Experiments with Corrugated Skins
In my early teenage years I was attracted by a drawing in Book 1 of Aircraft of the Fighting Powers, the Polish PZL P – 24. It was a fighter with the attractive Pulawski ‘gull’ wing. A prototype with an open cockpit was exhibited in plain dural finish and red trim at the Paris International Air Show in 1934 (Figure 1).
At this time it was ‘the best armed and fastest interceptor in the world’. The Aeromodeller magazine some years later published plans for a small free flight ‘semi-scale’ model with a 1.5 cc diesel motor. After an interval I built this, following the plans exactly. The wing struts were attached using carpet snaps, a kind of large press stud. Even in my inexperience I thought this odd and my doubts were confirmed when the wings came off in the air during the first flight.
The Polish airforce was never equipped with squadrons of the P – 24 but continued to rely on its predecessor, the P – 11, which was no match for the Messerschmitt 109 in the early days of World War 2. Almost all of the P – 24G version, with more powerful engine, enclosed cockpit and heavier armament, had been unwisely exported to Greece and other East European countries. A three-view drawing of the 24G was included in Jerzy Cynk’s fine book, Polish Aircraft 1893 – 1939. (Figure 2).
After more than sixty years I am thinking again of building the Paris Exhibition showpiece P – 24 but this time with functional struts, radio control and scale of 1/4.
All the Pulawski wing fighters had corrugated metal skins on wings and tail surfaces.
Corrugated skins were much favoured in early years by the Junkers aircraft Company. Modellers sometimes imitate these, on such types as the Ju 52-3M, with corrugated cardboard. This is a crude material. It adds little strength, has to be painted and does not stand up very well to general handling. (I am told by Tony Dowdeswell that John Menhennet built a Junkers 52 model with boiled spaghetti for the corrugations. I hope the mice didn’t get at it!)
I want my P – 24 to be like the Paris prototype, in unpainted metal.
The obvious material for the skins is printers’ litho plate. It is easy to obtain this cheaply from lithographic printers. After use the metal plates are usually sold for scrap and a few sheets may be obtained at very low cost, on request. The metal comes in different thicknesses, the thinnest being the most useful for flying model aircraft. Although litho plate is not usually thought of as a structural material, it adds greatly to the stiffness of a foam wing (with internal spars) when glued on under vacuum pressure, and is not very vulnerable to occasional rough treatment.
The problem with the P – 24, indeed with all the Pulawski series of fighters, is that the corrugations were very fine. Figure 3 is a photograph of a tail unit copied from a Polish book (number 15 in the Pod Lupa series from Ace Publications of Wroclaw, 2002). The reproduction here is not perfect but enough can be seen to illustrate the difficulties. This photo is also interesting because it shows that the full-sized corrugated skin was a little wobbly here and there with small gaps and other minor imperfections.
I examined all the references I could find. There are differences between the various publications. Often the corrugations are described as ‘fine’ without exact clarification of the meaning. One of the best sources is a readily available drawing of a P – 11C by the Polish draughtsmam, Z A Datkiewicz. (The wings of all the Pulawski fighters were similar but not identical.) Datkiewicz states, in imperial measures, that the corrugations were 8 to the inch but the lines representing them on his drawing are not consistent with this.
It is not clear if 8 per inch refers to distance between two ridges of the corrugations, ie., the distance between the crests of the ridges, or half this, from ridge to trough. In the latter case the corrugations would have been close to ¼ inches wide. On the model they should be 1/16 in ≈1.6 mm.
These points were checked by study of photographs of the only surviving P – 11, which is in the Cracow Museum, and from books and the internet. Very few show enough details to be sure. The internal wing ribs were not all the same distance apart, especially in that part of the wing which was bent to follow the complicated ‘gull’ dihedral. The best estimate I can make is that on the outer wing panels there were 44 corrugations for wing rib bay widths of 264 mm., ie , 6 mm ridge to ridge. A quarter of this, 1.5 mm, is thus a likely figure for the model. Another complication is that where the metal had to be riveted down to underlying ribs, about 20 or 25 mm of the skin was flat. Presumably these were lapped joints in the skins. (Some pictures show, on the underside of some of the P – 11 wings, that the skins had flanges standing up from the surface like crude boundary layer fences. I think this did not apply to the P – 24. The P – 11 tail unit in Figure 3 shows this on the rudder, but not the fin.)
If more certain information is available I would like to hear of it.
How can a modeller reproduce skins like this? Perhaps the question should be: is it even possible for an amateur to produce them? When I talked to other people some said I should just build the fighter with plain, uncorrugated skin. This was unacceptable. Anyone knowing the Pulawski series would be aware of the error.
Architectural modellers and model railways enthusiasts can buy small sheets of plastic marketed by various companies such as ANRI and Metcalfe, with corrugations moulded on one surface. Some of the available range of products would come close to the suggested dimensions but I do not think they would give me what I require. The plastic sheets are small and to skin a whole wing with them would require many separate pieces, all which would have to be butted together without overlapping. The result would look like a sort of jigsaw puzzle and every joint would be a line of weakness. A metallic appearance would also be hard to achieve.
A solution could be to employ a professional engineer to make a rolling mill, a quarter-sized version of the ones that must have been used to make the original PZL skins. With such a mill it would be possible churn out sheet after sheet of finely corrugated litho plate. I made enquiries, learning immediately that the cost would be quite prohibitive. Indeed, I could not find an engineer willing to do the job for the sake of one small production run. I was proposing to build only one model, not marketing hundreds of kits. To develop a mill would not be worthwhile for any professional.
I decided to try to make my own rolling mill. I did not expect much success but had an interesting time with this idea and the results were better than I expected although, in the end, it did not give me exactly what I wanted.
Hardware stores stock lengths of threaded rod in various sizes, with nuts to fit. I thought it might be possible to use two such rods one above the other, turning in opposite senses to pass sheets of plate between them. I recognised that the threads were not the ideal shape but the idea seemed worth trying. If my mill worked I might be able to refine it later but I needed first to test the concept. Two lengths of the threaded rod, each with a handle welded onto one end, were mounted as shown in plain wooden bearings, well oiled (Figure 4). The rollers had to be close enough together so that the threads meshed.
To make the bearings was done by first drilling holes through the hardwood end blocks clamped together, so that the rod would just go through without slop. The blocks were then separated and cut away just far enough to expose the thread, allowing the rods to engage with one another (Figures 7 and 8 shows this more clearly). The spacing was critical and several attempts were necessary to get it right. If the rods were too close to one another the metal sheet would not enter the mill at all. Too far apart, the plate would pass through almost unmarked. After a few trials and adjustments, small strips of plate were fed in between the rods, the handles turned in opposite directions, and the metal passed through, emerging with corrugations as Figures 5 and 6 show. The corrugations were about 2 mm apart, ridge to ridge. This was too coarse but the result was quite encouraging and I persisted with further experiments.
The strips I had made were not wide enough to cover a single rib bay on the model wing, and were far too short to go right across the chord except at the wing’s narrowest part near the tips. To try to make longer and wider strips I made a wider mill (Figures 7, 8 and 9).
A further improvement, I supposed, would come if I made the mill adjustable, allowing me to tighten up or slacken slightly the distance between the rods. With vertical bolts and wing nuts I was able to place washers of varying thickness between the upper and lower bearing blocks to change the pressure and depth of the corrugations.
I learned a lot! If the two wing nuts were screwed down a little too much, what came out was aluminium spaghetti (Figure 10. Pace John Menhennet). The threads on the rods simply tore the plate into narrow strips. If not tight enough, the sheet would pass through without impression.
Another difficulty was that if the sheets were cut exactly wide enough to pass between the blocks on each side it seemed impossible for them to go through straight. The sheets would slew slightly out of line and shifting slightly from side to crumpling the edges or worse (Figure 11).
These faults are more evident in the enlarged picture (Figure 12). With a little more persistence, always leaving a margin on each side for the slewing, some fairly good results were achieved but corrugations usually emerged with slight sideways wobbles.
A further difficulty arose when I tried to make strips of rolled metal long enough to go right across the wing chord at the broadest point. During the rolling process the threaded rods inevitably screwed themselves inwards and after a certain number of turns, could go no further (Figure 8.) The obvious answer, to use longer rods which would allow more turns before the handles ran into the wooden blocks, allowed longer strips to be rolled but the wobbling and slewing tended to be worse. Sometimes everything would go well and a good result, almost usable, emerged. More often, the outcome was a failure.
I had hoped to be able to make wider pieces of skin, to minimise the number of pieces in the jigsaw puzzle. To attempt this I made a much wider mill with the long rods. This did not work well. When the butterfly nuts were adjusted to produce satisfactory results at the sides, the middle of the sheet was not touched. The rods themselves were distorting by a small amount, enough to miss the centre of the plate as it went through. Tightening the butterfly nuts produced spaghetti at the sides and still the middle was barely touched.
I tried using some softer aluminium to reduce the distorting force on the rollers. A simple cooking tray of thin aluminium was cut up and a piece fed through the mill (Figure 13). The manufacturer’s trademark survived, showing how little pressure was reaching the middle of the plate while the edges were tearing and crumpling.
Figure 13. The base of an aluminium cooking tray, after rolling. The manufacturer’s mark was not removed!
Figure 14. Mill with restrainers and long rods
For the next trial I added a series of adjustable plastic restrainers (Figure 14. The restrainers were adjustable ‘feet’ intended for the legs of furniture). I hoped that the slight bowing of the rollers under load would be prevented by screwing the restrainers down slightly in the middle. This had some effect but the central part of the sheet still was not adequately corrugated when the edges were beginning to tear. The effort required to turn the handles was much greater and maintaining a steady rate of progress quite difficult. Even with relatively small pieces of plate I was still a long way from producing long and wide corrugated sheets for wing skins (Figures 15 and 16).
Figure 15. Narrow sheet, Edges corrugated, middle untouched
Figure 16. Sheet still barely marked in centre. Note also how the sheet rolled up as it went through.
At this stage I decided to abandon the threaded rod mill. I do not regard the time and effort wasted. I had enough success to suggest that if a modeller needs to manufacture a small and fairly narrow section of corrugated metal skin, perhaps one rib bay wide, or for an engine cowling, inspection panel or hatch on an aircraft, it can be done this way. The thread of the rods used should be carefully chosen and after a few trials satisfactory results can be obtained for a small section of skin.
To skin the whole wing of a fairly large model requires something different and more predictable. I decided to try another very simple experiment. I raided my box of spare bits and pieces and extracted all the lengths of heavy gauge hard wire I had. After a brief cleaning these were glued side by side with epoxy resin to a flat piece of chipboard. I laid a piece of litho sheet on this ridged base and, holding it firmly down, pressed the rounded end of a spare piece of the wire firmly into the metal between two of the underlying rods and still pressing, dragged it firmly along, forcing the plate into the furrow as it went, a sort of ploughing action but not so deep that the plate could be torn.
After the first pressing and dragging, the plate was moved over using the recently formed corrugation to align it properly. This was repeated. As more corrugations were formed, the plate curled up (Figure 17.)
With the help of a straight edge and gloves as protection (Figure 18), in a few minutes I had produced the sheet shown in Figure 19, pinned down.
The result was surprisingly good. To check one further point, I used epoxy resin to glue this piece of as it would be on a wing, to a suitably shaped piece of white foam (Figure 20). Since this was merely for demonstration, I did not use vacuum bagging or take a lot of trouble over the exercise, but merely weighted the sample down while the epoxy hardened.
There was no difficulty of adhesion. The plate, despite its tendency to curl the wrong way, bedded down to the foam and showed no sign of lifting or peeling off. To complete this trial I drilled tiny holes and pushed pins in to represent rivets. Each pin had a dab of glue to keep it in place. For an actual model more care would be required for a good finished appearance but I regard the corrugations themselves as quite successful.
Although the corrugations were too gross for the P- 24 model, the simple method used here could be used for a Junkers aircraft model or any other with this type of skin. The spacing of the ridges and furrows can be determined by choosing the correct wire gauge for the baseboard.
A further trial was made using lengths of much thinner wire, to make PZL fine corrugations. As before I was trying at first only to test the procedure. I made a base board consisting of more than 200 lengths of 1.2 mm gauge piano wire, each 450 mm long glued side by side to a flat wooden baseboard, edged with two thin straight pieces of steel (old steel rulers) on each side. These edging pieces, just visible in the photograph, were intended only to prevent any of the wires shifting outwards under pressure. (Later I found they were not necessary.) Two lengths of 25 mm square tubing with nuts and bolts, as shown in (Figure 21), were to clamp the litho plate down.
Figure 21. The 1.2 mm wire baseboard. The appearance of humps and waves is an illusion caused by reflections on the wires, which are actually quite flat to the board and tight together.
(The first difficulty was to obtain sufficient numbers of the steel wires. I raided three model shops and bought all their stock of 1.2 mm gauge, arousing a good deal of puzzlement among shop assistants andproprietors.)
Figure 22 Close up of the baseboard
Figure 22 shows a close up of the wires as they were placed tightly together and glued down flat. I used thin cyanoacrylate adhesive which spread out under the wires and hardened quickly. Afterwards, the assembly needed cleaning and polishing to make sure there were no unwanted pieces of grit or glue that would interfere with the next operation. The size of the base was determined by the standard plates provided by the local lithographic printer. I was fortunate in that these are long enough to cover the whole chord of the wing and wide enough to skin several rib bays.
Figure 23. The baseboard with a litho plate clamped.
(A note about lithography. The aluminium plate is coated on one surface with a thin layer of light-sensitive emulsion. The items to be printed are photographed and transferred to the plate for development. The plate then goes the printing machine. Ink is taken up by the image but repelled by the rest of the plate, which transfers it to the printing roller and thence to the paper or card.
One side of the plate is bare aluminium and may need cleaning and polishing, since this is the side which will be visible on the finished model. It is best to do the polishing before making the corrugations. The emulsion does not yield to common solvent fluids, such as methylated spirits and paint thinners, but if required the images can be removed with meths, leaving the emulsion untouched. It is sometimes possible to obtain lithoplate before the emulsion is added.)
Figure 24. First signs encouraging
To make the furrows, the tool used was a simple scriber obtained from a hardware store. Grasping the tool firmly and holding it at a shallow angle, it is pressed down into the plate at one end between two of the base wires, and drawn with steady pressure along the groove. The intention is to push the plate down between two of the wire guides, all the way along the full length.
The angle at which to hold the tool can be established by trial. The extreme tip of the scriber must not touch the plate, because this would split the thin sheet. Working on the emulsion side seems to help, perhaps because the emulsion acts as a lubricant. It also helps to add a few drops of machine oil.
Move on to the next furrow. Progress can be rapid. Not a lot of force is needed, but it is very easy to let the scriber jump over into the next furrow, rather like a railway train becoming derailed. This is not necessarily disastrous. By returning to the track and trying again with more concentration, the slip can be corrected. Stop, rest, and start again with care. (It is a good idea, I found, to protect the index finger with a stick on plaster or a finger stall!) The next furrow then can be done, and the mark left by a derailment usually disappears. If the error is more serious, turn the baseboard round and work from the other end. Attention to the job is essential.
The corrugations were at the lower end of the scale for the P – 24, ≈1.2 mm apart, the gauge of the baseboard wires.
Figure 25. First attempt litho plate fully clamped, work started on one side. The scriber used for the process is visible, top right.
First results looked good but problems emerged before the first sheet of plate was finished. The metal began to lift and develop waves, some of which became locally severe enough to crack it as the work progressed. The first few centimetres of the sheet were finely corrugated and were useable, but the ‘ploughing’ became more difficult as the process continued and became impossible.
The reason is not hard to understand. Forcing a metal plate to corrugate causes it to curl. With rolling the expansion is on the crosswise axis (Figures 5 & 6). When the distortion is along the length of the plate, the metal is stretched laterally as each furrow is made (Figure 15 & 16). Being clamped down, as it pushes out sideways it has nowhere to go but up. To force it down as more and more furrows are made, makes matters worse and the plate begins to crack, which shows in several places in Figure 25
My first naïve reaction was to change the order of work, starting the first furrows in the middle of the plate and working outwards from there. This halves the problem and it was possible to complete a whole sheet. There is a little more room for the plate to expand and cracking was avoided. When all pressure was off it lifted as shown in Figure 26.
One further change was necessary to produce consistently acceptable results. It was a mistake to clamp both edges of the litho plate down hard. Instead, one edge was clamped firmly, the other was restrained very gently to prevent curling until the job was done, the butterfly nuts slackened right off on that side allowing the plate to creep outwards under the aluminium bar as the work progressed. The furrowing process began at the clamped side and as the scriber was drawn along each furrow, firm pressure was maintained with the other hand and a cloth, so at no time was the plate allowed to lift as new furrows were formed. In this way, good and consistent corrugations were made across the entire sheet, with a small but acceptable flat margin on both edges. The total expansion totalled about 13 mm. On release, the sheet curled up, but was fully acceptable (Figure 27).
Figure 27 On releasing the plate completely it curls up but can easily be unrolled.
At this stage I consider the problem solved. Neither corrugated cardboard nor boiled spaghetti will needed in future.
Using 1.2 mm gauge wires for the base, the final result yielded 50 corrugations for a rib bay width of 62.5 mm. I was aiming for 44 corrugations, not 50 (See paragraph 9 above).
I doubt if anyone, except myself, will go over the model with a micrometer but I intend to go up one wire gauge. This will require a new baseboard with heavier wires. 1.4 mm is a stock size. This should get me very close to the desired 44 in 62.5 mm (62.5/44 = 1.42 ).
Another raid on the model shops is necessary; 85 of the stock length (914 mm ≈36 ins) cut in half. Another point is that I must leave a flat across the litho sheet at each underlying rib, for the rivets.
I seem to be building the P – 24 from the skin inwards and there is a long way to go! Don’t hold your breath.