Tuesday, August 13, 2013

Postblogging 1939, July, 1, Technical Appendix: The Sky's the Limit

Not the Nemean Lion: A Napier Sea Lion, posted at Enginehistory.org by New Zealand's own Bill Bishop

And just by way of contrast,

The Sabre has appeared on these pages before. That's because it's arguably the most complicated production piston aereoengine of World War II. (Talk about running prototypes, though, and the sky's the limit.) Although I've already thrown up this contrast once before in this series, it is still worth a reflective pause to think about the company that was still representing itself to the world through the Napier Sea Lion, above, while it was running a prototype Sabre in its workshop. Also, for reasons that I will not get into, because I'm a little ashamed at how slow has been my progress, I need to acknowledge the assistance of Dr. Stephen Miller of Berkeley, and he has asked for a plug for the effort to revive the Nemean Games at Athens.

Here it is:

Now it's time to talk about planes and engines.

First, though, apologies to the German aviation industry, which did, indeed, send a model of the Do 215 to the Brussels Air Show, and to C. G. Grey, who blew the secrecy off whichever combustion turbine engine appears in the cartoon reproduced here in August, not July. So he let a month go by between letting slip about the Youngman flap here.

Some other booboos crept in. You'll notice that I have my narrator comment on the novelty of the "mature economy" thesis two posts running. That's because I have a bee in my bonnet about the correlation between falling birth rates and the current economic slowdown, and find it fascinating to see the same argument playing out in 1939, with an incipient baby boom set to render it obsolete. But why the baby boom? Blah blah wank wank, more later.

And, more importantly, I completely missed the gonzo way that magic aeroplanes took over the stage at Paris and Brussels. by way of introduction, here's uhm, yeah:

I haven't the foggiest idea how to credit this image, which was uploaded by a suspended user to a "mystery planes" thread at RC.groups*

Some old-timey newspaper serial comic pilot is about to takeoff in an actual Payen Flechair. Or, rather, a development, since that looks like more metal than two 100hp Samsons could pull through the air. As near as I can tell, the Flechair was purely a paper airplane flitting through the aviation press ether at the end of 1938, at least before the modelling community got to it, and as for the idea that it was proposed as a (pulse-jet plane before being reinvented as an almost equally premature tandem prop, well, your guess about the accuracy of that is as good as mine.

To get a sense of the milieu out of which the Flechair comes, you can look at Lieutenant Colonel Kirkland's long-ago defence of the performance of the Armee de l'Air in 1940, which I really will return to at some point. In the mean time, I snip:

"Modern French day bombers included the 307mph Lioré et Olivier LeO 451 (18 squadrons, 392 sorties, 98 losses), the 298-mph Amiot 354 (4 squadrons partially equipped, 48 losses), and the 304-mph Breguet 693 (10 squadrons, 484 sorties, 47 losses)."

Kirkland's point is that the French had lots of modern bombers, which is a bit overargued, but not my point today. I could add the Bloch 174 and have an interesting discussion about why this particular multiplace de combat ends up being described as a reconnaissance type. (Hint: Robert Farley is wrong, not that he cares what I think.) The actual point is that the French were very late in reaping the rewards of an aviation defence spending boom, but at least that meant that the cutting edge vapourware and near vapourware that they were showing off at the Paris Exhibition of December 1938 and on into the summer of 1939 reflected reasonably cutting-edge thinking. Months before the RAF finally got around to thinking about a high-altitude bomber version of the Wellington, the Mark VI, which spurred the initial development of the Rolls-Royce Merlin 60 with its epochal two-stage supercharger (epochal in the sense that it was a two-stage supercharger that actually worked), the French were working on a specifically high-speed, high altitude strategic bomber, the Amiot 354. 

As premature dreams of 1940 go, the Amiot does not rank very high. High altitude penetration did not really pan out even in the post-nuclear age. As a means of conventional bombing, it is just goofy. One can see its attractiveness, though. The specification takes technology to the place where it wants to go, anyway: stratoliners. I link to the Boeing 307, although I could also go to its more successful rival, the Constellation, which had a long and painful gestation going back into the 1930s as well. The point, of course, is that high efficiency stratospheric cruising was clearly the solution to achieving transatlantic ranges, and technological progress across a whole range of aeroengine ancillaries combined to deliver a practical piston engine stratoliner at just about the same moment that parallel developments of the same technology delivered the first viable jet (oops) (and turboprop) airliners and rendered complex monstrosities like the Sabre all-but dead on arrival.

Yet real it back in a bit, and look at those technologies, and we can see glimmers of an actual, usable, militant future. Specifically, a typical runaway wartime fighter performance race. Practically all of the elements that will be seen on the battlefields of World War II are there in embryo at Paris or at Brussels. This will not come across as anything but raving egotism, but it seems to me that if I need to be reminded of the extravagances of three-speed superchargers driven by auxiliary engines and the like, then the world, too, deserves to be reminded. Reading John Terraine's long-ago Right of the Line, readers may be impressed by a block quote intimated to be from an RAE assessment of the Bf109E done some time after the capture of one such aircraft in March 1940, lifted from one William Green's 1970 Aircraft of the Third Reich:

"...[I]nsofar as the Spitfire was concerned, when fitted with the two-pitch airscrew --and at that stage of the war, virtually all Spitfires were fitted with such airscrews, as priority in the supply of constant-speed units had been allocated to bombers --this was bested from virtually every aspect of the Bf109E-3. . ."[165] 

The point here is that the Germans were better than the RAF, and Hugh Dowding totally won the Battle of Britain, you guys, despite the Air Council being totally gaga.

In fact, constant speed units were not being issued as priority to bombers. The Air Ministry had only approved CSU airscrews for bomber use for technical reasons. Fiddling with your airscrew control is not what you want to be doing in a dogfight, and the new, completely automatic, fighter-suitable Rotol airscrew was being phased in along with a number of other improvements, most notably a much faster starter unit, on the Spitfire II, with which Fighter Command re-equipped in the second half of 1940 and first quarter of 1941. Although by that time Rotols had been refitted to all British planes in a crash programme in the early summer of 1940. An exploration of the issues behind this delayed implementation (specifically, weight, root fairing and power solidity) will suggest why the two Air Ministries might have taken different tacks without resorting to "RAF lurves strategic bombing LMAO."

Leutnant Erich Bohendiek, quoted here, by  Mike Williams and Neil Spurling, who are very shy about putting their names to their website gives us a brief account of what it is like to be flying behind a fighter engine with a CSU unit that isn't completely automatic. 

I was not flying my usual plane but, as I was the Technischer Offizier, I had to fly a plane with a new automatic propeller just to test it. That was my bad luck, having that bloody plane on that day for the first time because that 'automatic thing' turned that angle of the propeller so that an average speed was always maintained and not a kmh more! That meant trouble when starting and trouble at high altitude as the plane was nearly always unmanoeuvrable and swaggered through the air like a pregnant duck.

To be sure, Daimler Benz and VDM got together to fix this pretty quickly. It's just that it wasn't fixed until it was, and until it was, there's an argument for not putting CSUs on fighters. 

But why am I arguing with John Terraine? He's not listening, and he's dead, and I'm just quote-grabbing. Most of the people who object to this extract note that it compares two aircraft flying on 87 octane, while Spitfires were by this time rated on 100 octane. The comparison, whether it was realised in the spring of 1940 or not, would be endemic. The Germans could not afford synthetically reformed avgas, because synthetic reformation reduces avgas output (although you do get more road tar!), and the Germans did not feel that they could afford to cut their avgas output.

But here, at last, is my point. On the one hand, we have the incipient dream of stratospheric flying. On the other, we have three apparently disparate technologies: the adjustable pitch airscrew, supercharger, and high octane test gas. (I could throw in various incarnations of the slotted flap, but they don't go on engines, unless you're some crazy inventor who wants to put them on airscrews. Which such people existed, but I think that the Flechair will be enough chasing after aviation innovation craziness for one day.)

Yet, in fact, the three disparate technologies go together, each bootstrapping the other towards the goal of, again, stratospheric flying. Superchargers heat up the fuel charge. Therefore, the higher the octane rating, the higher the altitude at which a supercharger can give full power. But highest speed and lowest speed are fixed to a range. The faster your airliner cruises, the faster, and more dangerous, the landings. So if you want to extend your flight envelope, you need aerodynamic innovation, like CSUs. But now you have a system piggybacking on your engine that needs an automatic control input from an engine that has a supercharger that profoundly changes the engine state. Put another way, you need an airscrew control unit that can respond to supercharger boost increase by adjusting the airscrew pitch in such a way that the screw can absorb the increased power, and do so in a smooth, invisible way. As opposed to lurching through the sky like some under-engineered monstrosity.

Speaking of which, the reason that the Bf109E3 was faster than the Spitfires that it was flown against in March is that it was a smaller, lighter plane. The reason that the Spitfire was bigger was that it had a bigger wing, to give a lower landing speed. The only reason that the Bf109 hadn't an even bigger advantage was that it was so sloppily designed. (Seriously, check out the externally braced tailplane!) The Spitfire's big wing was one solution to extending the speed range. A more extreme one was the afore-mentioned slotted flap in its various incarnations, which introduce new performance curves into which you have to fit your supercharger and your airscrew all over again. What happens in the CSU when you hit maximum boost near the ground with your slotted flaps fully extended? This is starting to get complicated!

So what seems like a simple expedient for improving fighter performance turns out to be linked to, and dependent on, a range of other developments in engine and airframe technology. 

So all of this was going on in the spring of 1939. The approximate goal was the Fairey FC-1, a ludicrously obscure plane on account of it never being built.  (Though a French rival, delayed by the war, did appear in 1945 and probably gives us a look into what the FC-1 might have been like. Unfortunately, I can't find the damn thing on Wikipedia today, and have given up looking. I could have sworn it was a Latecoere product...) 

Instead, we got a developmental convergence on fasterer fighter planes. It is quite likely that, had WWII never happened, there would have been no Spitfire IX, no DB605, no Merlin 100+ series and no late-series Bfs. These were monstrous machines. Fighter development would have run through twin-engine types and Sabe-powered freaks, which, of course, were monstrous too, but monstrous in a single-silo way, in that it was left to Napier to come up with the Sabre in its full craziness, and not to an intersecting research and development community that has to run from wind tunnel to De Havilland to Rolls Royce to RAE to Imperial Chemicals and back again and again every time someone comes up with a new tweak. The next step in the evolution of flight would have been defined by an Auckland-Vancouver Empire Air Mail service.  Or, rather, because that would have been flown by the G boats, by high speed Atlantic flying by stratospheric Short airliners. 

It is also very much worth wondering whether we would have been slower to get to computers had control not emerged as the key limiting factor on performance.

Or, at least, that's the kind of story I would be inclined to tell if someone were to give me a big cheque to write a counterfactual history about what would have happened if WWII didn't happen, or were delayed a few years. Anyone? Damn.

*This blog does not endorse smoking in hangars.


  1. You're going to love Boris Chertok's memoir of the Soviet space programme, Rockets and People. All four volumes of it! and completely free from the NASA History Project.

    Especially as a major theme is how Chertok became a cybernetics/control systems engineer without ever intending to do so, and how the controls guys ended up heading the effort to reconstruct the V2 and exploit the German industry behind it because a) nobody else realised space rockets/missiles were a thing rather than engines, warheads, fuel, or controls separately and b) there wasn't a People's Commissariat of Control Systems, so unlike everyone else from the Ministry of Aviation, the Armaments Commission, the Artillery Corps, the Guards Rocket Mortar Force, the NKVD, SMERSH, and the People's State Commission of Applied Tree Climbing, his gang were able to make it up as they went along rather than referring to Berlin or Moscow, by sheer Amtsanmaßung and cheek.

    Also, big priority: lay hands on some of those schweeet German kreiselgeräte and messtechnik of all kinds, making good and goddamn sure someone you trust implicitly personally delivers it all to your lab back in Moscow...

  2. The whole "control systems theory" thing of the late 40s is so weird. You go into it with the basic problem that feedbacks are mucking up the partial differential equation that is supposed to describe the induced stable state.

    Or, as a simple analogy, the automatic steering device takes a heading from the compass. When a wave wacks the rudder, a little motor pushes it back to the proper heading. The force being proportional to the whack. Which is code for the simple harmonic equation, one of three partial differential equations that we know how to solve analytically.

    So, when feedback gets in there, suddenly we can't solve the system analytically. We fall back on numerical solutions. This is pretty well developed field of applied math, and has as its mechanical or electrical analog the insertion of other circuit elements, mostly resistors, into the loop.

    So we're doing math to figure out how big the resistor should be. It's approximate and iterative. The more calculations we do, the better the fit. So we say to ourselves, "Hey, why don't we automate this calculation?" We end up with a simulation of the original system that runs again and again, each time with a resistor-analogue element set to a different value.

    Vannnevar Bush built one of these in the 30s, to simulate the long-distance transmission lines bringing hydroelectric power down to New York. But there's not many problems that justify building a simulation. And it's just a math problem! So why not build a universal simulator?

    Okay, sure, 'universal simulator.' Impossible, but a cool idea. But wait! We're already in the realm of approximation here. One way of looking at the math is as just a series of executable instructions: yes/no. And there's these gadgets around that do that....

    The upshot is that the problem of solving complex analog problems leads to the widespread use of digital devices. Somehow, we went into it as systems engineers and came out of it as computer scientists.

    1. A bit more Chertovblogging; the Soviets had a political split about it! Moiseyev at the Air Force Academy, Zhukovsky has his theory of technical stability, and the support of the powers that be, but unfortunately this doesn't actually work! Whereas the NII-88..then NII-885...then Korolev's OKB-1 (not the earlier OKB-1, the other one) gang have been reading this stuff from MIT! Which does work, but is heretical! Fortunately everyone involved is Jewish and therefore already banned from becoming too prominent, so nobody got shot in the head!

      Also, the aviation industry people discover that the artillery industry insists on technical drawings being detailed enough that series production can go ahead without ever consulting the original design team, while the aviation people tend to both change everything all the time, and also to rely on craftsmen sucking their teeth and if-it-looks-right-it-is-right a lot.

      And they keep drinking the rocket fuel. 70% ethanol coloured blue with manganese crystals.

  3. Next time I take a volume of the MIT Radar Handbook out, I'm going to have to get myself some cheap, blue vodka to go with it.