Re: carbon

Robert Cringely

Lighter is always better.

I built two Glasairs -- #102 (the first kit sold) and #123 -- and brought both in under 1000 lbs empty, which is rarely done. It helped that I built 102 right next to the prototype in Washington with a lot of input from the designers (I was their guinea pig). This was after I built an earlier composite aircraft, a Rand KR-1, which had skins made of Dynel, a synthetic fiber known mainly for its use as artificial hair in wigs. Dynel is NOT good. Here are some interesting things I learned from building those aircraft, both of which had substantial wing spars and did not take skin strength into account in ANY stress calculations. That is, the skins were not required to carry anything other than air loads and -- in the case of the Glasair -- to keep the gas from leaking out. If that's the case with the Quickie, then ONE layer of glass on the 45-degree bias should be enough. That's all Kenny Rand used on the KR-1, he didn't even use the fabric on a 45-degree bias, and it was crappy Dynel to boot, which was great at following compound curves but also stretched so much it transferred the loads completely to the spars anyway.

"Yes, but this Q is a Rutan design we are talking about," you say. "It's different from those others."

Not according to Burt Rutan when I asked him the specific question of whether his wing designs required the skins to carry other than air loads to meet strength targets. Rutan told me he designed the spars to take the loads in all his homebuilt designs because it was simpler that way from a computational standpoint. And his spars were designed with ultimate loads double the service loads as opposed to 1.5X in a metal structure. That is to meet a 4.4-G Utility category or a 6-G Aerobatic category where an aluminum structure would have been designed to fail above 6.6-Gs or 9-G's. respectively, Rutan designed JUST THE SPARS to carry 8.8 or 12-Gs.

An important thing to remember about Burt Rutan is that he worked at Edwards AFB as a FLIGHT TEST engineer. His design process was always minimal while his testing was rigorous, because testing is what Burt likes best. And there is nothing wrong with that. Look at the original VariEze prototype, which had a VW engine and no ailerons. Burt thought he could get away with the VW because HE REALLY HAD NO IDEA WHAT HIS PLANE WOULD WEIGH. It was under-powered so he threw another 70 lbs of engine on board without changing ANY of the structure except to put some lead in the nose, further increasing the weight. Ailerons, too, were added because of testing nightmares (canard elevons turned out to be a bad idea). That same airframe today carries up to an O-320 and is STILL plenty strong.

These planes don't BREAK, we BUST them.

These aircraft use standardized cloth that doesn't so much take into account the needs of the design as it does the inventory of the supplier. So on wings of this type (where the spars carry the majority of loads) a single layer of cloth on the 45-degree bias offers enough strength to carry any torsion loads and keep the insides of the wing on the inside. But because we live in a real world, most designers make the skins two layers thick simply so they will hold up better to manhandling. Others may opt for three layers of UNI (0-degrees, 45-right, 45-left) because that's the textbook way to get the highest performance wing -- performance that wing IS NEVER ASKED TO GIVE because the spars are carrying the loads.

Here's a dirty little secret about composite aircraft design: most prototypes aren't "designed" at all -- they are just built. Two layers on the outside and one on the inside if it is a sandwich structure. And most are left just that way. Or the prototype is too heavy so they redesign it a little lighter, which usually means making it lighter then seeing if it breaks. Because of the variability in materials and production techniques there is a lot more make-and-test here than there is Finite Element Analysis. Maybe that will eventually change but not until computers get better or composite materials get -- like aluminum -- a lot more standardized.

Martin Hollmann did the structural design on all the early Lancairs -- AFTER they flew. There was a guy who wanted to fly his Lancair 320 around the world so he asked Martin to take as much weight out of the airplane as he could while, at the same time, increasing the gross weight and adding extra fuel tanks. This may have been the first time a composite aircraft was scrupulously (re)designed for minimum empty structural weight. Hollmann dropped the strength target to the Normal category (3.8-Gs), added exotic materials where he could, accepted a 1.5X safety factory (instead of 2X) and managed to take 200 lbs out of a 320, dropping the weight to around 850 lbs, of which 400, remember, was engine, accessories, and prop. Those wings had single-ply outer skins.

There's another very interesting aspect to this. We go to a lot of trouble and waste a lot of materials putting two or more layers of 45-degree cloth on wings and fuselages because that's what we're told we need to carry the torsional loads. Just roll the fabric out on the wing (0-90) and the torsional strength is reduced by almost half. But wait -- aren't those two layers twice as strong as they really need to be? Sure. So in practical service you can meet the same strength with one layer at 45 or two layers at 0, with the two layers being sturdier. Well, heck, why would anyone do it any differently than that? Kenny Rand didn't (he was a TERRIBLE engineer, by the way) and hundreds of KR-1s flew without falling out of the skies.

Which brings me back to my point: these airplanes are under-designed and over-built to compensate.

Look at the Qs in particular. The only real structural change was adding carbon spars to the canard and that was in an attempt to keep bad pilots from breaking their planes during Pilot-Induced-Oscillation (PIO) on landing. It's questionable whether that really helped, by the way, but other than carbon spars and improved tail springs (also a PIO issue) there were no significant STRUCTURAL changes.

It's your plane, take some weight out of it if you like, being sure to keep the CG where it belongs. Heck, most builders go the other way and add weight back in for the darnedest things. DON'T DO THAT.

But here's my caution. While composites are easy to make, they don't fatigue in a traditional sense, and they are generally so over-built that you can make lots of changes with relative impunity, THEY WILL FLUTTER if you try to go too fast or too high, which is why I am looking at all-carbon (and 100-percent mass-balanced controls) for my Q1.


-----Original Message-----
From: L.J. French [mailto:LJFrench@...]
Sent: Saturday, December 19, 2009 09:20 PM
To: Q-LIST@...
Subject: RE: [Q-LIST] Re: carbon

I am building a Q and would like to make it as light as possible. Do you
think I could get by with 2 ply as long as I am careful to not hit my hanger
with it?
LJ French

-----Original Message-----
From: Q-LIST@... [mailto:Q-LIST@...] On Behalf Of
Robert X Cringely
Sent: Saturday, December 19, 2009 2:36 PM
To: Q-LIST@...
Subject: Re: [Q-LIST] Re: carbon

What are the constraints on this design? On most composite designs the
deciding factor in how many layers to use on the wing skins, for example, is
hangar rash. Yes, hangar rash. Less glass could always be used but it would
be damaged too easily. So we sit around speculating about the implications
of changing materials on a design that's
overbuilt in glass OR carbon. What is the history of structural
failures on Quickies? Zero. Change it to carbon and it will still be zero.


On Dec 19, 2009, at 1:50 PM, "rick_nordgarden" <grnordgarden@...>

--- In Q-LIST@..., "Robert X. Cringely" <bob@...> wrote:

An ALL-CARBON Q1 should carry its loads exactly like an ALL-GLASS
This is an example of how risky an untutored "common-sense" approach
to engineering can be. Make two identical foam wing cores, then skin
one with fiberglass and the other with carbon fiber. With the same
number of layers of cloth of the same weight per square yard the
carbon-fiber wing will be stiffer -- and therefore weaker. Uniformly
stiffening a wing's skin alters its spanwise load distribution,
shifting load away from the tips and toward the centerline; the
wing's load-carrying capacity and its g-limit at a given load are
both thereby reduced. Stiffness and strength are not necessarily
complementary properties; to a great degree they're antagonistic.
This problem can be overcome by altering the number and/or weight of
the plies when switching materials, but that means re-engineering
the structure -- the fuselage and bulkheads as well as the wing-
skins and spar layups. Before you can build a *safe* all-CF Quickie
you'll have to design one.

Rick Nordgarden
Council Bluffs IA
Dragonfly MkIIH under construction


Quickie Builders Association WEB site

Yahoo! Groups Links

Join to automatically receive all group messages.