Thread Tools
This thread is privately moderated by BAF23, who may elect to delete unwanted replies.
Feb 04, 2016, 04:58 PM
The sky is the limit
BAF23's Avatar
Thread OP
Build Log

Piper L21B based on E-flite PA18 Super Cub 25e


Part 1: The assessment

The Cub was never on my list of favorites to buy, but after a PC9 got ruined during delivery due to exposure to a storm on the Tarmac of Vienna, the E-flite Super Cub 25e was close to the reimbursed value and I took it as a compensation. Somewhere it fitted my desire to learn aerotow and fly the Super Cub as still used nowadays by the Belgian Air Cadets as towships . Pilot in picture below is Polle, former F104/F16 colleague who amassed 3000hrs of towing in these Super Cubs and presently flies a civilian one restored by his father in law in the yellow early colors of the Belgian Army AOP's. Although I performed some towing (gliders and banners) in a 135hp Super Cub (OO-VZL), most of my 200 hours Cub flying was done as an instructor (from the back seat).



Being scale 1:6,2 it will be a good match to tow my 1:6,6 scale Parkzone Ka8b that I also vinyled in period Air Cadets colors https://www.rcgroups.com/forums/show....php?t=1959173 . Although a quarter scale Super Cub must be better suited for towing, I really didn't feel like assembling and disassembling a strut equipped airplane on the field each time I'd go fly. This 187mm span model can be stocked home and transported fully assembled in my camper, a major advantage. It also is a nice field companion for my recently restored 1:6 scale Fournier RF5b Sperber motorglider in period Air Cadet color scheme https://www.rcgroups.com/forums/show....php?t=2565923

a) A bit of history

When I flew with the Belgian Air Cadets at the end of the sixties, they towed us with Stampe SV4bis biplanes, which I also fly as a model https://www.rcgroups.com/forums/show....php?t=1904282. In 1975 the few remaining ones were sold and replaced by 6 ex Netherlands Army L21B Super Cubs. Attrition reduced this number over the years but in 2015 they still had 3 of them in use. As in 2014 I still had access to their maintenance hangar in Goetsenhoven, I made tons of detailed pictures because at that time I still was thinking about ever making a model that could participate in the standoff scale Benelux championships. In the meantime I relinquished that idea in favor of a practical scale looking aerotow model that also can be used for fun flying from grass fields. Out of the 9000 Super Cubs built till 1981, I choose s/n 18-3844, ex KLU 54-2444, R-154 (Klu) and since 1975 registered LB-05. Over the years it sported a wide variety of color schemes, registration letter type and size, spinners and most distinguishable air intakes under various cowling's. I settled for the 2014 configuration and had Callie produce a set of scale stickers for me.

b) Model configuration

As usual I wanted an electro powered model but figuring that out for a towplane is not a simple task because all those engine and prop calculators exist only for single aircraft. Adding the weight of a glider and its wing area just isn't possible in those charts and doesn't work as such. Whilst I initially envisaged towing up to 5kg 4m span gliders with the strong Axi 4130-20 motor that I had purchased to power the PC9/PC7 with 8s batteries, I quickly limited my objectives to a maximum of 3kg gliders behind the 3kg Super Cub. I figured the drop-in E-flite power32 with a 4S4000 battery and 13x6,5 prop and a 85amp ESC would be sufficiently powerful for such gliders. For the scale aspect I certainly had no need for the fast almost vertical tows you often see nowadays. This power combo doesn't augment the wing load much above an acceptable 70gr/dm² so that Super Cub type scale flying and landings still would be possible.

Before starting any work I read all 268 pages of the RCG thread https://www.rcgroups.com/forums/show....php?t=1445753 and copied some notes and pictures for reference during the actual build. It soon became apparent that the ARF wooden kit had shortcomings. People complained about the inadequate tailwheel assembly, the alignment of the main wheels, the failing rubbers of the gear bungees, lateral imbalance between the wings, the limited battery area, tail heaviness and weak engine mount glue. Some people reported sudden tip-stall tendencies and difficulties in taking off. After opening the box I not only found poor application of the Ultracote on every single part, but also that one wing had a wing warp that certainly would aggravate tip stall . Having to iron the wings anyway, doing it with the wing pushed down against a wooden wedge to create subtle washout rapidly corrected that major flaw. The box came with a full complement of (heavy) hardware and scale-like details. The quality of packing and transport protection was worth the platinum stamp. The already applied decorations rather limited the choice of final decoration to just that particular airframe, which makes is very recognizable but all too common on most model fields. The lack of pilot and spinner in the kit are a real shortcoming. The illustrated instruction booklet is superb but the lack of a basic tools in US sizes seriously complicates correct assembly for rest of the world buyers. Better take care of every bolt and nut because not a single spare is included and those measures are unavailable in Europe. All this makes me conclude that this kit doesn't merit the Platinum series denomination.



I knew the real Cub had differential ailerons, but certainly not the 80% recommended by E-flite. After analyzing the situation it became clear that this grossly exaggerated correction had been caused by poor design and use of very cheap badly placed aileron hinges. Pictures of a real Piper with full up aileron deflection clearly show that the differential is limited but the adverse yaw largely reduced by the use of Frise type ailerons. The principle is simple, use low placed hinges deeper in the aileron so that with up aileron application, the front of it is lowered into the airstream below the wing and thus augments drag. The airfoil shape of the aileron moving down keeps the extrados of the wing smooth without creating any additional area drag, only higher induced drag due to higher Alfa at that point. If the parasite drag on one side compensates the induced drag on the other, there is no adverse yaw anymore during aileron applications.

With none of the moving surfaces in the kit assembled at the factory, this became easier to correct. Use of the same type of hinges as the flaps in the kit, but placed differently was the way to go. The flap hinges were correctly drilled out to create scale flap movements and the plastic fairing in the gap was not only a nice visual scale touch, but also created a good venturi effect during deployed flaps. During a ground refueling of the LB05 I positioned the stick full to the right to photograph the throws and respective aileron angles, but looking below the wingtip also shows the leading edge of the Frise aileron descending well into the airstream whilst the port aileron shows a smooth curve over the extrados, also note the upper pull pull cables disappearing rather flat into their bulges whilst the lower ones are clearly exposed to the airstream. That is also a result of low placed pivot points.



What first looked like a nice scale touch is definitely not a common sight on Super Cubs and were only installed on the very last produced models. I'm talking about the metal looking ailerons and flaps. As they look more appropriate on Cessna's I decided to get rid of them and revert to fabric covered surfaces. As the L21B's that the Air Cadets have a non standard larger air intake filter, the molded polyester nose will require substantial alterations. As the shape of the rudder on the model is not conform to the original, that will also require alterations. A non-scale towhook will be installed just aft of the cockpit top window and a fake antenna can help hide it when not used for towing. The tow cable passing along the tail surfaces requires the tail bracing to be very solid and the elevator and rudder servos to be of the metal gear variety. As the model has limited pre-drilled space for wing servos, I choose for the recommended Spectrum 4,8kg at 6v A5030 digital servos for the wing and their metal gear A5040 variant for the fuselage mounted servos. The weak bungees for the suspension will be replaced by real springs and the leather cover around by metal airfoil shaped fairings. As I read that most builders needed substantial additional weight in the nose for balance, I will position my thicker battery from within the cabin in the engine mount box and move the fuselage servos forward. The twin metal pushrod for the elevator will be replaced by a single lighter one (idem for rudder) and both elevators linked rigidly to each-other. That pretty well sums up the major works I fist envisioned. The E-flite produced scale interior and external lights were not purchased because they were expensive and non-representative for the aircraft I want to portray, plus I never fly unless I have good natural light conditions and my interior will have to be adapted for the large battery and towhook mechanism.

Part 2: stripping the model

Mid-January 2016 I had collected most the required parts and moved the box from the attic to my hobby room to undertake this modification/assembly process, but first I had to strip it to either gain access or replace kit provided stuff by other materials/parts. The first job was to detach the flight controls and access panels that had been taped to the larger kit parts for transport. Some glue traces from the tape remained on the Ultracote and had to be removed by benzene. I then used a long flat sharp cutter to separate the fake metal plastic from the ailerons and flaps. With much delicacy I was able to separate two complete panels but these were so brittle and well glued to the wooden substructure that it was taking forever and too much wood remained on the plastic , causing me to have to use filler and sanding to get the wooden structure ready for future covering. I then experimented softening the glue with the heat gun, but that also caused the plastic panels to loose their intricate shape. Seeing no further use for those panels, I found out that by blowing 120°C air on them, I was able to slowly rip off the now soft plastic, without any wood coming along. Although neither ailerons nor flaps had any internal framework resembling that of a real Piper, I nevertheless am convinced that applying opaque Scale white Oracover over them will embellish the model compared to the simulated aluminum covers. The stripped surfaces have torsion stiffness that is greater than what I could achieve by making new surfaces with scale type rib spacing. Real Super Cubs are fabric covered but have a metal structure underneath.



Because I wanted to change the shape of the rudder, I removed the Ultracote and that was relatively easy with the help of 80°C air being blown between the wood and the cover to loosen the glue. For the scheme I choose, rudder and elevators are orange so I also removed the Ultracote from the elevators so they later would show the same coloration (on my RF5 I saw the difference when the orange Oracover had been applied over a nude wood frame or over already white covered surfaces. The trailing edges of the tail surfaces looked much too thick and blunt, removing the Ulracote provided the means to sand those trailing edges a lot thinner, improving both the looks and the the aerodynamic properties without losing too much strength. I didn't touch the fixed tail feathers because I wanted to keep their solidity so they could endure eventual slamming of the towline. This picture shows the difference between the upper standard and lower trimmed elevator



A sharp knife and pincer were all that was needed to lift a corner of the many red fuselage decorations, the rest of it coming off with a gentle slow pull. The kit being at least 2 years old, the red outline of each decoration remained visible until it was rubbed down with acetone. Very close scrutiny under certain light angles still reveal a subtle change in white where decorations had been applied but it is so minimal that I don't see the need to cover the fuselage again for that. I also completely removed the red and shiny black Ultracote on the curved deck between the instrument panel and the cowling.



The polyester cowling/nosecone is something else. Here they used a very thick paint that had to be rubbed down using abrasive paper. Failure to do so would have the outline much visible under the new paint. Not only that, but what others describe as cooling vents underneath are in fact just bulges and instead of being cut-open, need to be filled and shaped. The lines for opening the engine access panels didn't match and neither did the fake fasteners nor thick rivets. All these protuberances and depressions also had to be completely sanded away. Furthermore, the specific version I wanted to model had a much enlarged non-standard air intake under the nose and no small flat intake under the prop. For this I found no other alternative than to grind these original areas away completely so they could be replaced by custom produced shapes. Following picture shows the extent of works on the cowling and the amount of balsa coming along when attempting to cold remove the fake metal aileron plastic in one piece. Also visible but described further down is the already reshaped rudder and replacement of too weak suspension bungees by metal springs.



Part 3: Major modifications

1) Rudder

By comparing the following pictures of an unmodified model (left) and real aircraft (right), it is obvious that for whatever reasons, E-flite elected to produce a rudder shape that is not conform at all to the original. Also note the tailwheel assembly of the model being too flimsy and inadequate even for taxiing.



Some corrected the rudder top by gluing extra wooden inserts within the top arch so that the trailing edge could be sanded in a banana line from the top instead of first following a circle. The difference can be seen on previous pictures taken on my green mat. That off course reduced the total rudder area, but that could be compensated by augmenting the area at the bottom of the rudder. Why they decided to abruptly break the lower fuselage line at the rudder is a mystery and has no reason, so I glued additional wood to the bottom of the rudder to reproduce the correct shape and augment the rudder area at that point (also visible when comparing the above composed picture). At that point I also slimmed down the (too) blunt trailing to make the looks more scale (the real one has fabric just over a single tubular shape). The bottom corner was then reinforced to form a solid base for bolting the side-horns to which the springs for the tail wheel steering and rudder pushrod were going to be attached, eliminating a separate rudder horn but more about that later in the landing gear paragraph.

2) Cowling

As mentioned during the dismantling, the nose cowling ended up with two serious gaping holes (see second last picture). The one under the prop shaft was difficult to close because it had to conform to the curves in the 3 axis and any error would be very visible. Because of possible nose-overs I also wanted this to be relatively solid so it wouldn't separate easily. To combine those requirements I took a block of 1cm thick balsa that I painstakingly sanded around the marked hole pattern until the front protruded out of the cutout while the back hugged the remainder of complex interior nose shape for about one centimeter. This necessitated many trial and error dry fittings and when satisfied I used liberal amounts of PU glue to ensure the whole surface would be in contact and all gaps closed. The plastic scrap box also provided the material for the larger intake.



After drying and initial sanding it was clear that the amount of wood sticking out still wasn't sufficient to obtain the correct rounding, so another plaque of 3mm balsa was glued in front of the block. This time lots of careful sanding produced a seamless nose contour that cannot be discerned from a factory molded one. Impregnation with PU glue solidified the balsa and 600 grit emery paper made everything smooth as a baby butt. Balsa blocks were also shaped to conform and fill both depressions at the bottom. Some think they are exhausts and even cut them open, but I have sufficient pictures to prove they are just plain bulges so that is what I reproduced. With those bulges finished I now could carve the new lower panel lines passing at an angle just above them.



A small plastic recipient in the scrap box was given a round nose with filler, painted white, mesh from plaster sanding material glued around it to reproduce the typical air filter containers, a piece of black wire binder glued in between to portray the rubber gasket, and a strip of balsa shaped to conform to the curves of the simulated air filter and the lower intake to fasten it. The intake was formed by cutting open a plastic tablet container and heating it up until it got the exact shape and dimension, including the side lips with which it is attached to the underside of the cowling. This prominent air filter intake under a smooth nose is what sets this variety of L21B apart from the mainstream PA18's. All the sanding had taken away so much material that at some spots the almost clear polyester became much apparent and the differences in underground would make uniform orange paint application difficult. I therefore used two thick coats of white Revell plastic paint as a primer to produce a uniform base, but also to fill the small creases left by the extensive sanding. After every coat I used 400 grit emery to further equalize the surfaces and guarantee better bonding of the next layer. I use that Revel paint because the correct orange color was Revel nr30 and I know from previous experiences it sometimes reacts in a strange way to various subsurfaces, especially over Humbrol paint. A few relatively thick coats of orange were sufficient to obtain a near perfect finish (even using paint brushes because I cannot spray in my apartment) and after sanding away the overlaps on the top of the cowling I applied flat black for the anti-glare panel between the hinge lines of the engine access panels. I must say that I am pretty happy and proud of that nose, few people will recognize an E-flite model in that. Here is a picture of the air filter awaiting to be inserted in the intake lip after a few more rubbing and paint layers to get the cowling smoother. The new panel lines also show well.



3) The wings

The main wing was left as it was when it seemed unfeasible to take out the useless plastic tubes because they had used glue on some ribs to maintain them in position. The forward tubes were for the electrical wiring to the position lights. In the port wing they had installed an additional two almost parallel tubes, one for each adjacent landing/taxi light in the leading edge! This had been cut out in the factory and planked. All this had its weight penalty and when I put the scale under each wingtip (I didn't care so much about the complete wing weights but more about the imbalance) it indicated a 14gr difference. I then made a cut through the underside of the fabric and glued two 7gr lead weights to the outer rib at the approximate CG position. That would take care of the previously reported roll tendencies under wing (G) load .



After ironing the white cover back into position I tried to apply orange Oracocer over the tips but that proved impossible to stretch without air pockets over the factory applied white Ultracote. I thus resorted to cut away all the white cover over the outer rounded wingtips, and left the skeleton nude for as long as the rest of the wing wasn't completed. The flaps nor their mechanism didn't need any modifications so after a trial fit of their slightly shortened pin hinges I put the new fabric on the flaps. Because neither the ailerons nor flaps had their ribs representative of the original, nor lined up with the ribs on the wings, I had decided to cover them with Scale-white Oracover. That is a slightly more expensive variety that is completely opaque because of the underlaying aluminum coat. The underlaying structure cannot be seen anymore, even against sunlight, but unfortunately also absorbs part of the radio signals. I figured with this Piper having a high wing, and no aerobatic maneuvers on my program, the narrow aft area of the wing being a bit more RF absorbing would cause little problems if I kept both inner antennas separated in length.

After covering the flaps I glued them in the kit-drilled orifices (which in this kit-run were perfectly aligned) and by carefully pushing them down differently in the flaps and wings, I was able to obtain a mechanical lock against the wing in the full up position (relieving servo strain under flaps up conditions). When programming the transmitter I found that the flaps throw was excessive and I slightly bent the pushrods so they could be connected (with some difficulty) in the second instead of third hole of the servo arm, the outer hole being used at the horn. Be sure to do such adjustments (as well as the ones for the ailerons) before you attach the struts because these stand in the way to loosen the servo panels if you have to modify the arm angle on the servo.

As described in the previous model configuration paragraph, the kit ailerons not only lost their plastic fake metal cover, but also had to be modified into Frise type ailerons. Medium sized pin-hinges with the pivot point well into the ailerons seemed feasible, but very much weakened the leading edge part because of the gap needed for the pin with the aileron up. I therefore decided to scrape away some balsa over nearly the complete width on the upper leading edge so I could glue a carbon strip . Furthermore, triangular reinforcements were glued along the ribs where the pin hinges would be glued in carved out channels. Because Frise ailerons only work (by increasing drag) if the lower front end is nearly square, I reversed the ailerons to have the shorter edge on the bottom, and another balsa strip was glued over that so it could be sanded square at the lower front.



Having the ailerons upside down meant I also had to cut new beds for plywood reinforcements for the control horns. Although initially I thought of installing a bellcrank through the aileron with the actuator on the bottom and a fake pushrod on the highly visible top, this was not feasible because the aileron servos could not be installed on the same spot as the wing struts (the latter being used for the pull-pull aileron cables on the real aircraft). I thus decided to keep the kit's solution with the actuation horn being placed well inside the strut width. The fake upper cables will be installed at the correct more outer position during a later stage.

The recommended (expensive, noisy and Parkinson-prone) Spectrum A5030 servos were a straight drop in their pre-drilled mounting brackets, and the flap linkage could also used as provided. With my aileron horns mounted further back, the provided long pushrods were too short, especially because I also wanted the servo arm to be tilted forward by about 30° at the neutral point to create the desired mechanical aileron differential and augment the effect of the Frise type ailerons. For sturdiness and less play I elected to use a stretch of 2mm threaded wire with metal clevises at each end. The treaded wire can be relatively easily bent to provide better alignment and clearance for the full down position (servo side). The servo electrical wire had the correct length to be pulled by the rope in the wing into the space for the flap servo. From there on I made a JR servo connection to a short extension which was soldered to a Multiplex wing connector. I then cut half the length of the flap servo wire and also soldered that directly to the same MPX plug.

By cutting away just a few millimeters from the wing opening for the wires, the MPX plugs can be passed in and out through the thick first wing rib. I had read that many builders had problems hiding the servo wires and connectors in the very visible cockpit area so my solution is to have a MPX connector sticking out of the intended fuselage opening. After connecting it with the wing connector (both sides) they then can be pushed into the wing where they remain invisible in the balsa covered cavity between ribs 1 and 2. In case I have to change a defective wing servo, I kept their 3-wire leads intact till the wing plug. On the fuselage plug I connected both negative and both positive pins with each-other so I only had 4 wires going towards the receiver. That way these 4-twisted wires in the fuselage towards the receiver can be glued to the rooftop and painted, making them nearly invisible. I was forced to already make those fuselage wires at that stage because I couldn't connect the servos to my tester anymore because of the MPX plug, and I wanted to adjust all the linkages and neutral points in symmetry before further cover and decoration works.

I then used orange Oracover on the tips but although the flatter lower side was easy, the complex shape on the upper half made it very difficult to apply, especially around the position lights area. After many attempts I decided to cut that upper part in parts to facilitate the application over the rounded edge part, but the result although acceptable, is far from perfect. I found it easier to apply all the wing decoration and stenciling before going any further because this was the last time I could lay the wings flat on the table any way I wanted without any interfering struts. Caliegraphics did a good job but I now regret having cut the price down by having the many “no push” , “fuel drain” and “no step” letters printed on a white background instead of having her print every single letter on its own. This won't work on the orange rudder or elevators so I will make those “no push” on transparent transfers myself on my printer.

Although I had no plans to install lights, the recess in the port leading edge was so vast and visible (even after a coat of flat black) that something had to be done with it. With nothing supplied in the kit, I browsed through various stores an finally found cheap flashlights with 1,5cm reflectors, not that easy anymore because most such lights are now composed by a multitude of smaller LED lights instead of a single light bulb. After extracting the reflector (the rest went directly in the bin) it was too deep and had to be Dremeled off to the limit at its back, plus large angled recesses drilled into the wing so the reflectors sat deep enough to clear the clear plastic cover. That realistic cover (visible in next to last picture) was then glued with canopy glue straight on the Ultracote. This seemingly minor work took much longer than expected but even without light bulbs the result is stunning and looking very scale.



It then became time to tackle the wing struts to finish the wing sub-assemblies. I preferred to do that at such early stage because it has to be done with the model upside down and at that stage the empty fuselage without rudder was easier to handle, and the fake upper aileron cables and fuel caps had not been installed. The critical items are the fittings for the jury struts that have to be adjusted in depth in the wing without piercing the upper surface, and the many screws and nuts to keep all those struts rigidly together so only four have to be used for the final fuselage mating. After I loosely mounted the gear/strut attach tabs on the lower fuselage I slid the wings over the main aluminum spar but one of the dowels didn't pass through the factory drilled holes. I got my precision measuring tool out and that indicated the dowel diameter as 5,2mm. It obviously was not metric but when I selected inches it gave a decimal reading whilst my US drill bits are labeled in fractions. It somewhere must be a drill size I do not have because 3/16th is 4,76mm and ¼ is 6,4mm. When will the USA standardize with the rest of the world on metric because at least it is easier to measure and classify in the tool box. In the manual the screws for the gear mounts were supposed to be driven-in using 1/8th hex wrench but that was too large and 7/64th was the required size, how confusing for a European mind used to just take the next millimeter number up or down, instead of having to guess a completely different fraction of an inch. To simplify things I used a round file to slightly augment the diameter of the hole, case solved.

Attaching the struts to the wings (and fuselage) was not straightforward, just following the instruction booklet didn't work because the jury strut openings were not correct to allow these to be mounted at the required angle to the wing strut fittings. Bending them would have damaged the wing so everything had to be bent and adjusted before fixing it to the wings. The recommended flush mounting of the strut fittings also didn't work, the main struts would have been in a bow, but by leaving almost half of the thread exposed on 3 of the 4 fittings I finally got everything aligned. There was sufficient longitudinal play in the eyelets of the long struts so that washout could be forced by fastening them into the wing so they pushed out. Here again E-Flite expects everybody in the world to have a 3/32” hex wrench and a 1/4” nut driver in their toolbox. Don't they realize the rest of the world went metric since more than a century? It is extremely difficult to assemble those struts without the proper tools (which can hardly be found in Europe). I finally found and bought one of those larger tool kits which includes SAE sockets but go figure, the 1/4” one was a tad too small for the nuts, and next size 9/32” too large, did they use Chinese nuts in this kit? That was another 20 euro wasted on tools. These multitude of little problems add up making the denomination “platinum series” inappropriate for this kit.

The wings were then permanently bolted to their respective struts and after sliding them away from the fuselage I undertook the last job of installing the fake upper aileron pull cables. Position was dictated by the availability of a wide aileron rib close to the strut axis. Cutting remains of servo arms and gluing them into slits I cut in the aileron top was the start. The top holes had been widened to accept nylon clevises with short stubs of carbon. I then made a single hole in the lower part of the linkage cover plastic bulge, used a black pen to mark the rest of the slit, then used canopy glue to fix the plastic on the Ultracote on top of the wings. The length of the carbon is then cut so that it can slide through the hole whatever aileron angle. It is totally nonfunctional but adds a nice visual touch to the upper wings. Another nice scale touch was gluing fake inspection access panels at the specific spots. Those plastic caps, that are intended for covering screws after furniture assembly, were Dremeled flat at their bottom and also attached directly to the Ultracote with canopy glue. I then used a pen to stencil by hand the symbols for fuel drainage on the respective caps, and earth connection marks on the leading edges.

Following picture shows and upper and lower wing with their markings, but the other sides are blank. Flaps and ailerons are at angles that are not representative for flight throws. Landing lights had not been glued yet. Lower aileron actuator dictated by strut limitation, linkage had to be bent twice to obtain a straight push on both the servo arm and control horn. Upper fake actuator at correct scale spanwise position. Green MPX plugs sticking out of the wings to facilitate plugging before they are buried in the wing during final assembly. Kit provided white nylon wing fasteners are huge. Struts are mounted permanently to the wings.



This completes the wing sub-assemblies and cowling modifications, the fuselage parts will be added in a few weeks
Last edited by BAF23; Feb 15, 2016 at 03:14 PM.
Sign up now
to remove ads between posts
Feb 15, 2016, 03:06 PM
The sky is the limit
BAF23's Avatar
Thread OP

Aft fuselage


4) tail surfaces sub-assemblies

In Part 3 chapter 1 I already described how I changed the overall shape of the rudder so I won't repeat that here. Before covering the reshaped rudder I had to find a way to attach the tailwheel spring actuators from the newly purchased tailwheel kit. The plastic dual-side arm that was part of that kit (more about that in chapter 7 undercarriage) seemed very solid but had been designed to grab rudders of about 10mm thickness. As I also wanted to connect the servo linkage to that same arm I started measuring and figuring out lines towards the existing fuselage actuator-guide and expected position of the tailwheel arms. It was immediately clear that to ensure proper spring connection and actuation, the rudder arm had to be mounted as low as possible, and preferably at an angle towards the tailwheel arm.

To combine all those requirements and end up with an aesthetically acceptable result forced me to cut off part the middle rudder arm so the side plaques could be mounted flush into recesses that I then carved on the forward underside of the rudder. On the port side it then rested partly against the plywood that was there to fix a horn, and partly against the strong wooden vertical backbone. The starboard plaque only rested against balsa which would not be sufficiently solid, especially if the rudder got hit by the tow cable during flight. As simple gluing also wasn't enough I reverted to drilling two 2mm holes so I could pass two screws through and bolt both halves of the rudder arm together under the correct angle whilst sandwiching the rudder structure. That still only would provide about 1cm of grab so I decided to make a slit along the bottom of the rudder to accept a length of carbon strip to further spread the forces backwards. Following picture shows all these pieces before final assembly with strong PU glue and finishing with lightweight filler.



Although I toyed with the idea of using pull-pull cables for the rudder (as a real Piper does), this would have required me to undo the existing satisfactory fuselage Ultracote to route both cables front to back. I thus abandoned the idea and decided to use a single pushrod for the rudder through the existing port routing (the photogenic storyboard side with open doors wouldn't reveal any exterior pushrods that way). The kit-provided elevator rods were dual split at the single servo side and each one then actuated a single elevator. Previous builders reported difficulty for the heavy rigid rods to follow the existing fuselage curved guide tubes, resulting in high friction that the servos needed to overcome, not what you want for precise elevator and rudder control. As I also wanted to move the servos 15cm more forward for CG reasons, I decided to relegate those steel rods to the spares bin and use lighter single more flexible pushrods through the existing sleeves. On a real Piper, both elevators are rigidly linked and actuated from within the tail end. Pictures from builders who had stripped the fuselage showed that this system could not be duplicated due to the concentration of plywood/hardwood/balsa in the tail section of the fuselage (to offer good support for the vertical stabilizer).

After much thinking I settled for a rigid link through an additional hole in the tail (as per real Piper) and a single horn actuator on the port side because the fuselage mounted guide tube was closer to the horn on that side (thus less susceptible to bending when pushed hard against a possible dragging tow cable over an elevator). The kit had a clever lightweight system to rigidly attach both fixed horizontal stabilizers to the fuselage with just two carbon rods and four functional metal tail braces. It would have been even better if they had routed the leading edge junction through their scale like opening that is intended for trimming the horizontal stabilizer on the real aircraft, a deviation I cannot find any rationale for. Joining both elevators during the assembly thus required a solution that would be torsionally rigid but still be able to pass through the minimal additional fuselage hole. The position and size of the extra hole on the model was dictated by the existing vertical hard balsa aft end of the vertical stabilizer all the way down to the bottom of the fuselage. This also better be solid because model Cubs are prone to overturning during landing roll-outs on grassy surfaces, and most forces thus end up on this single vertical wooden member. With the chosen hinges pivoting in the middle of the gap, whatever junction on the model would also require some limited up and down freedom.

Because of the long arm aft from the CG, the preferred classical U shaped metal was ruled out. Real Pipers had a tubular junction that that would not meet my requirement for torsional stiffness because it just had to be glued inside the balsa elevator leading edge. For ease of assembly, a square full carbon piece protruding from one elevator gliding into another square hollow carbon glued into the other elevator seemed a good solution, but I couldn't find such small dimension things that didn't allow any play between both (and that was essential for elevators). In a regular household hobby store I stumbled upon rectangular 6x4mm solid plastic rods that could be engaged in U shaped rails so I both one of each. Back home it was obvious that the U-shaped plastic was too large to be glued to the relatively thin leading edge of the elevator (although that hadn't been sanded). I finally settled on sanding the rectangular full plastic in a triangular wedge shape that could be immediately glued in one cutout leading edge, and at a later stage in an already prepared custom triangular cutout in the other elevator after being passed through the fuselage hole.



Alignment during gluing was essential but difficult because I insisted using PU wood glue to get a complete grip between the plastic and the wooden opening, but a triangular shape is prone to slip out of a triangular opening if none of the sides are horizontal or vertical and the glue only settles after a couple of hours whilst expanding in between. That is another of those moments when you loose a lot of time just to monitor the possible blocking and reaction of the pieces during the process without being able to do much more. After that I made the slits for the hinges. E-flite suggests using their paper type CA glued hinges but because a real Piper has quite some gaps between the fixed and movable surfaces I preferred to use real hinges. Those are heavier but require less force and therefore allow more precise control of the elevator and rudder, both very important during tow or landing. I also elected to use three hinges on each surface because of the possible hit forces of the tow cable. If you only use two and one is dislodged or destroyed, you bite the dust, losing one out of three leaves two to keep the surface aligned and allows control for a return to terra firma.

E-flite already made very wide all the way through slits for their paper hinges and that weakened those spars so much that I first cemented them close with thin and medium CA before making my much smaller cuts at the chosen spots to join the 3 stabs with their flight controls. After the hinges had been glued in the elevators and rudder, I carved out some balsa on the top of the port elevator to allow the counterpart of the horn to be embedded and with a flat top. E-flite says these can be discarded for this kit but with the forces aerodynamic forces of two elevators and possibly the tow cable on a single horn, I found it wiser to use the two parts sandwiching the elevator with two screws and nuts instead of just relying on two 2mm wood screws that grab in a mere 2mm of plywood. I then covered those surfaces with orange Oracover but without pressing the the area down where the elevator junction still had to be glued after the rest was already in position. The tail assembly will only be made after all other fuselage works are completed as to allow the fuselage to be turned around during the rest of the modifications/assembly.

5) towhook

Contrary to real Pipers that have their hooks aft of the fuselage below the tail, model towships have their hooks on top of the fuselage not far behind their CG. The reason for that is that unlike real gliders that can be kept precisely centered and slightly above the tug, model gliders tend to wander around because their pilots just can't judge their position when steering from hundreds of meters away, often completely from the side or straight below. That can result in the rope getting serious slack and if the glider isn't in the exact good spot when it tightens and snatches, might pull the tail in any direction if the hook is not close to the CG. In such case the limited aerodynamic forces of the rudder and elevator might not be sufficient to keep the tow straight or under control.

Neither real Pipers nor their models are structurally strong enough to install a towhook somewhere on the top of the fuselage or the wing. The position of the overhead rooftop glass and fabric thereafter doesn't make choices easy. Most model towhook systems are designed to be bolted vertically against a strong fuselage former end rest on a flat part of the turtle deck. A vertical pin slides in and out so the servo has to be positioned somewhere below the towhook. On the model you can do that behind the cockpit but the distance behind the CG becomes rather long and the servo inaccessible whilst there is nothing but a balsa stringer for grabs under the rooftop shaped fabric. Next more forward former is at the back of the cockpit access doors and just behind the rooftop clear panel, and also coincides with the passenger's position in the cockpit, making any modifications there very visible through the cockpit windows. On the other hand, the BAC Pipers have an unusual antenna at that precise spot, making it possible to completely hide the towhook (with a fake antenna assembly) when not towing.



After careful considerations I decided on the forward option because it had the advantage of using some existing structures with various custom made reinforcements that would not be too visible. The aluminum hook had two slits (for two towlines?) but as I only would use one, I decided to allow the lower one to grab around the top flat plywood crossmember just behind the top window. That way the hook wouldn't stick out so much and after grinding away some of the ring that is supposed to sit on top of the fuselage, the flat part of the hook system rested on the fuselage former. As that was only 2cm high and 2,5mm thick at that spot, it wasn't large nor strong enough for the forces that can act upon a towhook. It thus became paramount to lengthen and reinforce that former with a 2mm birch inset in the existing open arch, another 2mm birch plate covering the whole width of the former and 3,5cm high to lay against the existing former, plus a 3mm plywood strip along the width at the back to anchor the lower screw of the hook. Slits had to be made to saddle the central top balsa stringer, and on the sides to allow insertion and passage of the wiring for the wing servos.

That made for solid anchorage of the towhook, but the (snatch) forces still had to be spread on much more of the fuselage. To make this conspicuous I elected to cutout 2mm plywood panels that were then glued on the inside against the existing frame between the middle and aft cockpit frame. I first made trial shapes with paper, then with cardboard before finally cutting out the plywood. The window cutouts had to be slightly larger because the vacu-formed window shape could not handle thicker frames. I might have used just the top half of those panels but torsional forces on the middle cockpit former made me decide to go all the way around. That resulted in a very strong boxy structure all around the back of the cockpit, with hardly any traces of it through the windows. Compression to the back behind the hook was taken care of by two birch longitudinal inserts glued flat against the top balsa stringer and against the middle and aft cockpit framework, the second one through the kit's servo access on the stringer between the aft cockpit and the next frame.

All those individual panels were finally refined with much trial and error until their fit was tight and all could be inserted in proper sequence through the relatively small magnetic cockpit doors and form one strong unit. Following picture shows the paper and cardboard patterns in front of the various plywood and birch plates before insertion. The longitudinal reinforcements are temporarily placed on top of the fuselage between their frames. The modified aluminum hook lays on top of the new traverses that will sandwich the existing plywood one. To the right of the model the finished cowling with air filter in place, in front you see the one-side mated half horizontal stab with its elevator and horn, the other side can only glued after final mating through the fuselage. The rudder is also covered and has the control arms in place and permanently fixed with both glue and screws. All this extra wood and towhook parts add an extra weight of 58 grams.



Hook operations are driven by a single pin moving up and down, this limited the position of the servo arm to a single vertical line. With the passage of rudder and elevator pushrods to their servos this meant that the hook servo needed somewhere be suspended in a void, preferably below the waistline of the aft windows. It is preferable to have servos that can develop torque at least the weight of the glider. That is an overkill for normal conditions, but then you don't release the cable, the glider does. The tug only releases the glider when there is something going wrong, and chances are that the flight dynamics between both vehicles during those uncontrolled gyrations results in a lot of strain from whatever direction on the towhook (mechanism). Expecting to pull up to 3kg gliders, I thus opted to use a 4,5kg(cm) metal gear servo. As I happened to have some Hitec HS225MG in stock this looked like a good choice.

Servo power is only as good as its attachment and in this case all the forces would be pulling the servo up from the floor. For that reason I first bolted the servo horizontally to a couple of hardwood stubs, which in turn were bolted to a plywood floor with triangular hook pieces as reinforcements. That assembly can then be bolted/glued to existing traverse plywood making up the aft floor. Eventual servo replacement can later be performed with a screwdriver through the cabin doors. That assembly (including servo) weighs 44gr bringing the final total for the complete hook system, bracket and cabin reinforcements to a grandtotal of 102gr. Following picture shows trial dry placement of subcomponents for position and alignment.



6) flight control actuators

As already mentioned earlier I had removed the longitudinally positioned servo tray from the back of fuselage and after shortening it a bit and grinding away some floor members (for the deep servos) it fitted nicely transversally between the hook servo pedestal and lower door half. I then temporarily fastened the servos in position to experiment with their orientation versus the possibilities to route the sleeve extensions for the pushrods from where the kit sleeves stop, via custom hold-downs and along (even through) the hook servo bracket, towards the servo arms. Following picture shows the original servo compartment where sleeve extensions were joined to the existing ones (on the right) and new support panel to the left. Quadruple wires are the aileron and flap servo ones that are routed from the wings via the top of the fuselage before coming down towards the receiver.



It took me some time to get all those items nicely lined-up for minimum friction. All 3 servos were removed again in order to paint the floor, supports and brackets flat black (for low vis), and the cockpit side interior panels gray. After drying, the 3 servos were bolted into place again and their arms fixed into the desired neutral servo positions. Access to these can later be done by the the transparent rooftop that is simply bolted to the wing center. The rudder and horizontal tailplane on the side of the horns were temporarily attached to the aft fuselage for the next step.

I then cut the lengths of the Bowden cable inner components. These were 1,8mm nylon hollow tubes through which a 0,8mm wire was inserted for strength. As these formed a single unit they moved freely within the 3,2mm fuselage fixed sleeves, but could only be equipped with their adjustable clevis endpoints when already in the sleeves. At the extremities I used 2mm connectors that are normally soldered to metal rods. In this case I used a knife to make multiple incisions in the nylon at the extremities so the PU glue would get a more solid grip. On top of that I also used pliers to squeeze the assemblies tighter and as such disturb the nice circular shape so the expanding glue couldn't slide out later. All the actions in the last two paragraph had to be performed within the cramped confines of the aft cockpit. The back of my hands quickly became very sore during the two days I continuously was fumbling through the window openings to get everything done. In the end it all turned out well and besides saving substantial weight of the servos and the heavy kit provided metal rods all well behind the CG in the kit, the friction became almost nihil which enhances control precision and avoids straining the servos and associated electrical consumption (elevator and rudder are Spectrum A5040 digital servos which by design already use more power). At that stage I gave everything concerning those servos, pushrods and hook a flat black coat. This was still a first coat because further scale interior detailing will cause most of those things to be hidden even more in the end.



7) landing gear

After having read about the landing gear problems most guys experienced, I immediately ordered a stronger tailwheel assembly. US guys moslty use the Sullivan 859/860 tailwheel assy but I couldn't find them in Europe. Lindinger had a MBL size 20-60 nbr HY025-01101 in stock that seemed to fit the bill well. On the next picture you can see this kit on the left and compare it with the flimsy E-flite tailwheel parts on the right. As described in the part about the tail surfaces assembly, the rudder control arm needed to be cut in half to fit tighter around the rudder.

The main landing gear was very sturdy but when rods were inserted to check the alignment along a square-pattern sheet, it was obvious one of them had a major toe-out angle and the other just tracked straight. These gear legs already had factory applied Ultracote on them which I wanted to save. By temporarily attaching the assemblies to the fuselage and inserting a long solid hex drive tool through the axle position, I was able to slowly twist the axle holders so they both had some toe-in in their normal fuselage angle and become straight when fully spread. I didn't hear any inside welds break but as can be seen in the picture, the Ultracote had wrinkles that I later was able to iron away. The tires were nice but I hated the all too visible Hangar9 Pro-line and size letters standing out. The hub also wasn't representative for the BAC L21s but were of solid plastic and could be easily made flat without the CUB logo.



Whilst the original tailwheel assembly was supposed to remain in place by two minuscule screws into a plywood plate, the MBL kit foresaw two long screws that had to be engaged into blind nuts for wood. That seemed a much better solution but unfortunately the aft of the fuselage of the Piper was so narrow and already cramped with balsa/hard/plywood panels that there was no way to insert those blind nuts, even after an eventual removal of the Ultracote. A thorough inspection of that lower section revealed that E-flite already had some installed blind nuts for the tail-braces at the bottom of the tail.

That gave me the idea to first mount the tail bracket on a separate piece of plywood, but because this had to remain slim I first had to slice away half the depth of the blind nut metal center. These nuts have their own thickness which forced me to create some recesses in the fuselage in order for the new plate to be flat against the existing fuselage. I also drilled some channels for the long screw to penetrate the vertical beam for added lateral strength. The new plate can be bolted tight together with the brace supports in the factory installed blind nuts, allowing the assembly to be removed in case I break the bracket. If on the other hand I experience too much play during flight ops, I may glue the complete plate against the fuselage bottom for added rigidity. The two intended screws were too short for the new job and I swapped them for a longer pair foreseen for the strut/fuselage so they penetrate the total depth of the fuselage mounted blind nuts. I had no other choice because of the US SAE type of tread.

In the need to create maximum distance between the rudder arm and the tailwheel arm (for the springs), I decided to mount the latter backwards. That implied the screw for tightening pointing forward an thus be less visible when taking pictures of the back of the Piper. To save weight I cut the remainder of that pivot axle and totally discarded the ungainly metallic knob that was supposed to go on top. As that probably had been designed to minimize arm slip, I compensated that by grinding away the curve in the axle at the spot where the screw bites into it.

Last edited by BAF23; Mar 03, 2016 at 05:41 PM.
Mar 03, 2016, 03:28 PM
The sky is the limit
BAF23's Avatar
Thread OP

Part 4: Forward fuselage


1) constraints versus necessities

Before doing any work on and in the forward fuselage I carefully considered the available options. These were rather limited and still a bit of a gamble because I had no idea how the weight distribution would end up. The internal work on the nose would require a lot of turning around of the fuselage so neither tail nor landing gear could be assembled yet. In addition with the cockpit interior still a big question mark, I just had no way to even partly assemble the model for CG evaluation. I took it for granted that with all the modifications my model would still require nose weight to balance (as most other builders reported). Because I have 7 RcPlus 4S4000 batteries those had to be the ones to power my Piper, but they are too thick to fit in the kit's belly compartment. I also wanted power redundancy and some overhead for the 55+Amps I expected during takeoff. I therefore used a 85Amp Rcplus Skymaxx pro ESC with 5Amp internal BEC that can be adjusted to 6Volt, the voltage at which the installed servos develop their maximum torque.

For backup I usually install a 2S950mah battery whose output is reduced to 5,3V through a Castle Creation 10Amp external BEC. Both power sources feed a Schottky diode to ensure continuous electrics to the receiver en case of heavy demand or failure of one battery or BEC. I program my telemetry on the receiver voltage so that I get warnings whenever my voltage drops below 5,5v (indicating primary power was lost). The motor might be perfectly running if only the ESC BEC failed, but at least I'll be able to land with just slightly reduced voltage. Because I also like to be informed about my power consumption I install a FAS100 sensor between the battery and ESC. That sends telemetry data about Amps and Mah consumed making it easier to judge how long I still can fly, or to ask a towed glider to disconnect. The FAS100 info converts through a variometer sensor so I also have altitude telemetry. That makes a lot of items that have to find a place in the nose. I temporarily connected that setup for a practical test of all the sub-items functioning in unison, and transmitting back to the transmitter for numbers and audio warnings with partial power failures, but also was surprised how many items and wiring would have to find a (cooled) space in the confines of the nose.



The plastic cowling is only attached by magnets but cannot be opened without first taking off the prop, so that space will remain free. Due to the plywood-former shape in the cowling it also doesn't allow for anything thick to be attached to the sides nor top of the motor mount. The engine mount is just large enough on the inside to accept my main and backup battery next to each other, but those then have to be inserted through the cabin (doors), affecting the possibilities for a scale interior (pilot on seat and stick between legs). It also means that I cannot modify the lower cockpit door so it opens completely flat against the fuselage (as per real Pipers). Placing the battery within the engine mount allows it to slide 5cm further forward, and it still rests mainly on the strong fuselage and not on the weaker protruding engine mount. With that solution, all the rest of my electronics can be grouped and easily accessed through the forward belly hatch (which has two large cooling holes). To keep the servo wires short I elected to place the receiver in the back of the cockpit (through the aft access panel in the belly aft of the cockpit) with the power and telemetry wiring running invisibly beneath the cockpit floor.

2) power train placement

I started by taking care of the reported weak engine mount by first running thin CA along all the wooden factory-glued truss parts, then a liberal amount of PU glue over that to provide additional strength. The e-flite Power32 motor was then bolted to the mount, but with a double washer top left and single washers in the diagonal to obtain sufficient down and side thrust to compensate for the propwash during the high power/slow speed side effects during tow (later modified see semi-permanent assembly under chapter 5). The 3 engine wires were then routed through the spacer room to make sure they couldn't interfere with the rotating motor body before being joined with 3,5mm bullet connectors to the ESC. EC5 plugs were then soldered to the battery side, ESC and both sides of the FAS 100 telemetry power sensor. I did this mainly for standardization and interchangeability with the rest of my fleet, but when plugged-in these are rather long and with the 12AWG power leads take up quite a bit of space and can't take sharp corners.

The best practical solution seemed to be attaching the ESC with nylon binders flat against the bottom of the engine mount, exactly in the opening between the motor compartment and removable bottom hatch with cooling holes. The ram air from the intakes thus will ensure sufficient cooling for the motor through exhausts at the sides of the cowling, and for my electronics through the escape in e-flite's intended battery hatch. The FAS100 sensor doesn't develop much heat and was inserted transversally under the cockpit floor aft of the hatch. That could have allowed the battery leads to be inserted directly into the FAS EC5 plug, but would have required serious fumbling inside the cockpit for each battery change. I thus made an extension wire that allows me to connect the engine battery just outside of the open cockpit and even more important, allow me to disconnect both outside of the cockpit because they usually are rather hard to separate.

Power and data to and from the FAS100 was then connected to a normal precision altimeter/vario for signal conversion. I installed that one about halfway but know altitude readings (essential for not exceeding max allowable during towing) could be off a bit because of the variations caused by cooling airflow close to the sensor. The latter was fastened by means of Velcro to the side of where the front stick passes through the floor (more about that in next chapter). The signal wire to the smart-port in the receiver was then bundled with the throttle wires from the ESC before being pulled under the floor to the aft lower access hatch behind the cockpit, where I can plug them in the receiver at will. The orange throttle wire is the only one coming directly from the ESC, the red and black power wires originate from the Shottky diode for power redundancy. The diode is connected on one side to the BEC of the ESC, and from the other side to the Castle creation 10A BEC just behind the ESC. The other side of the latter ends in a JST connector that also can be pulled to just outside of the cockpit for connection with the standby 2S950 battery. When all wires had been cut to length and everything was in place I was surprised how neat and spacey this elaborate setup ended up.



3) scale interior aspects

Because of the wooden construction, the cockpit had a lot of larger than normal formers at various places. As they were needed for strength. I couldn't remove them or make them slimmer (as aluminum tubes) but used flat black paint to tone down everything that was not scale. The parts that are visible in the real aircraft were painted interior gray. Not being happy with the still visible servos (even in toned down flat black paint), I decided to play dressup. In my box with puppet figures I found one that I experimented with in the front seat of my quarter-scale Blanik glider. Because he was too small for that I later swapped him for a larger pilot, but its light weight and size made it an excellent candidate for a backseater in the Piper. The young Ken head was removed and a more mature head transplanted. I then undressed him and cut him in half at the torso. With some more judicious cuts I was able to saddle what remained of his torso around the hook release servo and its mount. It barely adds weight and looks rather good.



The height was limited by the head under the reinforced transverse hook frame. His shirt had a Velcro closure at the front and that allowed me to embed the complete hook mechanism behind it, leaving only a small piece of wire going into the hook visible in front of his face. The bottom of his shirt rested on the elevator and rudder servos, and their servo arms and linkage got covered by his arms. Balsa was used to make a baggage compartment surface above the control rods, thus making the receiver invisible and allowed me to glue a part of the original seat to it, behind the back of the pilot. Another plate was made to separate the cockpit from the back of the fuselage. After painting the interior of the window frames, the ones on the port side were glued in place. I then used cardboard and balsa to make the interior panels and long throttle boxes. A wooden floor was then made in parts that fitted between the existing frames. A new instrument panel was cut in plywood, actually ending up lower than the real one in order to straddle the battery and prevent it from wandering sideways or up in the cockpit. Foam glued within the top of the motor truss achieved the same purpose.



The Ripmax iceman pilot buste seemed a perfect scale and came well painted with a cap and headphone. It isn't really lightweight but worth the expense. E-flite provided two seats that are held by magnets to a plywood frame that has to be glued in floor slits. Unlike in real Cubs, the kit's floor doesn't sit on the bottom of the fuselage but much higher. Assembling everything as per instructions , you'll end up with a seat-back much too high in the cockpit. My simple solution was to completely eliminate the seat frame, save the magnets and use the tall seat-back as the horizontal brown cushion and the short side as the new black seat-back. A balsa plate formed the new back and I glueed a plywood plate under the front for the stick. The position of that was dictated by the end of the main battery. The carbon-rod based stick is used to prevent the battery from moving aft during flight. The stick passes through the plywood and then through the wooden floor into a balsa block underneath to provide solid anchoring.

The pilot buste was then glued to a 2cm foam baseplate that was then cut and carved roughly in shape to form upper legs and the right arm holding the stick. If you ever flew a Cub you probably remember the unusual position of your left arm and hand to operate the little black knob forming the throttle. I used a buildup of foam blocks with toothpicks in them to create such a left arm. At that time I also was toying with the idea of attaching the legs of the backseat pilot to the sides of the front seat but had to discard that because of the height of the floor and frames within the cockpit.



After finer sanding of the pilot's arms and cutting fingers out of the foam I made many installation trials because once on his seat, things became cramped especially with yet another frame above the pilot's head. Playing with the angles (easy when you just can sand the foam at the base) I finally got it right and painted the pilot as a military one wearing a David Clark headset before gluing him onto his seat and to the stick. In between I glued an additional layer of green transparency under the rooftop panel and made a series of decals of the cockpit interior and instrument panel on my printer, based on sized-down pictures of the real LB05 Super Cub that I had made on the flight-line. I also started to modify the tailwheel assembly to portray the towhook as per real Piper. Decals were then applied on the dashboard, side panels and upper inside wing panels but everything had to be adapted to fit around the non-scale but structurally essential kit frames in the cockpit.



Before gluing the pilot to his seat, I attached the magnets on the left and right aft bottom of the seat, and their counterparts onto the wooden floor. To provide even more solidity, I then glued another plywood plate in between the magnets under the seat, drilled a hole in the plate and the floor again to receive the stick for the back-seat pilot. That formed another anchoring point and prevented the back of the seat to move sideways.



Many trials were made with and without pilot, with and without batteries, to ensure this complete subassembly could still be inserted and extracted through the cabin doors, and batteries connected and inserted with the pilot assembly out. When assured of that I then glued the rest of the windows to their frames. The complete front pilot subassembly with seat and all weighs a very reasonable 71grams. This part was the last subassembly for the model, all that is left now is assembling all previously prepared sub-assemblies to fuselage.

Last edited by BAF23; Mar 03, 2016 at 05:47 PM.
Mar 03, 2016, 05:34 PM
The sky is the limit
BAF23's Avatar
Thread OP

Part 5: Final assembly and flight preparation


1) permanent assembly

It was easier to apply the fuselage registration marks and tail emblems with the model on its side so I did that before mounting the tail feathers. Getting the registration straight is not easy because of the angle versus the tapering fuselage with its non parallel stringers. Comparing it with pictures I got the angle right, but the letters seemed too small, did I order those incorrectly? I applied them the best I could but definitely wasn't satisfied about the proportions. Looking back and forth between pictures of the real aircraft and my model I quickly discovered my mistake. It already happened weeks before when I applied the large numbers on the wing.

It's highly unusual that markings on the fuselage are larger than the ones on the wings and that's why I instinctively applied the larger ones on the wings. I was able to remove those self adhesive letters with a scalpel and pincer without harming the Ultracote, but in the process they stretched a bit and I was unable to apply then correctly on a second location. Luckily Callie graphics keep records of our artwork and within hours a replacement order was arranged. I wasn't going to wait for that to continue and decided to apply those after the model is finished, I left the remaining letters in place until then, just in case then don't get delivered for a while (postal services sometimes take 100 years to deliver a letter, but they eventually arrive at their destination).

Before mounting the gear I assembled the tail because this could better be done with the belly flat on the table to get the angles right. Both fixed horizontal stabilizers just slide on two thin carbon transverse rods, that's all ! The book tells you to just epoxy those rods (and rely on the tail bracing rods for solidity), but I also CA'd the stabs against the fuselage for precaution. I know that is not scale, but there just was too much play to my like. I then had to perform the most critical gluing of all, sliding the second elevator into the hinge slots and then glue it against the triangular plastic bar that joins both elevators in my single servo modification. That was when I really needed the tail flat on the table so I could perfectly align both elevators. I only got one shot at it with slow drying PU glue but it ended up very well. The kit's system to get every surface at square angles by changing the length of the wire rods works very well but takes some experimentation and eyeballing.

The instructions are not very clear for that because using a square to verify the angles doesn't work on the curved surface vertical tail, and don't assume your model was built straight. Following order of adjustments seemed to work best. First insert the aluminum wing tube (main wing spar) so you'll have a a main horizontal reference. The 4 tail bracings then can be roughly adjusted in length and bolted to the vertical stab and lower aft fuselage, but not yet to the horizontal stabs. Start by adjusting the lower wires so the horizontal stabs are perfectly parallel to the main wing tube. Working on a perfectly flat surface helps because you can measure the distances from the surface to each outboard elevator point. You can force them up or down by varying the length of the rods with the screw just loose through the horizontal tail. Now turn the model till you can line the vertical tail up with an adjacent wall corner. In my case it was seriously crooked and considerable adjustments of the top wires was necessary before I got that tail completely square. That caused the factory applied complex curved Ultracote to slacken and wrinkle port and bottom of the tail, requiring the heat iron to stretch back to into nice curves. Continuous eyeballing and measuring from front and side were necessary to obtain a square tail assembly that is also aligned with the wings. Only put the lock nuts of the horizontal tails on when everything is straight and settled for a few hours. Don't be surprised if that seemingly easy assembly takes you hours before getting everything straight (pun intended).



Because the rudder had not been glued yet, it was possible to turn the model upside down to rest on the vertical stabilizer so I could connect and adjust the elevator pushrod and then use force to screw the tailwheel assembly tight against its second plywood plate. The model was then pulled back from the table so the tail hung in the void and the rudder could be glued into the hinge slots. When that was dry the pushrod was connected with the radio on to get the neutral servo position. I wasn't happy with the little thread length engaging the quick-link. Shining with a flashlight in the cockpit I saw I had ample thread left on the servo side and was glad I could access it with pliers through the rooftop to loosen the locking bolt. My backseat pilot torso was slid from its location for better access and I was glad that I had foreseen such possibilities into my crowded cockpit setup. The springs between the rudder and tailwheel arms were then made to fit and inserted. I am very pleased with the visual aspect and functionality of that colorful tail assembly.



With the tail work completed and all adjustments made to the throws and neutral points, and the scale cockpit items installed or checked, I saw nu further need to be able to stick my hand through the frontal opening and glued the lateral wing supports into place at the windscreen spot. The large rear-view mirror was glued to the port diagonal wing support, this is used by pilots to monitor the cadets attaching the tow cable and watch the signals from the guy holding the wings of the glider level. He is the one indicating when the cable is stretched and the glider starts moving, only then will the tow-pilot apply full power.

The main landing gear came next. As mentioned early in the build log, I already worked the main triangles for the wheels to have some toe-in tracking. What I hadn't noticed at that time was that the starboard aft fuselage attachment also had been welded at an angle. I was able to bend it in alignment for the mounting screws to fit, without breaking the weld nor causing the Ultracote to wrinkle. As nothing was really well aligned I had to loosen some of the screws and apply lubing oil to get the triangles to move freely, a prerequisite for correct operation of the suspension.

After having removed the low quality o-rings and replaced them by real springs, I temporarily mounted the hole system so I could have an idea of the spring force. I checked various springs from my stock but none were sufficiently stiff to my like, except the larger ones which first had to be substantially shortened. New attachment eyes then had to be made at that side and it was quite a work to get the two pair of springs fitting correctly with some pre-tension, but I was very happy not hurting myself in the process. The eyes could be used straight around the existing screws and if a bit careful with the orientation and alignment, produced a nice flat suspension system that will last forever compared to rubbers.



Because the leather sandow bags were also discarded, I had to make a system that looked like the super cub's aluminum airfoil that cover the springs but allow the suspension to work. I made some aerodynamic patterns that looked scale and were sufficiently thick to accommodate the springs without touching them (so they could stretch freely). I then cut 4 of these end plates from 5mm balsa to have sufficient area for the 0,75mm white styrene covers to to be glued to. The end plates were then drilled for the guide tubes to pass through, and temporarily pressed whilst they all were sanded together to have the same identical curves.



The end plates nearest to the fuselage were then glued to the thin axle, the ones closer to the wheels allowed to glide freely around the thicker sliding part of the axle. The styrene rectangles were then extensively rolled around a pencil till they achieved an almost teardrop shape, at which point they were epoxied to both end-plates at the same time. Rubbers were used at the extremities to press all of the styrene flat against the odd end-plate shapes, and thin CA along the entire aft junction that had been sanded at an angle. These assemblies are final and if springs have to be changed, these sub assemblies will have to be cut open.



All cross-members were then assembled and bolted to each-other and to the fuselage mounting points. I previously already attempted to remove the letters “hangar 9 pro-lite” from the tires but after a first try had to abandon because soft rubber got exposed. Sanding the hubcaps almost flat was easier, but after assembling the wheels on their axles I had no confidence just snapping the hubcaps on and used a bit of hot-glue for safety, without loosing the capability to pry it off if necessary. That complete main landing gear is very sturdy and even if the springs were almost impossible to stretch on their short stubs before final assembly, they feel a bit too soft to my likes now that the model is assembled. Furthermore, it is only after the final assembly of it that I saw that even with the spring system fully at rest, the gear is open too wide, causing the toe-in tracking to be almost eliminated in a 3-point attitude. Further gear flexing during harder landings will cause momentary toe-out, not what you want for directional control after touchdown. I can only imagine that e-flite intentionally adopted this wide stance to augment lateral stability, but I'm not happy with it., nor the looks . I'll check it out during the maiden and adjustment flights, but intend to rectify ithat problem soon after. This will require completely disassembling the cross members, cutting the bungee cover assemblies open, shortening the small diameter axle in the sliding tubes, and probably installing even stronger slightly shorter springs. The last of the permanent assembly was the placement of front windscreen with canopy glue all around and some tape and pincers to keep it fixed during settling all night.



2) semi-permanent assembly

It finally became time to install the “removable” items. I started with the nose because this was easier with the wings not yet attached. After sliding the cowling over the motor and into the pins and magnets, I slid the prop holder over the motor axle and the backplate of the spinner as well. Next came the previously balanced and decorated APC 13x6,5 prop and 2inch spinner. This setup could work but was looking awful because there was a gap of more than a centimeter between the spinner backplate and cowling. Because of the side- and down-thrust the spinner was really offset a lot versus the hole in the cowling. I would have liked to just mount the complete cowling a bit more forward at a slight angle, but e-flite's system with magnets made that impracticable.

The only workable solution was to eliminate the washers for the down- and side-thrust, and use the Dremel grinder to remove unequal lengths of the motor spacers (used for the power32 engine). The difference in remaining lengths (see previous picture with all parts on the green mat, just to the left of the prop) ensures side- and down-thrust whilst substantially reducing the gap. I then drilled the hole in the backplate of the spinner further out from 5 to 9,5mm diameter. This allowed the spinner and prop to be mounted deeper on the prop holder. All that allowed me to reduce the gap to a mere 1mm on the starboard side, but the side-thrust still causes the prop axle to be offset in the cowling center.




On top of that I discovered that those 4 magnets really aren't strong enough to keep the heavy cowling from tilting up or down during hard landings or nose-overs. To prevent this I just used one additional small screw on top of the cowling and two at the lower back to augment the magnets (so the cowling can never contact the spinner backplate).

3)weight and balance

With the model completely assembled I put it inverted on the scales and noted an empty weight of 2750grams. Take off weight with the two batteries becomes 3200 gr, and divided by the 43,3dm² wing area results in a totally normal wing loading of 74gr/dm². E-flite recommends a CG between 70 and 76mm behind the wing leading edge. With all the struts and wheels interfering, I had to put the SIG balancer in a forward tilt angle to be able to make any measurements. This was my first looks at the CG outcome after much of my guesswork and the many modifications. I was relieved to see it balance at 65mm, that was not too far off and required only 40grams in the tail to correct.

Not being in favor of dead weight, I checked for other solutions and saw that removing the standby battery from the nose and inserting it through the aft belly hatch also brought the CG to 70mm. The battery could stay connected in there all day, a wire below the cockpit floor would allow the connection to be made in the cockpit as before. On the other hand the 65mm CG was only slightly forward of the recommended bracket and would not cause problems for the maiden. How far forward? Time to get out the measuring tape and make some calculations. Wing depth is a constant 26cm, meaning that a 25%MAC position corresponds to 65mm and a 33%MAC to 81mm. These extreme MAC positions are mostly applicable to normal full size aircraft and I see no reason why they wouldn't apply to my traditional models. This meant that I could leave the batteries as I first thought, the model flying very stable at the forward CG limit , not a bad quality for a towship. The weight of the towcable over the tail will probably cause the CG to move further back after glider release. I thus decided to try the model with the present setup.

More calculations with the available numbers on a prop calculator showed very promising figures for towing. With this engine, battery and prop the Amps will max at 56A producing 832Watts in and 662Watts net power. The thrust to weight ratio is 1.07 (hovering!) and 6 minutes of mixed throttle use possible. The 260Watt per kilogram should be sufficient to tow 3kg gliders. That combo choice really worked out well, for power and for CG. The heavier AXI engine wouldn't have worked, its prop would touch the ground in 2-point attitude.

The new markings, fake antenna over the towhook, pilot footstep, ident labels and canopy fuselage gap filling are delayed till after my ski holiday.
Last edited by BAF23; Mar 21, 2016 at 06:00 PM.
Mar 03, 2016, 06:31 PM
Registered User
shinpaku's Avatar
This is a nice piece of engineering. Can't wait to see her in the air. Well done so far!
Mar 06, 2016, 06:04 PM
Registered User
jetpropdlx's Avatar
nice project ,
Great result !!!
good luck with the maiden flight.
Mar 21, 2016, 05:58 PM
The sky is the limit
BAF23's Avatar
Thread OP

Part 6: Final touches


After my week in Italy I had to delay the works again due to serious (health) problems of my 91 year old mother. When I finally got the new markings from Callie I decided to take my time and correct all shortcomings before attempting to fly it. I dismantled the wings again so I could reapply all markings on a horizontal surface. Instead of applying the new markings on the wings and risking of ruining them when replacing the adjacent orange wingtip panels that didn't look the par, I first replaced the latter. Still not perfect but a vast improvement over what I removed.



With the fuselage free I turned it upside down and dismantled the gear suspension x-brackets. I then used a knife to separate the aerodynamic suspension pod body from the fixed former. That worked relatively well because the epoxy didn't have very good grip on the styrene plates. I then was able to slide the fairings away from the spring assembly so I could work on them. On the previous pictures it was obvious that the gear was spread out too much and the stance looked awful. I measured the suspension assembly to be 1cm too long on each side. This could be corrected by sawing off that length either from the extremity joining the wheel axle, or from the female side of the sliding tube. After balancing the pros and cons I choose for the latter, knowing that implicated shorter springs and shortening the fairings. Shortened assembly in front of original one.



Because I didn't want to risk my prop striking the runway with the wheels spread out due to absorbing a bump in takeoff or 2-point landing attitude, the reduction in length of the springs helped reduce the max elasticity. I first installed two springs per assembly, 25mm long and although of slightly smaller diameter (6mm), stronger than the ones I first installed. With the fairings still removed it was easy to just put a few screws in (without bolts) to restore the x-frame end test the elasticity of the new assembly. It still felt a bit spongy and I was able to modify the length and offset attachment on the springs to install a third one on the other side of the M3 screws, and using bolts to keep then close together. Trial fitting confirmed all the springs stayed nicely together within the fairing, the spring action was now sufficiently stiff and restrained in elongation to still ensure shock absorption without spreading the gear legs too much. Best of all, The angles of the gear look much more realistic and the toe-in tracking guarantied whatever the (limited) spread. The right one on following picture already sports 3 springs.




After also cutting off 1cm of the fairing, I glued these assemblies back to their intended balsa upper side-plates and after some painting got good looking functional bungee fairings containing maintenance free springs for longevity and strength. These fairings might not be exactly scale, but the proportions look realistic and if anybody is interested in duplicating them, they are now 7cm long, 4cm deep and 1,8cm thick. Before mounting the assembly in place I took the opportunity to apply the side markings with the fuselage sideways on a towel, and used canopy glue to fill some voids between the windshield and top of the instrument panel bow. Other paint touch-ups were performed all over the plane, but before mounting the gear I used shaped piano wire to fashion the hand-holds on the aft lower fuselage and the pilot step just behind the starboard struts. Such details don't cost much time nor money, but add tremendously to differentiate your model from the similar ARF you encounter at the fields.



Last job was to make a fake antenna that could be inserted in the towhook to partly hide it from view. These antennas are not common and are rather big for a Piper Cub (see pic under Part 3: 5) towhook). I settled on building it up from scrap pieces of 4mm plywood. The vertical antenna was glued to horizontal plates that were cut/filed to rest on the top fuselage and engage around/in the towhook to keep it in place when required for show or flight. After rough shaping of the parts they were epoxied together and left to dry before further aerodynamic shaping with the Dremel took place.



It took quite a bit of sanding to get the antenna and fairing to look good, but I like the final result and it seems practical just slip into the hook mechanism and simply move the pin up. Because of that I'll have to think it over again what the fail safe position of that servo has to be. So far I had it on hook open, but that way I might loose my fake antenna in flight during a momentarily loss of signal.



Is my model scale: nope, is it standoff scale: nope. Although this e-flite ARF comes as a very attractive product, there are too many things on it that are way off from scale to even consider it as one, even with all the mods I performed. A real Piper has a welded steel structure , very visible at the cockpit area, but impossible to duplicate in wood regarding necessary strength, except by using exaggerated wide plywood and fake canopy pillars. Just comparing the model's top green window that had very wide (bolt on) surrounding panels, to the wing to wing perspex on the real aircraft, demonstrates that whatever and how much work you do on this kit (in my case about 350hrs), you'll never get a result that is even remotely scale.




Would I do the effort again, probably not, but what came out of my many modifications is not only aesthetic, a lot of the modifications are also structural or aerodynamic improvements to a model that has often be assessed as having many shortcomings. At first sight, my model looks rather convincing to the full size example, but I'm sure that throughout the build I was able to eliminate/bypass/counter a lot of the previously reported shortcomings. It was both challenging and fun to turn an off-the-shelf kit into a one-off pseudo scale that will not win prices, but not only has potential for use as an attractive practical model, but also allow my first steps into learning the difficult job of towing.

Last edited by BAF23; Apr 11, 2016 at 05:27 PM.
Apr 11, 2016, 05:26 PM
The sky is the limit
BAF23's Avatar
Thread OP

[B]Part 7: flight test[/B]


1) Maiden and ensuing minor modifications

The first weekend of April 2016 the temperatures rose and I prepared the model for its maiden. Range checks confirmed the reception was strong even with the aluminum backing on the aileron and flap surfaces. Taxi tests demonstrated positive response of the modified tailwheel, and the model tracking straight with the pronounced toe-in of the main gear. The third suspension spring seemed to eliminate any tendency of premature or asymmetric sagging. All this was performed on the hard surfaces of Zwartberg but that day the wind blew completely cross so I didn't dare to fly it.

Next morning there was almost no wind but because it was predicted cross again, I drove to Tongeren where they have a grass field that is oriented 90° to the hard one at Zwartberg, the main reason why I became member of both clubs. Being early meant there were only about half a dozen flyers and I got space. In fact it was my first flight between those members so I would be watched. I checked with their president if they wanted to check my model over prior to the maiden but my reputation preceded me and they told me that they trusted me. With the wind slowly coming up I lost no time and placed the model on the field for a takeoff into the wind.

With the grass not having been trimmed after the winter and aeration holes punched all over the field, the surface was less than ideal for 2-3/4” wheels. Being a tail dragger the prop pulled it nicely forward and up, and I got airborne without flaps after a 10 meter roll with the throttle gently opened to a little over half. After climbing to about 100ft I leveled off, reduced the throttle and made a turn. With the model perpendicular I had to give a serious amount of up elevator trim to hold level flight. I turned back overhead and pointed into the wind, observing no crab so the rudder trim was ok and I only needed one click of aileron trim for it to remain straight and level. I made another orbit and when outbound again, I made some aileron inputs to observe any adverse yaw. There was almost none so my “Frise” ailerons, aileron servo-arm offset angle plus 30% aileron differential did the trick, even using no rudder nor mix at all for this test.

With the flight controls set, I climbed to about 300ft for a couple of CG dive tests from which it auto-recovered, proving this was within limits. Next I reduced the throttle and waited for the stall, which it persistently did with a sharp port wingdrop and almost vertical dive. This was a serious warning for not allowing the speed to drop off when close to the ground. The stalls with ½ and full flaps produced the exact same results, and occurred at speeds I thought were still flyable. Operating the flaps near cruise speed produced some serious ballooning, but at slow speed and little power, my lack of programmed mix did not bother the longitudinal stability that much. I took the model up again and performed a spin from which it recovered as soon that I neutralized the controls.

Time for some low approaches to check the glide angle at the various flap settings. Feeling everything rock solid with a ½ flap final, I flared and made a very good landing without tendency to bounce or nose-over. Flaps up again only to discover I needed ¾ power just to taxi back through the grass. That 4 minute flight showed that my modifications resulted in an easy to fly model with good stability and controllability. A quick inspection showed nothing had come loose or apart, but a look at my transmitter showed the elevator trim to be near the maximum and the elevator anything but streamlined. I made a mechanical adjustment to the elevator linkage so I had full trim capability for the second flight on a new battery.

This time I took off with ½ flaps and opened the throttle faster, and after liftoff in about 5meters I gave full throttle and climbed away at a 30° climb-angle. Directional control inputs were necessary because the model almost wanted to make a left stall turn during the climb. Once at altitude I trimmed the model out more precisely and performed 2 more CG dive tests. Now the model pulled out fairly positively which I anticipated with the amount of (neutral) up elevator. This combined with the stall results and the strong longitudinal stability made me conclude that the CG was too much forward and had to be adjusted before further refining or flap mixing could take place.

I thus skipped that part of the tests and concentrated on the directional stability each time I flew outbound into the wind. The rudder was very effective but the vertical surfaces were incapable of absorbing the torque forces of the power combinations, even with the serious side and down trust that I had incorporated. Playing with the throttle caused serious heading changes but even with 20% expo on the rudder it was nearly impossible to smoothly counter those gyrations by hand, that will have to be done by experimenting with mixing values of the rudder on the throttle channel. It is important to get that right during go-arounds and while towing. Power 32 is definitely an overpower for that Super Cub, but the excess climb capability is less than anticipated so I expect 2kg gliders to be the limit to start with. During a go-around with full flaps it was obvious that with the present lack of mixes, the strong engine and prop blowing over the flaps made it a handful to handle if the throttle was opened quickly. The full-flap landing didn't feel good so I made a few takeoffs without flaps, and landings with half-flap, seemingly the easiest configurations. Lots of the problems were probably caused by the up elevator instead of streamlined position, but there was no way to adjust the decallage of the horizontal stab. I also suspect that running out of up elevator with the forward cg caused the stall and full flap landing problems.

Back home I put the model on the balancer again and started to experiment with lead and the position of the small backup battery. There was no place where I could hide the lead without cutting in the fabric of the tail so I dropped that option and decided to mount the 50gram backup battery in the aft hatch instead of next to the flight battery. Being a backup battery it doesn't discharge much during a day and can be mounted and connected all day long. I added a switch hidden deep in the cockpit, and that allows me to switch it on or leave it off at will. I then made a custom battery cradle that I glued in the aft bay (originally intended for the tail servos). I used 3mm plywood because weight was a necessity but came out at only 13gr. A balsa plate was glued on the hatch to prevent the battery from dropping through the fabric. Back on the balancer I now obtained a 70mmcg, E-flite's forward CG limit in the manual.



Another thing that bothered me was the lower cockpit access door. It was too high and therefore rested on the strut when open (see last picture of previous entry). The too big kit door handle just amplified the problem and caused the door frame to separate from the cover material because of torsional forces after a couple of battery changes/cockpit manipulations. Those torsional forces are eliminated if the door can lay flat against the fuselage as on a real Piper. The most practical way of achieving that was to make a horizontal cut through the cover material 7mm from the bottom of the door, then remove the door and using a 1,5mm wide circular saw-blade, cut the remaining 5mm of the door-frame so it could be glued to the airframe again.



As I didn't trust the cover material as a durable hinge, I looked for proper hinges but that wasn't so simple because the door had to be flat when closed, but had to swing outboard for the full 180° when open. As in my extensive stocks I found no hinges that could be attached to 5mm panels and be capable of such movements, I found an alternative solution by using the small hinges that could be found on old Cuban expensive cigar boxes. Luckily my father kept his cedar boxes to stock nails and screws so I offered him some modern plastic boxes instead and dismantled his wooden examples. The hinges on the Davidoff boxes were unsuitable because they were fastened by very rusty screws, and new screw-heads were too thick and prevented complete closure of the door. The box of Porto Plata Deluxe had hinges that were afixed by triangular 5mm extension pins and those proved ideal and were epoxied into the wooden frames. The remainder of the lower door frame now was too flimsy to handle the torsion so I removed all of the door cover material and glued a 5mm reinforcement on the lower inside of the door to offer solid anchorage for the hinge pins. I then changed the door handle for a smaller more scale example in curved metal. This door modification is not only more scale, but reduces the strain on the door during my multiple engine battery insertions and extractions or operation of the backup battery switch.



The elevator was then mechanically brought back to a more streamlined position. A mix was programmed between throttle and rudder channel and the neutral position of the rudder adjusted for idle. This will require more adjustments because it is directly linked to the tailwheel steering, thus affecting directional control while taxiing. The model was then ready for undertaking further development flights, starting with trimming for the CG and performing new dive tests and stalls.

Following subjects will be covered soon

2) development flights

3) towing

4) resume and action shots
Last edited by BAF23; Apr 11, 2016 at 05:31 PM.
Dec 17, 2016, 09:25 AM
The sky is the limit
BAF23's Avatar
Thread OP

Crash at Zwartberg during October 2016


With the end of the flying season approaching and an Indian summer weekend ahead, I made arrangements with a fellow clubmember for trying out the Piper with a glider in tow. He would bring a 2kg 2m glider and a lighter foamy one for my first tows. He was already airborne with a motorplane when I arrived and I told him not to hurry because as I still hadn't performed any of the development flights and hadn't flown the Piper in months, I prefered to make a few flights before starting to tow. This was on the hard runway and as there were many models flying I had to remain behind the safety line during the entire flights. The safety line being 20 meters from the centerline seriously complicated maintaining the takeoff-roll heading. I got him airborne and performed some general flying before undertaking some takeoff and landings, not just rollers but full stop and go's. The moderate wind being about 60° cross from the right didn't facilitate things so I settled for no-flap takeoffs and take-off flaps landings with power. The first two were not beauties (deviating about 20° left and right from centerline) but were passable. Next takeoff I opened the throttle too fast after liftoff and probably augmented by the wind blowing inbetween two hangars at that spot, I encountered serious trouble maintaining the runway heading. I opened the throttle fully to quickly gain some ground clearance but this destabilized the situation even more and using ailerons to recover and turn wasn't my smartest move. I was able to recover (with rudder) from a first wingdrop, but as often when unable to release backpressure (due to ground proximity), the other wing quickly dropped together with the nose. I almost got the wings back level again but couldn't avoid the 45° nose-down slow-speed impact, luckily in the adjacent grass surface.



As I approached the model I saw the cowling sat at serious angle and when I picked the model up it became obvious that not only the cowling, but everything in it (including the firewall) had separated from the fuselage. Both sub assemblies were only held together by the wiring loom of the ESC and battery. I must admit that the combined weight of the motor, ESC, battery and plastic cowling was more than I thought and excessive for the little plywood of the engine mount plate that continued into the fuselage floor. The firewall had held only with minimal glue onto rather delicate fuselage sides. The bottom had no structure because it was the removable cap over the original battery compartment. The top was only held by five thin balsa stringers. No wonder this much too weak assembly couldn't cope with the least increase in forces and I was already surprised the nose hadn't broken off during previous hard landings or nose-overs. Maybe the latter already created unseen cracks and partial separation before the last takeoff and could have cause the the unexpected and difficult to control nose swing to the right, but that is only guesswork.



I still consider this another major design flaw on this so-called platinum e-flite model. Why produce a kit with such understrength nose-section when you know that the model needs lots of extra nose weight to balance correctly? To me it looks more and more as a model that went straight into production (with all the flaws we all encounter during the build and flights) without having had prototypes and work the bugs out first. No wonder many (even experienced) modelers are very disappointed by this model in most of its aspects except aesthetics. A thorough inspection back home surprised me by showing almost no other damage to the rest of the parts, including the gear, engine and prop. The only things I discovered were minimal damages to the fake navigation light units and a de-lamination on the upper port fuselage wingrib (which was corrected by allowing PU woodglue to inbetween before clamping everything shut). Looking at pictures of other crashed e-flite Pipers it is obvious that the plywood reinforcement plates I glued in the cockpit to cater for the forces on the towhook, definitely helped dissipate the for-aft g-forces on the wings during this impact (the wing joiner tube still was as straight as an arrow).

crash zw 5

The actual repair was postponed till winter period but seemed straightforward. After careful removal of all the loose parts in the nose assembly I got a better picture of the situation and saw the possibility of using almost all of the collected parts for performing the repair. Many parts hadn't cracked but just became unglued. Removal of the former glue was thus the first task on hand but was just too easy with a knife. Even the broken lower engine mount plate came off without trouble.



I first made plain plywood 2mm plates that I glued to the hollowed ones on the inside of the fuselage nose between the instrument panel and the firewall sides. Those will take the burden of the forces caused by the heavy nose and spread them over a much larger surface through the fuselage. I then cut the broken forward end of the fuselage floor plate some way back as to be able to glue a new custom fabricated engine mount bottom plate to it. This plate was now plain and cut is such way as to get the fingers engage the original fuselage plate cutouts for added bonding surface. Being again only 2mm (restricted by fit into the existing remaining pieces) would not provide sufficient positive-g strength for the heavy battery and engine cantilevered at the front of the firewall. I therefore produced an additional shaped plain plywood long plate that could be glued underneath the previous assembly and thus create a very solid and rigid mount of not only the engine, but of the entire nose. That plate went from side to side and grabbed around the adjacent fuselage formers. All these pieces were often dry fitted against the remaining side parts of the original firewall (the latter being kept because of the particular magnet attachments and holes for forward and bottom cowlings. Except for the added bottom plate, the assembly of all the parts had to be done in one go because of the multiple lips in each part. This was even done with the motor still attached so I could ensure correct side- and downtrust with the motor axle to measure. The plywood ring with the magnets inside the cowling was refurbished and glued into position again. This was critical because any misalignment would cause the spinner to touch the cowling later. Cowling and spinner also got repaired and repainted.



All the pieces of the puzzle came together well and are now welded solidly with liberal amounts of expanding PU glue. I have no doubts that this assembly easily is three times stronger than the original setup. To compensate for the added weight I moved the ESC further aft (but still within the cooling air of the compartment, and all the rest of the electronic train further aft under the cockpit floor. After final assembly, the balancer indicated the same 70mm as before the crash but I added 7gr under the tail to bring the CG a bit more aft for following flights. Hopefully this will reduce the nose-over tendency when landing on soft surfaces.



The model looks brand-new again and ready for duty but this time I will not cut corners but methodically follow the development flights routine, making all necessary adjustment and mixes before moving into the multiple landing phases to get completely trained and satisfied with this tricky model. Only after ticking off all those squares will I attempt my first tows but that will not be before spring 2017. Stay tunes for further reports.
Dec 17, 2016, 01:44 PM
"I will return" Federico
rclad's Avatar
Hi BAF,
Sorry about your crash. The nose section is poorly designed. It's doubtful any stress fractures contributed to yours coming apart. Mine separated on very first impact - only second flight in total! Yours will be much stronger after your reinforcement.

This cub is tricky to take off and land successfully, and there is no more proof needed than the fact that someone with your experience and knowledge could also crash it. Did you see the mod I made to reduce dihedral? I've only got four flights on mine since then, but I think it improves take offs by reducing yaw induced wing rolling. A rudder gyro also helps a bit.
Last edited by rclad; Dec 17, 2016 at 02:29 PM.
Dec 18, 2016, 04:58 AM
The sky is the limit
BAF23's Avatar
Thread OP
Yes RClad, I saw your wing mod but as long as mine do not need a rebuild I'll keep the dihedral as such. I might just add a gyro for comfort but only after the final balance and trimming of development flights. Seaspned greatings, BAF23
Jun 18, 2017, 04:33 AM
The sky is the limit
BAF23's Avatar
Thread OP

Final crash




17 June 2017, I decided to fly the Piper again from short grass and took off without problem. Because another loud Monsun RC aircraft was doing his things up and down the runway I first flew conservative large traffic patterns at altitude around the field and after he landed I made some uneventful no flap low approaches before performing the landing. Everything looked steady but although I have more than sufficient up elevator travel, application of it failed to raise the nose and break the rate of decent during what had to be a flare. The suspension did what it was supposed to do and after a check of the Cub on the spot after the relatively hard landing, I decided to takeoff again from the same spot, blaming the lack of flare capabilities to insufficient airscrew power to blow airflow over the elevator.



After getting airborne again it immediately pitched up and dropped a wing. I picked it up but was unable to control the model and shortly thereafter it stalled and hit the ground nose first from a height of about 5 meters. Needless to say, the nose got seriously crumpled and one wing had torn away its fuselage attachment. I checked the controls and those were still functioning properly. I flipped the engine safety switch on my TX and took the model to the parking lot. It was a mess at the nose and while disconnecting the batteries I noted that is was not at its correct place, the pilot and stick holding it in position were broken but I could not yet make up if that was the cause of the crash or if the crash had caused this. Analyse of the parts and after talking to other pilots later made me conclude that the probable cause of the crash was that my improvised "scale" method of keeping the battery in place must have taken a blow during the hard landing, breaking the stick and probably lifted the pilot/seat/floor combo from its anchor points (only kept down by magnets) and allowing the complete assembly of battery plus pilot assembly to move backwards during the second takeoff. This explains why the model rotated much faster and ailerons could not keep the wings level anymore in the near stall condition. It also made me think of the first crash I experienced in nearly identical conditions, had the cause be the same as well?



Although this crash was a result of my poor design of the battery holddown, I decided I've had it with that model that never flew to my expectation (nor to the one of many other disappointed RC pilots). This had been a model I never really wanted, I chose it as a substitute of the same value for the box of another model that had been ruined during delivery. I spent hours embelishing that poorly designed/fabricated e-flight so called premium quality kit, but except for looking at it, never experienced satisfaction from it. Although I could repair it, my mind says to abandon it, remove all usable stuff and maybe cosmetically repair it to offer it to the Air Cadets to hang somewhere in a bar or so. No more Piper Cubs for me, I liked the real aircraft but not the available scale R/C models of it. Fun and Carbon Cubs are ok though and can be messed with, but most scale renderings are much more difficult to handle than the real aircraft so I'll leave these type of models to others.
Last edited by BAF23; Jun 20, 2017 at 01:35 PM.
Jun 27, 2017, 05:09 PM
The sky is the limit
BAF23's Avatar
Thread OP

Restoring it for static display


Because the disassembled model took more space in the room than as a complete assembled one in the rack, the day after the crash I already started the restoration to static standards. I definitely didn’t want to make extra costs and after carefully removing all servo’s and the rest of the electronic drivetrain, I removed the heavily damaged portions of the nose. Motor bearings also took a blow because although it runs on battery power and does not appear eccentric, it sounds and smells rather funny. The crumpled cowling looked totally ruined but after applying judicious pressure with the fingers and gradually building up one area at a time by allowing controlled amounts of PU glue to saturate the straightened bends, I got the initial shape roughly back after a couple of days. I then glued the broken wooden fillings back into the front lower part, repaired the dummy air filter intake and reassembled the pilot’s arms and hands, fixing the stick back to the front seat’s mounting plate.

345

As all this looked promising, I carefully removed the front windshield and used canopy glue to fix the cracked parts back into their original shape. I temporarily applied tape to the bottom frame part and copiously glued the back side of it. I knew that repair would still be visible but ordering a new window was beyond my desires. In the meantime I used balsa to reform the front top panel between the straightened instrument panel and the cowling. Here you see a first attempt at getting the basic curve. At the same time I straightened out the top of the cowling so both could match again.

344

I used flexible filler on both the fuselage top and cowling to obtain the final shapes and a smooth surface for direct painting on it. In the meantime I glued a polystyrene insulation block on top of the remains of the base fuselage plate and whatever was left of the straightened firewall. This block was then shaped to attach the cowling which was glued along the front against the block, and the back directly on the repaired top of the fuselage nose. After painting the fuselage nose flat black and glossy white, I used canopy glue to fix the windshield into position. Two coats of orange and flat black restored the cowling to a condition you even couldn’t identify as ever having had damage. I drilled a long hole into insulation block and glued a 5mm carbon rod into it. I then painted a similar sized old prop and put it in the spinner assembly. This was then pushed onto the carbon rod allowing the prop to be moved without really being free to turn, just a precaution for transport damage. With the 7 servo’s removed, I used small wooden plates snapped in between the linkage adjusters to keep the surface actuating rods in their original locations. That allows the flaps and ailerons to move, but with sufficient friction to keep them in any desired position. I then hammered the wing spar tube straight and after gluing the wing attach frame above the port cockpit side I reassembled the complete model. All this took me only 1-1/2 weeks before the model looks like new again.

346

Looking closer you would find sufficient indications on the aft fuselage and the wings to make you decide never to risk that model in the air again, but it will be a beautiful static model wherever it ends up.

Final note: During the annual party of 23 Devil sqn at Kleine Brogel on 6 October 2017 I handed the static model over to Polle, the guy flying the LB06 on the first picture of this long entry. I figured with over 3000hrs on such Pipers he was the pilot who would feel the most emotions looking at that model at his home and he certainly will take good care of it. In an Air Cadets bar it probably wouldn't have survived for long.
Last edited by BAF23; Oct 08, 2017 at 11:52 AM.


Quick Reply
Message:
Thread Tools