came down after 45 minutes of soaring because it was noon and I was hungry. I selected gear down and full spoilerons in downwind so I could perform a powered approach. The RF5 glides so well that you really have to come in low to do that, otherwise the least bit of power breaks any rate of descend. I got it right and touched down on the mainwheel 1/3rd down the runway.

During the afternoon I made a second flight but the thermals were not as strong anymore. After a takeoff and 4 climbs at throttle, and hour in the air, I flew the last part of my test program. I came in for a touch-and-go on the grass, kept the gear down for the next pattern and made a full stop landing after intentionally stretching the glide to just past the hard runway. It came to rest without nosing-over and even was able to turn around and taxi back till on the runway. Careful inspection did not reveal any gear anomalies so the Sperber can now be flown either from tarmac or from grass, both with the outrigger wheels in position. Back home I adjusted the spoileron throws and the throttle range to limit the power to 40Amps.

In June I took the Sperber along to a large-glider gathering in Wey (Germany). Because their rules prohibit flying petrol engines during the noon hours, the electro RF5b would have been perfect to fill the gap when no towships were in action. Unfortunately, the club used those "quiet" hours to make the extensive noise measurements of all the towships in the middle of the runway, elevated on a one meter platform with their noses pointed at to the adjacent village. I'll never understand the "logic" of some people.

I mostly flew the Sperber in the mornings while towships were still being readied. Flights of over half an hour were common, even performing mild aerobatics to lose height. The soggy ground had prevented the club from flattening the terrain and after a couple of flight my landing gear again caused trouble.

Back home I discovered an outer wing panel also had separated all around where the aileron started (aerobatics or torsional forces of the outrigger wheels?) and although still attached around an invisible wing spar, the outer segment pivoted a few degrees either side. I cut two spanwise 15cm slits in the lower wing, pressed PU glue between inner and outer wings and into the slits where I embedded carbon strips, filled out the remaining depressions and applied white vinyl over the repair. First flights will be without aerobatics until I'm sure the wings are solid again.

Being upset about the recurring gear problems, I decided to strengthen it. The too soft articulated main leg was replaced by a straight aluminum leg taken from the discarded B25 nosegear. As these were too long I Dremeled off some of the thick leg and the sliding sprung portion. The memaining bits were too short for a spring to be inserted in between, so I used a longer M3 bolt to keep both parts rigidly together. That leg had a 5mm hole instead of the 4mm pin of the retract servo. Replacing the servo altogether was no option because the size of the larger one wouldn't fit in my reinforced gear support plywood withoud extensive cutting and drilling.

The easiest solution was to drill-out the trunion to 5mm and insert a new stronger pin to join everything. These few words do not reflect the 10 hours of work with a variety of tools to get everything to work properly. One geardoor hinge also had separated from the fuselage and gluing it back was not easy within the tight space. The gear doesn't look as scale anymore (to a very critical watcher), but hopefully it now will be capable of coping with surfaces other than smooth Tarmac.
Last edited by BAF23; Jul 07, 2016 at 03:28 AM.
Jul 28, 2016, 11:12 AM
The sky is the limit
BAF23's Avatar
Thread OP

The 1:6 scale RF5b Finally on video


Second half of July 2016 I attended a week-long gathering of BIGGS towed gliders along the French-Belgian border at Pottes. Besides my blue Blanik and yellow Ka8b (and the recently restored Twin Astir II), I took the Sperber along for morning or evening flights. On one morning, low clouds prevented serious towing and I quickly assembled the RF5b and brought it online whilst clubmember Bert volunteered to make some pictures and a video shots.



The wind blew cross about all week but didn't hamper the flying much. After I told the cameraman what my flightplan was (takeoff, curved top-view pass gear up, straight pass gear down, low approach, full stop landing, all during square patterns), I noticed two members who flew a tow despite the low ceiling. There had been no event briefing yet, and although the lines of the pilot box etc had been marked on the ground, the vertical elements of the box had not yet been erected. With the cameraman preferring to be at the side mid-fields, I walked to the two flyers and asked them if I could make a short camera flight, describing my flightplan.



They had no objections and asked me if I minded them being in the air. I told them it was OK for me, took my motorglider and carried it to the normal takeoff point for the towships. At that time I saw the Wilga land and thought he would make just another short tow till reaching the clouds. I took off, expecting them to get airborne and climb, allowing me the 3 short patterns and landing before they would return. Instead they performed a low pattern with flyby in tow so I had to adapt my flightplan. I stood halfway between the cameraman and the two other fliers (who were in the future box), definitely not the best best position for what we were doing. As my cameraman asked for lower and a bit further away so the grass and distant trees could be in the same frame to show aircraft movement relative to the ground, I obliged but got complaints about flying low over the 2 pilots on the field. I landed out of the next pattern and although I was convinced I had flown controlled with positive vertical separation over them, I felt embarrassed it had been perceived as dangerous by others.

In the afternoon the cloudbase finally lifted and intense glider operations took place. Towing stopped at 6pm because the local club offered aperitif drinks before the meal (ham smoked on the field). With almost everybody on the drink, Bert and I decided to make another movie in much better light conditions. After the last man left the box, I lined up in front of it and took off for the same flight profile.


Although the lawn had been cut extra short, during the takeoff roll the right tiny outrigger wheel caught extra drag and I had much trouble resuming the takeoff axis. If you look at the following picture it is clear that anything but a flat surface will cause a swing when one or the other of those tiny outrigger wheels are dragged in any soft surface.



The remainder of the flight went as planned and ended in a good landing with spoilerons fully extended (up). All this can be seen on the youtube movie that my friend Bert composed and mixed with a happy sounding tune.

BiGGS Pottes Fournier RF5B Sperber Laurence Adriaensens (3 min 23 sec)


Because I was not happy about the takeoff and still had battery power available, I taxied back to the hold and took off straight this time. After a short pattern I set up for landing again but after initially flaring too much, it dropped from less than a foot with too high a sink rate and the landing gear leg appeared to have sheared off between the straight and curved part. That also caused half a geardoor do shear off but no more damage could be observed on the field.



Back home I discovered that the single too short conical M3 screw only grabbed the two last treads of the aluminum machined leg and just had taken those out. After substitution by a longer screw everything felt tight again, but I will purchase titanium screws before flying it again and use it as a spare in the meantime. After some simple replacements of some bolts and swapping the leg for a refurbished kneeling old one with stiffened suspension, plus minor repairs to the geardoor, everything works fine again but I decided this model is better suited for hard surfaces and will refrain as much as possible using it on soft surfaces during future use.
Last edited by BAF23; Jul 28, 2016 at 11:21 AM.
Dec 06, 2016, 01:03 PM
The sky is the limit
BAF23's Avatar
Thread OP

Crash in high trees


On a subsequent flight at the airfield of Zwartberg I was thermaling pretty well and reduced the airspeed by trimming back quite a bit to remain in the narrow updrafts. Another pilot asked if I minded him flying some aerobatics over the runway and I told him that was no problem. I was slowly loosing altitude above the downwind area and kept one eye on the other guy. At one time I really started losing altitude fast and decided I'd apply climb power as soon as I rolled out away from the other guy. I couldn't hear my engine because of his' but already opened my throttle halfway (forgetting that on this regulator/inrunner combo it only started running about halfway the throttle range. The inevitable happened, 45 before rolling out, the backtrimmed Fourier with still no or little power, tip-stalled and plunged down vertically in the trees before I could recover.

It took 3 of us a lot of time to find the RF5b and only after applying some power we saw a glimpse of it in the crown of 60ft trees that were too thin to climb into. The model looked intact from that distance but terribly stuck fully nose-down with each wing into a relatively thick branch. The trees were close together and after felling of the one with the glider in it, it tipped over 20 before getting stuck against another tree. Just before sunset the second tree also was cut and both fell together in a very cracking and dusty fall. We walked around to the fallen crowns and initially couldn't even find the model but after taking away some branches we discovered it, laying flat and apparently still in one piece. More delicate branch and leaves removal freed the model, with the gear down (failsafe position) and no apparent damage except for both indentation in the leading edge of the wings during the initial impact. Outriggers, geardoors, fabric covered elevator and rudder, none showed any trace of damage! I dismantled the model and took it home where following picture was made next morning.



I still couldn't believe that the model survived the 90 fall of the 60ft entangled trees just resting somewhere within the thick canopy crown. November 2016 I undertook the repair of the wings which wasn't too difficult (after I purchased a length of suitable balsa leading edge). Minor cracks in the Abachi planking around the aileron of one wing were glued with PU glue because the inner foam core seemed intact. I couldn't find any trace of a main wing spar in the entire foam core.



I then removed the vinyl cover and Abachi around the damaged leading edges so I could inspect the foam core. This was only damaged at the leading edge but no cracks were seen further in the wing and bending the wings produced no flex nor funny noises. I therefore cut away square parts, glued the new leading edge in with a carbon strip extending into the old intact leading edges, then cut 1mm plywood and formed it over the remaining foam on the bottom of the wing, but flat with the surrounding Abachi. When all was dry, I used the cut-away foam bits to fill the voids and applied a liberal amount of expanding wood PU glue to create a new rigid core behind the new leading edges. When all was dry, I Dremeled the new surface to contour a perfectly shaped new top that was then capped with shaped 1mm plywood overlapping on the old solid foam.




The result before after filling out and before vinyling was very pleasing and hopefully as strong as before the mishap.



Assembling the model I discovered the 10mm steel wing-rod wasn't completely straight anymore. The port and central parts were still fully straight but then the starboard side showed a few degrees of displacement. At a friends place we were able to straighten it out but I still found it unbelievable that nothing on the fuselage nor wings was damaged during the bend that probably happened when the model ended up smashing ground in-between the thick tree foliage after the fall.

With its space in the camper now being occupied by the custom transport cradle for the large Ka2b glider, I had to eliminate the wooden cradle for the Sperber and produced a new compact one out of a scrap Isomo transport box (for in the camper's garage). This was necessary so the model could be safely and easily transported with the gear down (easier for assembly) and all the individual components together. Here you can see how compact it all is with one wing not in its slits for viewers clarity. Note both leading edge repairs are invisible from a distance.




Last but not least, during the final mechanical and electronic operations check I reprogrammed the transmitter with a throttle curve with engine response starting as soon as the throttle is opened a crack, and wide slow response till half throttle (for better glideslope control), and linear response for the higher power outputs. This will hopefully allow me to control the model even when I cannot hear the engine.


Quick Reply
Message:
Thread Tools