Japanese Long-shaft Rarity - the Atsuta 19
In this article we unveil a model engine marque which as far as I can tell has completely escaped previous notice among model engine enthusiasts outside its country of origin. We’ll be discussing the Atsuta glow-plug engines which were manufactured in small numbers in Nagoya, Japan during the early 1950’s.
The very existence of the Atsuta engines never appears to have become apparent to contemporary observers outside Japan - despite an intensive search, I can find no mention of them at all in the English-language modelling media of the day. There’s little doubt that the engines were never marketed outside Japan, although a few may have found their way to other countries through such routes as personal purchases by members of the post-WW2 forces of occupation. Even in Japan itself, the marque never appears to have achieved anything approaching mainstream status, possibly being marketed on a strictly regional basis. For this and other reasons (to be discussed below), the Atsuta engines are extremely rare today even in Japan and are almost never encountered elsewhere.
This being the case, there’s less than usual to say about these rare but interesting engines. It has however been possible to learn a little about them from contacts in Japan, in which regard I’d like to record my indebtedness to my friend Rudy ni Balinia (from whom I obtained this example) as well as Akira Fujimuro, one of the key figures in the post-war Japanese model engine and model aircraft fields. Despite the combination of distance and the language barrier, both of these gentlemen have greatly assisted me in my attempt to learn more about the Atsuta marque.
Even in Japan, detailed information about these engines appears to be rather scanty. Let’s see what we can do to draw aside the veil a little ……
Connotations of the Atsuta Name
The name Atsuta is very closely identified with the City of Nagoya in Japan’s Aichi Prefecture. Like all major Japanese cities, Nagoya is organized into a ward system, with Atsuta-ku (Atsuta ward) being one of the more centrally-located of the 16 wards into which the city is divided. Today it is one of Nagoya’s major commercial and industrial districts.
The Atsuta district has long been associated with the famous Atsuta-jingū (Atsuta Shrine), a major Shinto shrine traditionally believed to have been established over 1,900 years ago during the reign of the Emperor Keikõ (71-130 CE). This is one of the most universally-revered shrines in Japan, being familiarly known as Atsuta-Sama (Venerable Atsuta) or simply as Miya (the Shrine). For this reason alone, the name Atsuta would have been instantly recognizable to any Japanese citizen during the period of which we are speaking. Moreover, it would undoubtedly suggest the City of Nagoya.
Given the instant recognition factor attached to the name, the application of the name Atsuta to the model engines with which we are dealing would immediately imply to any contemporary Japanese citizen that they were manufactured in Nagoya. There was a full-sized precedent for this in the form of the WW2-era AE1A Atsuta aircraft engine, which was a licensed version of the German Daimler-Benz DB 601, a 12-cylinder liquid-cooled inverted-vee aircraft engine. This unit and its derivatives were manufactured between 1938 and 1944 by the Aichi Kokuki KK (Aichi Aircraft Company) at its Atsuta Engine Plant in Nagoya, being named for that location. It was used to power a number of well-known Japanese warplanes during WW2.
Since the Aichi Aircraft Company was dissolved following the conclusion of WW2, it appears that the Atsuta model engines had nothing whatsoever to do with that company or its descendants. The latter-day descendant company of the Aichi Aircraft Company is the Aichi Machine Industry Co. Ltd. (no mention of Atsuta), which manufactures automotive parts and light trucks for Nissan. Nonetheless, the name Atsuta would have retained a strong aviation-related connotation among air-minded Japanese people as of the early 1950’s.
For both of the above reasons, the use of the name Atsuta was entirely appropriate for a model engine range manufactured in Nagoya. It combined the name of a nationally-famous and widely revered location with a very direct connection to Japanese aviation history.
Production of the Atsuta Model Engines
According to the sole advertisement for this engine which my Japanese informants were able to provide (in Japanese, naturally!), the Atsuta engines were developed and manufactured by an organization calling itself the Japanese equivalent of “Atsuta Internal Combustion Engine Research and Development”. It’s interesting to note that the engines were evidently not manufactured in the Atsuta ward for which they were seemingly named but rather in the adjoining Minami ward lying immediately to the south towards the harbour from Atsuta ward. The manufacturer’s address was given as 25 Ichinokiri, Aza Kafuku, Kasadera-cho, Minami-ku, Nagoya. It was the instantly-recognizable connotations of the Atsuta name that clearly lay behind its adoption.
During the years following WW2, the Minami ward developed into a major centre for heavy industry, including a Mitsubishi Motors plant which remained in operation until 2000. It’s thus scarcely surprising to find a precision model engineering venture being located there.
The advertisement begins with the rather sweeping statement that the Atsuta 19 is “the most advanced engine” of them all! It goes on to claim that this engine is “the world’s only (rotary valve) engine that can perform in either forward or reverse rotation”. We shall have a later opportunity to evaluate this statement. The advertisement notes that this makes the engine ideally suited for use in a pusher installation. The manufacturer is said to be providing “innovative design and sound manufacturing technology from the point of view of the modelling enthusiast”.
Information received from Japan suggests that the Atsuta 19 was the only model ever produced by this manufacturer. According to Akira Fujimuro, the Atsuta engines were manufactured on a modest scale in a small workshop as opposed to a major factory.
At this point in time over 60 years on (as of 2014), the identity and background of the individual(s) responsible for the Atsuta model engines remain obscure, probably being lost forever at this stage. It would be nice to think that some of those who had been involved with the full-sized Atsuta aero engines had some involvement with the model engines which came later, but we have no evidence for this.
The precise production period appears to be the subject of some uncertainty. My Japanese informants tell me that the engines were produced over a relatively short period of time at some unspecified point during the 1950’s. Looking at the engines in architectural terms, I would hazard a guess (which is all that it is) that they most likely date from the early fifties. By the middle of that decade they would have been considered very “old-fashioned” indeed by emerging contemporary model engine design standards.
Another circumstance supporting the idea of a relatively early date is the previously-noted fact that no mention of the Atsuta engines is to be found in the English-language modelling media of the 1950’s. The activities of Japanese model engine manufacturers went completely un-reported in the English language until January 1953, when the first article in which Japanese engines were covered appeared in “Aeromodeller” magazine. Thereafter, English-language model engine commentators paid ever-increasing attention to Japanese products. Despite this, there is no mention of the Atsuta engines to be found anywhere. The clear implication is that the range was already history as of 1953. The architectural evidence cerrtainly supports this dating.
There is also considerable uncertainty regarding the issue of production figures. Akira Fujimuro stated his belief that the numbers manufactured were very small - perhaps between 300 and 500 examples. However, if this impression is accurate, then we face a serious difficulty with respect to interpretation of the known serial numbers. My own illustrated example bears the serial number 0548, but thanks to Rudy ni Balinia I also have images of similar engine number 3250. If there was no break in the serial number sequence, this would clearly imply the production of at least 3250 examples. Only the uncovering of more serial numbers would help us to resolve this question. Given the engine’s extreme rarity, I’m not holding my breath …….
The Atsuta 19 - description
In writing this section of the article, I am at last on firm ground, since I have an actual example of the Atsuta 19 to examine and test as well as a contemporary image showing the engine in component form. This is just as well, since I was unwilling to disturb this near-pristine example beyond the removal of the backplate to examine the induction and crankcase arrangements. My reasons for this decision will become clear as we proceed.
The Atsuta 19 is a short-stroke plain bearing lightweight glow-plug motor having a number of features which place it well out of the rut in design terms. We may as well start with a few vital statistics. The officially-quoted bore and stroke were 17.0 mm and 14.6 mm respectively for a displacement of 3.31 cc (0.202 cuin). My example checks out at exactly these figures. The weight of the engine (with plug) is a commendably light 140 gm (4.98 ounces).
This displacement figure is interesting insofar as it puts the engine just above the displacement limit for American Class A competition. This is another indicator of a relatively early date - by the early 1950’s almost all Japanese manufacturers were designing their engines to conform to US competition classes. This was becoming important even in Japan itself, where the ongoing presence of the predominantly American forces of occupation led to many domestic contests being flown to US rules.
The Atsuta 19’s bore and stroke measurements result in unusually over-square geometry for an engine of this vintage, although there are precedents. Even so, it seem possible that the designer(s) may have been retaining the option of adding an under-bored 2.5 cc (.15 cuin.) version at some point using the same bottom end. A bore of 14.7 mm combined with the same 14.6 mm stroke would yield a displacement of 2.48 cc with square internal geometry. However, my Japanese informants believe that only the 19 ever actually appeared on the market.
At first glance one might be forgiven for believing that this engine was a radially-ported design - the absence of an exhaust stack together with the general style of the cylinder would appear to imply this. However, in reality the engine is basically a conventional cross-flow loop scavenged unit featuring a plain bearing along with rear rotary disc valve (RRV) induction.
The crankcase is a clean pressure die-casting which incorporates the main bearing in unit. The most obviously out-of-the-rut feature is the unusually long main bearing. It seems likely that the idea was to make the engine more amenable to tidy installation in a long-nosed scale or semi-scale model. The use of RRV induction would also support this goal by creating a front end which is completely free from any protuberances. The combination of a plain bearing with a rather basic disc valve design argues strongly against any notion that racing applications featured in the designer’s ambitions.
The engine features a steel cylinder of typical Japanese form by late 1940’s/early 1950’s standards. The cooling fins are machined integrally with the cylinder itself. The cylinder is located on the top of the crankcase casting by a relatively thin location flange which seals with a gasket. The cylinder head is a separate aluminium alloy component which is produced by pressure die-casting, a technique with which the manufacturer was clearly very comfortable. As with the cylinder, a thin gasket is used to create a head seal.
The head and cylinder are both secured to the crankcase by a pair of long screws which pass through clearance holes drilled through the cylinder head, cylinder cooling fins and location flange to engage with two tapped holes in the crankcase casting fore and aft. Two additional shorter screws are placed at the sides of the head to assist in the maintenance of a head seal. These thread into tapped holes formed in the relatively thick upper cylinder flange.
This assembly appears unusually open to distortion during the assembly process, particularly given the combination of unevenly-distributed hold-down stresses, very thin cylinder walls and the extremely wide exhaust port which removes much of the support from the upper cylinder wall and cylinder installation flange on the exhaust side. There is no trace of binding in the illustrated example, and I was keen to keep it that way! It was largely for this reason that I elected not to disturb the cylinder assembly.
The cylinder porting is somewhat unusual, albeit not atypical in an early post-war Japanese context. Both exhaust and transfer ports are milled through the relatively thin cylinder walls to create rectangular openings of unusually large area. The exhaust opening is cut above the cylinder location flange, while the transfer port is cut below the flange. This would preclude the creation of any overlap between the two ports were it not for the use of a step cut into the crown of the cast iron piston on the transfer side. As can be seen in the attached image of the engine in component form, this step is crescent-shaped rather than being cut straight across - a good design feature, since the width of the transfer port would otherwise require a step of unusually large width and volume, to the great detriment both of transfer gas velocities and combustion chamber shape. As it is, the narrow crescent-shaped step provides excellent directional control over the incoming transfer gas, along with enhanced gas velocity.
Even so, the presence of the step does compromise the shape of the combustion chamber at top dead centre by creating a pocket of gas on the transfer side well away from the ignition source. To mitigate this, the plug is displaced to the transfer side to reduce flame propagation distance to this pocket and thus speed up its involvement in the combustion process. Another very intelligent design feature ……. The engine’s measured geometric compression ratio is a rather modest 7 to 1, suggesting that high operating speeds were not envisaged.
As a result of the presence of the step in the piston crown, there is considerable overlap between the exhaust and transfer ports, without the need to open the exhaust port unduly early. Timing figures are thus reasonably conservative, the exhaust being open for some 140 degrees of crankshaft rotation, while the transfer period is some 110 degrees. These numbers are entirely appropriate for the relatively modest speeds for which this engine appears to be designed.
The transfer port is fed by a pair of internally-cast bypass passages located in the visible bypass “bulge” seen in the attached images. This “bulge” is internally provided with a centrally-located partition to provide better support for the lower cylinder. The two parallel bypass passages formed in this way appear to have more than ample area.
The piston drives the crankshaft through a pressure die-cast connecting rod of quite substantial dimensions. The one-piece solid steel crankshaft features a full-disc crankweb having a crescent-shaped counterbalance of modest dimensions machined onto its rear face. The crankpin has an extension of reduced diameter to drive the disc valve in the conventional manner. An aluminium alloy spacer is fitted to the crankpin behind the con-rod big end bearing to prevent the rod from fouling either the disc valve or its central mounting pin.
At the front of the engine, the prop driver is another clean pressure die-casting which is driven by the shaft through a self-releasing taper. Another feature which seems to date this engine to the early 1950’s is the use of two integrally-cast pins to prevent prop slippage instead of the more conventional knurling of the prop driver face. The prop is secured in place by a steel bolt which threads into a hollow portion of the shaft at the front.
Looking now at the backplate, we find that induction is controlled by a pressure die-cast aluminium alloy disc which operates in conjunction with a pressure die-cast backplate incorporating the intake venturi cast in unit. The disc valve mounting is well thought out in that the steel mounting pin is threaded through the backplate and secured by a lock-nut. This allows easy adjustment of the disc valve clearance.
The actual induction port in the backplate is a simple round hole, no attempt being made to contour it for enhanced gas flow. In order to achieve the desired timing, the window in the disc itself is unusually long, giving a full 180 degree induction period (30 degrees after bottom dead centre to 30 degrees after top dead centre). This appears to be an entirely appropriate induction timing period.
The most interesting feature of the disc valve configuration is the fact that it has been designed to allow the engine to be assembled for operation in either direction, exactly as claimed in the previously-noted advertisement. As illustrated, the engine is set up for conventional rotation. However, if one simply removes the backplate and re-assembles the engine with the backplate rotated by 120 degrees in a clockwise direction (viewed externally from the rear), the engine will then be timed for opposite-hand rotation with exactly the same induction timing. Note the presence of small piston clearance cutaways on the rim of the backplate installation “spigot” to accommodate either configuration.
To maintain access to the needle valve control if the re-configured engine is to be used in an upright mounting orientation, it will be found necessary to remove and re-install the needle valve assembly from the opposite side of the venturi. In this form, the venturi will be located at the bottom of the backplate as assembled, with the needle valve assembly in a near-vertical orientation. Considerations of torque balancing could potentially make this a very useful option in a multi-engine scale or semi-scale model context.
The needle valve is of the externally threaded variety, screwing into an internally threaded spraybar which is secured with a nut. The very fine needle thread might be expected to do much to improve the precision to which needle settings could be optimized. Needle tension is provided by a coil spring, which appears to be highly effective. The non-optional alignment of the spraybar when set up for normal rotation is a little unusual, but it does have the advantage of placing the needle control very conveniently for the fingers as well as avoiding the need to modify the engine bearers in the model to accommodate the needle control.
All fits and finishes in the illustrated example are well up to the best contemporary commercial standards. The engine appears perfectly fitted throughout, turning very freely with the plug removed but with no detectable play anywhere in the system. With a plug fitted, compression seal is excellent, as is the base compression seal. I would judge that this example has seen little if any use in a model, although it does appear to have had a certain amount of bench running.
A point to note in connection with this engine is the fact that the crankshaft end float appears to be rather critical. The axial clearance provided for the connecting rod big end and spacer to fit between the crankweb and the disc valve is minimal. Consequently, if the shaft and rod are not to bear on the alloy disc when set back during starting or when operated in pusher mode as suggested in the advertisement, then end float has to be kept small enough to ensure that the rear of the prop driver absorbs any set-back forces. Otherwise, a condition would be created whereby the disc valve would in effect become the thrust bearing, which is bound to cause high levels of friction along with accelerated and/or uneven disc valve wear. Many examples of the E.D. Bee diesel and its Hornet relative suffer drastically from this problem.
On the other hand, there needs to be enough end float to ensure that lubricant has free access to (and passage through) the extended main bearing, for which lubrication is likely to be a rather critical issue. My example has about 0.2 mm end float, with very little leeway in either direction as far as I can tell.
If the engine was to be used in pusher mode, I would definitely recommend the use of a thin steel thrust washer between the rear of the prop driver and the front of the main bearing. Otherwise, the situation would be an alloy-on-alloy thrust bearing - not a recipe for longevity. It appears that the engine may have been fitted with such a shim as supplied - engine number 3250 seen in the earlier component view clearly had such a thrust washer, since it’s present in the photograph. I suspect that my example is simply missing this component. If the engine is used in normal rotation mode with a tractor prop, this washer is of course unnecessary.
The other point worth noting is that this appears to be one engine which is best left undisturbed once settled down and run in. The cylinder design with its two main hold-down bolts, thin walls and oversized exhaust port is such that the cylinder is almost guaranteed to take a “set” when first tightened down. If it is subsequently disturbed, there is little chance that the exact same “set” will be re-established upon reassembly, with negative consequences on both performance and longevity. Best left well alone …………..
Apparent Design Problems
We've noted the unconfirmed possibility that over 3000 examples of the Atsuta 19 may have been made in total. If this is indeed the case, the next question must be - where are they all now?? Why is this engine as rare as it appears to be, even in its native Japan?
A large part of the answer must surely lie in the engine’s apparent vulnerability to damage of one sort or another. For starters, the mounting lugs appear unusually and indeed unnecessarily fragile. The fact that they are of “mouse ear” form seen in plan view would clearly make them highly vulnerable to crash damage, as would their relatively minimal vertical thickness. This issue would be exacerbated both by the unusual length of the main bearing and by the fact that the mounting surfaces are vertically offset from the engine’s central axis. Both of these factors would contribute to the very effective transmission of front end crash stresses to the mounting lugs, far more so than in the case of an engine having a shorter main bearing located coaxially with the engine’s mounting surfaces.
For this reason, I would expect that a substantially higher than usual incidence of crash damage most likely bedevilled the Atsuta 19. If the engine was marketed at a competitive selling price, this may have led to damaged examples simply being discarded rather than being seen as having sufficient monetary value to be worth repairing.
However, it appears that the real culprit may have lain elsewhere. According to Akira Fujimuro, most of the Atsuta engines failed in service because of inadequate lubrication reaching the front end of the very long plain-bearing shaft. This resulted in damage to the shaft and bearing, sometimes extending to catastrophic failure. Such an issue would quickly have become apparent to contemporary users, a fact which would have done nothing to enhance the engine’s chances of marketplace success. It may have been this issue as much as anything else which led to the seemingly speedy withdrawal of the Atsuta 19 from the market.
If this is how things went, then the supply of spare parts would also have been terminated early on. This in turn would explain why any examples that subsequently broke their cases were discarded - spare replacements were simply not available.
Either way, it appears that an unusually small proportion of the engines that were manufactured have survived down to the present day. Those of us lucky enough to own examples will need to take very good care of them! In particular, anyone (like myself!) who is planning to run one would do well to lubricate the front of the main bearing before each and every run and also make sure that there is a little clearance at the rear of the prop driver to provide some end-float, thus allowing the free passage of lubricant into and out of the bearing.
In view of the likely identity relationship between the Atsuta model engines and their full-sized Daimler-Benz based Atsuta predecessors from WW2, it’s interesting to note that the full-sized Atsuta engines also experienced serious crankshaft problems. The design of the Daimler-Benz originals from which the Atsuta engines were copied was such that the fit and alignment of the crankshaft in its bearings was critical given the greater than average length of the engine. The Japanese manufacturers appear to have had difficulty building the engines to the required tolerances. Postwar evaluation by the US Air Force’s Foreign Aircraft Evaluation Center found that the standard of workmanship displayed in the Atsuta aero engines was not as good as that of the Japanese Army's Kawasaki Ha-40 and far worse than that displayed by competing engine makers Mitsubishi and Nakajima.
There was also the issue of the quality of steel available to the Japanese during the war, which fell somewhat short of that specified by Daimler-Benz. The overall result was that the full-sized Atsuta engines proved to be somewhat prone to crankshaft failures. Despite this, the Aichi company managed to produce some 873 Atsuta series engines during World War II before repeated Allied bombing of the Atsuta engine plant eventually brought engine production to a standstill.
The rarity of the Atsuta model engines, particularly outside Japan, is such that few collectors will ever see one, far less own one. Despite this, these engines seem to me to be well worth noting, since they undoubtedly constitute a hitherto-unreported piece of the overall picture of Japanese model engine manufacture during the early post-war period. Moreover, their unique combination of design features places them well outside the rut, reflecting a refreshing willingness on the part of their unknown manufacturers to think outside the box as it then existed.
If time permits at some point in the future, I will subject my example to a test and add the results to this article.
Article © Adrian C. Duncan, Coquitlam, British Columbia, Canada
First published December 2014
Updated September 2018