BALSA DUST VERSUS THE DIGITAL DOMAIN
A Review of Compufoil
by Chris Williams
Let’s assume you were able to hire a Time Machine…If you were to scoot back to the time of Neanderthal Man, grab a likely looking subject, haul him back to the present and plonk him in the cockpit of a mission-capable Lockheed Blackbird, it’s highly likely he’ll be well impressed. Eventually, that is, once he’s come to terms with Fear and Confusion, and cleared up the mess brought on by Terror.
Back home, around the campfire (did the Neanderthals have fire?), he’d wave his arms expansively and say, (this is a rough translation by the way)
“Wow, you won’t believe wot they’ve done in the future, they’ve only gone and killed off all those damned tigers.”
I often think that I know what this feels like when I try yet another programme on my PC to find that what was once the work of a team of skilled chaps with slide rules can now be achieved by yours truly with a mere mouse click.
Take for instance the design of a glider wing. I have always made up the ribs for my wings by the good old ‘Sandwich Method’, whereby a load of balsa blanks are sandwiched between two metal templates, and the whole assembly sanded smooth to the accompaniment of a rising mushroom cloud of balsa dust.
It works pretty well too, unless your wing has a sharp taper, in which case things do tend to get a bit divergent and much scratching of the bonce is required to keep things on track.
If you are designing the thing yourself you may not get too many problems if the wing has a simple straight taper, but consider, if you will, the design of a fairly complicated wing, such as is to be found on the Minimoa. Starting out at the root you will be forgiven for thinking the job a doddle, after all, the inboard panel is of constant chord, no problem there then. Now to the gull-join and whoops the chord length suddenly increases, rather like the Bird of Time, but, and this is the mind-bender, the thickness has to stay the same.
This means a major modification to the chosen section, which can then transition back to its original form way out there at the tip. The Sandwich Method can cope with this, just, but what did they do in the 30’s.
Just how did a planform drawing of the T21 wing, for instance, with its semi-elliptical shape get to become a set of wing rib drawings accurate enough to fit the chosen planform out there in the workshop, where the workmen the were rough and tough, and their tongues liable to wax caustic.
All day, all bleedin’ day it took me to fabricate this rib at station twenty-six, and I’m tellin’ you mate, it don’t bleedin’ fit!) If you take the shape of your airfoil section at the root and reproduce that shape at fixed points down a tapering wing mathematically, this is something, and I hope I’ve got this right, this is something called ‘Lofting’. How this was achieved I have absolutely no idea, but I’m fairly sure it involved pointy hats, runic symbols and an industrial strength magic wand.
Right, back to the present day and the all-powerful PC. After a bit of needling from one of my column readers, I went to the Compufoil web site and downloaded a demo of this neat little programme.
Ah, I’m getting ahead of myself here I can see, what, you may ask, is a Compufoil? Back in 1992, Eric Sanders, a software engineer for General Motors, (who lives in Ohio and flies sailplanes), released his new baby as a DOS based programme, specifically designed to ease the design of model aeroplane wings. In ’95, he released the Windows version of Compufoil and reckons he has worked on the programme for an average of three hours a day since ’92! A copy of it now nestles snugly on my hard disc, and every now and then, as time permits, I poke it with my mouse to see what happens.
What happens is this’wow’, I say, as I sit with the others around the campfire, ‘you won’t believe wot they’re doing in the future..’ To put it in simple terms, Compufoil does far more than your average bloke can take in all at once, and to start with, as with any new programme that you have to come to grips with, the mind is simply overwhelmed with the available information. Gradually, it starts to make sense, and the excitement mounts, well, it will if you’re into designing model aircraft. Then you get to stage two, where you try to force the programme to fit its results into a fixed set of parameters, as with a scale model, and you realise that things are going to get a bit more complicated. First, though, let me give you some sort of idea as to what Compufoil can deliver.
1. A rib profile of specified length of which you can define the shape from a library of airfoils. (Quabeck, Eppler, NACA etc)
2. A set of controls that allow you to define the thickness of the sheeting material; the location and depth of any number of spar slots, LE stock thickness, TE stock thickness etc.
3. Compute a set of lofted wing rib templates for a variety of planform shapes. You can define the planform in a variety of ways, elliptical, semi elliptical, straight taper etc. You can join any number of wing panels together, although for a shape as complicated as the Minimoa, it’s better to treat the wing as two separate panels. You can define the number of ribs, their spacing, and even plot odd ribs that are placed at non-equal distances anywhere along the wing. (That would have been most helpful with the Petrel)
4. You can generate a set of jig holes, if you prefer to build this way, and even dial in the amount of washout that you might need.
5. You can plot reverse templates anywhere along the wing, which will allow you to check the accuracy of the finished structure.
6. You can play with the airfoil to your heart’s content, changing thickness, camber, etc. which is ideal for that little problem with the Minimoa wing.
7. Print it all out…? My first question to Eric was how? All most of us have is an A4 printer, how can you print out on that? What happens is this, your drawings are tiled over however many pieces of A4 that it takes, and there are alignment marks on each sheet to allow you to accurately join them together. Neat, huh?
So far I have only described the stuff that applies to us balsa bashers, there’s loads more for those that like to build their wings from foam. The templates are generated in a similar way, but with lead in and lead out shapes applied, and all sorts of esoteric stuff that allows for such things as the thickness of the cut and such like; information at which I haven’t yet had time to look. Compufoil comes with an extensive library of airfoils as I have already mentioned, and also features a copious on-line help file and tutorial. (On-line means you have to scroll it up on the screen or print it out yourself, well, what do you want for the money, a three-colour printed manual?) Also included is a handy little calculator which will convert metric to imperial and vice-versa and as well as the usual functions allows you to convert areas as well.
It was during this initial familiarisation period that the inspiration for my next project arrived and it was time to see what Compufoil could do for real. The wing in question is the one that supports the diminutive but, by all accounts, very lively Vogt LO100, and I would invite you now to take a shufti at it.
As you can see, the planform starts out parallel and then converts to combination of a swept straight leading edge and an elliptical TE, just the combination that gave me such a headache when designing my version of the T21.
So, to establish the design parameters…the main reason for building this sailplane is to indulge in aerobatics (the coming thing, they say, on the scale scene) and so for once my favourite Quabeck section is going to get the order of the elbow. A few years ago I was flying a dog-eared fifth-hand F3b Calypso, (sentimental fool that I am, I named it the Collapso) and I was mightily impressed with its smooth performance in aerobatic mode, so this looked like a good place to start.
Selecting the airfoil library in Compufoil and looking for RG14a was the work of a few seconds, and there she was on the screen, not just the shape but lots of useful info besides, such as camber and thickness percentages. Hmmmm8.47% thickness, now that’s really isn’t going to look very scale is it. So, let’s try 11% as a reasonable compromise, click, click job done. While we’re at it, we’ll set the sheeting thickness at 1.5mm and the LE thickness at 13.5mm (12mm for the balsa LE and 1.5mm for the false LE). You may think that it’s a step into the unknown modifying a section this way, but I did do this before on my last project, the Petrel, where I thickened up the root to achieve a more scale appearance, and I was still more than pleased with the model’s performance.
When we get to the point where the wing planform changes, so does the thickness of the airfoil as the spars taper out towards the tips, and it is from here that we can revert to RG 14’s proper thickness, transitioning from 11% down to 8.47% out at the tip, hopefully the best compromise between appearance and performance.
At this point it is important to realise this fact: a simple, straight taper wing planform can easily be computed, with a planform print-out and all the spars designed in. When it comes to an elliptical planform, however, things get a little complicated. If you are designing a sport model you can fiddle around to your heart’s content until you find a shape that pleases, but getting the shape to match that of a pre-designed wing is much more problematical. In the case of the LO100 it’s easier by far to consider the semi-elliptical part of the wing as a separate panel and the act accordingly.
Step1. Save the modified airfoil as a new file.
Step 2. Bring up the ‘Generate lofted full set of ribs‘ dialog box and fill in the required parameters, span, rib spacing etc. Set the modified section in the root setting box, and RG14a in the tip. Now for the really tricky bit. (I must admit, it was handy having Eric’s advice surging down the Digital Highway) To get the correct shape he advised thus ‘You need to use the “Modified Ellipse” shape. You would use a Leading Edge Chord of zero and a Trailing Edge Chord equal to twice the desired chord length. This will give you a straight leading edge and a true elliptical trailing edge. You will need to custom -create the ribs for the rounded tip…’
You know, the lad was quite right, proceeding thus gave a set of ribs that almost exactly matched that of my scaled up drawing, and the differences can be surely be put down to the small errors that are unavoidable when scaling up from a 3-view who’s antecedents could well be dodgy anyway!
Marvellous! Now we were really cooking on gas and it was with keen anticipation that I set to the task of creating the spar slot cut-outs, which, as I’m sure you’ll remember need to reduce in depth out towards the tip. Here things went somewhat divergent, as no matter how I tried, I couldn’t get the slots to match the placing of the spars on the plan. Eventually I discovered the problem; as Eric had warned, the ribs out at the tip, the last couple or so, were under-size, as the full-size planform deviates from a true ellipse at the tip, and this was throwing all the calculations out a more than a tad. My solution was this; lay out the plan, print a few representative ribs from both ends and the middle of the wing panel, cut them out and see if everything lines up. If not, fettle the parameters again and see what happens.
There are lots of reference lines in the rib drawings, and these simplify the task of working out just where he slots need to go. There are half a dozen different ways of calculating this task built into the programme, you can set them at the point of max thickness, or at a percentage point along the rib, to name but two, but for this wing panel shape you have to use the ‘Custom Panel Location’ box, which allows you to set the distance of the spars from the reference point of the LE.
Now maybe I’m a tad cynical, but I would advise anybody using this or any other programme to print out a few ribs first and try them out on your plan; not because I don’t trust the programme but out of a healthy respect for Murphy’s Law if it can go wrong, it jolly well will. This I found to be very true when I set Compufoil to work in the metric measurement scale, becoming very confused with centimetres and millimetres and initially starting out with wings with a thousand ribs in them and such like! You’d look a damn fool, wouldn’t you, cutting out a thousand ribs and then climbing out of the wood shavings trying to match them to your plan.
On large-scale models the wing sheeting tends to swallow your thumb because of the large unsupported areas, it is usually necessary to make up the riblets to go in between the ribs, just like the full-size. In practice this means that once you are satisfied that you have things right, you simply halve the distance between the ribs, thus doubling the amount of them. (You mustn’t forget to cut of the tails of the riblets, and not get confused between the two)
OK. Time now to fire up the printer. Nestling in a dusty cupboard I remembered a box of A4 sized sticky label sheets; now what if I were to print on to those, all I would have to do is cut out the ribs, peel off the backing and stick them to the balsa, right? For once, theory and practice matched perfectly, and at two or three quid (£) per box of 100 labels it’s not even going to break the bank! The ribs for the LO100 took some fifteen sheets or so, with about forty minutes of time to stick the pairs of labels together and cut out all the ribs, paper ribs that is, the balsa ones took longer.
What, I hear the ones of you still awake ask, about the runt-sized ribs attip?
Well, for me it’s back to the scanner/photocopier; simply enlarge them until they fit lengthways, without increasing the thickness. Adjust the slight differences in the spar slots by offering the rib up to the plan and marking out accordingly, well it can’t all be Star Wars you know.
What then can Compufoil offer you? If you buy your models ready-made, or second hand, I guess you won’t be that interested, although you could look up some of the more popular airfoil’s statistics and amaze your flying friends with the depth of your knowledge.
‘…Selig 3021 eh, hmmmm, well it shouldn’t fly too bad with a camber of 2.96 degrees provided you’re flying the wing at 1.08 degrees at which, as I’m sure you know already, is the ideal angle of attack for that particular section’
Later, in the pub, fellow club members will nudge each other and point at you and say; ‘Y’see that chap over there? Stay away from him, he’s a smartarse…’ If, on the other hand, you’re in to making up (or getting others to make up) your own foam wings, you will no doubt find Compufoil a godsend, offering as it does, such a wide range of parameters for the production of your templates. (Including of course the ability to plot the templates at various stations along the wing if you are making the foam cores in sections) If you are any sort of designer, not just of scale models, but of any model, then once you have sampled the forbidden delights of this programme, you will wonder how on earth anyone manage to design anything before the advent of the computer.
Speaking for myself, I have only one misgiving; scale models take a long time to build, I’m pushed to manage two in the course of a year. This means that only once every six months will I need to fire up Compufoil and persuade it to do the business… will I ever remember how to work it?
Back to the campfire now… ‘The truly amazing thing is that you don’t even have to step outside the front door to obtain a copy, you simply download it and pay with your credit card with your details simply encrypted over two E-mails. Even a blockhead like me can do that, here, shove another log on the fire, will you, don’t you know there’s still some tigers out there?’
If you would like a trial copy of Compufoil goto http://www.compufoil.com/mainwindow.html
COMPUFOIL 2000 UPGRADE
In the beginning there was dust…
For me, History is divided into two parts: pre-Compufoil, a dark and savage age (and dusty too, if you were using the sandwich method), and post-Compufoil, an age of enlightenment, a true renaissance period wherein model aircraft designers were truly set free. For those who have not yet seen the light, let me recap…Compufoil is a computer programme that allows you to set the parameters of your wing design, whether it be built-up or foam, and the resultant wing ribs/templates come merrily spewing out of your printer ready for instant use. (For a fuller explanation, read my review in above) Not content to sit back on his laurels, the programme designer, Eric Sanders, has been busy modifying and tweaking the thing and has produced the next upgrade, Compufoil 2000.
So, what’s new? Hah! Wouldn’t you like to know, but for those of you not yet used to them, there’s this thing about computer programs…they can do an awful lot of stuff. How can I explain this? Let’s imagine, for instance that a bunch of you were to load me up with the design details of the wings to be attached to their new projects. Let’s take fifty sets of wings as a random figure, and three months to design them on Compufoil. At the end of this marathon I reckon I would have a pretty fair idea how to usea out fifty percent of the utilities in the program,but then maybe I’m not as smart as I think. Normally I would design two or maybe three sets of wings in a good year, so by the application of simple arithmetic, you can see that in the normal course of things it would take a goodly while for a chap just to get to that fifty percent figure. This is completely normal. The word processing program by which means this is being written is also full of wonderful utilities, but I only know and use about five percent of ’em, and that’s on a good day.
Let’s look again at the program’s core functions…First you can call up a profile from the program’s extensive library. (Or enter your own co-ordinates). You can modify the section in any way that you like, of course, thicken it, thin it, change the camber etc. Now you set the section’s main parameters; chord length, sheeting thickness, LE thickness etc. Then you generate a set of ribs for your wing based on this section. You can have the same section throughout, or specify a different section at the tip, either way, the work up to this stage is very easy. Now you need to specify the placement of the spars, and it’s at this point I find I need to shut the door and take the phone off the hook. There are any and various ways to set the spar positions; you can define the distance from the LE, or the TE; you can snap them the to ‘foils thickest point, or specify an angle in degrees. Then you can decide whether you want then to come to the surface, sit below the sheeting, or reference them to the datum line for internal spars. You can set the rib slots to accommodate any sort of spar taper and set the wing planform for any number of wing panels for planforms like the Duo Discus. That’s about as much of the program that I use for my wooden machines, and like my word processing program, it’s just the tip of the iceberg.
So, what is new?
On the lofting side it is now possible to swap .lft files with your club-mates without the recipient needing the co-ordinates, which might not be in their library. (i.e. if a .lft file is given to someone, they no longer need to supply the airfoils used in lofting. (What’s lofting? I’m going out on a limb here, but I think it’s the process of taking two different sets of co-ordinates and calculating all the user-defined stations in between, after all, it’s what computers are good at…calculation, right?) This could make you a popular figure within your club, but it might eat into your building time.
Ribs can now have a different airfoil at each panel juncture position, perfect for those who truly believe they can design the perfect glider wing!
Support has now been added to use negative chord lengths for modified elliptical lofting to create scimitar shaped wings. See what I mean? So much functionality and so little time, if only I could lose the day job… On the spar slot set up side you can now loft the spars through multiple panels (instead of on a per panel basis), for taper, custom placement and depth criteria.
You can now specify a horizontal reference line for those who like clamping the foam template into the holder before cutting the core. For the traditionalists you can now specify for LE building tabs to set the wing level with the board.
And so it goes; space and time do not permit delving any deeper into the subject because of the dangers of Compu-Stasis, the
process by which submerging yourself into a fascinating computer programme causes a withdrawal from reality and a reluctance to get in the car and go to work. Even before the upgrade I had always regarded Compufoil as a genie in a bottle, waiting to spring to life and render the process of wing design and construction so much easier than before, when balsa dust ruled the workshop. Do you need all these goodies, above and beyond the basic stuff? Well, you never know when you might want to design the perfect wing, do you, and you might well want it to have a scimitar planform, too…?