Tuesday, December 4, 2012

Running Away to the Air, 7: Nowhere Near Willow Run

It was supposed to be a car for the ladies, with its push-button transmission and warning lights. Specifically, it was for suburban ladies, with space for kids, and a big engine that let it pull out of one traffic light after another on the way out of town, and swallow up the interstate miles. If failed, perhaps, because Robert McNamara's number-crunching analysis said that it was the wrong kind of car for Ford, and because a flood of Beetles was closer to what the market wanted. 


It's B-24 "Witchcraft," as featured at Warbird Alley, available for your air show through the Collings Foundation, and featured at Jay Leno's Garage, here. The first B-24s erected at Ford's purpose-built Willow Run plant were delivered to the USAAF seventy years ago last week. I've talked about this before in the context of what I called VLR ("very long range") magic, and I later threw off links to some compelling monographs that study the actual history of the American aviation industry,(1) but, as the sneer word suggests, the myth-history goes on. I was reminded of the date by Brad Delong's "Liveblogging World War II" feature, and, as usual, some commenters popped up, still deeply immersed in myth.

 Wikipedia's actually pretty good these days --I know, my constant refrain before piling off research on it, but there you go. Anyway, on 6 March 1887, one Reuben Fleet was born in Montesano, Washington. That's 39 years after Washington became American territory, only 44 since the much-hyped first pioneers travelled what was erroneously described to them as the "Oregon Trail."** Through means not advertised in the official biography made available by the Reuben Fleet Science Center, the Fleets had come into possession of substantial amounts of forest land around Grays Harbor, Washington, and thereafter prospered as American pioneers were wont to do: living off rents through their own hard work.

In perhaps the most amazing turn of events ever, a then-thirty years old Reuben Fleet, already a captain in the National Guard with a sterling combat record against the Wobblies, was commissioned major in the United States Army after enlisting out of his realtor's office. He became an aviator, commanded a training section, and then  joined Hap Arnold's staff in the signals branch of the Army Air Corps headquarters in London, returning after the war to run first the Army's brief experiment with air mail, then the Engine Experimental Section at McCook Field, Dayton, Ohio, as it then was, where he played what the Science Center describes as a "key role" in the development of the turbocharger for air operations. We shall pass over tedious facts about metallurgy and rotational speeds in silence. "Key" roles may be found in the history of a technology even twenty years before its first successful application, after all. Notwithstanding these.

In 1922, Fleet resigned from the Army Air Corps, along with his supervisor, and went into private business, with a chain of acquisitions that culminated in the creation of Consolidated Aircraft of San Diego, which won its first contract for a Navy seaplane in 1928. From there, the road went clear to the Catalina, ordered as a "patrol bomber," notwithstanding the fact that, like all flying boats of its era, the hull could not be pierced and the wing was wetted on takeoff, so that munitions had to be winched out of the cabin and onto the wing before action. The Navy, eager to play "strategic air force," received more than 220 between 1936 and the outbreak of the war. To put this in perspective, Holley notes American military purchases of 552 units in 1926; 1,219/1928; 747/1930; 593/1932; 337/1934; 1,141/1936.***

Fleet opted to sell the majority of his shares in Consolidated to an entity called Vultee Aircraft in 1941, a firm started by Gerry Vultee and Vance Breese in 1932 in some kind of in-no-way-can-anyone-be-sued-for-anything relationship with American Airways, the firm that emerged from the Air Mail Scandal  holding many of the contracts. He was by this time rich enough to found a Science Center, and so we remember him.

Meanwhile, in the vaguely real world of aviation contracting, Consolidated's success with the Catalina led to aspirations for an army contract for a "heavy bomber." Asked for speed and performance, Consolidated's men went to the moulding loft(2) and adopted some fairly typical expedients for American designers of 1938: reduced wing area to bring down drag and a tricycle undercarriage to increase safe landing and takeoff speeds, which, in turn, allows for better performance in the air at a given speed range, which is determined by the nitty-gritty of aerodynamic design and can go disastrously wrong as one plumbs the world of flaps and rudders. More to the company's credit, it plumbed its experience with flying boats to give its new Model 332 a remarkably good aspect ratio, a drag-at-speed reducing design feature then in the news. 

Even less to its credit, Consolidated somehow let "freelance aeronautical engineer"** David R. Davis into the picture. Wikipedia has the "official" version, apparently sourced in a 1987 article by aviation technology history doyen Walter Vincenti, in which the extraordinary performance of the Davis wing in some trials in the California Institute of Technology  wind tunnel are the result of Davis somehow lucking into the "laminar air flow" design later seen in the P-51.**** The official writeup by Consolidated engineer J. H. Famme describes it as a "modified Davis section," basically geometrically similar to the wing of the previous Consolidated flying boat design.(3) The history of the University of Washington's wind tunnel programme, which was also involved in testing the Davis wing, offers the saner (with all due respect to Vincenti, who did not, if my recollection of the article is correct, endorse the idea that the Davis wing actually worked) explanation that the tunnels were having startup problems and giving glitched results.

So the Davis relationship was weird, and poor high altitude performance was basically baked into the B-24 cake by the choice of wing geometry. Like the Short Stirling before it, the B-24 had the misfortune of flying bombing missions against Germany's crack AAA at a lower altitude than its stablemate, in this case the B-17, and so suffering a higher loss rate that led to its being phased out of service in its main operational role.(4)

That's not the point. The point is that the B-24 was chosen for production at Willow Run, and it was a disaster. Just to put things in perspective, as official historian Irving Holley, does, the aviation and automobile industries were completely, utterly different. In 1937, aircraft workers worked an average of 42.3 hours at an average compensation of 67 cents/hour. Automotive workers did 35.9 hours at 89.1. The difference in pay is accounted for the difference in output value, of $4400/year vice $15,000, and that difference by the relative capitalisation of  $800 versus $2600. This doesn't necessarily mean that the aviation industry was undercapitalised, however. Much of the spending in the automotive sector was on jigs and gauges that made it possible to build automobiles on production lines. Aircraft are big, so they can't be built on production lines, making jigs and gauges pointless.

Unless you build the aircraft in parts, and then erect them in a final stage in a semi-production-line method. There's some weird disconnect in the literature where if you get a work that's early enough, or out-of-touch enough(5), this is the future of aviation, since mass production is better than anything. Then someone noticed, perhaps about the time of David Hounshell's deevastating indictment of the "American system," or more likely those exciting days when Japan was going to rule the world real soon now with "flexible production," that the world didn't actually need that many of a lot of things . 

It's not that the aviation industry wasn't pretty heavily capitalised by that time itself, but it was mainly with machines for making  high-temperature high-stress high-strength thingies. They weren't doing that with the B-24. It was about the last generation of planes that were built with raw materials ordered out of an industrial catalogue and arranged into a final structure using the traditional methods of boat-building.

All the skilled metal working involved in that leaves me scratching my head a little about the difference in pay as between automotive sector and aviation sector employees. I'm not sure that it would have been possible were it not for the fact that Consolidated was off in San Diego, far away from Detroit. I'm just not entirely clear where Consolidated was getting skilled metalworkers at the edge of the American west, and how this reservoir of skilled labour was holding on. On the other hand, we are talking about the late 30s here. It's not clear that the Consolidated labour model would have continued to work through the war years.

The other big problem was that mass production requires long runs. Long runs mean "freezing" a design. Which means getting a great many of a design.  Here we stand at an impasse. Holley offers the example of the C-47, the militarised DC-3, and here we have the problem. Everyone thinks that the DC-3 fleet was awesome, but to the extent that it was built up at the expense of more C-54s, one has to wonder. 554 C-54s were built compared with 10,000 C-47s. Yet, as Holley points out, the C-47's lack of range made it useless for island-hopping in the Pacific. Having twenty times more of one design than another isn't as much of a selling point as might be hoped when it can't do the job.*****

It's not like this did not go unnoticed in the lead up to the opening of Willow Run. In the official historian's account, Henry Ford was a sneak who erected Willow Run ostensibly as a parts plant, but with the aim from the first of "mass producing" America's best bomber. It was something that, given the WWI fiasco of the Eagle Boats, might have been expected, but the Administration let him get away with it, and, with plant in hand, production contracts fell into Ford's hands (520). Holley describes the resulting transfer of concrete knowledge from Consolidated San Diego to Willow Run as involving the creation enough production drawings to employ a "legion" of drafters for "years." One could wish for more specific numbers considering the likely impact of learning-by-doing on Ford's Detroit labour force.

Besides drafters, the transfer involved vast amounts of work for tool-and-die makers, toolmakers and production engineers, the human resources that Ford did have. Holley notes the miller used to work the  centre section jig, built for Ford by Ingersoll Milling Machines at a cost of $168,000, more than ten times the average cost of a military aircraft in 1939, and about the price of a finished B-24, if I recall my Goldberg correctly. For Holley, the issue is one of cost, particularly because constant modifications interrupted production series. My eye tends to focus on an increase of empty weight from 41,000 to 60,000lbs. Compare this with the 44,000lbs of a Lancaster. Unlike the C-47/C-54 tradeoff, which was at least between a more useful and less useful type, Ford had contrived to build the biggest factory in America to build the largest run of aircraft in the history of the American industry. 

Which were, if not useless, significantly inferior to the B-17. (Note that this is only entirely accurate for Willow Run. Production at other plants that stuck to traditional manufacturing methods had much better weight control.) To be fair, no-one saw that coming in 1939. The B-17 was by then already an old design. The first B-17 was delivered to the air corps on 1 March 1937. That's a pretty old plane to be reckoned combat worthy in the last months of World War II! Except, of course, that it's a myth. The prewar short production runs of B-17Bs were nothing like the "mass production" aircraft beginning with the B-17E that appeared in service in 1942. 

Externally, the aircraft were very nearly identical, but internally, the transition to largescale production forced Boeing to reexamine its production choices. Seven years is a long time, especially in an industry that has seen a dramatic research investment in that period, and especially in one that has been the recipient of a massive technology transfer. Consider that if, in 1935, you chose to use off-the-shelf sheets of production aluminum to make an open girder as a strength part, and in 1941 there is finally aluminum tubing cleared by the army for aviation use, problems with internal corrosion having been solved in the interim, why not use the inherently stronger tubing? Choices like these, which substituted stronger substructures, and often stronger materials, into the same dimensioning, produced an aircraft that could not but have been colossally stronger than the original. Combine this with an airframe not crippled in its conception by the choice of aerofoil, and you get a product that is going to be better than its newer rival. 

Too bad, Consolidated. The moral of the story, as I read it, starts with the relationship between Fleet, Davis and the men who bought Fleet out. From this distance, and admittedly through the lens of  Wikipedia, it all seems almost unbearably amateurish. These men fell into aviation, and entirely lack the technical background that we expect to see in leaders of a high tech field. But that's okay, because the aircraft that they are producing are not particularly "high tech." They're basically (and literally) flying boats, and people have been building light, high performance boats since long before educational credentialism was a gleam in the eye of SCIENCE! 

As the planes got better, the technology got more advanced. It's just not obvious where the locus of this improvement is. It's perplexing, at best, to see the engineers who would be building spaceships in 1959 being taken in by a flim flammer like Davis. It's hard to imagine them, in retrospect, doing the crucial work at the fledgling Cal Tech wind tunnel and then going back to work to lay out their planes with cords on pegs in a mold loft. Yet that is how it was done in 1939. Which is why I think that it is at least more interesting,  perhaps more important, that 67 cents an hour men built the planes,  until they laboriously transferred their skills and knowledge to the 89 cents an hour men who tried to build them at Willow Run. 

The transfer, I take it, was not to the untrained, unskilled migrants  on the assembly lines that Ford had to recruit from Tennessee to Texas with what Holley describes as their chaotic lives and chronic absenteeism.****** Ultimately, Willow Run came to be about transfer of knowledge from boatmakers to toolmakers and diemakers. This couldn't be privileged, high tech knowledge in a useful sense. There was none at either end to be transferred or modified. The American postwar "high tech" sector was emergent from this interaction, driven by the need to turn the crap at hand into weapons that America could use to, well, fight a war.  It's a mysterious process. 

Or I say that. Privately, and here, since this is just a blog, I can suggest that the key transfer of knowledge is buried in those aluminum tubes, to the under-enlightened ways in which American industry came to be able to deliver tubing that could be guaranteed to be pitting-free on the inside. This transfer, I suggest, was from Britain to its American ally. The implication is that >gasp< Britain was "higher tech" than American in 1939, and to the extent that this was true, it was because of defence spending beforehand. I come away from this meditation  conscious, again, that for my story of World War II to be true, the transfer of knowledge from Old World to New in 1939--45 must have been of significant extent and with significant consequences for the real American economy. It might even be a missing element in the story of that transition from 1939 to 1959 that we're all looking back on  now with the hope of replicating

*Lifted From Palm Spring Automobilist, Here.

**More typically described at the time as a "San Diego sportsman." Either way, he has vanished from the ken of history.

**Scare quotes because, unlike the branch of the trail that led to California, there was no wagon road through to Puget Sound, and thus no trail in a meaningful sense. The first three years' migration straggled into Fort Vancouver more as refugees than as pioneers, and had to be re-equipped for farming. Things tailed off after that, but mainly because everyone was going to California instead. As to the question of how "white" settler society established itself on the far side of the coastal ranges in these years, I think, as in the somewhat parallel case of Western Australia, we're being discouraged from inquiry.

***Can I point out that Corelli Barnett can suck my member, or has that boat already sailed?

****And, yes, that may be taken to imply that Consolidated had no wind tunnel of its own in 1938.

*****Unfortunately for economy of argument, the official historian can't put a discussion of aircraft range in there and note that we're only talking about useful loads, and note that the C-47's range wasn't somehow organically limited by its design. It's a question of useful range. Anyway, Holley, 514.

******I think that Holley might be using code for something there, but I'm not sure what.

1. Here, again, including some interesting articles as well as monographs. Note that the section authors in the USAAF official history are often ill-served by bibliographies:  Broadberry, Steven E."Manufacturing and the Convergence Hypothesis: What the Longrun Data Show." Journal of Economic History 53 (December, 1993): 772-796; Goldberg, Alfred. “Equipment and Services.” In Wesley Frank Craven, and James Lea Cate, eds., The Army Air Forces in World War II (6 vols. Chicago: University of Chicago Press, 1955): 6:171–426; Hounshell, David. From the American System to Mass Production, 1800-1932. Baltimore: Johns Hopkin, 1984; Holley, Jr., Irving Brinton. Buying Aircraft: Matériel for the Army Air Forces. A work in the series US Army in World War II. Washington: GPO, 1964.s University Press, 1984; Mensforth, Eric.  “Airframe Production.” Proc. Inst. Mech. Eng.  156 (1947): 24-38; Zeitlin, Jonathan. “Flexibility and Mass Production at War: Aircraft Manufacture in Britain, the United States, and Germany, 1939-1945.” Tech. Cult 36 (Winter, 1995): 51-79

2. I do have an article from Aviation from 1939 celebrating Consolidated's awesome mold loft, as well as a design history of the B-17E, published in a 1943 number of the same journal (see below). It's just that my notes are in hand in a box, and it's 1:30 already and I have to work tonight. If you want the citation, email me. 

3. Famme, J. H. “A Design Analysis of the Consolidated B-24 Liberator.” Aviation, July, 1945, 121-143.

4. Roger A. Freeman, Mighty Eighth War Manual (New York: Cassell, 2002) is a good source. If I recall correctly, there's a discussion in Craven and Cates, too, but there's no way that I'm going to the library this afternoon.

5. Cairncross, Arthur. Planning In Wartime: Aircraft Production in Britain, Germany, and the United States. London: Macmillan, 1991.


  1. One interesting difference between Oregon and New England is Chinook Jargon, a simplified trade language created by Indians but widely learned by the settlers: nothing comparable is attested in New England (much less California). That should tie into the question of "how "white" settler society established itself on the far side of the coastal ranges in these years."

  2. Lameen, your link is broken. Here's the Wikipedia article, with shoutouts to the Royal Engineers* and to the Chinook Wawa Wikipedia. http://en.wikipedia.org/wiki/Chinook_Jargon

    *God wished to be a creole aristocracy, and lo! it was done.

  3. Yako yiem halo kliminawhit.

    This is a true story.

    Waum illahie klip sun, kopa Byrne Oakut,

    On a late summer evening on Byrne Road,

    kimta tenas wahm snass chako,

    after a gentle summer rain,

    spose hyack cooley konamokst chikchik, Ford pe Chevrolet, 

    in a race between a Ford and a Chevrolet,

    spose Ford tolo kopa tenas-sitkim mile

    if the Ford won on the quarter mile

    pe Ford man mamook klahwa,

    and the Ford guy slowed down

    kopet cooley, yaka halo mamook y

    soon enough to avoid going airborne

    oakut opoots, 

    at the end of the road,

    Ford, yaka skookum chikchik.

    then the Ford was the skookum car

    --From Terry Glavin and Charles Lillard, A Voice Great Within Us.


  4. I'm not sure - if I remember my aircraft performance modules correctly, estimating climb performance and aircraft ceiling's quite a complicated process - but I'd not have thought that the Davis wing itself would have been responsible for the B-24's inferior ceiling; the weight issue (and possibly the engines?) would seem to be the culprits. IIRC the B-24 and B-17 had a similar cruise L/D (I think this is in Loftin's _Quest for Performance_).

    I read that U.Washington page as suggesting that the Davis wing's exceptional experimental performance was probably a wind tunnel artifact (I'm guessing due to Reynolds number effects that didn't show up on real wings with manufacturing defects and surface roughness), but if Loftin's right it shouldn't have been notably inferior aerodynamically to existing sections either.

    (Also, I may be misreading your note 5, but I wouldn't have described Cairncross as particularly out-of-touch; I don't recall _Planning in Wartime_ being very enthusiastic about mass production, or indeed any part of the planning process...)

  5. Hey, Jakob!

    I should really clarify to say that I don't think that the B-24's problems came from the Davis wing. Overweight is, of course, an issue. I'm probably at risk of over-generalising at the expense of poor Ford, but it was mainly Willow Run-produced B-24s that were critically overweight. Some later production series from Consolidated San Diego were quite good, but the high altitude performance issues remained.

    It is my understanding --intuition, really-- that this is what you have to expect from any design that sacrifices wing area to performance. What looks like drag at low altitudes and near-sea level pressure becomes lift at higher altitudes. It doesn't help that what we call a "laminar" wing profile is just a design that seeks to retard the break down of "laminar" air flow across its surface to turbulent. This happens, at least in part, because the speed of the air over the surface reaches the transsonic range (Mach 0.8--1.2), and this will occur closer to the wing edge at high altitudes.

    As for Cairncross, it's quite possible that I've maligned him unfairly. I should really look into that....

  6. I'd point out that the rate at which the technology was changing and at which bugfixes and improvements were being generated is only really comparable to either a big software project or else a Japanese kaizen engineering project. Agile (in the software world) and quality-through-lean-production are all about keeping the runs short and making a virtue of implementing change requests quickly.

    It would be ironic if the UK had actually invented lean production in the second world war, and then spent the next 40 years trying to make our industries more like the Americans' by upscaling and bashing the craft unions at any opportunity. Wir siegen uns tot!

  7. Erik: again from memory (ie the book's all the way over there somewhere...) the point about early laminar flow wings is not so much that they have higher peak speeds but that the point of maximum thickness is further back along the chord, thereby maintaining a favourable pressure gradient for longer, giving more laminar flow.

    I'd also have thought that the B-24's high aspect ratio would have gone some way towards counteracting the higher wing loading, as it would have generated lower induced drag at higher lift coefficients (this is the basis of the U-2s design, though this also has a low wing loading), but perhaps the draggy fuselage outweighed these benefits.

    Alex: the lean production point is one that is I think made by Jonathan Zeitlin in some of his work on post-war British industrial efficiency; I'll see if I can dig out my notes somewhere.

    1. I can't find the precise Zeitlin reference, but I think it's "Between flexibility and mass production: strategic ambiguity and selective adaptation in the British engineering industry, 1840–1914" in _World of Possibilities: Flexibility and Mass Production in Western Industrialization_ (eds. Sabel and Zeitlin)

      Otherwise it's this piece, Zeitlin's "Americanizing British Engineering? Strategic Debate, Selective Adaptation, and Hybrid Innovation in Post-War Reconstruction, 1945–1960" in Zeitlin and Herrigel _Americanization and Its Limits_

  8. This comment has been removed by a blog administrator.