The Reeves H18 diesel
by Adrian Duncan
Click on images to view larger picture,
hover over the images for a description.
At the time when we published our detailed article on the Reeves 3.4 cc diesel, we were writing about the sole Reeves diesel model in our possession. Despite this inconvenience, we were able to piece together a fairly complete history of the Reeves model engine range in general, in addition to analysing the Reeves 3.4 cc model in detail. Readers wishing to become familiar with the broader outlines of the Reeves story are invited to refer to that article, which has been brought up to date since its original publication.
All things reportedly come to those who wait, and eventually we were fortunate enough to run across a highly serviceable example of the smallest Reeves diesel model, the H.18 of a nominal displacement of 1.8 cc. This allows us to present an in-depth review of that model to go along with our earlier study of the Reeves 3.4 cc design. Let's get right to it!
Readers of our original article will recall that the Reeves story began in June 1946, when Edward Reeves began to advertise both complete engines and sets of castings and pre-machined components with blueprints from an address on Church Street in Shifnal, Shropshire. The initial advertisements were low-budget affairs which appeared in the Trade section of the Classified Advertisement feature in Aeromodeller magazine.
The Reeves venture appears to have remained pretty much a classic small-scale "garden shed" operation throughout much or all of its working existence. At some point prior to 1949 the business was relocated to Victoria Road, Shifnal, where it was to remain for the remainder of its active period.
Initial products of the Reeves enterprise were confined to a utilitarian but highly serviceable petrol engine which was initially offered in both 5 cc and 6 cc forms, although the 6 cc version quickly seems to have become the "standard". In 1948, the spark-ignition Reeves design was replaced by the first of a series of diesel models in the form of the 3.4 cc design (actually 3.18 cc!) which was the focus of our earlier study. This was followed in late 1949 by a few examples of a 4 cc diesel model which was in effect a development of the 3.4 cc model. By this time, Reeves was trading from the Victoria Road address under the name of Reeves Model Power Units.
Up to this point, all Reeves products had been designed around the standard plain bearing crankshaft front rotary valve (FRV) layout. The engine with which we are now concerned, the Reeves H.18, represented a complete departure from the previous Reeves designs in that it was a disc-valve plain bearing engine of relatively sophisticated design. It first appeared in the latter half of 1950 and remained in small-scale production for somewhat over a year at a selling price of £3 2s 6d (£3.12 in modern money). Present-day collectors, dream on.
The Reeves H.18—Description
The Reeves H.18 bore little or no resemblance to any previous Reeves product. It was a far more up-to-date design featuring rear disc valve (RV) induction as well as short-stroke internal geometry. As events were to prove, it was to be the only short-stroke design ever marketed by Reeves. Bore and stroke were 0.510" (12.95 mm) and 0.500" (12.70 mm) respectively for a calculated displacement of 1.67 cc (0.102 cuin). These figures are confirmed by actual measurement.
This brings us to one of the oddities connected with this model. When Aeromodeller magazine published Lawrence Sparey's test of the H.18 (see below), the above bore and stroke figures were given correctly, but the metric displacement was cited as 1.77 cc! While more in keeping with the engine's advertised nominal displacement of 1.8 cc, this figure simply cannot be reconciled with the actual bore and stroke figures featured in the design. Try it for yourself!
To add to the mystery, the displacement in cubic inches was given correctly as 0.102 cuin!! Again, there's no way that this converts to 1.77 cc—instead, it is exactly equivalent to the engine's actual displacement of 1.67 cc! Oddly enough, no-one at Aeromodeller seems to have noticed this discrepancy at any time—the above incompatible figures are repeated seven years later in the entry for the H-18 in the "World's Model Engines" table which formed an appendix to the 1958 publication Model Aero Engine Encyclopaedia. A mystery which must forever remain unresolved, along with the reason why the Reeves H.18 (and the later Goblin 2.45 cc model) were included in that table at all when they had been out of production for at least 5 years!
Aeromodeller did get one thing right—they cited the engine's weight as 3 ounces exactly (85 gm). My example checks out at precisely that figure.
The fact that the H.18's true displacement was somewhat shy of the 1.8 cc figure by which it was both named and advertised represents the extension of a tendency by Reeves to overstate the displacement of their products. As we showed in our companion article on the Reeves 3.4 cc diesel, the true displacement of that model was actually only 3.18 cc!
Turning now to the engine's construction, the attached exploded view extracted from the Aeromodeller test report shows the major details very clearly. Beginning with the crankcase, this is a gravity die-casting which incorporates the crankcase, main bearing and cylinder barrel as a single unit. The cooling fins above the exhaust ports are turned into this casting, and the barrel is internally bored out to accept a drop-in cylinder liner. The beam mounting lugs are of generous proportions and are set well above the thrust line.
The crankcase incorporates a single bypass passage at the front—the external "bulge" which accommodates this is clearly visible. Two exhaust ducts are formed in the casting at exhaust port level, one on each side. Three 6 BA tapped holes are provided at the top for retention of the turned alloy cylinder head, which in turn carries the steel compression screw on a 3/16 x 24 Whitworth thread. This control is generously proportioned and very comfortable to use, although a finer thread might have been preferable.
The main bearing is plain and of unusual length. It has an outside diameter of 0.375 in, which combines with the 0.250 in. diameter shaft journal to create an unusually thin wall, particularly for such a long bearing. In my view, this is one of the weaknesses of this design—the long thin-walled main bearing is completely un-braced. Overall, it appears to be highly vulnerable to crash damage. But perhaps experience proved otherwise; the long unencumbered bearing would have made the Reeves quite attractive for use in scale models.
The cylinder liner is of hardened steel. It is a light push fit in the bore provided for it in the upper crankcase casting, locating on a flange at the top of the bore. The working liner is thus unstressed by any assembly forces.
The cylinder has two large rectangular exhaust ports, one on each side, which feed exhaust gas into the two ducts in the case mentioned earlier. A single transfer port in the form of an upwardly-angled drilled hole is provided between the two exhaust ports at the front. This transfer port overlaps the exhaust almost completely, giving a very short blow-down period. It is fed through the previously-mentioned bypass passage at the front of the crankcase, making the cylinder a one-way fit in this model. From a gas flow standpoint, this arrangement is less than optimal—the lower entry into the bypass passage is encumbered both by the full-disc crankweb and by the piston skirt at and near bottom dead centre. Combining this feature with the single-hole transfer port, I would expect this design to be transfer-limited.
Both the piston and the contra-piston are of hardened steel, a material specification which was carried over from the earlier 3.4 cc model. This combination can be less than ideal for contra-pistons, since a closely-fitted steel contra tends to stick in a steel bore when hot. However, the fit in the case of the Reeves was seemingly optimized from this standpoint and no trouble is experienced from this cause, on the present example at least.
The use of a hardened steel piston in a hardened steel bore is also a problematic combination, as the makers of the mid-1950's JB engines were to learn to their cost. The problem is that running-in wear is minimal, so the fit has to be pretty near perfect as constructed. Fortunately, if there was one area in which Edward Reeves was an expert, it was lapping pistons into cylinders! The fit on this engine, like that on both of my Reeves 3.4 cc models, is beyond reproach—no sign of "stiction" at any point in the stroke, and yet the engine seems to hold its compression forever despite having obviously received a fair bit of use in the distant past.
The piston is of relatively lightweight construction, with fairly thin walls. The fully machined hardened steel con-rod is mounted on a silver steel gudgeon pin which is pressed into the piston from the rear in order to prevent fouling of the transfer port at the front. Con-rod length is relatively short, accounting for the engine's somewhat squat appearance. In fact, the piston skirt is relieved at the front and rear to clear both the disc valve and the crankweb at bottom dead centre. The engine does not feature sub-piston induction.
Once again, the use of a hardened steel rod can cause problems with premature wear on the gudgeon pin and/or crankpin, but the bearings at both ends of the Reeves rod are both relatively long and extremely well finished. Accordingly, wear should not be as much of a problem with this engine as it is with some others. Both rod bearings remain very well fitted in this example despite the use which it has clearly received.
The crankshaft is a one-piece hardened steel component having a counterbalanced full disc crankweb. The main crankshaft journal is solid throughout with no central drilling to save weight. It is an outstanding fit in the plain bearing which supports it. As noted earlier, journal diameter is 0.250 in. The prop mounting thread is 2 BA, with a conventional spinner nut used to secure the prop.
The prop driver on this example is of steel and locates on a 30 degree included angle taper machined onto the front of the crankshaft. Images of other examples show what appears to be a cast alloy prop driver very much like that used on the earlier 3.4 cc model. The driver on my example has the appearance of a "production" component and is very well matched to the engine, having both exactly the same diameter as the spinner nut and the correct taper at the centre. However, it may be an owner adaptation of a component from a different engine altogether. It's also possible that Reeves ran out of the cast drivers at some point and that some examples were fitted with a steel item which was easily manufactured in-house without castings. At this late stage, we'll probably never know for sure.
Turning now to the backplate, this is another gravity die-casting which is secured to the rear of the crankcase with three countersunk-head 10 BA screws—a very neat approach. The wedging action of the countersunk heads must place considerable stresses upon the relatively insubstantial backplate mounting "mouse ear" lugs through which they pass, and I would not recommend tightening them too hard in case the lugs are split by these wedging forces. With countersunk heads, overly-snug tightening is not necessary in any case—the wedging action discourages loosening.
A conventional 1950-style disc valve of cast aluminium alloy is employed. This component is machined all over after casting—the sole evidence that it is cast is the kidney-shaped aperture for the actual gas entry, which retains the as-cast surface finish. The disc is conventionally driven by an extension of the crankpin which locates in a hole drilled in the disc. It is mounted on a steel spigot of generous diameter which is either pressed or threaded into the backplate—I didn't try to remove it to check! There is no retaining lip at the outer end of this spigot, and the disc is maintained in contact with the backplate solely by the con-rod and (when the engine is running) by internal gas pressure within the crankcase. This is of course a perfectly sound approach—it was used on the Mills 2.4 cc model, among others.
Another underlying design compromise is to be seen in the form of the location of the integrally-cast intake. This is centrally positioned at the top of the casting rather than being displaced somewhat in the direction of rotation as in the more usual case. The problem with the location adopted in the Reeves is that even with the drive hole in the disc located as closely as possible to the trailing end of the disc aperture it is not readily possible to delay the closure of the disc valve sufficiently to incorporate an adequate period during which the intake remains open at and around top dead centre. In fact, with a kidney-shaped disc aperture and a round intake bore as employed here, the intake will actually close before top dead centre unless additional measures are taken—not a good situation from a gas flow standpoint!
On this example, a considerable amount of careful hand work has been done to create a wedge-shaped cavity which extends the intake register opening in the direction of rotation and thus delays the closure of the system. This has the effect of postponing the closure until past top dead centre. The size of this hand-made cavity is about as large as the casting dimensions will permit without risk of breakthrough, but even so the closure delay past top dead centre is quite small—only around 15 degrees or so. Still, it's far better than no delay at all or even worse, having the system close before top dead centre as it otherwise would!
It's unclear whether or not this was done during manufacture or whether it is the work of a knowledgeable owner, but there's no doubt at all that it would improve performance significantly, to the point that it would not really appear to be an optional feature if the full potential of the design is to be realized. At the other end of the induction cycle, the disc valve opens unusually early—around 15 degrees past bottom dead centre. The timing of the induction period is thus a full 180 degrees.
The integrally-cast intake venturi has a measured bore of 0.201 in. It is equipped with a conventional spraybar having two jet apertures, one on each side. The spraybar is internally threaded 8 BA to accommodate an externally-threaded needle. Tension is provided by a single-leaf spring clip of sheet bronze which bears against the edge of a serrated brass disc soldered to the needle. This is very positive in action. The long needle extends above the top of the cylinder and is thus very easy to reach as well as being well protected against crash damage.
One feature upon which we will comment further in due course is the use of an unusually coarse taper on the tip of the needle. Admittedly the 8 BA thread is very fine, which should facilitate the use of such a taper, but even so it appears that the taper could with advantage have been made a little finer.
Overall, the engine has a very neat and quite pleasing appearance. The external finish of the castings can most fairly be described as "utilitarian"—the castings themselves are a bit rough, with a few blow-holes in evidence, but appearance is significantly improved by the application of a light polishing treatment following a certain amount of hand-filing to remove burrs and flashing. This is typical of Reeves products—they were always best finished where it really counted—inside. This example of the H.18 cannot be faulted in the latter regard.
The engines were identified solely by having the sequence "H 18" stamped upon the outer end of one of the mounting lugs. The illustrated example displays this identification on the end of the right-hand lug (facing forward), as did the example tested by Lawrence Sparey (see following section). However, the one illustrated in Mike Clanford's useful but often unreliable "A-Z" book clearly shows this marking on the left-hand lug. It appears that either orientation may be encountered.
The engines also bear a serial number which is stamped on the end of the opposite lug. The illustrated example bears the number 235, which is currently our highest known serial number for one of these engines. Kevin Richards reports having owned two of these units in the past which bore the serial numbers 161 and either 173 or 178 (the strike of the last digit was a little ambiguous). If anyone can provide any more H.18 serial numbers which might help us to estimate production figures, we'd be most grateful. At present, all we can say is that at least 235 of them were made.
The Reeves H.18 in the Modelling Media
The Reeves range seems to have remained pretty much below the radar of the mainstream model aeronautical media of the day. This was no doubt due in large part to its "cottage industry" origins and associated modest production rates which prevented the range from achieving any real prominence in the national marketplace.
Consequently, the H.18 appears to have largely escaped the attention of the model engine commentators of the day. One particularly odd circumstance is the fact that the engine was completely omitted from the 1951 revised edition of Col CE Bowden's book, Diesel Model Engines, despite the fact that its then-discontinued predecessor, the Reeves 3.4 cc model, was covered in some detail.
Writing many years later, the late OFW Fisher mentioned the H.18 in passing on page 40 of his 1977 publication Collector's Guide to Model Aero Engines, noting simply that he had used one in 1954 along with an Elfin 1.49 BB to power the prototype Meson Mk. 6 free flight model which subsequently appeared in plan form in the March 1955 issue of Model Aircraft. However, Fisher did not include an illustration of this or any other Reeves model.
Apart from the maker's periodic advertisements, the one contemporary appearance of the engine in print that has come to my attention is the published test by Lawrence Sparey which appeared in the March 1951 issue of Aeromodeller.
Sparey had little but praise for the engine's design and construction, commenting particularly upon the outstanding piston-cylinder fit. He reported that starting was "very good under all conditions" and that running was "smooth and consistent at all speeds, with good flexibility of needle control". In the latter context, he noted that the rear-mounted carburettor with its extended needle valve was very convenient from an operational standpoint.
Sparey reported a measured peak output of 0.1034 BHP @11,700 rpm. While the indicated specific output of 0.062 BHP/cc is nothing for us to get excited about today, it was a perfectly acceptable performance by 1951 standards for a compact 3 ounce engine of 1.67 cc displacement. Writing in the context of his own times, Sparey actually characterized this performance as "extremely good".
Apart from this, the engine exhibited another useful characteristic upon which Sparey felt moved to comment. This was an unusually flat power curve. Sparey pointed out that power fell off below the peak far less markedly than with many other engines that he had tested—0.08 BHP (77% of the peak output) was available as low as 6,800 rpm. In fact, the power curve derived by Sparey indicated that the engine was operating at 95% or more of its peak output at all speeds between 10,000 and 12,400 rpm, giving an unusually flat peak to the power curve and hence a wide range of practical airborne operating speeds and useable airscrews. From the standpoint of a practical aeromodeller, this was a very positive characteristic.
Sparey reported no mechanical troubles during the test (which he usually did when they occurred) and summarized his findings by stating that the engine "seems satisfactory from all points of view". About as positive an endorsement as Edward Reeves might reasonably have expected!
As far as the record shows, the rival Model Aircraft magazine never ran a published test on the H.18.
The Reeves H.18 Revisited
Having a well set-up example on hand that appeared to have already received quite a bit of use, I decided that nothing would be lost by giving it a few more runs myself to see how far Sparey's results could be confirmed by present-day experience. So I headed to the local model air park on a sunny but very cold afternoon to give it a go.
Sparey's test report cited the recommended airscrews as 8x6 or 8x8 for control line, and 9x4 for free flight. I therefore elected to test props lying more or less in that range as well as some slightly smaller sizes to allow the engine to reach airborne speeds on the bench.
Not having any idea about the correct settings at the outset, I set the controls by feel and guesswork and relied on port priming for the initial start. As events proved, I guessed about right for the compression but had the needle set way too rich to begin with. This quickly got the engine pretty wet, and I had to shut off the fuel supply and reduced compression to clear things. But eventually I got things sorted to the point where a start was achieved.
Once the engine was running, I soon learned a few things about its control characteristics. The contra piston was just about perfectly fitted for a steel item—a bit on the loose side when cold but perfectly fitted for adjustment when hot—the common problem with a steel contra piston of freezing in the hot bore was not apparent at all, yet the engine held its compression settings firmly. Response to the compression control was quite lively, due no doubt to the relatively coarse 24 tpi thread used on the compression screw. Despite this, the optimum setting was easily established.
The needle valve proved a little more problematic. Contrary to Sparey's reported findings, I encountered some difficulty in establishing an optimal needle setting for a given load. This was no doubt due to the very coarse taper applied to the needle tip—a few clicks on the needle (a mere fraction of a turn) made a significant difference to the running. Once the setting was established, the needle held that setting firmly, but getting to the optimum point was not as easy as it would have been with a finer taper.
After the initial run, I had no further starting difficulties whatsoever. I found that priming was quite unnecessary at any time—a couple of choked flicks got the engine going almost immediately every time. A very easy-starting engine indeed!! When the engine was cold, it was necessary to pull the needle out 3 or 4 clicks and raise the compression slightly, restoring running settings as the engine warmed up. Hot restarts were easy at running settings.
In performance terms, the engine proved to be no world-beater, but it did shift a meaningful amount of air and ran very smoothly at all speeds tested. Vibration levels were quite modest—certainly well within acceptable limits. Results achieved on the day were as follows:
The power curve derived from the above figures indicates that I missed a few of the settings (that touchy needle, no doubt!) but generally confirms the performance levels reported by Sparey. The main observable difference was that my engine didn't reach 0.08 BHP until around 8,000 rpm as against Sparey's reported figure of 6,800 rpm for the same output.
As can be seen, I never actually reached the engine's peak, but the above curve implies that this particular example probably peaks at around 0.100 BHP @ 12,000 rpm or thereabouts—near enough to the figures reported by Sparey. This is very nearly the same output as that obtained during my previous test of the Reeves 3.4 cc diesel, albeit at over double the rpm. So this design represented real progress by Reeves in performance terms!
It's also apparent that Sparey was quite correct in stating that the engine has an unusually flat peak to the power curve. Having established these points, and not wishing to risk damage, I decided against pushing the engine any faster on an even smaller prop than the 7x4 since such a load would take it past its peak in any event.
The engine completed this test session with no problems and seemed willing to keep on running as long as required. Overall, I found it to be a more than acceptable "sports" performer by the standards of its day. The two things that I would have changed would have been the unbraced main bearing and the taper on that needle.
It must be said that the performance figures cited above are difficult to reconcile with the maker's reported airscrew recommendations. In a control-line context, it would take a pretty "fast" 8x6 to get the engine up to the approximately 9,500 ground rpm which would be required for the achievement of peak performance in the air. An 8x8 would completely kill the engine in performance terms. I'd probably go for an 8x6 cut down to around 7-1/2x6 for best results. If the engine got this up to 11,000 rpm in the air, as seems likely, a theoretical airspeed of 50 mph should be achievable in a reasonably "slippery" airframe—perfectly acceptable for a 1.67 cc engine of 1950 vintage!
In a free-flight context, where airborne pickup is considerably less than in control-line, the recommended 9x4 again seems pretty excessive based on these tests. For maximum performance, I'd go no larger than an 8x4.
This said, the unusually flat power curve indicated by the tests reported above would mean that operating even 2000 rpm below the peak would represent less than a 10% power shortfall below the maximum. In most practical non-contest applications, this would probably be scarcely noticeable. Viewed in this context, the maker's recommended prop sizes seem a little less unreasonable.
The Rest of the Tale...
...may be quickly told. Following its introduction in the latter half of 1950, the H.18 remained in production for some time and was still being advertised in Aeromodeller in November of 1951. However, by January of 1952 the company was no longer advertising either the H.18 or any other specific model. Instead, they were hinting at new models yet to come. The implication is that production of the H.18 ended in late 1951.
By April 1952 the hints dropped by Reeves had become reality in the form of the 2.5 cc Goblin diesel, another plain-bearing disc-valve model which evidently replaced the H.18 while retaining many of its design features. Unfortunately, the Goblin got off to the worst possible start by failing to complete its test in the hands of the resident Model Aircraft tester, almost certainly Peter Chinn. As noted in the published test report which appeared in the magazine's October 1952 issue, the engine broke its crankshaft during the course of the test, and Chinn also reported that he had been unable to establish contact with the manufacturer to discuss the issue, nor was he able to obtain a second example for comparison purposes.
This unfavourable test outcome probably sealed the doom of the Reeves range. Actually, the fact that Chinn could not establish contact with them in connection with his then-unpublished test of the Goblin implies that Reeves Model Power Units may have already been in the process of winding up their affairs at the time of Chinn's test. They must have had some already-manufactured engines still to liquidate, because they continued to advertise sporadically, the final advertisement appearing in the December 1952 issue of Aeromodeller, still promoting the Goblin.
This was the last that was to be heard of the Reeves model engine line. A sad end to what had clearly been a very sincere effort on the part of a skilled model engineer to turn his passion to commercial account.
Following the initial publication of this article, I was delighted to hear from a former owner of my newly-acquired Reeves H.18 serial number 235. This was none other than Graham Podd, who is well known to collectors and engine builders alike. Amazingly, Graham still had the original box in which this engine was supplied! He had acquired the boxed engine from a non-modelling individual who had found it languishing in the roof space of his house and was going to throw it out before being advised (correctly) that it would fetch a good price on eBay! Graham had kept the box strictly to preserve it but was kind enough to pass it along to me, thus re-uniting the engine with its original packaging.
The engine had clearly spent many years in the attic and the box has suffered accordingly. But the label remains in good enough condition to be fully legible. It confines itself to an illustration of the engine along with the model and makerís address as well as the selling price. On this label, the price has been amended from an original figure of 65s 0d to the 62s 6d price at which the engine was generally advertised. It seems that the original selling price proved to constitute a market deterrent.
Graham also advised that he has what appears to be a late version of the Reeves H.18, the existence of which had escaped our attention until now. This variant has a more substantial cylinder head and thicker cooling fins. In addition, it has an aluminium prop driver. It bears neither a serial number or any form of model identification. This example appears to be unused. Graham advises that the backplate does not feature any additional internal machining to improve the induction cycle, making it seem likely that the cut-away in my own engine is the enlightened work of an owner rather than the manufacturer.
Our sincere thanks to Graham for making us aware of this seemingly-rare variant.
We've noted that the H.18 appears to have remained in production for a little over a year. There's no way of knowing how many were produced during that period, but the extreme rarity of the engine today suggests that the number can't have been that high. We'd need more serial numbers to get a better handle on this figure, and in the absence of such data we can only say that the engine's present-day rarity indicates either that many of them were damaged in service (that seemingly vulnerable main bearing?) and subsequently discarded or that there weren't that many of them made in the first place. Perhaps it was a combination of both.
There's no question that the engine was extremely well-made where it counted and that if carefully treated it was well able to give good service. In addition, it is a refreshingly out-of-the-rut design in the context of its time and is thus a most interesting model to add to any collection of early British diesels. Just be prepared to pay somewhat above the odds if one does turn up! Finding one is not easy, but anyone acquiring an example will be adding something of great interest to their collection!
This page designed to look best when using anything but IE!
Please submit all questions and comments to firstname.lastname@example.org
Unless otherwise expressed, all original text, drawings, and photographs created by
Ronald A Chernich appearing on the Model Engine News web site are licensed under a
Creative Commons Attribution-Noncommercial-Share Alike 3.0 License.