Thread Tools
This thread is privately moderated by BAF23, who may elect to delete unwanted replies.
Apr 24, 2014, 08:03 AM
The sky is the limit
BAF23's Avatar
Thread OP
Build Log

Flair Ka8 quarter-scale restoration

The result in July 2015, but it took a year of blood (yes, like everybody I sometimes cut myself during work) sweat and tears to get from the picture below to the one on top. I don't mind, like the result, and it keeps a retired aviator busy all year through.

Model as advertised and bought second-hand in April 2014

A week after getting a secondhand rare 30 year old WIK Twin Astir II glider, I stumbled on an advertisement for a quarter-scale Flair Ka8, fully assembled with servo’s and towhook, and advertised as “in good condition”. It was expensive, but a few hours after a half price bid I jokingly made, I got a mail that I could come and collect it. A few hours later, after having flown my Parkzone 2m25 Ka8b foamy to regain currency after the winter, I arrived at the seller’s home and spotted the assembled burgundy Ka8b in his open garage. It was love at first sight, the shape of that Ka8 nose was exactly scale, as was the front skid and the rear rubber skid. The wing had spoilers that extended on top and below the wing, as on the real glider. This model was still of traditional build, made of no less than 600 balsa parts (Flair manual predicts 400 assembly hours), and covered with fabric. Negative points were an atrocious outside towhook, a very roughly applied paintjob, and numerous “patchwork” repairs with inconsistent materials and colors all over the model.

I had been lounging for a nice Ka8b but was not happy with the looks of the 3m seagull nose, the lack of spoilers on the Phoenix 3,5m and 6m models, and none of them being in the practical (for size and detail) ¼ scale. I heard about the Flair model but knew they stopped production around 2008, and so far not many ended-up for sale in good condition within driving distance. After having spent much money on a ski holiday in Italy, and the purchase of the rare Twin Astir II, the purchase of this Ka8 was not a rational choice. Chances of finding another one in that condition were slim, and the price being right, I bought it with the idea of getting it airborne in the condition I bought it, fly a season with it, and then restore it over winter to make another true scale beauty out of it.

I’m still undecided about lots of factors. How far can I destroy the nose to get an invisible tow release mechanism installed inside (I first have to find out where and how the lead is positioned and fastened between the balsa). Which real Ka8 will I portray? Again the Belgian Air Cadet scheme I used on the Parkzone Ka8? Or some other attractive schemes I have detailed pictures of? Anyway, the covering on the complete glider will be replaced. There is just too much patchwork after local repairs, and there is no way I can apply a light color over that dark fuselage. I have no idea how easy or difficult it will be to remove the cloth from the balsa. But how will I cover it then? Traditional fabric material again? (feels nice , soft, strong and real) or modern Oracover type stuff? Tough decisions and lots of work ahead during the winter of 2014-2015.

A few weeks later I already had made research and found out the model had been produced in the UK between 1997 and 2008. An illustrated build and flight review was still available online and partly compensated for the lack of plans and assembly book of my model.
I found a lot more assembly pictures and tips in a dedicated R/C group tread.
and on
One of the builders published some general specs that looked promising for good thermal characteristics and suitable figures to be towed behind some of our stronger club towships.
Wing span : 3,75 m - 147 in
Length : 1.80 m - 70 in
Wing area : 90 dm2 - 1400 sqin
Weight : 4.4 - 6 kgs ( 9.7 - 13.2 lbs. ) manufacturer's spec, mine came out at 5.0 kg ( 11 lbs. )
Wing loading : 49-66 dm2 - 15.6-21.4 oz/sqft manufacturer's spec

After some e-mail exchanges with the previous owner/builder, I found out that he had covered it with Solartex without preparing the wood for better adherence. That is excellent news regarding ease for future removal of the complete covering. The second servo in the nose is a servo that was too lightweight for hook opening but has been left glued in. The NiMh batteries in the nose might have had their best time but perform their ballast job well. Although he only could find the plans of the wing, and none of the fuselage nor the assembly and rigging instructions, he said the CG was perfect when he had last flown it in the exact configuration I got the model (without receiver).

First detailed look and inventory of jobs to be performed

My first job thus became assembling the model and finding out the CG position so after all my modifications I knew where to balance it. Assembling the model starts with sliding both wings together over a solid 2mm thick 14mm wide steel bar measuring 29cm. Trying that on a carpet at home already proved difficult because nothing is straight on those wings, dihedral and slight sweepback complicating the assembly so I’ll have to make a jig for field assembly after the model will be fully refurbished. I can slide the steel bar 16cm in each wing, so I will have to take care to get it centered to be equally supportive in both wing halves. At the back of the wing, a copper tube accepts a 3/16th steel rod (missing on my model). Seeing a rod already had previously perforated the bottom of the port wing between ribs nr2 and 3, I got puzzled by the length of the rod. Carefully looking at a picture on the RC Group build log, I could read the instruction to cap the rod at the outside of rib nr2. Further in depth analysis revealed no caps have been applied, and the angle on one side is different from the other, causing the rod to perforate the fabric.

I am very surprised these 4meter wings are just held together (and providing the dihedral) by a mere 29cm steel bar in a joiner box between the top and bottom spar, and a 5mm steel rod (of only 16cm) at the back, exact length tbd. after exact tube measure. To be honest I must say that 4 strong bolts hold the wings in position on the fuselage, but I find it minimal for dihedral solidity, and definitely will avoid pulling too many G’s during flying (Real Ka8s have no aerobatic capabilities).

The plan indicating the position where the copper guide had to be capped.

With each wing having its dedicated aileron and spoiler servo, I don’t like the myriad of wired coming out of it, each having to be pulled through the fuselage to connect into a cockpit mounted receiver. During the rebuild I definitely will replace those plugs with green Multiplex 6-pin connectors just below the wing saddle. The present wire extensions are of various plugs or solder joints, ending on stepped on JR type connectors (because when the wings are stowed vertically they lay on the floor for about a foot. After having pulled the wires in the cockpit, I was able to position the joined wings on top of the fuselage, sliding it under a block holding down the leading edges (mod by previous owner?). The front screws are metal and protrude too much above the wing, the back ones are in nylon and much too long. The turtle-deck is kept in position by a single ugly screw into the wing. All of these wing fastening articles will be replaced during the major winter overhaul, but are adequate for the summer proving flights.

The assembled model is BBiiiiigggg and barely fits in my dining room, but that was the only place I could balance it on one piece. Initially guessing at a CG near 25% MAC proved correct, and I was glad I had opted for the large SIG balancer because this is a large, wide and heavy model with little hard support under the wings. After some adjusting of my balancer (reverse operation) I noted the CG to be at 75mm from the leading edge (with similar steel bar to simulate the one that will come at the back of the wing). I then weighed the glider and that revealed a rather hefty flying weight (NiMh batt included) of 5,4kg, more than I expected for a traditional balsa glider, but possible when you look at the massive use of (ply)wood during the build phases .

The model moved forward from a temporary custom foam stand onto the SIG CG balancer

I took a more detailed look at the overall fit and looks of the model, and it revealed that a lot more work would be needed to bring this model to an acceptable scale standard. Starting on the nose there is that horrible external nosehook assembly.

Atrocious (home-made?) external hook

This might work fine in normal conditions, but is not a sight. It is also common sense on heavier gliders to use pivoting hooks, instead of sliding ones, to be able to disconnect even under tension in emergency situations. The use of a stronger servo still cannot guarantee the pin to be able to slide out under heavy loads. Suspecting the hook was solidly glued to the vertical fuselage frame behind the one piece polyester nose cover, changing the hook for an internal friction free release mechanism will be a major undertaking without breaking the nose.

Internet picture of another Flair Ka8, this is the type of hook I want to flush mount in the nose.

Actuation of the (middle) pin was initially done by a left lower forward installed servo that proved too weak, but could not be removed anymore because it had been strongly glued into position and became inaccessible. A stronger servo was then installed further behind and higher (about between the pilots’ legs). The actuating wire rubs under a tray where the old (now unreliable) NiMh or NiCad were fastened (don’t know how) along pieces of ballast lead. The incorrect angled instrument panel and solid covered nose prevents a good visual assessment and corrective work. The only way to have a peak at how things had been installed was by making a picture with the camera under the instrument panel.

Looking forward towards the nose it does not look like good craftsmanship

Looking at an internet picture of another Flair Ka8, nose access for modifications seems very difficult.

One piece polyester nose-cap should flow seamless into solid balsa sides

On the right of my model’s nose it is not seamless at all, and that might be a probable place to cut out an opening through the planking to access the hook mechanism and battery tray. The instrument panel will also be cut away because in real Ka8s they are positioned totally vertical and recessed deeper in the nose. After the removal of those obstacles, things will become more clear about what and how to improve the model.

The cockpit is cluttered by non-scale stuff affixed anywhere

The canopy is another disaster. It doesn’t fit well and a huge step is visible between the nose and the front of the clear canopy.

Definitely not very aerodynamic, and not easy to correct with plywood formers everywhere around.

The crack was caused by forces not being absorbed by the gap in the wooden parts in the corner.
Note the ineffective grey spongy material between the fuselage and wing, and patchwork repair. I have no idea what those metal household metal hooks were used for. The canopy rail engages the front of the bow with two pins that will have to be relocated to get rid of the step, and at the back it is screwed on into the fuselage top, hardly a practical system. The ungainly main power switch at the outside will be eliminated when cockpit access becomes more practicable. A sliding scale side-window will be incorporated after the clear canopy will have been adapted, and cleaned off the grossly hand painted bow. At the back of the wing, a triangular turtle deck is supposed to cover the back wing bolts, but again this piece is held by a dowel at the back, but has to be screwed visibly on the wing at the front. A new system will be made with a hook or so at the front , and magnets on the back. The main wing bolts are non-original metal ones protruding way too much above the wing.

Fuzzy image of the shortcomings on top of the wing, patchwork over wing joint perforation.

One last problem regarding the front is the scale nose sledge, very well built for grass fields, but on our clubs’ tarmac runway this will quickly destroy the wooden bottom, and as on the real Ka8, metal strips will be bolted on the bottom and replaced at regular intervals. A simulated leather cover between the sledge and fuselage also has to be produced for the scale look.
Moving to the tail I discovered excessive play in the archaic surface control hinges. These metal parts have relatively large eyes that are kept together by a kind of split pen. The latter being much smaller in diameter allow so much play in all directions that I’m afraid of possible flutter on the elevator and rudder. Contrary to the standard Flair model, the elevator is not actuated by a long rod through the whole fuselage, but by a single servo in the port horizontal stabilizer. A U-shaped wire connects both elevator for equal angular deflection. I had read on the build log that some people had reinforced the U-shaped wire because it was too weak. Mine looks solid but later after removing the canvas, I will inspect that better and already have various solutions in mind, ranging from replacing all original tail hinges, to using separate servo’s for each elevator for redundancy.

Simple split pens look vintage, but cause excessive play

The horizontal tailplane is removable by the use of two large screws (again of various types and head-height, and they allow tailplane incidence adjustments to be made by inserting shims if necessary. The rudder is actuated by Bowden cables that can be adjusted at the tail, and are connected to a well-accessible servo just behind the cockpit. This seems to be the only control surface I will leave “as is”, except for the replacement of the rudder top hinge, and covering the rudder to get rid of yet another patchwork repair with red electric tape ! Movable surfaces are fabric covered hollow wooden structures, but the fixed stabilizers are shaped hollow structures fully covered by balsa, before getting their fabric cover. The latter is in bad shape, either time allowed bubbles to develop, or the cover had been applied too slack and needs tightening by heat or so. Because next winter I intend to remove all the fabric, I will leave it this way during the familiarization flights in order to facilitate the subsequent removal.

Sad looking covering of the balsa planked horizontal stabilizer

Besides the aforementioned problems regarding the wing-join metal bar and rod, there are no serious shortcomings. The problematic trailing edge distortions mentioned in the build logs are minimal on mine, and I can see no signs of wing wrap. I still haven’t connected the electrics so the spoilers operation and condition were not assessed, but covers around one spoiler is fully detached and will have to be temporarily glued before test flying. The ailerons are hinged by fabric (and these still look solid) so the play I feel has another source. The bellcrancks and pushrods seem adequate, and the play seems to originate at the servo arm. The aileron servos are deep in the wings and have no access panels, so I’ll have to wait till I uncover the wing to find out the problem. Upon covering again, I will have to ensure the concave bottom part adheres well to the ribs, because now some surfaces just tightened up in a straight line instead of following the compound concave trailing edge intrados shape. One of the wing leading edges has a bad dent into it but that is not structural and will be easy to repair. The model having been flown from a grass field also lacks the protection skids at the outer wing panels, and those will have to be fashioned in order for me to operate my model from the tarmac Zwartberg model runway.

The leading edge dent on an otherwise perfectly constructed pair of wings

So how happy am I about my purchase? Still very happy thank you very much. Chances of finding an affordable complete and almost ready to fly Flair Ka8b on the second-hand market in Belgium are very slim. Although the inspection so far revealed already revealed a serious list of shortcomings, most are easy to correct and wouldn’t keep most other modelers awake. It is just that my standards are high, especially regarding the scale aspects and finishing touches. I now am confident the model is sound (an overkill for solidity when looking at the amount of wood into the kit) and was assembled with sufficient structural integrity where needed. The rough quality of finish won’t allow me to fly all summer “as is” (as I had hoped), but still good enough to perform some proving flight and get accustomed with the flying qualities of that nice model before I tackle the major rebuild.

It now sports a similar paint job of a Ka8b that was purchased new (together with 2 others) by the glider club of Schaffen during the 60’s, and I even don’t find it worth applying the registration letters on the model for the short time I will be flying it in that configuration. The model as purchased certainly is worth the money I paid, but not much more in that condition.

Present scheme based on “charly” in Schaffen during mid-eighties OO-ZLD 3346 (accidented)

I have two other schemes in mind, one of them a Ka8b I flew during the 60’s as a Belgian Air Cadet, and as of 2014 still lingering around in a sorry state in the back of a hangar at Goetsenhoven.

Air Cadets PL54 was one of the few Ka8s sporting orange ailerons, making it much more visible.
The other scheme I consider very seriously is that of a Ka8b that flew at my present airfield of Zwartberg as from 1975, but after hailstorm damage in Germany at the turn of the century, got repaired and now flies weekly in the same colors, but with German registration.

Possible scheme OO-ZAU: cn 8906 3/3/75 Limb Vl 30/5/99 hailstorm in DBR, scrapped, now D2945

I will add further posts regarding the initial test flights and winter rebuilt under this thread, please be patient.
Last edited by BAF23; Feb 03, 2018 at 06:13 AM.
Sign up now
to remove ads between posts
May 14, 2014, 04:59 PM
The sky is the limit
BAF23's Avatar
Thread OP

Preparing for my maiden flight

With the calm spring weather at the end of April, I took out a few days to mount a Spectrum AR10000 receiver in the cockpit. This might be considered as an overkill by some, but I had one in stock and liked the 2 satellite receivers mounted well-apart, a safety precaution against possible interference of a future Sky-Assist variometer transmitter. For fine tuning and mixing purposes I prefer to assign each servo to a single channel, so I already use 2 for the ailerons, two for the spoilers, one for the elevator, the rudder, the towhook, the variometer battery power check, and reserve one for a second elevator servo to be installed during winter. This totals 9 channels in use and warrants the use of that type of receiver. Another advantage of that DSMX receiver is that besides having a standard SmartSafe (reducing the throttle to idle and thus one spoiler to out, but all controls holding their latest position in case of a signal failure), it can be programmed with Preset Failsafe, allowing me to program the failsafe position for each individual channel. In this case I choose for towhook open, spoilers both fully deployed, ailerons neutral but slight left rudder, elevator will be programmed to be trimmed for the full spoiler deployment (after testing). This effectively caters for signal failure even under tow, or for the model getting out of range being sucked up by excessive thermals. I now hope the model will spiral down until I regain control, or comes back to earth in a relatively controlled way. I programmed my Spectrum Dx10t transmitter with differential aileron, full mix between both spoilers, mix between aileron and rudder, and mix between spoilers and elevator. Servo travels were adjusted for correct spoiler operation including the overcenter lock to keep them flush when not in use, and other surfaces to eliminate any servo buzz. Power for the receiver is provided by a single quality 2S2100 battery temporarily squeezed behind the cockpit, then via the exterior power switch to a 10Amp Castle creations BEC before the (adjusted) 5,5v enters the receiver.

I also made a cradle for the fuselage, with on the sides dedicated places for the elevator, wing joint spars, turtle deck, and the screwdriver and keys necessary for field assembly of the model. I screwed standard underwing nylon protectors under the wingtips in order to eliminate damage operating it on our tarmac runway. I felt ready for test-flying it in that condition, but unfortunately during the two wind still days, no powerful towship was available. With a forecasted two windy weeks ahead, I took the model home again and decided to already start correcting things I didn’t like. In order to allow faster assembly on the field, I cut and soldered all wing servo wires on shorter multiplex green plugs. I use one plug for each wing, connected both positive and negative wires of ailerons and spoilers together on the middle pins, and used the outer pins for the signals. In the fuselage I made new wiring lumps from the receiver to multiplex connecters glued with epoxy to cross members under the wing support. The receiver does not have to be touched anymore during model assembly, and it is easy to plug in the respective left and right multiplex plugs with the assembled wings resting on their leading edge before being lowered into the wing bed. The deteriorated foam between fuselage and wing has been cut away and sticky surface cleaned with benzene, before being replaced by brown rubber adhesive used for eliminating drafts around home doors and windows. The too long nylon bolt was shortened, and wing assembly is performed in a fraction of the time it used to be.

Left MPX plug joined, right one still separated, brown new wing rubber supports in position.

The wing aft turtle deck has an awful gap, but that complete system will only be changed during the full restoration. Seeing the bubbles on it, I figured this would be a good test piece to figure out how effective my model heat-iron would be on the Solartex for bonding, removal and stretching. I quickly found out that with the knob at about 75 percent open, the bubbles quickly disappeared and the fabric bonded solidly to the wood again. Only drawback, the dark red paint had a tendency to stick on the iron protective cloth, rapidly discoloring it and making it useless for the white surfaces. With that in mind, I started working the iron on the lower side of the horizontal stabilizer, and was surprised how well I was able to eliminate the numerous bubbles. I then tried it on a lower wing, and was able to stick the cloth to the concave part of the aft wing again, and take out the slack of the Solartex cloth between the wing ribs. I then tackled the spoilers, ailerons and elevators,. All four wing surfaces and tail got treated that way and amazingly not a single bubble remains, except on some of the patchwork repairs made with different materials. The fuselage got the same treatment, and after two days of work, including washing off the dirt in the Solartex pores, the model’s appearance improved 100 percent, and probably the glide number increased by a couple of units as well. On the other hand it accentuated the many rough transitions between the wing planking and the ribs, indicating an enormous work over the winter to smoothen everything out with lots of wood filler and sanding paper. The clear canopy has been made structurally sound around its the frame again, but aesthetics will probably have to wait for winter. I already received 3 sliding small windows, one for this Ka8 and two for the new Twin Astir canopy that I ordered in Germany.

I then removed the patchwork on the right aft wing joint and discovered the end plate had been made of light balsa and had separated from itself. I epoxied a stronger plate outside of the rib, before temporarily covering the opening with white vinyl. I also opened up the patchwork covering the servo space of the spoiler actuator on the lower left wing. I had the possibility to adjust the length of the actuator to correspond with the normal extreme throttle positions, taking into account the overcenter lock. I used a knife to cut a narrow access slit on the right wing to do the same. Changing both servo arm lengths, my spoilers now work perfectly in unison, without use of a servo reverser, and without having to dial in (funny) subtrims, travel or offset figures, and respecting the throttle idle point for TX/RX arming during powering up. I did the same with the elevator, and adjusted the lengths of the Bowden cables towards the rudder to also obtain a straight rudder position without electronic trimming. Fuselage patchwork was removed and replaced with much smaller repairs with nearly identical colored vinyl or transparent tape strips. The result is a model that now is aesthetically acceptable to fly for the rest of the season, something I couldn’t have imagined when I first inspected the model. I then made bubble plastic protective covers for all the individual glider parts, hopefully avoiding transport damage, especially on those very large fabric covered wings. To facilitate sliding the wing in those bags, I screwed spare parkzone Ka8 aileron-link plastic covers over the actuator links. I also made control locks for the rudder and elevator to prevent wear on the servos during manhandling.

Aileron actuator covers, and wingtip nylon skids bolted on, fabric now tight and clean.

My excitement for 4 meter gliders grew so much I became member of a dedicated Benelux club for larger gliders, Biggs. The club president wrote me he also has a Flair Ka8 and he graciously provided me with the all-important figures of CG (82-92mm behind the leading edge), and control throws: elevator 20mm up and down, rudder 90mm left and right, and ailerons 25mm up and 10mm down. The throws I found on the model were very much different and I used mechanical and electronic means to obtain the factory numbers. Because the previous owner’s CG was too much forward, I opted to remove the fixed NiMh batteries in the nose. The full pilot puppet I have from the Ka6e very well fits in the Ka8 and with its yellow coverall will be very well matched to the Zwartberg Ka8 colors. Making a seat for him covering the receiver and lipo battery in the process will be easy, but the towhook servo stood in the way for a scale forward cockpit. He weighs 200gr with the main part being rather close to the CG. With that in mind, I intend to balance the model at 82mm with the pilot in position, and without also fly it without him at 92mm to feel the difference, and decide on final CG figures. I plan to install the autonomous variometer sensor/transmitter under the wing, so flying with or without won’t affect the CG, and the transmitter will be sufficiently far from the Spectrum receivers that they won’t interfere at maximum range.

Removal of the battery pack proved no simple task. Everything seemed glued solid together, support frame, battery pack and several lead shapes. I was able to cut away the instrument panel without damage, but became no wiser as to how dislodge thinks deeper in the nose. Poking here and there with a long screwdriver to separate the plate glued to the forward frame, I was able to extract the assembly weighing 675 grams, and gain access to the rest of the nose without cutting in the polyester outer shell.

Top right: NiMh battery pack and U-shaped lead, old and new servo, single screw holding the latter

I separated the battery back which was a Sanyo 4,8V 3000mah NiMh pack of unknown date scaling 250gr. There seemed to be lead blocks everywhere in the nose. The new hookrelease servo was a Graupner C5077 with 5kgcm torque which is adequate for normal tow releases. The old servo was a Futaba S148 of only 2,5kgcm (makes me wonder what servo is used to actuate the elevator from inside the horizontal stabilizer). After testing, that cheap 44gr servo went into the scrapbox for future applications. The strong Graupner servo had been mounted on a balsa cradle that was only held with a little glue at the front (I separated it with one finger pressure), and a single screw at the back. No use installing a strong servo if it isn’t anchored down properly, so I also removed that assembly. I now had a good view and access to the nose bulge, and could find no traces of hook anchoring in the wood. On the outside, I used a small model plier, and without efforts was able to separate the pin from the hook, they just had been glued together at the tip! (see closup picture of this in the first part of this article). The hook was then easily unscrewed by hand from the wood, and the short pin just slid out with pliers. I was glad I hadn’t testflown the Ka8 with such a weak hook assembly. This was a totally unacceptable mechanism for an 11lbs 4m glider.

Note the household type hook I even wouldn’t dare to use to hang a towel on, and the slippery pin.

Upon delivery of a 10mm proper towhook, I could see no way how to secure it strongly to the frame only from the inside. Drastic surgery was required, and after seriously thinking it over, I decided to peel off some of the patchwork fabric from the right bottom nose quadrant, then used the Dremel to make a horizontal cut through the hard polyester nose (along the future paint separation line), and another one along the bottom wooden spine, the back being free.

Peeled off fabric, dremeled aff polyester part, quality hook prepositioned for evaluation.

The open space now revealed how two 270gr lead weights had been bolted left and right of the frame, I liked that but temporarily removed the right one for more working space. To do that I had to make a small hole on the left side to insert a 3mm Alen key to stop the screw from turning. I then was confronted with another dilemma, following the example of a few other builders who cut all the way through the strong plywood and glue the hook assembly in the middle (but weakening the structure as shown on an internet skeleton picture in the previous part), or installing it next to the frame, just cutting in it where the cone sits for added fore-aft strength. I ended up installing the towhook backwards one third recessed in a custom carved out half-thickness plywood nose, in the already weakened and drilled-out channels the previous owner used for his kitchen hook. I mounted it in reverse so that the flat part of the cone rests against a longer and stronger plywood portion, with the bonus of not having to push the nylon loops so deep when hooking up. The hook tube ends up just slightly right and below the centerline, pointing down at about a 15° angle, to be in line with a normal tow position slightly higher as the airplane in front. It also will be hardly visible that way.

Reversed hook in channel, custom reinforcement plywood parts ready for buildup along the hook

There is just sufficient space for the activation arm to be mounted 90° to the right, so it will be accessible for maintenance if I later have to cut the nose loose again along the same lines. A buildup of four purpose cut thick plywood supports epoxied on top and below the hooktube ensures the completed assembly will be as strong as can be, spreading the forces along a large strong area, incidentally also forming a support for the polyester nose to be fixed on. The piece of lead also had to be shaped in order for the tube and its rod activating mechanism to pass unhindered. All the pieces were glued together in one go with 30-minute epoxy, in order to form a single very strong unit.

Completed hook assembly, lead and actuator arm in position at a later stage.

I then had to take a long break to figure out how to continue. My primary concern of installing a reliable strong towhook was over, but I now had to juggle with options of how to install the servo, the ballast (and how much), whilst freeing as much as possible of the cockpit area so I later could make a scale version of it. The route I choose was to find a way to lock the essential large heavy U-shaped lead in the nose (squeezing and blocking it without gluing), keeping it in position by a 30° angled and modified hook servo tray mechanically kept rigid at the front and epoxied at the back to a crossbar that is bolted to the frame, making everything removable in case of maintenance. The bottom of the servo tray pushed the large U-shaped weight against the bottom spine. The crossbar will also serve as the base for the later to be installed recessed scale instrument panel. I redistributed and affixed some lead I had found, in a more forward position, reducing the total weight (because all ballast is tightly squeezed in the extreme nose). All these modifications now keep the complete pilot’s space (including legs) free of R/C related stuff, and allows adjustments after the trial flights.

I used 2 metal adjustable quick links on a stiff rod, one to permanently connect the hook mechanism (not accessible anymore when the nose is closed again), the other accessible and removable to the servo arm. To find out and adjust the throw of the rod I kept the nose open and free of the modified leadblock in order to have adjustment space.

Custom prefabricated and adjusted elements before final assembly in the nose

The inside of the nose is now packed with stuff, still leaving some space for additional lead if required, but everything is now concentrated more forward allowing the scale instrument panel (and pilot) to be located at their normal spot. Everything is accessible and removable with only two crews retaining the back of a complicated structural but rock solid assembly.

Previous pictures might have suggested a huge gaping hole, but an overall picture of the fuselage just before gluing the polyester quarter back on the supporting foam and wood supports, shows how local the surgery took place. You might expect large gliders to have ample space to work in, but keep in mind such models need a lot of ballast and larger servos and hooks, making the space as cramped as in smaller gliders.

The patient calmly on her side, still under anesthesia just before closing the surgery area.

After the lead was bolted tight again, I glued some support blocks and wood strips, to which the polyester nose could be temporarily affixed with hot glue. Next the cut-gaps and joint around the hook opening were filled with epoxy to create a strong rigid assembly. The flaps of original Solartex then were then ironed back on in the reverse order they were had been temporarily bent out of the way. The Solartex fabric then was cut away from around the hook opening, so any friction of the nylon ropes would not be transmit to the fabric and risk ungluing it, the airspeed under it making segments peel off like a banana. I did not want to glue the fabric in at this point, because over winter it has to be easily removable again. Painting the orifice dark red also was a no-go, because its final color will become white, and the wear of the nylon over the white would allow the old red coat to show through. I’m happy the way this delicate operation came out, structurally very strong and sound, and aesthetically not worse as when I got the model.

Nose closed up, offset tow orifice, aluminum skid plate, custom cradle unsuitable for inverted work

Because at our club we invariably take off from tarmac, the wooden sledge had to be protected from excessive wear and I purchased a 2cm wide 2mm thick aluminum strip which I shaped and files smooth before fixing it to the wooden skidplate using the existing screws at the ends. I have a same length ready for replacement if necessary.

With everything in place I turned the model over and completely assembled it again for a weight an balance after all the changes. It had not been rocket science, but my (gu)estimates couldn’t have been better. With everything including the pilot in position, it slightly dipped the nose at the 82mm recommended foremost CG, and without the pilot it balanced at the recommended 92mm most aft CG position.

In balance at 82mm, note hardly visible nose modifications

The maiden will be made with the pilot in place, and after trimming and adjustments to the glider have been made, I’ll fly without pilot to check out the difference in handling and thermal behavior. Without need for further short term changes, I cleaned out the complete cockpit area to accept the pilot seated on a makeshift seat separating him from the electronics, but also keeping battery, receiver, BEC and their respective connections from moving around.

Lots of space now, wiring will be further rerouted during the winter modifications.
After more than a week’s work, the Ka8 is again ready for her maiden flight (with me at the controls now). I am convinced the model is now in a much better condition to undertake those first trial flights, and my confidence grew accordingly. My to do list for the Ka8 is already half completed, and I have good prospects of making not only test flights this year, but numerous normal fun flights before dismantling the model next winter for a full restoration. Transportation is easy because except for the two wings, everything else including the necessary assembly tools is on the transport cradle. The cheap interim bubble plastic covers at least provide some protection enroute.

Nicely packed awaiting its first field trip.

The results of the flight testing will be posted here soon.
Last edited by BAF23; May 14, 2014 at 05:12 PM.
May 15, 2014, 05:50 PM
The sky is the limit
BAF23's Avatar
Thread OP

Flight testing

The day after the repairs, I got a mail that a powerful tug would be at the field for the maiden of another 4m glider. Shortly after lunch I arrived at a cold and windy airfield with the sky mostly covered by dark clouds. Luckily the wind blew steady in the runway, so I started assembling my Flair Ka8b. During the assembly I asked my towpilot (and experienced glider builder/pilot) to examine every detail to make sure I didn’t overlook anything. Such buddy-buddy checks are often deemed a waste of time and perceived as a lack of trust, but I’ve seen so many (maiden) mishaps that easily could have been avoided if somebody else had made a final (thorough) check. It is so easy to overlook something or forgot to make a final connection (through temporarily distractions).

The play in the aileron mechanism was excessive (as I already had found out), but he agreed that if I refrained from high speed flying, I could fly it like that till winter. My single (metal) quick-link to the double elevator should be kept shut by a mechanical restraint, any flutter (by the still to be replaced hinge clevises) could result in the quick link opening and the total loss of the model. He agreed I could fly it cautiously that day so I could quickly change the rod-length between flights if necessary (which was proved later). The hook mechanism was extensively tested under heavy strain, and passed with flying colors. He not only checked the correct deflection of the controls, but also applied counter pressure in order to test if the servos produced sufficient torque to move the flight control surfaces against strong higher speed airloads. I must admit that was a test that I had overlooked back home, and not a bad idea on seven year old second-hand glider with unknown and invisible servos installed within the wings and tailplanes. During all this, about six clubmembers gathered around the model and had animated conversations unrelated to what we were doing, not exactly the quiet environment I had hoped for to conduct that last-chance inspection. We then moved onto the field where the mixes were demonstrated, and a range-test performed untill twice the book distance. Such large gliders can easily be flown much higher and further as regular planes, and I didn’t want to take any risks. Final test was shutting down the transmitter to check the failsafe servo positions. Benny declared my model airworthy and started his towship while I hand carried the Ka8 to the runway end to hook up and release the cable a few times. He briefed me on his intentions regarding the climb, and we felt ready for the maiden.

The glider surprised me by being airborne even before I got the wings fully level, the distance must have been about a meter. It reacted quickly to any aileron input, but I soon saw my glider pumping behind the towship. Benny reduced the speed and I decided to hang on, but with the finger on the release switch. I never expected that model to be so nervous around all axis, but managed to follow during the two turns to a safe release altitude. We agreed to separate, and as soon as I opened the hook, the glider pitched up to near vertical, and generous amounts of down elevator were needed to get her in a glide. I used almost all available down trim before I could establish a normal speed. Roll control was imprecise and erratic, but the model was capable of flying very slowly. Stalls seemed benign, needing exaggerated up elevator before the nose dropped sharply, unfortunately with serious wingdrops to the right. Using left rudder during the next stall, it dropped straight, and I forced it with full rudder into an early recovery to explore the spin entry. It took longer as I expected before the gyration started, and it recovered quickly when the elevator pressure was released, even with positive rudder still applied.

I liked that, but in the meantime my altitude got lower, and I still wanted to test the spoilers. I repositioned her into the wind with approach speed, but each time I pulled spoilers, it started a serious bank to the right. I was sure I had aligned the spoiler extensions points and opening speed against each other, and couldn’t understand what was happening. I took her around for another series of upwind spoiler tests at a greater speed, but the results were similar. I ascertained myself that I had sufficient aileron control to neutralize the roll tendency, but had insufficient altitude to check the spoiler to elevator mix. As I came around the pattern into the relatively strong wind in final, I was surprised how little altitude this 11 pound glider lost, and I had to use spoilers to help her onto the runway. Playing the spoilers each time required pitch and roll corrections, so it was anything but a steady glide, but the landing on the first few meters of the runway was relatively smooth and at a ridiculous groundspeed.

Whilst the other 4m glider got towed, I stood at the side and mechanically adjusted the elevator quick-link to attain the same landing position, but with zero trim. Whenever possible, I always adjust my neutral points mechanically instead of electronically by trim or subtrim. Next I checked the model over very carefully, but saw nothing unusual, so a second flight was possible. Extending the spoilers very slowly revealed the starboard one to extend just a tad earlier so that will be corrected back home, but still doesn’t explain the serious right turns. By that time the other glider had landed and I lined up again for the second flight.

This time I was prepared for the early liftoff, and noted the Ka8 stabilized not so high above the towship, and showed no tendency to pump anymore. After release, I flew towards a cloud to check for thermals, but found it difficult to notice because of the unsteady nature of the glider, was he making bank on his own again? or was it a thermal lifting the wing? I seemed to be holding altitude well, and I noticed during some intentional reversals, that even using identical aileron inputs both sides of the turn, and unaffected by use of simultaneous rudder, the bank rate was much different, certainly at the lower speeds. Dialing in a copious amount of left rudder trim, and removing a little left aileron trim, seemed to resolve that strange behavior. Stalls now were straight, and the two-turn left spin was recoverable with the remaining right rudder capability, but proved insufficient to initiate a right-hand spin. Spoiler movement didn’t cause undesired bank angles anymore, but all those tests had cost altitude and the second landing was performed, now steady in direction, but still very unsteady regarding glide angle.

Whilst the towship pulled the elegant retractable gear yellow glider again, he announced me he had to stop to refuel. Cake and a Coke seemed fine to me so I could think longer about this strange rudder position my now idle model assumed. Over tea-time I thought about the next steps, and asked another club member to make some video shots of my next takeoff and landing. After the break it quickly was my time because the towhook travel range of the other glider had not been properly adjusted, and after the second time the rope released after rolling a few meters on the ground, he pulled off the runway and got to work correcting that major deficiency, imagine it would have happened at 100ft. I quickly brought my glider to the start position, and under the eye of the camera, got airborne.
15 mei 2014 15:35 (1 min 7 sec)

The tow now was much more stable, and turns for thermals felt normal both ways, and during reversals. I then positioned for a CG dive test, but it recovered much too quickly. Being unsure of the elevator trim, I repositioned for another one after stabilizing the glide angle longer, but the nose again came up much too fast, indicating the 82mm forward CG is just that, the most forward position it should be flown. That made me wonder how the previous owner flew it on 75mm ! Especially for the camera, I performed a loop and some high turns right above the field, and therefore entered downwind a bit lower as desired. As usual, with cameras running, the landing was the worst of the day. I landed it too fast, and applying back pressure to hold the nose-skid off the tarmac, it got slightly airborne again and turned away to the left (left rudder trim?), then came down not aligned.
15 mei 2014 15:41 (1 min 11 sec)

Time to explore the recommended back CG limit. I took the pilot out of the cockpit, and awaiting my turn for the tow, discovered what was wrong and I had overlooked. Looking over the nose towards the tail to see how much left rudder was needed to fly normally, I discovered the fixed part of the vertical stabilizer to be anything but aligned to the fuselage line. From the back it also was obvious it was seriously wrapped, but that had been overlooked because its chord being minimal compared to the large rudder. The torsion angle of the vertical tail is what caused the glider to turn right in about all circumstances, and required an unusual amount of left rudder trim to compensate. I pointed this out to the guys, and took it to the runway for a check with the CG 1cm back at 92mm. The tow felt smoother, the glider seemed less nervous, and after trimming at altitude, seemed to fly steady as a rock at any speed. Turns were much easier to coordinate, and after carefully trimming the elevator, I made two consecutive CG 45°dive checks to confirm the very smooth pullout. That seemed the ideal CG for this glider, it now flew like a charm, and I flew twice as long as the GRP retractable gear 4m glider with whom I shared the sky. The landing was much better as well.

This called for more, and during the next flight I stayed up till long after all others had landed. Despite a stiff cold wind, no sun and dark clouds, I stayed aloft for 23 minutes after tow release. I could have stayed longer but started getting cold, and decided to use the remaining altitude to check the spoiler to elevator mix. I quickly found out my mistake. With these funny original Phenol material spoilers extending above and below the wing, there is a sort of over-center lock which I do not understand the mechanics of, but is used to keep the spoilers flush with the wings, even at great speeds or large lift coefficients. Because of that, half the throttle movement is used before the spoilers actually start to pop out, but due to my linear mix, the elevator compensation already happened for half the value, causing the glider to already pitch up while no compensation was needed. I reprogrammed the compensation only to start when the spoilers actually start to deploy in the airstream. That will probably resolve the pitch and glide angle problem in final approach and during the flare. I like to be able to randomly play with my spoilers at any altitude and even during the flare, without having to manually compensate the elevator. After a few flights I had it nailed on my Multiplex Ka6e, and am convinced I’ll get it working on this Ka8 as well. After those flights totaling more than an hour battery time, I measured and was glad it still showed 75 percent capacity. The Ka8 servos seem to consume about double what those of the Ka6 do, but a day’s flying during glider gatherings seem within limits. A thorough post-flight inspection only revealed the old forward skid screw to be loose in the spine, it got replaced by a slightly longer and wider wood-screw. The crooked vertical stabilizer is very solid and planked, it will not be easy to straighten that out during the rebuilt. During winter, while undertaking the full restoration and installing a second elevator servo, I’ll probably swap the receiver for an AR9100 with built-in double battery redundancy. This large nice model is worth some extra precautions. With the added weight of the second elevator servo in mind, I elected not to remove any weight from the nose to achieve the new balance, but temporarily attached weight in the tail.

I am more than satisfied with the model so far, its aerodynamic characteristics seem close to the ones of the real Ka8, and it thermals much slower (and tighter) as my Ka6e. With any kind of thermals, I’m convinced I will stay airborne much longer as anybody else. The model is majestic in the air, gracefully orbiting until she comes down and lands in a short distance, her wing falling down after standstill like the real glider. All that remains to be done before taking her along to the June BIGGS meeting, is to apply registration markings, correct the few discrepancies mentioned earlier, and rearrange the interior to allow flying at the 92mm CG with the pilot in the cockpit.

The day after the maiden I passed by the former owner and recuperated the drawings of the wings and the rolls of white and red Solartex. Two days after my maiden I made what I hope are the last intermediate changes. The electronic transmitter trim anomalies (mainly rudder) were translated into mechanical adaptations. The quick-link at the elevator got taped up so it cannot open in flight. The spoiler opening operation has been synchronized electronically and the elevator mix adapted. All vinyl repairs over the spoiler servo access and wing rod repairs were removed and replaced by proper Solartex. That material seems to be very easy to work with, and is extremely strong. I also applied the white registration letters on the fuselage as per an old picture I found of the original full size glider

To get the CG right with the pilot in his seat, it required too much lead in the tail, so I again took the forward cockpit assembly out and removed 95 grams of lead from the U shaped ballast. That still was not sufficient, but with the weight of the future second elevator servo in the tail, I used 28grams of sticking lead almost invisibly between the horizontal stabilizer and the elevator joiner. With the Ka8 now balancing perfectly on the 92mm mark, I weighed the glider to check the final figures. My precision scale being limited to 3kg, I first weighed the wing assembly. Both wings with the joiners, 4 attach bolts and turtle deck, indicated 1980 gram. The fuselage with the tail and battery installed weigh 2995grams. The pilot another 205gram, so total flight weight is now 5180 gram. My weight reduction program in the nose seemed successful, and If those 90square decimeter are correct for the wing area, this means my Ka8b has a wing load of 57,5 gram/square centimeter, a very acceptable figure in the glider model world. I also found a way to efficiently stow the Multiplex Ka6e, Flair Ka8b, Parkzone Ka8b and Hughes 500 helicopter on the bed of my camper, and that weekend drove to the Netherlands for a glider fly inn at Sittard. It was the first time I flew gliders outside my usual club, and was slightly apprehensive in between so many experienced pilots of large gliders and towplanes. Everything went fine, I made 5 flights with the better visible Ka6, and 3 with the Ka8, each time behind different towships. With much more traffic than I was used to, I flew conservatively and was not able to further test or adjust the spoiler elevator mix, but at least it didn’t balloon anymore. All the work had produced a fantastic glider which was easy to fly and land.

An abrupt end for my 2014 season.

The Flair Ka8b became my favorite glider because it flew so well and easy, staying aloft in the smallest of thermals, and precision landings a piece of cake with those effective spoilers. This glider provided me the pleasure and confidence to “play with the big boys” at the various BIGGS gatherings. In September I took it along to a 4-day fly inn at Bastogne, but on day two when performing my preflight, I discovered only one spoiler was moving. I heard the servo move in that wing, so it must have been a failure in the linkage. Having no means to access and no idea how it all functioned normally, even removing some wood to unscrew 2 of the 3 pivots on a spoiler to allow it to pivot out of the way did not provide me visual nor physical access to correct the problem. With hardly any wind and the field being long enough, fellow pilots suggested flying without using the spoilers, and to make sure I wouldn’t accidentally move them or one of them popping out because it wasn’t properly locked anymore, I applied some strips of tape over both spoilers.

For the first flight I approached at my normal glide angle, but due to the lack of spoilers or headwind, I used most of the field to get it down, even using sideslip to kill height more rapidly. For the second flight I came in lower and still landed halfway. After lunch, for the third flight, I decided to come in with an even shallower angle, but had misjudged the wind which had risen, and angled in the baseleg to cut corners. Because we only were few guys flying and we had Eric’s powerful Rans towship, the pilots stayed at midfield instead of the approach end. All these factors combined led to me to think I would (barely) clear the farmer’s electrical wired fence (they had unpredictable bulls on the other side), but I was wrong and saw my lovely Ka8 coming to a brutal standstill. Walking over there I saw the port wing had hit one of the square wooden posts at the height of the spoiler, and it had penetrated for about 5cm deep and 10cm wide into the leading edge box, but still an inch short of the main wing spar. The damage thus was not too bad thanks to the very slow speed the Ka8 had at that time. I thought I had crossed the wire and was flaring when it hit the pole. With the bull on the way to investigate what that red glider was doing on his side of the fence, I quickly disconnected my transmitter pult and carefully stretched over the electrical wire to grab my cherished property and bring it back to safety. With the DX10t on my belly it would probably have touched the wire and I really didn’t want to find out if high voltage on the body or metal casing of my transmitter would cause harm.

This had again been a case of me being pushed to fly with a model that was restricted in functionality. I made a mental note never to do that again. What would probably have been a spoiler repair of a few hours back home, ended in a major structural repair job necessitating much cutting in the wing and the purchase of a suitable balsa leading edge profile. With the season ending, I decided it was over and out for the Flair Ka8b in this form. It already had performed about 30 flights, many more than I had thought of before its planned restoration. I for sure would have found out about the spoiler problem during that restoration, but delayed it too long, hoping to fly till the winter stop.
Last edited by BAF23; Jun 17, 2015 at 03:25 AM.
Jun 18, 2015, 02:22 PM
The sky is the limit
BAF23's Avatar
Thread OP

The spring 2015 total overhaul

After having converted half of my fleet to FrSky receivers, I had sufficient models to fly around and undertook the task of restoring the Flair Ka8b in all its glory for the new season. Because I wanted my 1:4 scale model of a real Ka8b looking exactly like the full-size one flying almost daily on the other side of our airfield, I started to completely strip the strong covering from the glider. The iron used to normally apply this material was now used to heat it up sufficiently in order for it to come off without too much balsa sticking to the cover, or paint to stick on the balsa. This required just the right temperature and small areas to be treated at a time, beginning with the ends of the overlaps. I couldn’t believe how many overlaps were found at some places, like his wife prevented the builder from using her scissors to cut off the excess strips, and he had to apply all this material on his plane. It took me hours to get it off, because just as any other material, it sticks better to itself than on wood, and in this case also was covered with a very thick coat of paint.

When all traces of textile paper had been removed, I was happy to discover that the wing had been very well built, completely straight except for one flap and small trailing edge part. The glue was still solid and what I initially perceived as cheap plastic servo’s proved to be year 2000 high-end quality Volz WingMaxx 3kg 20gr metal gear and bearing equipped 10mm thick servo’s, the first of their kind for flat wing use. The pushrod operating it was almost identical to the real Ka8, and because it had no play I decided to use it. The ailerons had been hinged solely by cover material strips over the extrados, and this seemed like an excellent solution (when using less layers !). Even with the wing open, I didn’t get any wiser concerning the spoilers which had just been glued to wooden supports rendering the mechanism unattainable without major surgery.

As you can see on an internet picture of a virgin assembly, the only nuts and bolts are in the middle section, all other pivots are tiny hollow metal tubes that are free to slide out one direction or the other. The previous owner resolved that problem by gluing balsa strips to each side, forming a complete cap over the mechanism. That had been fast and handy, but to be able to cut the center airframe glued chassis necessitated those hollow tubes to be pushed out. I therefore had to drill balsa away on both sides of the tubes so they could be pushed out one by one, thereby freeing the cap so I could see the problem. When the slider at the right gets pulled by a servo, it causes the inner arm to pivot to extend upper and lower spoilers simultaneously, the other two arms just blindly following without being actuated. When pushed all the way in, the shape of the slit causes a kind of mechanical lock that prevents the spoilers from being sucked out by the negative pressure on top of the wing. This clever and simple actuator setup depends on a single extra tiny hollow tube to remain in position in the shaped cutout between the actuator arm and the pivot. That is where mine had failed in Bastogne.

Repairing it necessitated the complete assembly to be taken out of the wing, and with the top spoiler away, I could use a long knife to cut the balsa free from the Phenol. After studying the problem I replaced that tiny metal tube (which I found in the wing) by a screw which I allowed free movement through the slit, but glued into the pulling arm (there was no other possibility because there was no space for a bolt). I made sure that even if the screw came lose, I prevented it from coming out by gluing a small piece of wood on the head side. After reassembly I filled the holes I had made with balsa and filler, and after sanding all sides smooth, applied cover tissue around each spoiler, getting an idea of the temperature I had to dial into my iron to soften the glue and create a solid bond.

After the cause of the accident had been remedied, it was time to tackle the repair of the other wing’s leading edge. After cutting away the damaged 1,5mm broken balsa cover, it became obvious how to make a solid repair.

I first cut the damaged leading edge at an angle, enlarging the contact surface of the shaped new insert, and making sure both sides always touched the front of the solid plywood front ribs. Being relatively wide I also cut half of the planking above the outside ribs to have better adherence and lateral junction to the existing balsa. A new rib was then fabricated to replace the completely broken one, but was glued in a slight angle so it could be bonded with the remaining bit. I also made a slight cut in the old and new leading edge balsa to accommodate the new 1,5mm cap and providing added bonding surface while later to reduce the sanding process. The new cap was kept in place with numerous pins and weights to make it follow the curve during drying. A day later I turned the wing around and after applying some more glue on the interior joints of the repair, I closed the narrower gap the same way.

After a flashlight inspection of the still installed serviceable spoiler on that side, I saw that the tiny metal pin was allowed to move out too much to my like, and to prevent a problem similar to what had happened on the other side, I was able to use of a pincer to push it back in and glue a thin balsa strip to prevent it from moving out again. That way that spoiler assembly didn’t have to be removed from the wing. I then used a compete pot of lightweight balsa filler to fill all the voids and steps in both wings, and to create that perfectly curved surface to allow a smooth finish with the textile paper. It took me four rounds of filling and sanding before I got it right, but what a difference (also on the ailerons).

The picture clearly shows that besides my leading edge repair in front of the spoiler, the port wing must have had another encounter with a pole or so in front of the aileron, but that had been repaired very professionally so I left it alone. Having the intention to cover the wings in one go and keep it free from patchwork in the future, I decided to build some access panels for the servo’s in case I needed to exchange them in the future. The Volz Wingmaxx aileron servos lay in a plastic box and can be easily taken out of it, the previous owner just kept them in place by gluing a balsa cap over them and then permanently covered them with soft material. I cut those caps off and glued additional recessed plywood parts around the servo, not only for solidity (two ribs had been seriously weakened), but also to provide a good grab for the screws of the styrene plates that will be inserted and fastened flush with wing. Their role is triple: ensure the servo stays put in its tray, provide quick access to exchange a servo without damaging the Solartex, and form a solid base for the pushrod fairing which up to that time I had just screwed loosely into the balsa. A similar setup was installed to gain access from below to the spoiler servo’s.

With the wings ready for covering, I tackled the horizontal tail. Whatever the quality of the elevator servo, I prefer to have one for each elevator instead of linking the surfaces (I like redundancy on larger models). I therefore had to cut through the horizontal stabilizer to find out what was in there. I found the same quality servo as in the wings, but with that servo model being discontinued, there was little chance finding another one to install on the other side.

After extracting it (the blue one in the illustration) I decided to keep it as a spare servo for the aileron actuators, and choose a pair of smaller cheaper Hobbyfly metal gear analog servos as a substitute. As the elevator pivot points looked solid, I used a metal grinder to cut away the middle part linking both elevators, leaving just sufficient stub length to fit in the existing metal holes. After some balsa surgery and the installation of support blocks and production of styrene cover plates plus the soldering of the servo wires into a dependable Multiplex green plug (using separate pins for all the connections), the tailplane could be further prefabricated. I used this stage to gain proficiency in covering open surfaces such as on the elevators, with one go of the previous owner’s supplied white covering material. I quickly gained proficiency and experimented with the various temperature settings on either the application iron or the heat gun.

That gave me the confidence to tackle a complete wing in one go, but that was also the moment I discovered that my additionally purchased white Oratex paper was only 60cm wide whilst I needed 68cm at the wing root to cover the top and bottom of the wing in one go. What the previous owner had provided me was about 3 meters of 70cm wide white cover material, and in that width had to be Solartex. Both materials seemed rather similar in texture and appearance, and although having more recent Oratex material in stock, the width made me decide to apply the Solartex material on all horizontal surfaces, and the Oratex on the fuselage and vertical surfaces. Halfway the restauration I had most of the sub assemblies in various stages of assembly/dis-assembly and took a picture of how far I got at that time.

Previous picture pretty much shows the magnitude of the restoration work, but also the extremely strong (labor intensive) design of the Flair Ka8. The previous owner must have spent hours preparing and gluing all those small structural fuselage pieces, but had done a fantastic job getting it all together. I first had been suspicious after discovering his unacceptable finish and accessory integration, but this time was glad to see very few flaws in all in the important structural assembly.

I bought water based white glossy lacquer paint and applied it to the re-covered moving elevators as a test. Although the paint remains flexible, I had expected it to better flow through itself and as such produce a super smooth surface over the dull Solartex. It actually was more difficult to apply than expected, even with crossed strokes of a specially purchased quality soft brush. I then wanted to apply the Cadmium yellow shiny Oracover over the fixed horizontal stabilizer, but it was so transparent that even the wood nerves of the balsa sheets showed through, and also the mix of underlaying material colors. I had no other choice than to first cover the entire stabilizer in one go with white opaque Oratex, and then apply the yellow Oracover over it. The stunning result is well worth the 7 hours I spent completing those tasks. Although I’m an old-school builder, I previously never used Monocote or the more modern variants. Just as I had been able to avoid home computer use until 1995, I kept using silk and dope for much longer than others. I probably am still much slower than others with my application iron, but start getting a feel of temperatures to be dialed in with the various materials and shapes to be covered.

A second coat of paint on the elevators very much improved the result whilst adding only 2 grams per elevator. After covering the second wing I pulled the knot on how to finish them. Previous experiments had shown that Oracover over Oratex was working fine, but Oracover over painted Solartex easily developed bubbles because the trapped air had nowhere to go. Contrary to any logic I thus decided to apply the yellow Oracover leading edge area (in one go) over the unpainted Solartex, and only then paint the remainder white on the wings. Those 200x23x9cm yellow accents were huge and difficult to handle. I used the factory side cut as the straight line for the top of the wing, and painstakingly drew a 2m straight line on the protective cover and used scissors and a steady hand to make the cut. Trying to cut with a scalpel knife through the back of the protective paper and hoping on a clean cut through the very thin Oratex underneath was not tempting, the more I had to do it on the floor because of the length. I was surprised how straight the cut appeared to be, but small deviations were tolerable because it would be on the bottom of the wing. To get the line straight on the extrados, I first applied a strip of stretched straight masking tape as a guide, then ironed the Oracover touching it. I then further ironed the Oracover towards the leading edge, then around that and ending up where the material stopped. I then lifted the blue tape from the wing and pressed it on the adjacent end of the yellow Oracover to obtain the safe area for painting. The superfluous Oracover protective sheet was then taped to the blue tape to protect the rest of the yellow decoration during the ensuing wide streak painting. Following picture shows the situation at midnight on the kitchen floor, bathing towels underneath to protect the Solartex, horizontal tail/painted elevators/servo access panels/painted spoilers added in along the iron for grouping purposes. Yellow Oracover having been ironed next to its straight line blue guide tape, but still having to be affixed all the way around the leading edge and back on the intrados.

I voluntarily cheated a bit towards the wingtips in order to have more yellow material visible from below than from the top (for contrast in the air). After 2 hours of application (necessitating numerous small temperature adaptations on the iron), I was satisfied of the job, still not a single visible wrinkle on any of my horizontal surfaces! It took me another day to duplicate the operation on the other wing, and apply two coats of white lacquer to the surfaces. The water based white paint was very thick (but was supposed to be applied without thinning when using a soft brush). As with anything else, after some practice I finally was able to obtain an immaculate finish without streaks or drops, but it required a delicate touch for the last of each partial surface multiple direction passes to get it right. With about 6 hours between applications, and only being able to paint one side of one wing at a time, it was well past midnight when I was able to remove the masking tape. I hate to keep that on longer than 24 hours because it then starts sticking stronger, and the fully dried out paint then has a tendency to chip off at the edges. With the last of the surfaces being delayed for painting till the next morning, the first 3 masking tapes came off without chipping, but the recommended 45° lifting angle tended to lift the yellow Oracover from the balsa, and tear the not yet fully dried out paint with it. A very slow close-to-180° pull felt more comfortable and produced excellent results.

When first trying to insert the spoiler assembly back into the finished wing I was shocked it physically was impossible to squeeze the wooden blocks (which were glued to the fixed middle part) through the narrower slit in the wing. Taking it out had been possible after cutting the blocks away through the removed half of the spoiler cover, but with everything now fully assembled and covered/finished, I first saw no possibility to get it back in and certainly not to connect it or even glue it to the wing. After much thinking and modifying the spoiler blocks and their counterparts in the wing, I found a way to get them in (together with the actuator rod) after epoxy had been applied to the forward flat receiving block parts in the middle of the wing, but care had to be taken during insertion (in extended position) to keep the actual spoiler one millimeter from the 30min epoxy. When they were positioned correctly with the wing still resting on the leading edge, I manually pushed the spoilers into the closed position and used credit cards between the spoilers’ (top and bottom) and the wing slit to keep everything centered during the drying process.

Just before the 30 minutes elapsed, I carefully moved the actuator rod through the still open access panel to observe what and how it moved. As there was no way to reach the attachment without cutting one side of the spoiler off, this was the moment of truth and if unsuccessful, the last moment to attempt to break the spoiler assembly free before the glue settled. I got lucky, but a few hours later noted one extremity of one spoiler could possibly cause a jam during retraction, so I had to carefully cut a small slice of one the vertical caps to eliminate that problem. I then opened the spoilers again and was able to slid 1,5mm balsa plates between the side that had not been epoxied, and the wing. These plates were imbibed with white glue, easy to remove from the finished surfaces after the slide, but sufficient to make a solid bond at the trailing edge of the spoiler assembly. All this certainly was not my preferred way because I hate to make fixes that cannot be accessed later.
Wings and horizontal stabilizer were then fully assembled and all access panels, covers and skids adjusted. The large (two feet wide) OO-ZAU registration letters from Caliegraphics were then applied under the wing, taking into account the positioning versus spoiler operation and its servo inspection panel. Not easy doing that all by myself, because once the back protective paper is removed, as soon as any part touches the wing, there is no way to remove it anymore without stretching both the letter and the painted Solartex it came in contact with.

Detailed inspection of the fuselage dictated that the GRP nosecone was in need of serious seam cosmetics after my quick insertion of a quality hook in the nose, and that the overflow between GRP nose and plywood fuselage panels was anything but smooth. Furthermore the awful huge step between top fairing and canopy base had to be eliminated, and a mechanism installed to allow for canopy closure without having to use a screwdriver. I started by removing all the extra wood blocks that the previous owner added just to screw on cable loops or fasten electronics, because I intend to substantially improve the interior within the limits of the untouched strong stringers and plywood formers. Practical canopy operation also meant that the horrible outside battery switch could be eliminated and a more discrete switch placed inside the cockpit. That hole in the plywood got patched up with epoxy, and pieces of wood glued behind the cockpit to keep the battery in place whatever forces would be encountered. Because the nose section probably will encounter some blows during times ahead, I opted to use liberal amounts of epoxy glue to close all the gaps and depressions plus junctions to the wooden frame wherever I still could reach them. Additional weight in the nose never is a problem on scale gliders, and hoping never having to reach the towhook nor lead again, I wanted the entire nose to be one very strong unit. After much filing and sanding, the last smoothing was done with lightweight but strong filler.

Getting the canopy to fit without noticeable seams took more time than anticipated. A thick layer (about 3mm) of filler on top of the GRP nose was not tempting, neither structurally nor esthetically. Due to the structural integrity of the canopy and its shape forming wooden frame, lowering the front could only be done by lowering the entire cockpit side’s 7mm wide strong wood rails. The trapezoidal shape of the canopy frame also dictated that if the front goes down, the aft moves back and because that was formed by another strong (ply?)wood plate resting on a similar one on the airframe, complicated the problem. Furthermore, the iron pins keeping the front down couldn’t be repositioned, and drilling additional holes into the forward fuselage bow weakened it so much that I had to fabricate small metal plates to be glued into carved out recesses in the bow, to keep the front of the canopy snug without ungainly gaps. With the aft of the canopy extending into a wing fairing, it was easy to use the space underneath to fabricate an angled platform on which a spring loaded tab was allowed to move. A small slit was cut in the canopy/wing fairing to allow the slider to protrude for activation. After some experimenting, I found the ideal angle for the assembly so that when closing it, the angle of the canopy back-plate pushed the pin back until it slid into the hole to keep it secured during flight. It wasn’t easy to get most of the grossly brush applied paint from the canopy assembly, and with the previously mentioned and patched up Perspex problems, I knew it never could turn out nice, but good luck finding another clear canopy for this extinguished Flair kit.

Only when everything fitted correctly did I start the delicate job of cutting out a quarter-scale vent window. That job is by far the most apprehensive jobs I undertake: one slip with the knife through the thick canopy material and you will see it forever, 1mm too wide and the sliding window will fall through the opening, and the 1mm copper screws to fasten the clear rails are just a nuisance to get in with thick fingers. Each of the 6 screw holes first has to be drilled with a 0,8mm finger drill in the rail, then 1,2mm drill in the clear Perspex, then cone widened with a 2mm drill so the copper screws rest flush with the canopy when fastened. I acknowledge that installing such windows is close to pure masochism, but it adds a very visible scale touch to any glider and also allows much needed fresh air in the greenhouse when parked in the sun. A big bonus is that by judicious placement of both, I can slide the window open from the outside, and with one finger through the vent window of the fully locked canopy, move the battery power switch between on and off, limiting the opening of the canopy to only change the battery. I then made a temporary balsa inset cockpit floor to have a pattern to make a stronger removable plywood one. This had to be done before the fuselage was covered. All fuselage stringers were re-glued where necessary and rounded off by sandpaper in order to have more realistic “tubes” outlines instead of sharp lines or flat zones showing through the cover material. Loose cross members got glued and the bottom box reinforced where the rubber of the skid had dented it during a hard landing by the previous owner.

The skid under the tail had been a continuous source of problems. The flattened copper tube was only held by a single screw on the forward side, and at the back the rubber damper had just been glued to the cover material. Needless to say, every time the glider ground-looped even skidded just a bit after landing, the flat skid stuck into the terrain and became loose, preventing further flying. I resolved that by flattening the tube a bit more at the front, so two screws could be bolted in tandem in that reinforced plywood section. I then tried to drill a hole for a screw through the rubber damper, only to find out it was completely brittle by age and had to be replaced by a similar custom produced rubber block. Further filling and overall sanding in anticipation of Oratex corners was then made, but I had to leave the wheel in position because the axle had been soldered to metal side reinforcements and could not be separated unless using a grinder. Except for filling the huge structural gap on top of the gear compartment, I left the wheel the way it was, hoping never having to replace it later.

In order to get a feel of the Oratex application over open and close undersurfaces, I started with the vertical tail and rudder. As I encountered no problems, I had a long detailed look at the fuselage to decide how I would tackle that. I quickly eliminated the one-flap option because the 60cm wide Oratex was too narrow to pull around the complete fuselage at its thickest point. Being unsure how the material could be made to follow the nosecone, I used a leftover flap which I attempted to stretch over the bottom nose, the way the single yellow Oracover would be applied over the white Oratex. To my surprise, when the heat gun got pointed at the cover at 240°C and from about 3 inches, I was able to stretch the material quite a bit to get a nice smooth surface adhering very well on the GRP nose. The same heat got it back off, but the nose remained tacky from leftover glue after separation. After more measuring and using large paper patterns to see what was possible with a 30cm wide strip, I got bold and decided to attempt to cover the complete fuselage with just a left and right 175cm long stretch of Oratex.

I was apprehensive, and just as with the stretching exercise, I missed an additional pair of hands, but good planning for a median straight iron pass went well. The lack of hands and complex fuselage curves led to the intermediate situation depicted in previous picture, not the best surface to start heating up to get it smooth. I’m sure a lot of people would strip the whole thing and start again with smaller patches, but I never lost faith. After a complete afternoon of stretching, gluing, shrinking, more stretching, more shrinking with the heat gun, and so on, I was proud to have one half of the fuselage completely covered without wrinkles. I carefully rounded the Oratex over the top and bottom of the fuselage stringers behind the wing, then cut the excess material off with an extra sharp knife exactly over the fuselage extremities. I also made a sharp straight cut in the middle of the top of the nose, the only but very visible place where left and right side meet on relatively flat surface. I then started applying the other half of the fuselage, but this time stretching the Oracover of the edges over the first applied material before cutting the excess off. The nose bit again required quite a bit of dexterity and I often had to retract my fingers in panic when the heat of the gun blew along my nearby fingers stretching the material to its limits.

Up to the point of overlap, I was very satisfied and cut the excess material off shoulder-to-shoulder on the top of the nose. A bit of filler and 3 coats of white paint later eliminated all junction traces. Only when the nose was covered to my like, did I move further back to cover the rest. It was quite a serious work, but the first result was very encouraging, certainly compared to the patchwork application of the previous owner/builder.

With the model inverted I measured how wide the yellow strips had to be to cover the bottom and next cut a rough shape of it on one side, making sure to have a perfectly straight line for the division.

My hopes of doing it that way soon vanished, because the Flair Ka8 is not scale but semi-scale, and in a lot of previous pictures you can clearly see that unlike the real glider, the model’s outermost fuselage stringer doesn’t prolonge straight along the rigid/soft cockpit division towards the center of the nose. Furthermore, the compound nose curves and imperative separation line ending just below the middle of the nose end at the towhook opening, forced me to create a subtle bend to follow the contour lines, but still looking almost straight along the whole fuselage.

The result looking acceptable, I taped the yellow and applied two coats of white lacquer to the rest of the Oratex of the fuselage, then applied the remaining decorations and installed the pilot for a first weight and balance. This proved within limits, my model then grossing 5350gr or 200gr more than before. The last modification I did was using a combination of dowel with a magnet, magnet in the fuselage, and bent piano wire to hook under the rear wing joining spar, so the turtle deck can be just slipped into position instead of being bolted on top of the wings. With an important gathering of large gliders around Tongeren for the long Ascension weekend, I decided to cut short on the rest of the restoration and get it flight tested on my club field the day prior to the start of the long weekend. I installed a FrSky X8R receiver and adjusted all throws and neutral points to the values I had measured with the Spectrum equipment just before dis-assembly. I balanced it at the exact same CG that I had so successfully flown it before, expecting a straightforward new maiden.
Last edited by BAF23; Jul 30, 2015 at 06:19 PM.
Jun 18, 2015, 02:52 PM
The sky is the limit
BAF23's Avatar
Thread OP

Disaster during flight testing of the fully restored model

Benny came out especially for me with his Calypso 62, and after I assembled my gleaming Ka8b, I had Phaedra perform a buddy check on the model before its maiden. She declared it flightworthy for a maiden, but we got a few eyebrows raising regarding the serious down elevator neutral point (but that was how I had flown it before). She pointed out a few other shortcomings, the main ones being excessive rudder to tail friction (caused by the thick Oratex being pushed into the hollow part of the junction, and excessive elevator play at the servo side). After a last test of the towline release under high tension, we got it airborne. It left the ground after a few meters and it wasn’t very stable in pitch during the tow. At the release point it just staggered up and I had to use almost full down elevator to have it fly a normal glide. It didn’t fly well during the turns and dropped a wing at speeds higher than the previous stalling speeds. I got it back on the ground but it was obvious that some drastic changes had to be made. After getting the heads together the only reason I could think of was that the rubbers I had installed between the fuselage and wings plus horizontal stabilizer, had modified the previously flown incidence angles. With no way to measure it we thought the best solution to continue that day was to change the decalage angle by cutting off some rubber and asymmetrically tighten the bolts that held the horizontal tail in place on top of the tail section. I then carried the model for a second tow from the other end of the runway because of shifting winds.

The following flight was an indication that we were on a good thinking path because the tow was steadier and it didn’t stagger so much upon release. After trimming out, the CG dive test showed no signs of recovery, so it needed additional weight in the nose. I was really surprised about the fact that this Ka8 had to be re-rigged to such an extend after a primarily cosmetic restoration. With Benny’s towship out of service due to a rudder problem, the test day was over and the rest had to be done home with some guesswork, but at least I knew what to expect and was glad I made both those flights at my home field under variable wind and disturbing cloud background conditions. As Phaedra had time to shoot some images of those proving flights, I’m glad I can share these with you.

EBZW Flair Ka8 (2 min 11 sec)

Back home I glued extra balsa on the forward seating of the horizontal stabilizer, diminished the amount of down elevator in neutral trim, and added 41gr lead in the nose to bring the CG halfway Flair’s recommended range. I know this is anything but scientific, but this had to do so I could participate at the large-glider meeting. During the first day I made 5 flights, each time fine tuning mechanical servo arm adjustments to obtain better flying qualities with zero trim deflections. Weather and other factors made me skip a few flying days, but when most started to leave on Sunday (also due to a stiffer wind), I was able to perform another 5 flights one after the other, some of 20 minutes. On one of those, at 400m height, I got a telemetry warning of battery low immediately followed by battery critical. Not wanting to take risks, I pulled the spoilers and came in to land, only to see and measure the battery still had 60% capacity. Back home I exercised the servo’s to battery depletion to check if I could duplicate the warnings, and concluded that it probably had been the relatively loose connection between the battery balance plug and the voltage sensor that had been the culprit during maneuvers in tight thermals. I changed the programming of the warnings so that if both voltages not only passed, but stayed below the chosen levels, I would get repeat warnings every 15 seconds so that I could hear the difference between a momentarily drop and a definite drop.

A few days later I was flying at our club again and Chris pulled me up with his old twin-tailed towship. The thermals that day were unavoidable, and I often had to pull full spoilers to come back to the 300m region instead of being quickly sucked up to 500 meters. On the other hand, during that half hour flight, I had ample time to evaluate the pitch trim settings versus the different spoiler positions. With such unstable weather this wasn’t easy, but I soon found out and mentally noted the trim positions on the transmitter. After that flight I made some adjustments again and the second flight proved I was nearing a perfect balance. The position during the tow, the CG dive check recovery, the stalls and accelerated stalls, everything was like I wanted. These serious thermals also caused the wind to regularly and abruptly shift from one side to the other. After my wishful last mechanical neutral elevator adjustment, I awaited my turn for another flight before making the pictures along the identical full-size glider which had been purposely parked along the nearby hangar. After hooking up at the runway end, I looked at the windsock and changed the left wing low into a right wing low for the full crosswind. I then quickly walked back to the pilot square and told Chris I was ready.

Being in a hurry to vacate the runway, I forgot to check the windsock positioned behind us. I thus didn’t notice the wind had veered completely to the other side, and as soon as we were rolling it caught under my high wing and even before the towship was airborne, I was facing a hopeless 90° right bank despite the full left aileron and rudder inputs. Our former president was shouting “los” meaning to open the hooks, but on my new Taranis transmitter in the tray I still was flying with the original small sticks and buttons, and with full left rudder (mode 2) my little finger which I normally use to flip the release button forward with, was only finding air. With me being pulled further to the side of the runway and dragging the tail of the towship downwind after he lifted off. Instead of risking the release and flying at too low a speed into the safe area (but with trees), he cut his power and crashed into the grass just to the left of the runway, me still being attached and dragged inverted on the right side of the tarmac. His gear had been torn away and that had caused damage to his adjacent long exhaust pipe. Except for a few servo gears, not much damage was visible. He was a sports and concluded although not difficult, he wouldn’t repair it and now would concentrate on finishing his nearly completed home designed new towship (a much enlarged CalamatoW). When I turned my recent marvel around, it didn’t look too good. Amazingly, except for the vertical tail, there didn’t seem to be much structural damage, but about every single subpart showed very visible major aesthetic damage from being dragged inverted along the tarmac. Instead of taking the pictures along the full-size example, I carefully dismantled my Ka8b and drove home very disappointed.

I felt terribly disappointed for Chris, but also thought the work I had done the last months had been useless as I never would be able to repair that model to the same high standards. I couldn’t believe that I again had pushed myself in working interminable hours to (half)finish the model for an event, but delayed the essential modifications on my transmitter: longer sticks, larger and anti-slip switch for tow release on one side and engine power cut on the other side. Once home I took the time to assess the options and concluded this could become a feasible repair. Instead of doing it quickly, I decided to take my time but also ensure that all the deferred works had to be corrected or executed before I would take the air again, even if it meant skipping a few weeks of flying and nice large glider gatherings. First of all I made a new to do list which was longer than expected. The following pictures are an illustrated resume of the front to back damage. The top of the nose had been deeply fractured and my so well blended left and right Oratex halves ruined by the tarmac.

The thick clear canopy had very deep scratches in two directions and I first considered it as a total loss. My new way of attaching the canopy had been torn away and even couldn’t be dislodged on the field.

Both wing leading edge tops showed considerable tarmac scraping damage and even perforations in the D balsa capping.

The trailing edges also showed damage, some by stretching or even tearing the Solartex apart.

On other places the balsa caps had been compressed causing undulations in the trailing edge and depressions in the shape. The repaired spoiler had become dislodged again and had been dented. It also became jammed inside by torn cover material edges.

When falling on its back, the shock had caused not only a serious compression on the top of the vertical fin, but also the separation at a strange angle between the complete vertical tail and the aft fuselage.

One elevator servo seemed dislodged from the horizontal stabilizer. Last but not least on my list were the installation of dual power batteries by way of a Schottky diode, the completion of the scale cockpit, and the necessary modifications to my transmitter.

First job was to remove the Oratex from the tail section. The damage looked worse than I had imagined and required more sorts of wood than anticipated. It had been the Oratex that kept the parts together after the crash, once removed I could just twist the vertical stabilizer free from the fuselage, the latter also needing structural repair to the many stringers and reinforcements. When I removed the Oratex from the top of the tail, wood parts stayed attached to it, leaving a distorted and open fin to start with.

Every broken piece first had to be cut out at appropriate angles to allow a solid bond to the new parts. I used the ultra-strong PU wood glue and pressed or clamped each part together during the slow drying process.

It was a delicate work that took days to dry and smoothing out with filler, and I took the opportunity to better finish the capping over the rudder and increase the tolerance to avoid rudder servo strain. I also modified the seating for the horizontal stabilizer so the assembly would become stiffer at its experimentally found new angle. I then applied new Oratex material and blended it with the existing one under the shadow of the elevator so it wouldn’t show. I must say I was rather pleased with the repaired result. Here is a picture of it prior to the application of the other part of the club badge.

In between the drying of the tail parts, I tackled the wings, starting with that bloody spoiler. Although it was loose, I wasn’t able to get it out without again cutting in the blocks that were supposed to have kept the assembly in place. You might remember I had to glue them in place by applying expanding glue with a toothpick through the narrow gap between the open spoiler and wing. That obviously had not been sufficiently solid so this time I did it properly, even if it meant more cutting, custom gap fillers and more glue. The kit spoiler design doesn’t allow for a proper method to insert/remove a spoiler after the wing is assembled and covered, so I hope it now will stay put for the rest of the glider’s life.

For the repairs of the rest of the wings I was able to make small holes on the underside of the trailing edge, allowing me to pour expanding PU wood glue through them, with the wing resting vertically on the trailing edge. The glue then spread by gravity between the distorted or compressed upper and lower balsa strips forming the trailing edge. During the expansion process I used clamps over metal rulers that extended along the straight parts of the trailing edges. After drying I thus obtained a completely straight trailing edge with all depressions well filled with rock solid glue.

On the leading edge I was able to use the heat iron to remove the upper yellow Oracover. After measuring I found out that my remaining spare cover material was insufficient in length or width. I therefore decided to only cover the portion outboard of the spoilers, and only to the upper side. That complicated the repair compared to replacing the complete yellow leading edge in one seamless go, but remember I only painted the wings white after the yellow had been applied. Doing this in reverse would have required the purchase of new Oracover (shipping or fuel costs being more expensive than the required material), and a very difficult cut to be made to form the bottom straight line. After using a knife to cut off only the top of the leading edge yellow, I filled the balsa perforations with PU glue before sanding them flat in preparation for the Oracover. Note that I didn’t touch the original strong white unpainted Solartex. For the new yellow top leading edges, I started with a straight line on top, right along the painted white Solartex, fixing it with the iron and cutting it with a sharp knife a few millimeters below the leading edge, then pushing the rest of the material down with the iron, produced an invisible seam between old and new material, except for a small overlap forward of the spoiler. Wing repairs went fast and the result is also pleasing, visitors have to be pointed out where I knew there were flaws.

Because reaching the inside of the nose tip was impossible, the easiest way to repair the serious crack in the polyester under the painfully stretched and overlapped Oratex, was by Dremeling the crack into a V-shaped bed through the Oratex, then fill that with 30minute epoxy to obtain a solid bond between the split parts. After using the rotary sander to flatten the epoxy, I then applied lightweight filler over it and when dry, hand-sanded it to follow the compound nose curve. I then taped-off the white following the cockpit cutout, plus the yellow on the underside, and applied two thick coats of white lacquer over the complete front, which I then rubbed and buffed to cover the badly scratched Oracover and repaired crack. Note that I didn’t use a single patch of new Oratex or Oracover because the overlap would have been very visible, whilst the buffing allowed me to obtain an invisible seamless transition between the original paint and the extra layers.

The canopy was something else. Because I thought the deeply and heavily scratched example was only good for the trashcan, I ordered a replacement from the German Modelstudio who still produces a copy of the Flair Kit. When I got it I was disappointed, instead of the thick Perspex this was a thin acetate sheet that had been shaped over a mold. When I put it over the damaged one I could see that at one certain angle it hugged the old shape, and much extra material then could be cut off. That thin canopy was not suitable to install the sliding vent windows I had, and I was not sure I would be able to separate the frame from the damaged Perspex. I thus abandoned the idea of producing a scale side pivoting canopy, and because I had nothing to risk, gave it a try to eliminate the deep scratches with the rubbing disks in the Dremel accessories box. The Plexiglas was so thick that after about an hour of careful Dremeling, all the scratches were gone and there still was sufficient material to have a very stiff (but now opaque) canopy. Of course the machining had not been perfect, resulting in a sort of faceted surface.

I then used quite a bit of 400grit abrasive paper to get the smooth curves back. This took hours because although the scratches were gone, there still were multiple dots visible from the tarmac stones. When those also were almost eliminated, I switched to 600grit paper, and finally 1000grit with water to obtain a baby smooth surface. Unfortunately, it remained opaque and now was twice as large as the original scratches. Next attempt was with a rubbing and filling compound intended to repair scratches in car paint. After much hand rubbing of various applications, the canopy was translucent again but the pilot and instruments still looked a bit blurred. With the many lustering products I had in my garage (also to polish the camper’s plastic surfaces), I experimented and used my large rotating car polisher for a few hours on the canopy, which I held in position on a carpet between my two bare feet. By the end of the day, I again could read the instruments through the canopy and if not told about the problem, you probably would think it was the way it came out of the mold at the factory. I was very glad I could use the original again and didn’t have to swap frames or transfix the scale vent window. Nobody believes that what they see now, was that deeply double direction scratched canopy they saw after the crash. To me it was well worth the hours of rubbing and buffing because that canopy is very representative and visible in the original Flair model.

Having a nice clear canopy again, I decided to spend some extra time detailing the cockpit. With Flair’s heavy wooden crossmembers, it was impossible to duplicate the real aircraft’s tubular structure underneath the wood/fabric, but I used a picture I took of the real glider in the hangar, to duplicate as much as I could of the rest.

The Instrument panel still has very visible traces of its old OO-ZAU registration, but after hail damage in Germany, the original glider has been repaired and now flies in Belgium under the German registration D-2945. Because I preferred my model to portray the glider in its original Belgian registration, I just omitted the added German placards. I first took out the incorrect shaped balsa instrument panel with fictitious instruments (an artificial horizon but no airspeed indicator!) and cut out a new panel with the correct shape. Unfortunately, a structurally important crossmember (supporting the towhook release servo) now became exposed underneath the rounded pilot-leg cutouts, but by painting it grey as the rest of the interior, it doesn’t disturb the looks too much, but obscures the baby-blue rudder pedals I produced. As I always have a complete pilot in the cockpit, this discrepancy is visually acceptable and helps block the pilots’ legs and feet during turbulent flights.

By zooming the purposely made pictures of the original individual instruments until they came out of the printer in the correct size, I was able to insert them between Staufenbiel supplied quarter-scale cockpit instrument shapes and their plastic windows, but using the exact instruments with correct green and yellow arcing for our club’s Ka8b. The compass was a leftover of a scale cockpit supplied for the Blanik, but not used because it was not conform to the real Slovenian example. An ear piercing jewelry was used and painted to represent the yellow hook release knob, a blue similar spoiler lever installed at the left, a wooden floor with anti-slip mats on it at the bottom, a custom stick fixed to that removable floor (to access electronics hidden underneath) and not scale in order to fit in my pilot’s repaired right hand, and the yellow fake canopy opening handle to the left.

I then made a kind of generic seat to accommodate the pilot, but being also quickly removable without tools to change the batteries or work on the electronics. On the right side I fabricated the very visible movable green trim lever, but due to the Flair transverse was unable to install the map pocket at its correct place. If not told, even the pilots flying the 1:1 glider will not notice, as the printout scale map in the pocket catches all the attention. You have to admit that compared to the original it now can stand the test of anybody but competition scale judges. It is all a bit rough for the purists, but I never intended to compete with this glider, I want to use it as my primary 4m class towed model that is vice-less to fly and easy to operate and land, yet pleasing to look at from any angle or distance.

With such a nice interior, it became time to for cosmetic surgery for my battered pilot, because after the Ka6e crash in Germany during 2014 he looked more like Frankenstein in the face, was missing a complete left forearm and half of his right hand. Epoxy, recuperated hand parts, lightweight filler, paint and patience were all I needed to for him to pass his medical and regain flight status ;-) For his safety he gained a 4-point safety strap (that I glued to his yellow flight suit for easy removal during the battery changes).

Besides the visible work, I also modified the power supply with a second battery behind the pilot’s seat, only one of them being telemetry monitored by a cell voltage sensor through the balance plug, but both being connected a Schottky diode and the battery switch to the receiver located in the extremely strong plywood bottom cockpit box, together with the variometer and altimeter. The Ka8b is now not only cosmetically in near perfect condition, but has added redundancy built-in wherever possible (telemetry, dual battery, dual elevator servos etc). In the meantime I also modified my transmitter with longer sticks and wider release buttons with anti-slip rubber and other tactile sensitive materials for all the switches that have to be operated by feel instead of being visually identified first, or possibly missed when attempting to operate in anger. This whole episode set me back another two weeks in the flying season, but halfway June I’m ready to join the rest of the circus again, much better equipped and prepared compared to the rush a month before. I cross my fingers that I will be able to operate this superb and fine handling classic glider for years to come.

After such extensive repairs and modifications, I made a new weight and balance and now the complete fuselage in flying condition (including the dual 2S2100 batteries) weighs a hefty 3625 grams, the port wing 966 gr and the starboard one 975 gr. Adding everything up, the scale tips at 5564 gr for an 85mm CG, resulting in a wing load of 61,8gr/dm2. I’ll never forget the lesson to look at the windsock when in the pilot square, before telling the pilot that I’m ready to go, and under shifting crosswind conditions will wait until it blows from the low wing side before giving the go. At that time I will not only have my fingers on my longer sticks, but make sure that my left pink will physically rest against the hook disconnect switch, ready to actuate it by just a little forward push. Hopefully I soon will be able to complete these postings with pictures of my scale model next to the full size one, or some videos of the majestic classic shapes of this attractively painted model gracing the skies during the large glider meets somewhere in Europe.
Last edited by BAF23; Jul 30, 2015 at 06:24 PM.
Jul 30, 2015, 05:46 PM
The sky is the limit
BAF23's Avatar
Thread OP

Unlucky OO-ZAU

After the major repairs I took it to our field for a couple of checkflights, which were uneventful and I even was lucky that during that session the “real” gliders brought their Ka8 back to the hangar. My usual photographer grabbed her camera and soon we were able to make some nice shots of the real former OO-ZAU (sporting German registration now), and my scale duplication of it.

Sometimes I regret not having finished mine like its actual big brother, but I prefer to preserve a piece of Belgian glider history, and think the large blue callsign letters on the tail (that reveal his ancestry) spoil the lines of the original discrete finish.

The fun was of short duration because a few days later I took it to a BIGGS gathering at Schaffen. It was a bit more windy than forecasted but nothing exaggerated. The only other member of our club present was gathering the pieces of his 5 meter glider which he totaled after mistaking another glider for his, thereby unknowingly steering his straight into the ground shortly after release from the towship. After assembling mine next to his dis-assembly, I carried it to the lower laying field but experiences difficulty keeping it on my shoulder due to some gusts. I noticed several gliders to be idle along the grass strip and an unusual short queue for towing (that should have rung a bell). In the meantime I was talking to other pilots and inquired about the local rules. With the wind mostly blowing in the axis and few gusts, I asked somebody to hold my wing level for the start of the tow. The tow went well and I released at 300meters, but although picking up a zero left and right, never got a good thermal and soon joined the pattern. Another glider followed within a couple of meters so we agreed that I would land on the left side and he on the right.

My approach was super stable and well lined up for a touchdown without spoilers along the pilot-box at the beginning of the runway. At about 1-1/2 meters from the ground the wind picked up and rapidly sent my Ka8 up in the sky. I responded by pulling full spoilers which only leveled the model, but with decreasing speed and the wind shifting momentarily 90° to the right (as it often did later that afternoon), my Ka8 flew well past the pilot box but was veering to the left even as I applied full right aileron and rudder. I saw it hit the ground in an almost level attitude but still in a sideslip, stopping against the nose of a nice scale 6meter ASW28 which was parked there because the owner was towing with his Pilatus PC6. After carefully separating the models it revealed that he had no structural damage, but some paint scratches on his nose and canopy (which he wasn’t happy about because of his recent immaculate new paint-job). At first glance I saw no other damage to my Ka8 than an opening on the left fuselage under the wing, but a peek inside showed the two adjacent thick plywood fuselage formers to have substantial damage.

During the disassembly of my model I observed many other very experienced pilots experiencing trouble with even heavy polyester gliders from the catabatic winds flowing over and around the hilltop. I didn’t stay much longer and returned home seriously disappointed. During the trip back I lined up some facts, as can be seen on the previous you-tube film, during a perfectly normal landing the left wing inexplicably dropped a fraction before touchdown. The speed seemed normal but at that moment the shifting wind blew from a 5 o’clock position and I had the spoilers fully open. Also remembering the takeoff accident when a crosswind caught under my wing at the start of the takeoff run, I came to the conclusion that this glider cannot handle crosswind, and spoilers better be retracted during the last portion of the approach.

Whilst initially thinking of tackling other projects first, a more detailed inspection of the Ka8 revealed no other damage at all than that single hole in the fuselage and a wheel that seemed a bit loose. I couldn’t resist peeling some more Oratex cover away to have a better idea of what wood I would need to undertake the repair later, but what I discovered was a straightforward and relatively easy repair job.

The forward former was broken at the bottom, and the blue nose of the ASW28 had crashed against the wheel box assembly. The latter could be repaired after fashioning a 3mm plywood plate with the exact shape and shallow angles as to be glued over the missing part of the box. The former was 5mm plywood because it supported the forward attachment points of the wing.

The next former was even more damaged, and necessitated a custom made replacement for half of it. Strength was essential because the aft wing attachment bolts also were nearby and the wheelbox also passed a lot of force onto that part of the fuselage. After carefully cutting off some more fabric, I was able to glue piece by piece back into position with the help of clamps and other devices.

Having no 5mm plywood in stock, I glued 3mm and 2mm parts together before cutting them in the desired shapes.

Not thrusting pure head-on glue joints, I later added plywood reinforcements along the sides as well to cope with the forces of the wings and wheelbox in that general area. This means that at certain places the formers are now a whopping 2+2+3+2=9mm thick. After the addition of the side stringers came the difficult task of closing the gap. As you might remember, during the restoration the complete sides were applied in one go, touching (and thus glued) only to the side stringers and not on the formers. Applying a patch of new white Oratex was viable because it mostly would be hidden under the wing when on the ground. Gluing it over the surrounding fabric remains was no option because it wouldn’t hold well on the paint, and the latter would prevent even schrinking of the new and old material, last but not least the access through the opening for the wing wasn’t large enough to allow me to push against the back of the fabric. The only way around the problem was that whilst still open, I could glue 1mm thick balsa strips to the back of the existing still solid lateral sides of the gap.

Cutting the new insert as to form a shoulder to shoulder joint when heat glued to the balsa supports, and using small overlaps on top and bottom, I was able to create a frame for the insert, shrinking it independently from the old cover to make it tight, and using lightweight putty to fill the minimal gap. Next I applied the yellow Oracover with minimal overlap under and over the old parts as to not separate by the airflow, then painted the white with thick paint to blend into the remaining gaps. Invisible repairs of this type are an utopia, but I am satisfied with this quick repair but hate my model to slowly resemble the patchwork it had when I bought it, I now have a better understanding why. On the other hand, if a very qualified surgeon performs a surgery deep inside you, scars also will remain visible on the surface for the rest of your life...

With those lessons learned I will not treat that Flair Ka8 as a large glider trainer anymore, but only use it with moderate winds blowing straight in the runway. Traditional built old gliders are less forgiving than modern GRP reinforced ones, and have more stringent limitations that have to be taken into account, as was proven the day of that mishap when I saw a beautiful Habicht fold a wing later in the day. There is a learning curve in operating large gliders and I am slowly climbing it, and falling off the curve in the process is not pleasant but unavoidable. I hope OO-ZAU will now remain spared from further serious mishaps because I like that glider but get tired repairing it so often without a bunch of satisfying flights in between.
Last edited by BAF23; Jul 30, 2015 at 05:59 PM.
Jun 01, 2019, 10:19 AM
Everything's A Composite
Knoll53's Avatar
Just found and really enjoyed this thread. Thanks for posting in such detail. It's all about the detail!

I have the Flair kit to build AND the time to do it, so am currently researching what has been done. I am thinking of scraping the kit's fuse and scratch building a scale trussed know, for fun.


Quick Reply
Thread Tools