HOME

Mamod Steam Engine Conversion

A three cylinder Mamod? Well, why not!

The project described here is not up to the high standards expected by model engineers. However, I make no apologies for that. Even serious model engineers enjoy a little light entertainment occasionally. (I don't consider myself to be a serious model engineer but I do hope a few can bring themselves to read this.) Toy steam engines (Mamod, Wilesco, Bowman, Bing, SEL, et al) are good fun and I'm sure many readers have owned at least one in their formative years. (I have several and I'm well beyond my formative years.) As examples of engineering, however, they are variable and even the best are simple to say the least. But then, that’s part of their charm.

The most common configuration - single cylinder, single acting oscillating type - suffers badly from vibration; caused, not only by the lack of balance, but also by the single short power impulse per revolution. Twin cylinder models are better. They are sometimes quite well balanced and have two impulses per revolution but there are still dead points at the extremities of each stroke. Surprisingly, double acting cylinders are not much better. Both single and twin cylinder versions are still unbalanced. (Twin double acting cylinders are usually placed 90 degrees apart and cannot be properly balanced).

The obvious answer - three cylinders - has rarely been implemented at the toy end of the market. Three cylinder engines can, in theory, be perfectly balanced and, even with single acting cylinders, have three power impulses per revolution which can be a substantial percentage of the total cycle.

One day, looking at the two Mamod twin cylinder SE3 engines that I then had in my possession, I thought it would be comparatively simple to double up on the engine parts and produce a 4 cylinder engine. Unfortunately, without scratch building a complete crankshaft, the two halves would have to be joined at the existing crank pins. This would mean that the cylinders would still be at 180 degrees to each other and this would be no advantage over two cylinders. Effectively, it would be just doubling the size of the cylinders. However, by removing one of the cylinders and adjusting the remaining three at 120 degrees apart, the result would be (almost) perfectly balanced.

At this point, I wasn't seriously considering the conversion because one of the two engines I had was in excellent, almost original, condition and I wanted to keep it that way. Several months later, I saw advertised another similar engine up for auction in extremely rundown condition which was likely to go for a song. I immediately thought about my previous musings and put a bid in. In the end, it went for a bit more than a song but still quite cheap.

I then had two twin cylinder SE3 engines in less than perfect condition. One was in reasonable condition, especially the boiler. For the sake of argument, let’s call it engine A. The other had obviously just been dug out from the back of someone’s garden shed and was pretty corroded (especially the boiler). I'll call this one engine B.


The engines in original condition (A on the right)

A previous owner of engine A had, for some reason, replaced the flywheel with a completely inappropriate one. It was extremely large, cast in bronze and seems to be home made, with the central hole inaccurately drilled; (It wobbled when running). Also, one of the frames, holding one cylinder and crankshaft bearing, was loose and the whistle was a new replacement. The paintwork was not original anyway so I had no qualms about cannibalising it.

Engine B was in generally poor condition and incomplete. What was there, however, seemed to be mostly original. The boiler was beyond redemption. With the exception of the safety valve, all other boiler fittings were corroded solid (despite the fact that they were brass) and couldn't be removed. The water level plug had been replaced with an ordinary screw which was jammed solid. The whistle had probably broken at some time and had been soldered back (at right angles!) just to plug the hole. The chimney was missing along with the exhaust pipes from both cylinders. The cylinders themselves, however, seemed to be in reasonable condition just needing cleaning up.

All together, therefore, I had all that I needed to make one good 3-cylinder engine. The first decision I had to make was how to mount the thing. The existing pressed steel base was not big enough to hold the extra works. The alternatives were to build a new base from scratch, to extend one of the existing ones with a plate or to, somehow, join the two existing bases to form a larger one. The simplest would have been the extension. However, by this time, I was beginning to think that the new engine should be made entirely from genuine Mamod parts and in typical Mamod style. That meant that the only alternative was to join the two together to make one large one. My initial thinking was to cut off just the ends and lap-join the remainders in such a way that the final size could be adjusted as required.

The next decision I made was concerning the boiler. I originally thought that, because I was adding a third cylinder, I would need 50% more steam. Then I realised that it could use the same amount of steam at 2/3 the speed. This would be no problem because the three cylinder engine should run much smoother and should be quite happy at a much slower speed. I decided to make no changes to the existing boiler from engine A.

The final decision was how to join the two crankshafts. I wanted to make the joint dismantleable so I first thought of using a length of threaded rod in place of the two pins. This would have meant enlarging the big end of the con rod somewhat and would have looked wrong. I then realised that the joint didn't have to be rigid; in fact it would be an advantage to have some flexibility there. One of the pins could be drilled out and the remaining pin slid into the hole. The pin may have to be replaced by a slightly longer one but at least it would be, essentially, original. (As it turned out, the pin was only a push fit so it came out with a good tug. I had to enlarge the hole slightly to let the other pin slide in and out easily. The original pin turned out to be long enough as it was.)

One point noticed at this stage, though not a problem, was that the designs of the cylinders & pistons were slightly different in the two engines. Close examination proved that the differences were minor and the parts were interchangeable. In fact, the big end of the connecting rod was somewhat narrower on engine B which would make it more suitable for the central crank and would almost look deliberate.

The pipe-work would be a simple extension of the existing on engine A. The steam feed would be a combination of pipes from engines A&B. The exhaust for the cylinder furthest from the boiler would have to be a new length of copper pipe. I then had the new design complete in my head and started thinking about making a drawing. Unfortunately, I didn't, at the time, have a CAD package for my PC and I no longer have an old fashioned drawing board. The changes were mainly re-jigging of existing parts so, in the end, I kept the design in my head.


The engines dismantled (A above, B below)

I started by dismantling the engines. I had already had engine A apart on a previous occasion so I knew there would be no problems with it. Any ideas about salvaging any part of the boiler from engine B were abandoned completely at this stage. It went straight into the bin. I gave all the parts (especially of engine B) a quick rinse in degreaser at this stage, just to better see their condition. There was not much time spent deciding which parts to use. Where there was a choice, the choice was obvious.

I then made some quick measurements to decide where to cut the two bases. The first unforeseen problem then presented itself. Because there is a depression pressed into the base for the burner, it seemed to make sense to position the cut with the depression in one half on the waste side. This was a mistake because, having done that, the join on the top came at an inconvenient point. I had to cut the bases back even further so that the join came between the two engine frames. At first, this didn't seem to be of much consequence but later found it to be a problem.

The bases are made of approx. 1mm thick steel which is too thin to cut with a hacksaw and a bit thick for tin-man’s snips. Fortunately, I have the ideal tool for the job called a "Goscut". It is a cross between tin-man’s snips and a nibbling tool. It doesn't distort the metal like snips do but is faster than a nibbler on long straight lines. 1mm steel is just about the limit of it’s capability but it did the job. I don't know if the Goscut is still available. I bought mine many years ago when it was sold under the Eclipse brand name.

Having cut the bases to size, I was able to assemble the main parts, temporarily, to check for alignment. This was when I discovered the requirement for a modification. I found that the step at the lap joint in the two bases was enough to cause a problem. This meant that I would have to fit shims under two of the frames to line up the shafts. These could be cut from the waste material to guarantee the correct thickness. Making and fitting the shims appeared to cure that little problem perfectly and I was also able to prove that my initial calculations and the design in my head did, actually, work!

I then turned my attention to the pipe-work. This is where the second unforeseen problem showed. My original plan was to use the existing tee pieces and pipes, which would have been long enough without making any joints except at the tees. This meant unsoldering the pipes from the tees and re-soldering them in a different order. Unfortunately, the pipes seem to be brazed or welded into the tees rather than soldered. Heating them with a blow torch, even up to red heat, still didn't release them. (Maybe they are crimped as well). I had to rethink the pipe-work. I was able to use the pipes and tees without separating them but this required a joint in the pipe between the two tees. This I made with a bit of brass tube from the junk box. Not genuine Mamod but not too obvious. I may be able to re-do this bit later if I can get hold of some replacement Mamod tee pieces. The exhaust for the third cylinder, the one furthest from the boiler, had to be made up from a new length of pipe. At least it was genuine Mamod 1/8" copper pipe. I had to enlarge the hole in the funnel base to take the extra pipe. This was easily done with a needle file.

One annoying problem showed itself with repeated assembly and disassembly of the cylinder blocks. The small screws holding the blocks to the frames are totally inaccessible with any normal screwdriver while in-situ. This is not a result of this modified design; the same problem was present in the original. Hex heads to the screws would have been far better. By the end of the project I had found suitable hex head replacements in a suitable size. That just about completed the non-cosmetic work (so I thought). The next stage was to reassemble the complete works, with screws instead of pop rivets, and see if it ran.


Overall view of converted engine as first assembled for alignment testing

You've guessed it - it didn't! It didn't just run badly; it refused to go at all. There was obviously steam getting to each of the cylinders. Lifting each cylinder off it’s block against spring pressure proved that. After separating the three exhaust pipes I could also see wet steam exiting from all three when I spun the flywheel by hand. It felt very rough when I spun it so that was probably a clue. I applied plenty of oil but it didn't even start. One cylinder seemed to be giving some power but not enough to drive the other two, which didn't seem to be providing any significant power. I loosened all the frame screws hoping that it would allow some free play to overcome any slight misalignment. That made very little difference. The three cranks were not exactly 120 degrees apart but near enough. There was nothing else I could see whilst in steam. Disappointed in the extreme, I put out the burner, drained the boiler and left it to cool before dismantling it again. Engine A had worked fine and two of the cylinders were from that engine. The third cylinder and piston came from engine B (the one with the boiler I daren’t fire up) so that could be something to swap when it cooled. Had I partially blocked a pipe when trying to remove the tee pieces? Probably not; the best cylinder was the one furthest from the boiler, but that would be something else to think about. Thoroughly depressed, I went and put the kettle on. At least that worked, which restored my confidence in steam somewhat!

After I had dismantled the crankshaft assembly, I peered down the row of bearing holes and saw that there was a significant misalignment. The shims had ensured that the holes were aligned vertically but not horizontally. The problem seemed to be that the two bases (I must start thinking of them as two halves of the one base) were not exactly in line. I could see from the holes in the overlapping section of base that there was an error of just a fraction of a mm. Was that enough to cause the problem? Fortunately, I hadn't yet riveted the halves together so I loosened the screws and retightened them in perfect alignment. I rechecked the bearings and saw an improvement but it still wasn't perfect. To get perfection I had to force all the components affecting the alignment to the limits of the play available. Obviously, the holes were not drilled in exactly the same positions in the two halves.

This done, it all felt a lot better when turned by hand so, full of confidence, I fired up again. This time it started - eventually. My theory that the engine would run much better than a two cylinder at slow speed was well confirmed. In fact, at first, it was a good job it did because it wouldn't go faster than about 200 rpm. (I don't have any means of measuring the speed but, at that speed, you can estimate the revs per second). Confirming my initial fears, it did seem to be short of steam. It didn't really get going until the burner got hot and provided a good flame. As soon as the flame began to die, so did the engine. However, for a few minutes, it ran at a moderate speed and quite smoothly. The stall speed seemed to be around 100 - 150 rpm which was a satisfying confirmation of the theory.

Two things were learned from this trial run. Firstly; the bearings must be very carefully aligned. My method of "optical alignment" seemed to work but won't be very convenient if I dismantle it often. Secondly; I need a good meths. burner in good condition. I've found, in the past, that this type of vapourising burner can be improved by cleaning out the holes in the vapourising plate. An old wick type probably wouldn't have worked at-all well and a modern solid fuel one would probably be worse still. I had to do some empirical investigation (trial and error) to determine the best burner for the job from the few that I have. I may one day try one of the gas fired conversions available but that would be going away from my original Mamod-only policy.

Having proved that it worked, I thought the rest of the conversion would be mainly cosmetic. The half empty frame looked odd without a cylinder so I cut that in two leaving only the bearing end. Strictly, I only needed one flywheel so there was an unused rectangular hole in the base. I considered filling that in but decided against it. I initially put a second pulley there so it didn't look too odd. I later fitted a second flywheel, slightly larger than the original but still Mamod. It looks more intentional than the lonesome pulley. Being a larger size, it also offers an extra element of flexibility for driving arrangements. I fixed the halves of the base together with pop rivets and drilled some extra holes for rivets to improve rigidity. I considered riveting the four frames, making sure that they stayed properly aligned, but eventually decided to retain the nuts and bolts. The slot in the chimney had to be filed out to make room for the extra exhaust pipe. I replaced the whistle with a (genuine Mamod) pressure gauge. This was partly to reduce steam loss and partly for looks - I think it deserves it. The felt pads in between the cylinders and valve faces were all worse for wear; one had disintegrated completely. I cut some new ones from felt pads bought from the local hardware shop. Not genuine Mamod but they are out of sight and can be considered consumables if you're pedantic.

Finally, I cleaned, painted and polished everything as appropriate. I made no attempt to reproduce the Mamod colours exactly but I kept the general scheme - red base, flywheel inners and inside of chimney, black firebox, green frames and everything else unpainted. Having done all the cosmetic work I reassembled it, photographed it in pristine condition, and then fired it up again. Disappointingly, I seemed to be back to the same alignment problem again and there was still a shortage of steam. Also, I found that the red paint I had used on the base (an aerosol) softened in contact with alcohol! This was a problem when cleaning and also with any spilled meths. I later tried a paint advertised as heat resistant but this was no better so in the end I had to use good old fashioned enamel paint with a brush.


Overall view of converted engine after painting

The throttle in the live steam line was leaking quite a bit. I didn't have a spare so I replaced it with a direct connection (again, genuine Mamod - taken from one of their kits). This means I've lost all speed control but then, in most of the early (pre-reverser) engines, it was a luxury anyway, fitted only to the larger engines. If I find a replacement in good condition at some later date I'll restore it. The alignment seemed to be variable and I found it was caused by flexing of the base. To provide some rigidity, I fixed a couple of rigid bars out of sight under the base. This resolved the alignment problem but I opened up the hole in the pin-less crank to make it a bit more tolerant of future misalignment anyway. Fired up again, it worked but not spectacularly. It was still short of steam most of the time. Turning the flywheel by hand I could detect a distinct rough patch at a couple of points in the cycle. It would definitely benefit from being run in but if it doesn't run well for more than a couple of minutes at a time, it would take ages to run in. I don't have a suitable electric motor to hand but I do have several other steam engines. One of these is a large Bowman engine with more than enough power. It will also run for 2-3 hours between refills. I set the two engines up with a belt from the drive pulley on the Bowman to the flywheel of the Mamod. I couldn't leave the setup to chug away on its own because the Mamod was running dry. I had to keep it well supplied with lubrication. The way the arrangement was in the most convenient positions for the belt, the Mamod was running backwards, i.e. it was pumping air into the boiler. This was no problem in itself but it meant that it was also pumping lubrication fluid into the boiler. For this reason, I used water rather than oil. This could be conveniently dripped down the funnel into the exhaust! After a couple of 2-3 hour sessions like this, the Mamod seemed to be a lot smoother. Encouraged, I fired it up again and was amazed at the result. It ran better than it had ever done before and with no shortage of steam! This roughness had probably been the main problem all along.

So that was it - I now had an engine I was happy with. All in all, it was an interesting project both from the technical and aesthetic viewpoints. Considering the problems I had, perhaps it’s not surprising that Mamod haven't done something similar. But, then again, Mamod wouldn't have to join two half bases; they would make a base specifically for the job. I suppose one of the attractions of these toys is their simplicity compared with the real thing; and let’s face it, they were originally sold as toys, although no self respecting 21st century child would be seen dead playing with one nowadays!


Setup for running in powered by Bowman engine


The final configuration after subsequent modifications

More Photographs