Comments on: Imperial Airways: now with extra airmail

By: Erik Lund

Erik Lund — Thu, 05 Nov 2009 20:37:07 +0000

Okay, then:
Richard K. Smith, “The Intercontinental Airliner and the Essence of Airplane Performance, 1929-1939,” Technology and Culture, Vol. 24, No. 3, July 1983: 428–449.
How interesting –essentially a comparison of two planes that basically weren’t intercontinental airliners, the Short Empire “C” class flying boats, and the Sikorsky S-42.
That said, Smith has an interesting point to make. The Pan-American clippers did have their part to play in the development of civil aviation, and the fact that they, like all American flying boats before 1940 or so look a little old-fashioned doesn’t mean that we shouldn’t look inside.
Unfortunately, that’s where we go astray, specifically into the deep weeds of (once again) advertising. Taking published Sikorsky numbers seriously, the S-42 was a more efficient weight lifter than the great bombers of 1944. Smith concludes that American design offices were incredibly good at weight control.
And, presumably, lost those skills when they turned to military projects.
A more likely conclusion is that Sikorsky’s claim that “the bare weght” of the S-52 was under 20,000lbs was a little nose-stretcher. I’ve linked below to a 1941 accident report that debunks another of his claims (the 62mph landing speed at full flaps). Unfortunately, I haven’t turned up a structural analysis.
Can we cut through the p.r.? Both Empire boats and S-42s were issued British Civil Aviation Certificates of Airworthiness as modified for the Atlantic proving flights of the summer 1937 Atlantic proving flights. They were published in _Flight_ for 15 July, 1937, page 70, and are very different from Smith’s numbers.
The rest can be put down to a failure to grasp why it was easier to fly long distances in tropical latitudes in the 1930s, although the slam against the safety and sructural strength of the aircraft that became the Short Sunderland from a partisan of the S-42 is, well, kettle, pot, black.

http://www.webstart.com/jed/house/NC15376/nc15376-full-pics.pdf

By: Erik Lund

Erik Lund — Tue, 27 Oct 2009 18:19:14 +0000

Jakob:
Fearon and Robertson sounds about right. Truly an aertefact of another age. Nowadays you can’t walk through the lobby of a big bookstore without being handed a splatbook with a description of the B-18. (Okay, maybe I exaggerate.)
For political imperatives and wood construction, I was referencing Schatzberg. It’s just that I kept visualising Vicenti and thinking, “that can’t be right…” I may have seen the book, can’t recall now….
I missed Richard K. Smith’s article entirely. I’ll track it down it down the next time I’m at the library, along with Higham on Bede (thanks for the tip, Brett!) *
I’d be interested in knowing what aircraft he is comparing to what, though. The first British intercontinental airliner that can really be compared with an American is the Ensign. On the one hand, I’ll buy that the Ensign was structurally inefficient compared to any American aircraft you’d want to name. On the other hand, I’d make a lone exception for the (first) DC-4, its only obvious competition.

*By the way, you youngsters have no idea what it’s like being an “independent scholar” with a fulltime job. Why, in my day… Oops, sorry, wrong diatribe. I think I need to work up a classic “don’t make my mistakes” lecture instead.

By: Jakob

Jakob — Tue, 27 Oct 2009 17:15:54 +0000

Erik: Were you thinking of Peter Fearon and AJ Robertson in the Journal of Economic History in the 70s?

On wooden structures, the only stuff I’ve read is Erik Schatzberg’s 1994 article “Ideology and Technical Choice: The Decline of the Wooden Airplane in the United States, 1920-1945″ Technology and Culture 35(1); I haven’t read his book, never having seen a copy of it at a reasonable price, and my uni library doesn’t have it.

On structural efficiencies more generally, I don’t know how Richard K. Smith’s “The Intercontinental Airliner and the Essence of Airplane Performance, 1929-1939″ (T&C again, 24(3) – 1983) has stood up, but I recall it suggesting that US weight control in design was superior. Of course, whether this is evidence of UK technical inferiority or simply different contexts for their airliners is another matter entirely.

I can only say I heartily agree on the difficulties of visualising the effects on prop pitch of supercharge and gearing – I always seem to end up with a bad case of fighter pilot’s hands…

By: Erik Lund

Erik Lund — Sat, 24 Oct 2009 18:58:22 +0000

I thought I was being slimy enough bringing up Rutbah Wells! Borrowing a C. G. Grey talking point? I need a shower….

By: Brett Holman

Brett Holman — Sat, 24 Oct 2009 02:51:36 +0000

I can't believe nobody has brought up the American air mail scandal yet! :)

By: Erik Lund

Erik Lund — Wed, 21 Oct 2009 19:05:23 +0000

To begin at the end, Lord Keynes once said something about the world’s so-called practical men actually simply being followers of obsolete theories. Corelli Barnett (himself echoing Nineteenth Century Nonconformists) claimed that British culture made it impossible for Britons to innovate. Proximately, this justified the Thatcher-era culture war. In a deeper sense, his argument still informs social development theorising, vid Joel Mokyr or David Landes.
This brings us back, or ought to bring us back, to his technique, which is to dig up an exhaustive array of British patentees, subsidiaries, and secondary adopters, combined with often ill-informed critiques of specific British products. and throwing them at the reader like some mad old physical culturalist throwing a medicine ball. The very idea of fact checking _Audit of War_ is so exhausting that one is tempted to give up before one begins.
At least the DC-2 provides us with a grip. Was it as good as they say, specifically in one of Barnett’s favourite citations, a minor mid-60s controversy in the the pages of the _Journal of Economic History_? This duscussion purported to use published statistics (hence my concern with published information about the DC-2) over whether the Douglas airliner showed that the British aviation industry had comprehensively failed to innovate. (I want to say that Roy Ferron was involved, although I think that he was the “anti-,” and I cannot for the life of me remember the “pro.”)
This is an especially interesting conversation because we should never have been having it. The MacRobertson race demolished Douglas’ claims at the time, showing a 20mph gap between the firm’s claimed cruising speed and realised performance, and demonstrating what any passenger knew anyway by looking out the window: that the time for metal-skinned wings was not yet. (Had it been possible to make practical metal-skinned flaps and tail planes in 1933, there would have been a strong maintenance argument for them. As it is, the literature resorts to political explanations.)
Let’s be clear about something: KLM did not “choose” to lose. It was beaten by a plane that was faster than it was; a plane that was not supposed to be faster than it by contemporary publicity, or by Barnett’s theorising.
Still, we do owe Douglas due weight of consideration. First, the firm had no choice. There was a Congressional mandate for an all-metal plane. Second, plasticised Egyptian aeronautical grade cotton was very expensive in the United States in 1933, due to preferential tariffs on imported cottons and the cost of importing Cellon dope. Third, the structure of American air mail subsidies pushed American manufacturers in the direction of higher cruising speeds at the expense of disposable lift (What’s left over when you subtract engine and structure weight from all up weight). Fourth, the somewhat specious safe takeoff requirements for high altitude Rocky Mountain airfields, combined with the retractile undercarriage adopted to increase cruising speed, leading to the low wing configuration, practically forced Douglas in the direction of the two-speed adjustable airscrew, and subsequent variable pitch and, eventually, constant speed airscrews. As we all know, of course, had American aeroengine manufacturers been faster in adopting reduction-geared engines, the aerodynamic inefficencies and weigtht penalties resulting from the premature adoption of this technology could have been avoided, and the path to the smooth adoption of two-speed or even multistage superchargers made more straightforward. Did Curtiss-Wright and Whitney delay the adoption of reduction gearing because their main customers already used adjustable pitch airscrews of various designs, or was it just a matter of not having R&D money during the Thirties? An economically endogenous account of technological innovation will have its own bearing on the Barnett-Mokyr (“There is so such a thing as a free lunch” ) model of innovation.
(As I said, the development of the airscrew is a rather complicated one.)

The facts here should not be in dispute. Metal structures are more efficient than wood ones, but the short-run developmental potential of plasticised wood materials favoured by Russian designers and by de Havilland were not exhausted until the early jet age. That is why the RAF and DCA shifted over to metal structures in the 1920s, while de Havilland continued to explore “wood” planes until the Hornet. Aerodynamic cladding is a different matter. Again, the potential of plasticised material (“doped” cotton) was not exhausted until much later than is sometimes assumed. Not only the DC-2, but the B-24 and early Spitfires had fabric control surfaces. The presence of fabric surfaces greatly reduced the clear maitenance advantages of metal cladding; although of course the need for metal structure for tropical operations was obvious.
So, for a designer, the question is whether you favour payload, operability, however defined, or passengers. Fabric-covered biplanes had a huge payload advantage over metal-covered monoplanes and a lower landing speed, the major determinant of operability, especially on “Empire” routes. What degree of speed advantage offsets this? Clearly if the DC-2’s speed advantage over the DH-86 is more in the range of 30mph than 60, this is going to influence orders. If the DH-86 then turns out to be an unairworthy plane, then it is time to start talking, very loudly, about Rutbah Wells, and hope to get a replacement flying quickly.
One more word about airscrews: patents. De Havilland wanted to supply airscrews as well as engines. All very well, but the Hele-Shaw patents were already assigned in Britain. Hamilton’s patents were clearly sound, so De Havilland had good reason to go for them. As far as I know, the Ratier used on the Comet would automatically revert to the low speed mode if it stalled, unlike the Hamilton in use in the MacRobertson, a safety feature rather than bug, albeit a pretty expensive one if it happened in cruising flight and overspeeded an engine. I could be wrong, though. When I try to think about the relationship between reduction gearing, superchargers and propeller pitch change mechanisms, my head explodes and my face turns a funny shaped. If it freezes that way I might as well kiss my chances at the girl down at the game shop goodbye, so please use short words and easy similes to explain my errors.

By: JDK

JDK — Wed, 21 Oct 2009 07:43:39 +0000

Dear Eric,
I’m not sure what you are trying to say before ‘getting our facts right’, so I’ll pick up there. I’m also not Corelli Barnett, not trying to prove political or national theories, just to clarify some facts of aviation history and performance.

You misquote the crew of the DC-2. There were four, not two.

Your sentence on the DH86 repeats ‘monoplane’, and as a result perhaps, I’m not sure of your point. It doesn’t matter, because you’d never get me in one, it was, of the aircraft listed, the main one to avoid.

(I’ve examined examples of all the types we list apart from the 86, which is thankfully extinct, and I’ve flown in the DH89 Dragon Rapide, and the Lockheed 10, as well as the Junkers Ju52/3m. DC-2 and Boeing 247 flights are a bit harder to catch, but I’ve crawled all over examples and talked to operators as well as having a go running the engines on a Lockheed 12. I’ve also talked to the men and women who fly these aircraft today, and use their understanding and explanations for my writing. The development of the propeller isn’t a ‘complex one’ you just need to look for the explanation around vintage aircraft and operators, and their manuals, not in academic libraries.)

You persist in missing the point of the race. The DC-2 placed first in the handicap – that is the transport rather than the pure race category – but they chose to take second in the pure speed category. An achievement of some note done in a standard airliner flying an extended version of the KLM route.

Sir Macpherson Robertson, sponsor of the race, as well as wanting to boost Melbourne (and himself), said: ” This is just the result I wanted – to show that a transport plane could reach Australia in four days.”

The Comet racer crews were exhausted on arrival, the KLM crews hardly fresh as daises, but certainly not worn out by the pressures and tough environment the Comet crews had flown in. The soundproofing and comfort of the DC-2 was a world away from the British airliners, let alone a dedicated race aircraft.

“DeHavilland … leveraging its developmental effort into an attack on the airscrew market.” Not true in any aspect. They didn’t have a propeller to do the job, there was not a British design that could, and so they chose what can politely called an ‘interim’ design from Ratier.

“The Ratier was certainly a peculiar design. It was more aerodynamically efficient and more useful than the two-pitch on the Douglas.” I would be interested in where you get the data to suggest the Ratier was a better propeller than the DC-2’s Hamilton Standard. Not only was the Ratier inferior to the DC-2 propellers, it was actually compromised in performance and safety (due to the Heath-Robinson airflow pitch-change mechanism) and irreversible, so could jam, act asymmetrically, and critically, in the case of an aborted take-off or landing or a go-around, was unable to regain course pitch. They were replaced on the Comet type as soon as possible and when the Comet Racer was restored at Shuttleworth, there was no question of using this poor interim design.

Further, de Havilland (note spacing and capitalisation) did not develop their own propeller design pre-war, they set up de Havilland Propellers in 1935 (note date) by licencing Hamilton Standard designs from – America.

Next, the engines. The Comet Racer’s unarguably marvellous design and performance was achieved by using a dedicated race engine operating at maximum performance – one of G-ACSS’ engines having a partial failure before finishing the race. No-one would argue that was a remarkable achievement, but the point in terms of aircraft development and performance was that the KLM crew, using standard airliner engines at standard cruise power, oil and fuel came in less than a day later (you could argue the loss of eight hours at Albury covers most of that time anyway, in terms of aircraft performance as opposed to race achievement, but let it pass.) As I pointed out, the Boeing 247 was obsolete, and they came in third only by pushing the Boeing’s engines – by arrival, they were un-usable for further flight.

Why the obsession with brochure data? It has and remains an aspect of PR not reality. No airline did by its aircraft on the brochure performance, certainly not KLM nor the American airlines sponsoring the type. It’s irrelevant, and hardly surprising in the run up to a race. The performance fact was that no British airliner could touch the Douglas in 1934.

You say “Now, the DH86 was not the most successful of aircraft…” It wasn’t that – it was an unsafe, unairworthy disgrace, unfit for safe airline flight. It’s structural and fundamental design failures led directly to at least one fatal accident on introduction to service, almost certainly more, and I remain unconvinced that after modification it was a sound design, given the obfuscation de Havilland and the British Air Ministry undertook when challenged over the fatal structural failures initially. The reaction of the British to the Australian suspension of its certificate tells us a lot of Colonial attitudes of the time, and the change demanded by Qantas’ Chief pilot from a single pilot cabin to dual to address pilot fatigue tells us a lot about de Havilland’s poor understanding of the job, and aerodynamics – few if any other airliners were single pilot by this time. I don’t know and don’t care whether de Havilland’s brochures on the 86 were ‘accurate’, the fact that they argued that a two-pilot nose would be slower is telling of their aerodynamic understanding – it proved faster. (See MacArther Job’s writing.)

The DH-89 was indeed a good aircraft according to its category. (I’m vaguely curious as to what ‘disposable lift’ might be. Payload is what the industry uses.) But it was too small, too slow and too delicate for major use for main routes by airlines. It takes a one-eye British industry booster to see the 1934 DH-89 as ‘equal’ to the 1935 Lockheed 10. The de Havilland airliner also came in tenth in the Great Air Race, well after the obsolete Boeing and the brand new Douglas.

The DH89 Dragon Rapide was not in the same class for performance as the DC-2 or even the Lockheed 10 while the maintenance and care requirements for wood and fabric de Havilland types were significant, while the metal aircraft were simply more cost-effective, easier to inspect and could be left out in the rain. In case anyone assumes an American metal vs British wood bias, note that in Papua New Guinea, the world’s greatest freight operation of the inter-war period used German Junkers all-metal aircraft because they were the only type able to do the job. In Canada, de Havilland types were found inadequate for bush flying, and the record setter for the flight across Canada in the late 1930s was the Lockheed 12, while Junkers types being popular despite their extra tariff-driven cost.

Any airline that could switched from DH86 Express, DH84 Dragon and DH89 Dragon Rapide to metal aircraft as quickly as was possible, and only went back when metal aircraft were unavailable. Only war kept those types in production and use until 1945.

The main reason the Commonwealth bought British airliners in the 1930s was because they had to – through Empire Preference tariffs and bans on the importation of US (and I presume other nationality) types. I don’t know if that’s why de Havilland types were second rate, but it’s certainly how they managed to make enough sales.

I’m not sure where Correlli Barnet comes in – if he need to at all. I’m not pro- or anti- any group – there was great achievement and innovation by British aero industry, but not perhaps as much as their story would have, although the development of the Comet principle onto the lovely but ill-timed Albatross and then the truly great Mosquito is case enough of de Havilland success in particular. Meanwhile American designs (and German) were certainly not perfect either. But were I buying an airliner in the mid-1930s, I’d know what would be best. The race sponsor, Sir Macpherson Robertson, de Havilland, Douglas, KLM both Imperial Airways and Qantas saw what the race showed and drew their conclusions. Which is where we came in.

Regards,

By: Erik Lund

Erik Lund — Tue, 20 Oct 2009 20:07:25 +0000

As we know from the JEH debate of two generations ago, the DC-2/3 is the _most important thing ever_. The fact that it was better than all British interwar planes demonstrates that Britain, despite massive support for research and development and all-out official encouragement of airmindedness, could not match the achievements of American private enterprise. Later, historians like Barnett and Wiener deployed this argument to revive the old Nonconformist argument that being British was bad for science.
I know that digging up a minor forty-year-old controversy seems little more than antiquarianism, but this discussion heavily influenced an obscure British policy maker of the 1980s (Roofer? Hatcher?), and had some small consequences on the way science, academics and even financial regulation is done the world round. So it does behoove us to get our facts right.
Those facts are: i) the DC-2 was entered into the MacRobertson Race with a claimed 75% power (1400hp max) cruising speed of 196mph, 62% power cruising speed of 183mph, disposable load of 6,125lb (34% all up weight), landing speed 61mph (_Flight_, 1 March 1934, 189–91)
At a very conservative consumption rating of .5 lb/hp hr, it should have had a range of 429 miles per 1000lb gasoline carried. Indeed, it was originally intended to install long-range gas tanks in the KLM DC-2 and enter the new Fokker 36 in the race to compete for the handicap prize. (_Flight_, 1 Nov 1934: this is a PDF lift, so I’ll suggest searching for “Albury”).
The figures cited above are part of a gruesome battle for sales in the early months of 1934. _Flight_ offers them in comparison with figures for the DH-86B, a biplane with cruising speed of 145mph at a consumption rate of 36 gall/hour, a total disposable weight of 3675lb (40% auw), and, most saliently, a landing speed of 66mph. The implication of a monoplane with a higher speed range than a monoplane led _Flight_ to very gently editorialise the possiblity of noses growing. This was repeatedly denied by Douglas, KLM, and the media. Now, the DH86 was not the most successful of aircraft, but the DH89 has a far better claim than the DC-2 to be fhe first airliner to flly commercially without subsidy. Note the statistic for %disposable lift, not spurious claims to speed. Unless your business model incorporates US Post airmail subsidiies that will be paid even if you send the mail by train, so long as you operate a passenger service –an excellent Congressional initiave for _American_ conditions–, it is going to be disposable lift that determines your profitability. In turn, efficiency of structural design that is decisive, not aluminum skinning, of, as we shall see in a moment, dubious merit.
The MacRobertson Race was announced months in advance. Plenty of firms announced that they would enter experimental types, notably Lockheed. Only DeHavilland made the cut off, leveraging its developmental effort into an attack on the airscrew market. (The story of the adjustable>variable>constant speed airscrew is a complex one. Perhaps it will be written one day. The Ratier was certainly a peculiar design. It was more aerodynamically efficient and more useful than the two-pitch on the Douglas.)
When the planes reached the finish line, the tale of the tape was this: Scott and Campbell Black took 71 hours to fly London–Melbourne. Thanks to small engines on a small plane, they stopped only 5 times, spending only 9% of transit time on the ground, reaching an average cruising speed of 190mph. Carrying 2 crew, 3 passengers and 420lb air mail (about 1/6th of the claimed disposable lift at which the aircraft was supposed to achieve the figures cited above), the Douglas was on the ground 15.6% of elapsed time, turning in an actual cruising speed of 173mph, and for most of the race only barely ahead of the Boeing 247D.
This is does not prove what Correlli Barnett thinks it does.

By: JDK

JDK — Tue, 20 Oct 2009 05:12:10 +0000

Hi Brett.

Sorry. Here’s your thunder back, only slightly worn. :D

Regards,

By: Brett Holman

Brett Holman — Tue, 20 Oct 2009 04:49:24 +0000

How’s that, JDK?

I was going to post on the 75th anniversary myself, but not sure it’s necessary now :)