Taxiway Of Tears


Diamond Aircraft’s factory—a place with a storied history—is located quite some distance from the London, Ontario, airport where it’s sited. Access to the runways is via a long, tree-lined taxiway that would be a pleasant trip if I hadn’t been nearly killed on it at least three times.

For editorial titillation, I’m exaggerating, but not that much. My first brush with death was in 1996 when I’d flown our Mooney up there for a press visit. After I landed, the radar showed an intense line of thunderstorms approaching from the west so Diamond suggested I taxi down to put the airplane in the factory until the storms passed. I didn’t quite make it. The gust front arrived just as I turned onto that taxiway and I got my first experience in taxiing in 50-knot winds. It can be done, but I don’t recommend it. The hail started just as we pulled the airplane inside.

Thanks to the march of progress and the insatiable need for media consumers to have moving pictures, I again confronted the grim reaper there last month by getting out of the airplane to start a couple of exterior cameras. This is not, of itself, life threatening except it was about 20 degrees with 22 knots of wind. I live in Florida. For a reason. I don’t do chill factor. Addled by cold, when I got back inside the airplane, I dislodged the audio jack resulting in a radiantly perfect video of two talking heads conversing silently. I’m moderately resourceful in audiovisual arts, but lip syncing eludes me.

On my second trip a month later, all three audio sources worked perfectly. Proving that bumbling incompetence has its rewards, I got to spend another day in the factory shooting last week’s video on how they build composite aircraft. It was worth the additional time because I got insights I missed the first time around.

What, exactly? Well, this: When I report on new aircraft, quoting the price is part of the basic journalism and this is like cuing the choir: “$500,000?!! They could build this cheaper than that!” There are variations on this theme, but many arguments make comparisons to the automotive industry which are, well, inane. I share the anguish over not being able to afford a $500,000 airplane, but after watching them being built for most of a week, I left the place wondering how they build them for that little. Or at all.

Compared to a modern car, a light airplane is not a complicated machine. I’m sure it has fewer parts and given the criticality of weight, much of the structure encloses … air. The demands on the structure—especially wings—are such that they can’t be infinitely strong, but just strong enough compromised against weight. The strength-to-weight ratio of composites against metals is an evergreen circular argument. But it ends with the fact that in light aircraft, composites haven’t rewritten the physics book. Metal Skyhawks and Archers have similar useful loads as like-powered Diamond DA40s. If either composites or metal afford advantageous production economics, it doesn’t filter down to a meaningful difference in price.

As you can see from the video, composite construction involves many separate steps, but perhaps fewer than metal construction if you consider setting a single rivet as a separate step. High volume is the imaginary savior of general aviation, but I think it is and always will be an illusion hobbled by the chicken/egg paradox with automation hovering off stage as the great leveler. In composite work, automation might not be so magical.

When I visited the Daher factory in Tarbes, France, two years ago, they were using robots for some layup work and were beginning to experiment with robotic riveting for metal structures. And don’t forget Eclipse’s bromance with friction stir welding that sure sounded miraculous to those of us who didn’t understand it—me included. In practice, it achieved no meaningful production gains, or at least not enough to overcome the relentless economics that conspire against profit in building airplanes.

When I’ve visited companies building composite airplanes and structures, I’ve always been told that the process is stymied by throughput limitations; it doesn’t scale well. One reason for this the curing process. Some factories use autoclaves—a fancy name for a pressure cooker—and parts have to cook for a set time. Building more of them is expensive and if volume doesn’t support it, they end up being underutilized. Sleeping assets as the MBA wunderkinds might say.

Here, Diamond has a slight advantage. The resins they use don’t require autoclaving, just slight heating to accelerate the cure—pre-curing, they call it—so the parts are locked down well enough to be handled, allowing assembly to move forward. The resins themselves have improved, too. You may recall that when the Diamond DA20 first appeared, it had an operating limitation dictated by glass transition temperature on the spar. That’s gone, as is the requirement to paint them white with limited color trim colors.

On my first trip to Diamond, walking into the composite room was unpleasant due to the strong odor of the resins used then. Not anymore. The room is odor free and well ventilated. The resins have a four-hour working life, so workers don’t have to rush around worrying about the stuff getting too stiff to work.

When I was filming, it occurred to me that the strength of these bonded composite structures depends entirely on the integrity of the glue lines. When one wing came rolling out of the pre-cure, I had a good long look at that bond line. It was absolutely perfect; no gaps, sags or goobers. Chemistry rules.

I amuse myself wondering how the companies hoping to build urban air mobility vehicles will try to scale this technology. I’m a believer that the vehicles will happen, a skeptic that the volume will. Maybe some kind of extrusion process, like the Elixir project I reported on at Aero last year, will emerge as viable. Or something else none of us can imagine at the moment.

But if I understand how UAM is supposed to work, they shouldn’t need taxiways. I’m down with that.

Other AVwebflash Articles


  1. I was an early acolyte of Boothroyd Dewhurst.
    Their gospel: a small count of very complex parts is vastly preferable to a large count of simple parts. Integration works, believe me.

    Mandrel-wound robotic tape placement is a wonderful “lights out” production method. Software for managing gap, lap, and steer, has come a long way in a very short time. When coupled with resin spray heads, it’s like having a carbon fiber inkjet printer.

    Love the videos, Paul. Please keep ’em coming.

  2. Uber may have an advantage if things work out as planned. They can pay half a million. Their birds will be used constantly. Then, the price may actually come down with volume. Also, they can likely avoid insurance costs by starting out in places with less tort risk. Hire a few of the right nephews in some countries, and things go smoothly.

    Their problem will likely be when their builder goes broke because they will likely contract for too low a price. Then they will have to either deal for the type certificates or deal with whoever buys them. They seem the types to think ahead. Maybe they have thought this through.

    I wonder if Diamond ever thought of making boat hulls as a volume filler. If you haven’t tried a nice composite canoe, I highly recommend you do so. Light weight and easy to paddle.

  3. Paul, you make a very valid point about volume production and its elusive lure to bring aircraft costs down to the Everyman level. It would be interesting if one could retrofit the 1970s production lines for Piper and Cessna with today’s CAD/CAM and CNC production machinery. That was the only time (other than WWII) when aircraft manufacturing came anywhere near achieving true volume production. Modern machining and match-hole drilling have done wonders for the experimental industry, but mainly to make the end product easier to assemble, and not to increase production volume. The advantage of composit construction lies mainly in its ability to form highly complex shapes to enhance aerodynamic performance, and not replace metal for strength or ease of construction. As you observed, the gains in composit processes have come more through chemistry than automation. Perhaps if a company finds a way to automate production of composit sub-assemblies that can be quickly assembled into a finished product it would help. Sadly, as the LSA market has demonstrated, less expensive airframes have not kindled a wave of new buyers. It is kind of a chicken and egg dilemma; which comes first, hoards of prospective buyers or cheap(er) airplanes through volume production?

    • The LSA market has demonstrated that if one is paying Mercedes S-class prices for a tiny airplane, one should get S-class benefits. What one gets instead is a plane that cannot hold two prosperous individuals and full fuel at the same time, whilst costing up to [and over] a quarter of a million dollars.

  4. I don’t see the incentive for high volume production for GA airplanes. I understand the “if you build it (cheaper), they will come” argument, but I don’t think it’s true. As an ex-Advanced Manufacturing Engineer for a several large corporations, I frequently requested CAPEXs. It would be a special leader indeed who would see the capital required here as having an attractive payback. The market forecast for new small airplanes seems mythical at best, and if it were my company I would be more interested in using my resources to efficiently meet the current demand vs trying to upscale production to lower cost to create it. I know that’s more survivalist then growth mentality, but is any GA manufacturer positioned otherwise? Icon? I don’t know. It just seems like a fools errand with the current state of the industry.

    If the argument is to meet demand of the UAS urban-commuter market, I take issue with volume projections on those too. As a current airline pilot who operates a heavily automated airliner, there is no stinkin’ way I’m ever getting in one of those things without a human who can make it do stuff onboard. The amount of human decision-making and intervention required per flight in my heavily automated machine in the most efficient airspace system in the world gives me no confidence whatsoever in the ability for a UAS to consistently, SAFELY navigate a complex urban environment sans human. The NAS simply isn’t even close to being able to handle this. We may see airline crews go down to one person, but it will be a long time before we can fly around as helpless pax in any unmanned system without tragic results.

  5. What impressed me on the Diamond factory tour was their typically Teutonic precision and attention to detail. Their dedicated and amazingly skilled workers created a beautiful product. Woven stainless clad fuel lines (not your Cessna copper pipe) and 23G seats. I also toured Cessna when they opened their Independence plant. Their ambitious “welfare to work program” program, unfortunately, resulted in ailerons rigged backward and all kinds of mistakes through lack of supervision. Pay your money and take your choice!

  6. OK! What’s happening with the Icon A5? Would Diamond consider adopting the aircraft? Run it on the night shift.

  7. I enjoyed the video showcasing the build of a Diamond aircraft. I also enjoyed this blog post. I can sympathize with taxing the Mooney in a 50 knot wind. Trying to manage an 800LB TW airplane that flies at 26MPH in a 30 knot wind is a challenge to fly when you want and not fly when you don’t want to.

    An average for Diamond is 2500-3500 man-hours to build and deliver an airplane depending on single or twin. Cirrus build times are similar at 2800 man hours averages for their piston single aircraft. It has taken Diamond and Cirrus decades to get to this level of efficiency ( depending on one’s definition of what efficient really is) of hand-built aircraft. Diamond stated 23 man-hours to complete a Garmin panel, less wiring harness construction, and air-frame installation. There were lots of work stations dedicated just to build the wiring harnesses from start to tested finish. They did not say how many man-hours that entire process took. I suspect 150-200 man-hours total would be reasonably normal. But they did track percentages of efficiency as it related to each work station throughout the hand-crafted build.

    Looking at what Diamond has to do in body work on a completed air-frame is still, good old-fashion grunt body work that accounts for lots-o-man-hours to putty, block-sand, primer fill , static prime, and then hand-shoot the paint. None of the video showed the layout work required for a three color paint scheme, and the paint shooting sequence to complete the paint job. This is why custom car builds are six figures. An airplane has two to three times the surface area of an average car. That is why an average used single engine airplane costs $15K-35K at a refurbishment shop. Having experience in painting both Cirrus and Diamond Aircraft, $26-40K is the norm because of certified paint process required which in the case of both Diamond and Cirrus are proprietary blends plus the chemical/type data required sequence to strip and re-paint just as is done at the factory when new. If not adhered to in all phases of refurbishment, the airplane does not meet type data specs and therefor is not legally airworthy. Having the local PMI/FSDO/OEM lawyer get involved debating the airworthiness of the refurbed airplane makes for some interesting dialogue. Detailing an airplane just prior to delivery is also very time consuming both to check and then to fix discovered blemishes. Wet sanding and buffing is not quite so simple as described, let alone, repainting and blending.

    This is why the majority of the Cirrus and Diamond certified service centers refuse to strip/repaint, do warranty touch ups, or insurance work. This is why many customers of composite airplanes are stuck in 6-9 months waiting periods for warranty and/or insurance repairs let alone a complete strip and paint. Very few places want to do this. This is another reason why both manufacturers are in the refurbishment business is because of the technical/chemical, and man-hour requirements learned in the original manufacturing process that must be adhered to and applied when the air-frames get older requiring refreshing.

    Looking at the activity of employees in the background, at all levels shown in the video demonstrates a hand-built airplane at each work station, hand loaded on the rotisseries and jigs with a block and tackle, carefully hand lowered, picked up or turned. Many were looking at tablets and consulting manuals ( a good thing), with carpets and plastic taped all over the place as one climbs in and out of the airplane to complete the task. No sense of assembly line frenzy, speed, robotic automation because that automation is very limited for airplanes, especially composite airplanes, of which selling 100-400 per type per year is a banner year. Nine layers of composites including the protective peels to build the wing, tail components, and fuselage skins just to get to the vacuum bagging process. All of it hand laid up. Diamond and Cirrus look like a well organized customer assist build center with employees supplying the labor in lieu of the customers actually being there.

    I am very impressed with Diamond’s sense of quality, reliability, and craftsmanship. It is clear they have invested heavily in a trained work-force. They have to because composite airplanes are still hand-built, custom aircraft. I too, am surprised that Diamond can deliver such levels of engineering and quality at $500,000 dollars.

    The world market for $400,000 – 1 million dollar airplanes is about 800-1000 per year total. Those figures have not changed much since the resurrection of piston singles in the early to mid-90’s. When airplanes sold for modern dollar equivalents of $100K or less, sales were 15-17,000 per year in the late 70’s. Van’s alone has demonstrated those price points with 12,000+ sales and climbing. It takes 12-2000 man-hours to complete an average Van’s kit with total expenses less personal labor at $100K or less, many averaging $60-80K complete including paint by folks who never built anything let alone an airplane. Maybe not Oshkosh Lindy quality but solid, airworthy, well equipped, and appealing airplanes. That suggests that to reach any levels of remotely meaningful economy of scale manufacturing 15-20,000 airplanes per year, a completed airplane must have the performance, style, handling, performance capabilities, automation, and price similar to a Van’s product. Anything costing more and delivering less reduces aircraft sales to our current levels.

    It is pure fantasy, as far as I am concerned, that “urban mobility” vehicles, which presently depend heavily on composite construction can gain any economy of scale in manufacturing. Plus, FAA regulations to come will include the present acceptable composite construction the FAA currently has knowledge with. While carbon fiber technology exists that can gain some benefit of automation as demonstrated by Elixir, it is still foreign to the FAA and its understanding of acceptable composite construction throughout the majority of the air-frame. That means more regulation regarding certification testing which is required for any commercial flying machine that can carry more than 50 lbs, human or otherwise. That is not including certification tests of the multiple electric engines, ESC’s, and batteries. It’s one thing to certify a new composite air-frame with an accepted and certified power source such as a Diamond, Cirrus, Windekker, or Epic aircraft has done. And all of those completed certifications were extraordinarily expensive and time consuming. Now add at to that recipe, new power-plant technology.

    Uber, Amazon, FedEx,etc., seem to have a relatively naive conception of the expense, time, stamina, and sheer willpower it takes to get through the FAA regulatory certification process. Maybe they think their money and business power will supersede that of a bloated, tax-fed bureaucratic juggernaut like the FAA. If I were a betting man, my money is with the FAA as the stronger. They have an unblemished record so far. Especially for regulations that are getting written, debated, and reviewed as development of new technologies mature or die.

    The 3D world of aviation has safety challenges unlike any other mode of transportation. For some reason in our homogenized, uni-sex thinking society, flying machines, particularly those who carry humans, can be developed like a car. Airplanes are definitely not a one size fits all machine. But folks continue to seem to think the problems of flight can be solved by some sort of algorithm.

      • YARS: It sure seems that it’s “The way of the future” by what’s being penned by the media.
        Here are some topics I’ve read:
        * self flying commercial planes
        * boeing autonomous plane
        * how soon will planes be automated
        * autonomous flying
        * airbus pilotless aircraft
        * pilotless planes
        * airbus self flying plane
        * automated planes

        Why, I even bought a book on Autonomy Research on Civil Aviation. This thing ain’t gonna sneak up on me.