The above drawing (click to enlarge), which appeared in the 3 May 1934 issue of Flight, depicts an ingenious bombing simulator manufactured by Vickers-Armstrongs -- the Vickers-Bygrave Bombing Teacher. The basic idea is that an image of the area around a bomb target (which is printed on a glass plate) is projected onto the floor, scrolling along to represent the flight of the simulated aeroplane at 8000 or 9000 ft. The bomb aimer peers down at the image through a bomb sight, and sends course corrections to the pilot, who alters the flight path in response. An electro-mechanical linkage then moves the glass plate accordingly.
When the pupil has calculated the direction and force of the wind and has sighted on the target, he throws a switch which represents the bomb release. A device times an interval, equal to the time taken by the bomb to reach the ground, and at the end of this period the movement of the "ground" is stopped. Painted on the floor is a fixed "trail point," which marks the point on which a correctly aimed bomb should drop. Any error may be seen by the difference in the position of the "target" and this fixed trail point.1
Something very similiar seems to have been used by RAF Operational Training Units during the Second World War, though they were then called Air Ministry Bombing Teachers. (Presumably the Air Ministry's in-house version, perhaps improved over the Vickers-Bygrave.) Many former wartime airfields still have their distinctive two-story bombing teacher buildings, for example this one at Waltham. But I don't know how widely such devices were used before the war -- though 601 (Bomber) Squadron of the Auxiliary Air Force, at least, had one in early 1934, according to the article. Given the poor performance of Bomber Command in the early years of the war, one would think that the RAF could certainly have used a few more bombing teachers!
I was thinking that a few bells and whistles could have increased the realism of the Vickers-Bygrave dramatically. For example, dry ice could be used to simulate clouds over the target. You could use a negative, with most of the features painted over, to imitate night bombing. Hydraulics (or manpower!) could be used to buffet the airframe, as in turbulence or anti-aircraft fire (a few firecrackers could help with that too). Not surprisingly, I wasn't the first to have this idea. This interesting site on the history of flight simulation has a page on the Celestial Navigation Trainer (CNT), developed at the RAF's request by the makers of the Link Trainer. Though no mention is made of the Vickers-Bygrave, it's clearly a very similar concept, with the addition of what is effectively a planetarium above, so that the navigator could practice celestial navigation. According to the RAAF (which had one at East Sale), 'The CNT instructor could introduce bumpy flying conditions, changes of wind, create daylight or nightfall, scurry clouds across the sky, or arrange static to worry the wireless operator'. But development of the CNT was initiated as late as 1939, and the first one didn't come into operation until 1941 or later. (The RAAF's remained in operation until the late 1950s, so it must have been very useful.)
Perhaps it's because, as a Gen Xer, I grew up with simulations in a way that previous generations did not, but it seems incredible to me that it took five years or more to take the basic concept of the Vickers-Bygrave and add substantial degrees of realism to it. (Well, I can't completely exclude the possible that this happened sooner, but I have no evidence for that as yet.) Then again, one of the dangers of simulation is that it can reinforce preconceptions, rather than challenge them: to a large degree simulations simulate what is thought will happen, rather than what will actually happen. In other words, garbage in, garbage out. So, maybe the failure to develop a Celestial Navigation Trainer before 1939 is of a piece with the failure to practice bombing runs under warlike conditions in the same period, and the failure to set up a Bombing Development Unit before the start of the war. If bombing is thought to be easy, then there's no need to train too hard for it. Wartime experience was, of course, the ultimate bombing teacher.
See also: this American bombing teacher from 1940, with that wondrous war-winning Norden bombsight fortunately shrouded from public view.
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
Permissions beyond the scope of this license may be available at http://airminded.org/copyright/.
- 'Bombing instruction', Flight, 3 May 1934, 434. The drawing is on the facing page, 435. [↩]
Gavin Robinson
I live just a few miles from Waltham and I had no idea about the bombing teacher building! I assume it must be on the other side away from the main road. I think the runway's still there, and you can see a hangar from the road but I'm not sure whether it's original. They used to fly crop dusters from there in the 80s, but now it's just people walking dogs or practising driving.
There could be a whole other PhD thesis waiting to be written about the origins and development of simulation. It probably does have a lot to do with institutional culture as you said, but maybe it's also because it wasn't until the mid-20th century that there was any technology which needed to be simulated in order to learn how to use it effectively.
Kurt Niehaus
Gavin, Good point about the technology not needing simulation until recently.
I know there are "combat simulators" in the martial arts-- a wood pole with paddles.
What about the models the Japanese drew up for bombing Pearl Harbor?
(http://www.history.navy.mil/photos/events/wwii-pac/pearlhbr/ph-ja5.htm at the bottom)
Generally, were the plates supposed to be theoretical targets (like Berlin) or just random practice?
Chris's monkeys
The Link Trainer was a 1930s innovation., wasn't it?
A simulator is just a control room with an artificial set of inputs. I really need to get cracking on the history of the control room before some bright young thing beats me to it.
Another data point: by the early 1940s, the RN was using a two-storey building to simulate submarine control rooms for officers doing the 'perisher'. Periscope at the bottom, tracking table on the top, with sillhouettes, etc. Ref is Edward Young _One of Our Submarines_, I think.
Brett Holman
Post authorI'm glad that I can continue to help you learn about your own area, Gavin! You're probably right about the lack of technologies needing to be simulated. I can think of things which could have been simulated, like a ship's bridge, or a powerplant control room, or naval gunnery, etc, but I don't know that they were, probably learning on the job and maybe some classroom instruction was deemed sufficient. I guess there's not much point in simulation unless it's easier or safer than the real thing. As Kurt says there were military analogues, and of course a military exercise or a wargame is just a simulation on a larger scale. (Eg, taking Chris's point, a RAF sector ops room in an air defence exercise is receiving inputs which mimic the real thing very closely, so it's essentially the same thing as a simulated ops room. But were there also mock ops rooms for training purposes?)
Yes, you should write that history, Chris :) Link Trainers date to 1930, though there were precursors going back to ca. 1910, as the site I linked to shows. But it's also interesting that the Link Trainer was not originally intended as a simulator, but as a fairground amusement! So it wasn't created to solve a perceived problem in flight training.
Kurt: I don't have any information on whether real or generic targets were used for the plates. Would be interesting to know (I suppose generic in peacetime!)
Chris Williams
Ashmore was using simulated inputs and practice interceptions to develop LADA in 1918. I think that he used simulated inputs on their own as well, but I'm not sure and I can't find my notes right now.
I haven't got any evidence of the railways simulating inputs or doing dry runs when they introduced central control. But what about gunners practising indirect fire? I think that yet again we're being let down by the lack of a decent technical history of the Royal Artillery. Dan?
PS - sorry about the obsolete Monkey gag above. I left my facetiousness button down by mistake.
Brett Holman
Post authorYes, I wondered about railways too ... when was central control introduced? My father used to work as a train controller and I well remember the boards showing the lines, trains and signals. (And the manual plotting of the data on graph paper!) Even though the technology being controlled and used for control was not that sophisticated, the system/network as a whole was. And lives (and profits) depended upon it working. So simulation might have been useful for train control, which of course doesn't mean anyone thought to do it. But once you've abstracted the real world into the switches and dials of a control room (or a cockpit, for that matter), it does seem a natural step to remove the real world and introduce artificial inputs for training, as well as testing.
One thought I had about gunnery is that perhaps it is too mechanical to benefit from simulation much? You get orders to fire on a certain map location, so you do the number-crunching -- simple ballistic trajectory, correct for air resistance, wind speed, barrel warp, projectile spin, whatever -- and spit out the required altitude and azimuth for the gun. Load, fire. There's not much room for judgement and so playing out scenarios in a simulator might not be very useful. Or am I just showing my total ignorance of how artillery fire actually works in practice?
Chris Williams
Central control in the UK - which I think was a significantly speeded-up version of the US dispatching system - came in in 1909, midwifed by Cecil Paget of the Midland Railway. In April 1915, Major Paget was appointed to head the BEF's Railway Operating Division. But don't forget, children, that British generals in WW1 were REALLY STUPID and never learned anything.
For artillery, what we need is a war and society expert who knows WW1 technical stuff and used to be a gunner. Hang on:
http://www.imperialservices.org.uk/
Alex
I've always thought the RAF's sector control system was essentially a big computer, not to mention a significant example of decentral decision making - yeah, there was 11 Group HQ, and even Dowding's own FCHQ, but the really important bit was the sector level.
I recall reading, when Network Rail was set up, an interview with Ian McAllister, the Ford executive who was hired to act as chairman, in which he mentioned that he wanted to automate the timetabling process which currently relied on "two PhDs in mathematics". This was in 2002!
Gavin Robinson
I think timetabling is one of those intractable "salesman's problems" that has no perfect solution which can be arrived at by algorithms.
I was thinking about simulation yesterday as I sat in the British Library waiting for my manuscripts to appear, and railways did cross my mind, but I hadn't thought about signalling or timetabling. I was just wondering how engine drivers were trained in the 19th or early 20th century. Steam trains might have been an early example of technology which needed to be simulated and could be simulated, but I don't know whether they actually were simulated.
On reflection my earlier point was an oversimplification: simulation does depend on having systems complex enough to need simulating, but also depends on the systems being simple enough that they can be accurately simulated, and on the machinery being fragile and/or valuable enough to make training mistakes too costly, or mistakes being too dangerous to the trainee and instructor. Horses might meet some of those criteria, but there's still no adequate way of simulating riding a horse. In practice the only way to minimise the costs and risks is to start on more docile and less valuable horses where mistakes don't matter so much, and then gradually work up.
Chris Williams
yr engine driver in late C19th UK practice began as a locomotive cleaner, then a fireman, then a shunter (driver), only then was he let out on the main line, and he had an inspector with him each time he drove a route for the first time. Inspectors were senior drivers. Lots of rail crew training was done via mutual improvement classes.
Source for most of the above is Adrian Vaughan's _Grit, Grime, and Glory_
Dan Todman
Sorry gang, I've joined this party a bit late. Seems to me that we've distinguished between training men and testing systems - so I'd be interested to learn at what stage in training the bomb trainer was used - but we perhaps need to go a bit further. A simulation used as training might usefully be simplified for all sorts of reasons to begin with - not least because making it as complex as reality might completely demoralise the poor trainee. But perhaps (and here I bow to Chris' expertise and that of his monkeys) when you're testing a system, you want to find out whether it works or not, as well as improve it.
As far as British artillery goes - I think that Jonathan Bailey has written that book, but perhaps you want something _really technical_, in which case I'm definitely not your man. But my impression is that whilst systems for controlling gunnery got a lot better in WWI, there wasn't that much time to test those systems out of the line, for the reason that gunners were in action for longer: so developments were I suspect more organic. Probably worth looking at what happens over the winter of 1917-18 however: for all that our presumption is that modern armies didn't move into barracks after the campaigning system, that seemed to be when GHQ underwent the biggest changes in structure.
To go back to some earlier comments - I'm rather wistful for that moment that didn't happen in the 1930s when the RAF, realising its need to combine with the RN, and the RN, realising its need to cooperate with the RAF, practised bombing German battleships and sinking them with submarines, all in the same 3 storey building (I know, I know, scale etc. Whatever - that _would_ have been a landmark to British defence planning between the wars).
Brett Holman
Post authorWhat about after WWI, though -- how were artillerymen trained for indirect fire after the dust had settled? I suppose firing ranges were common enough.
That would indeed have been very impressive, had they done that! For that matter, both ASDIC and RDF would be examples of things which could have been simulated, but probably weren't. As far as I can tell, it wasn't done for RDF, at least by the early war period. It may well have been misleading to use simulation, anyway, given that there wasn't yet a thorough understanding of all the ins and outs of radio propagation and reflection.
Chris Williams
I'm just reading John Bushby's _Air Defence of Great Britain_ . As well as knocking my thesis about Ashmore's centrality into a cocked hat (bugger) it also mentions that in the early 1940s. Ground Control Approach fighter controllers were trained in Wemby Stadium, using two ice-cream tricycles, each with a radio, and the 'controller' sat in the press box. The pitch did duty as a PPI. Metronomes 'governed' the speeds of the trikes.
I've read elsewhere that CAMship fighter controllers and pilots were also trained in the same way, but the location was less auspicious: a croquest lawn in a country house somewhere.
Brett Holman
Post authorLOL. Reminds me of that scene from Dark Blue World where all the Czech pilots are on their bicycles learning how to intercept Nazis.
I must confess I'm not familiar with Bushby. Any good?
Chris's monkeys
Hard to say, cos none of it is sodding referenced. Bushby was a fighter controller in the 1950s, so there's a lot of RAF backroom tacit knowledge in there, though, and he's good on the technical side. The question is, which bits are reliable?
Alex
Chris, that Wembley thing is ridiculously cool.
Chris Williams
p. 146.
Not only that, but according to the story, the trikes still had 'Stop me and buy one' painted on them, which might be a gag too far. A little bit of pink and yellow paint could have given them a far more apposite slogan for the duration: 'Stop me or buy one'.
Alex
I like the retro-modernist tang of it - very J.B. Priestley.
Pingback:
Airminded · MONIAC and the warfare state
Martin Bolton
I know this is a rather late comment, but I just noticed the discussion on the Bombing Teacher. It was developed at the Air Ministry Laboratory (hence "AML") at Imperial College, and is described in British Patent Specification 289,267, filed July 28, 1923. The inventor was Leonard Charles Bygrave (1892-1935), and the manufacturer Vickers (hence "Vickers-Bygrave"). The Air Ministry Laboratory developed instrumentation and bombsights. Bygrave was killed in a riding accident. He was at the time of his death Head of the Instrumentation Physics Dept at the RAE. The AML Teacher was listed as one of 16 "synthetic training" devices in use in the RAF in 1940 (as recorded in the Minutes of the RAF Synthetic Training Committee).
The origins of the Link Celestial Navigation Trainer were in the USA (designer P.V.H.Weems). It was ordered from Link by the RAF at the outbreak of the war, but very few were installed in the UK due to the high demands on station infrastructure and installation manpower this trainer made. Its impact on training in the UK was thus very small.
Brett Holman
Post authorMartin:
Late comments are always welcome, especially when they are so informative! Interesting that the patent was filed in 1923, over a decade before the article discussed in the post. So bombing teachers could have been used even earlier. But it looks like the Synthetic Training Committee you mention wasn't formed until May 1940 (according to Tee Emm, April 1941, 3), so a systematic interest in training simulators might have developed quite late?
Yes, the Link Trainer was American. But while it may not have been used in great numbers in Britain itself, it was used quite a bit in the Empire Air Training Scheme (change name to suit nationality). The RAAF, for example, had about 210 of them. (I also see from that link that it had 3 Celestial Navigation Trainers.) The RCAF used them as a standard part of the pilot selection procedure. So many of the large numbers of pilots who trained in the Dominions would have had some time on Links before coming to Britain and Bomber Command.
Martin Bolton
The Link device I was referring to was not the Link Trainer, which as is well known was used in very large numbers from 1936 on, but the Celestial Navigation Trainer. At the time the Synthetic Training Committee first met on March 11, 1940, the Link Trainer was an established component of RAF pilot training. At this time the Celestial Navigation Trainer had been ordered by the RAF, but the first one had not yet been delivered.
JDK
Martin's comment (going round the misunderstanding) on the Celestial trainer's rarity and thus lack of effect in training is a good one; and there's a chicken & egg thing regarding the arguable over-emphasis on pilot training at the cost of other aircrew tasks in pilot driven air forces.
Two interesting references on this kind of training are "Man is Not Lost: The Log of a Pioneer RAF Pilot/navigator, 1933-1946", by Frederick Charles 'Dickie' Richardson, Airlife, 1997, and a recent podcast by the RAeS on the history of synthetic flight training by Dick Eastment OBE ( http://aerosociety.com/news/Podcast/1505/A-Journey-Through-Flight-Simulation ). Both look at gaps and shortfalls, as well as the cool kit thing.
Brett Holman
Post authorMartin:
Sorry, my mistake!
JDK:
I think I read somewhere that some squadrons used worn-out, non-flying fuselages for training. I'm not sure what that entailed but perhaps it might have been better for other aspects of training?
JDK
Would need more detail on that, Brett, but the use of knackered kit rather than bespoke is a factor in itself.
It's an interesting item to note that most air forces of the 1930s recruited for pilots alone and the other roles were filled by wash-outs from that stream. Not the best way of getting the best people with different aptitudes such as those required for roles such as navigation.
The argument goes that the pilot-rejects were already of a high standard, and that's true, but it also perpetuated the pilot-driven air force and the elite pilot command mindset.
Perhaps there's a parallel universe where it's a navigator's air force?
Errol Cavit
Wasn't the navigator's (observer's) air force the Fleet Air Arm? Being able to find the airfield again is obviously something you can't trust a mere pilot to do!
BTW, some USN types like the Avenger had a sliding map table for the pilot.
Brett Holman
Post authorErrol:
Or the Luftwaffe, at least up to the early years of the Second World War: the observer was usually in charge (and senior in rank), not the pilot. Though the observers seem to have received flight training too.
Errolwi
Brett:
That's right, that fact had slipped my mind. Still done with e.g. Maritime Patrol Aircraft in many Western air forces.
JDK
My thought was ore the 'it's all about the pilots' problem we have in public perception of aviation and military doctrine.
Anyway, I did some digging which indirectly shows the secondary status of observers - 'Observer', is both a role and a qualification, not necessarily overlapping. In W.W.I the first role of all air arms was to put an observer in a vantage point.
RFC/Br Army & Australian Flying Corps had Observer' as a qualification with the single wing 'O' brevet awarded. (It was known to its wearers as 'the flying arsehole'.)
The RNAS added 'Observer' to the officer's rank, and the role was, in a sense, a continuation of the RN gunnery task.
Dropped in the 1920s, re-introduced in 1937 with dedicated non-pilot roles being required in newer, more complex types. Replaced with the Navigator 'N' brevet in 1942, but the more 'multi-skilled' and senior Observers fiercely proud of their brevet generally retained it.
The Fleet Air Arm was properly the FAA of the RAF from 1924 until returned to the RN in 1939, so the observer role was both naval and air force in that period, and the RN FAA still has the role of observer, as do police force helicopter arms and the like.
There are claims that the British Army role of observer dates to the C19 kite and balloon era.
The German Navy had their observers as the aircraft commander in W.W.I, and that status existed in various forces up and into W.W.II, but I don't think beyond, when the cult of pilots became the orthodoxy. [~Cof~]
(Quick research, mainly via Ken Delve's 'Source Book of the RAF'.)
JDK
Regarding the simulation whys and wherefores, there was a dramatic expansion of such things in W.W.II due to the dramatic expansion of training on dramatically new/complex machines (~ of course). I'm saying that because of watching a film 'Pan Am Clippers at War' last night (thanks Airways Museum) which had a segment on training flight engineers for the USAAF & USN with a fuel & oil systems ground trainer, which 'saved time and money by not using the real thing in flight', as (it is implied) that was how it was done before.
Brett Holman
Post authorJDK:
That's an interesting way to look at it. Here we've been focusing on simulation in terms of greater realism, as being able to more closely approach combat situations without having to actually be in combat. But that assumes that air forces felt there was a lack of realism, or rather any significant lack of realism, in their training methods. If we think instead in terms of simulation as more of a shortcut, almost as simulating the training process itself rather than operations, then it does explain why it didn't really happen until the war or at least the late 1930s, rather than earlier.
Would there have been any interest in simulation during the other great training expansion, in the First World War? Perhaps not -- aircraft were much simpler in general. Maybe for the expansion of the German Zeppelin fleet? A Zeppelin was much more complicated to command and crew than even the largest aeroplanes; and individual Zeppelins cost much more too. There were training Zeppelins, but there was always going to be risk involved in real operations, so I can imagine that there was drilling in mock Zeppelin gondolas before/between flights, or at least with the airship safely sitting in the hangar.
Phil Vabre
Thanks to JDK for putting me on to this interesting discussion.
Speaking as a person who has had quite a bit to do with simulation one way or another, I think there are a few points to consider that don't seem to have been mentioned yet.
1. A matter that is often not well understood, even by people who are responsible for training systems, is that there is a cost-benefit equation involved in simulation. Generally speaking, the cost of the simulator increases exponentially the closer absolute realism is approached. Few organisations can justify or even afford the highest levels of fidelity and there is always a trade-off to be made. Often it is more cost-effective to use low-fidelity 'simulation' and we can see this in the many part-task training devices mentioned in the posts above.
2. All else being equal, with expensive (high fidelity) simulations/simulators the 'unit cost per person trained' is also obviously much higher when only small numbers of people require training. So it should be no real surprise that the use of more complex simulators increased during the Second World War when large numbers of students justified the investment. Of course, this is not the only reason: required throughput exceeding the available capacity of 'real world' training is another. As is the increasing cost or unacceptably dangerous nature of the real-world environment being simulated.
3. Another issue is the technological capability of the day. It's all very well to talk about simulating, say, 'the RAF bombing German battleships in the same 3 storey building as the RN learning how to drive submarines', but exactly how would you do that? Today we could relatively easily (if not cheaply) do it using computers, but what about the electro-mechanical world of the 1930s? Such a device would have been prohibitively expensive and, if it could have been built at all, probably impossibly unreliable.
When considering simulation, it's easy to get carried away by flashy toys (or lack of them) without stopping to consider what the actual requirement is/was and the most effective way of meeting it. Sometimes the most cost-effective means is not the flashy toy.
Brett Holman
Post authorVery good points, Phil. On the other hand, we now have the benefit of decades of experience with simulation, so we can think about things in those terms. I suspect that the period under discussion-- i.e. when people were just starting to think about how to do this and the costs and benefits involved -- is when the wildest claims would have had freest play. Or rather, that's what I'd expect, since that's a general pattern with the introduction of any new technology (electricity, radio, the Internet, aeroplanes themselves of course). So while I accept what you say about the practical aspects, and that these would probably have prevented any serious adoption of simulations until the Second World War, I still keep returning to the question I posed in the post, about the apparently general lack of interest then in even the possibilities of simulation. We're just so simulation-minded now that it seems an obvious thing to do; it evidently was not obvious then. Compare airmindedness -- people were trying to fly far, far in advance of the technology to actually do so. Flying seems to have been obvious; simulation was not. I think this is interesting.
Neil Datson
Just a couple of points to add to this discussion.
On simulators, as has been observed, the more complete the simulation the more effective it should be; yet some circumstances probably cannot be simulated. For men going into battle you really need to introduce fear. Brett suggests firecrackers etc, but while bangs might be a start they obviously wouldn't cut it. William Armstrong, a bomber pilot in WWI, wrote in his book Pioneer Pilot that neither he nor any of the airmen he knew used their bomb-sights, except to check that the bombs had actually dropped. Years of practising on basic simulators that only trained them to aim accurately (always assuming that the bomb-sights were any good) wouldn't have been a lot of use. The challenge to training bomber crews is so much greater than a rifleman in the front line, as a rifleman's survival depends on shooting enemy soldiers, a bomber crew's only on coming back without their bombs. To really train the bomber crew effectively you have to take out a part of the human element that the rifleman can be allowed to retain, which is probably impossible.
As regards comments kicked off by JDK about pilots and observers, it is my understanding that the German army air forces in WWI put the observers in command. Whether the RNAS started out that way I don't know, but the RN was so desperate for good observers by 1917 that they were even re-training boatswains - who could be ill-spared - as observers. It seems likely that the RFC was unusual in always putting the pilot in control, and regarding observers as little better than cannon fodder. Can anyone comment on the relative status of observers in other WWI air forces, particularly the French?
Brett Holman
Post authorInteresting points, Neil. I wouldn't claim that fire-crackers (etc) would lead to more accurate bombing; we've enough experience with simulators by now to know that there are always limits to what you can do without doing the real thing. As I say in the previous comment I'm really more interested in the development of simulation-mindedness rather than its effectiveness -- but other people can be interested in other things!