After 20 years of flying my second homebuilt, preceded by 9 years in my first, I wish I could say that most things that can go wrong already have; but the gods might think me insolent. At this point, however, I can at least say what has worked and what has not.

My first project, Melmoth, which was destroyed in 1982, on the ground, as I sat in it, by an errant Cessna, had several features that were unusual at the time.  Jane’s All the World’s Aircraft described it rather grandly, in 1976, as a “lightplane research prototype”. It had a single air intake below the spinner, wing-mounted airbrakes, automatic fuel-tank switching, a double-slotted Fowler flap, a T tail with a stabilator, and ailerons whose incidence could be adjusted in flight.

The landing gear was retractable. The association of speed with retractable gear was automatic then; today’s new breed of fast fixed-gear aeroplanes with powerful turbocharged engines was still undreamed-of. But there was also a considerable satisfaction in designing and hand-fabricating a successful retractable landing gear.

The official raison d’être of the “Experimental Amateur Built” category is “education and recreation.” In the course of making retractable oleopneumatic gear from scratch, I learned many things about O rings, dissimilar-metal bearings, heat treating, centreless grinding, heliarc welding, and the effect of foaming on the damping properties of hydraulic oil. I added to my mental thesaurus phrases like “hydrogen embrittlement” and “hard chrome plate and bake,” that I could toss off in later years to bolster my otherwise very slight engineering cred.

The performance advantage of retractable over well-faired fixed gear is now thought not to justify the added cost of manufacturing, maintaining and insuring it. As a homebuilder, however, I carried no hull insurance, donated my labour, and found the gratification of a smoothly-functioning retractable gear – which it was, most of the time – well worth the trouble of building and caring for it. I had a few mishaps – most notably an untimely genuflection of the right main on the ground at, of all places, the Bede factory in Newton, Kansas – but I never repented the choice.

The adjustable-incidence ailerons, on the other hand, proved to be a pretty useless feature. Their practical effects on landing and cruising speeds were imperceptible. They did have one virtue, however. Because of an unintended asymmetry in their actuation geometry, they could be used for roll trim.

Melmoth’s airbrakes consisted of a rectangular slab of aluminium honeycomb that rotated out of a vertical slot in each wing. Driven by the same ex-T-33 hydraulic pump as the gear and flaps – once you have a hydraulic pump, you might as well use it for everything – these popped out quite abruptly, killing some lift in the process and giving passengers an unexpected, and undesired, thrill. They were good for getting down fast, but not much else. On my second design, the four-seater that I imaginatively named Melmoth 2, I tried a different approach: a flat surface of about three square feet that swings down out of the belly, like the airbrakes of World War II dive bombers. The brake is very effective, but it is located too far forward; air pressure piling up ahead of it lifts the nose. At approach speed it helps trim out the nose-down moment from the flap and presumably even adds a little lift, but a complete lack of trim change with airbrake actuation would be more of a thing to be proud of. That would have required putting the brake farther aft, but the location of the wheel wells would not allow it.

Melmoth had a double-slotted Fowler flap that deflected 45 degrees; Melmoth 2 has a single-slotted one that deflects 30 degrees. Both increase wing area by sliding aft all the way to the trailing edge for takeoff before tilting down to their full deflection. My homemade hydraulic cylinders, however, have been a constant source of minor leaks, however, and of an occasional crimson tide. At times I think that perhaps I should have used electric actuators; but then I remind myself that all untried systems are trouble-free.

Melmoth’s control stick was between the pilot’s knees. Melmoth 2 has a sidestick, which provides much less leverage, and so control forces are more of a problem. After flying for several years with rather high roll forces, I made a second set of ailerons with more aerodynamic balance. They are better, though still not so light as I would like. I also initially had the elevator trim tabs doubling as servo tabs to lighten pitch forces. This proved unnecessary, and left me with limited trim authority besides; I eventually ditched the servo component, and never missed it.

Melmoth 1 gets a T – tail

Very long range was a design goal of both Melmoths. The first carried 155 gallons of fuel; the second holds 140, but, because its wings have more than twice the aspect ratio of the first, has about the same 3,000-mile range. Keeping track of fuel quantity is hard when the fillers are at the wingtips and you almost never fill your tanks; you can’t just look inside to inspect the fuel level. My float-type senders, rescued from the first airplane and now into their fifth decade, are not infallible. I have always had a totalizer, but a totalizer needs to be reset from time to time, and to reset it you need to know exactly how much fuel is in the tanks at that moment. In Melmoth 2, whose wings are 16 feet long and entirely wet from root to tip, I added small fuel-tight ports at the wing roots to allow dipsticking the fuel.

Despite being bigger, Melmoth 2, which has the same 200-hp Continental 360 engine as its predecessor, is slightly faster, cruising at 170 knots at 12,000 feet on a 60-percent-power fuel flow of 8.5 gallons an hour. It also climbs better. The principal reason for both improvements is wingspan. Melmoth started life with a 23-foot span and a wing area of 92 square feet. After adding a turbocharger in 1980 – a change that many people said would be troublesome, but wasn’t – I increased the span to 28 feet and the area to 112, regretfully discarding the cambered-and-beavertailed tip tanks that gave the plane much of its character. Melmoth 2 has a 106-square-foot wing of 37-foot span. When it comes to range and climb rate, span is magic. For roll rate, not so much; Melmoth 2 rolls in such a stately way that I no longer do rolls at random moments just for fun.

Melmoth 2 is a long winged composite

On Melmoth 2 I kept the T tail, which I liked, but replaced the stabilator with a fixed stabilizer because of some memorably unpleasant minutes when ice accumulation on the leading edge of the all-flying tail made it bob up and down ominously.

Both Melmoths are roofless, with canopies completely transparent except for a narrow spine in the middle to which the gull-wing doors are hinged. Entry and exit has been a disaster in both, Melmoth 2 particularly, because it is even more difficult to get into the rear seats, which face aft, than into the front ones. Most passengers, however, are too polite, or too apprehensive, or too busy figuring out where to put the next foot, to complain. I, on the other hand, have perfected my technique, and slide into my own seat with athletic grace.

Or so I like to think.