Section 2.1.3 The design of a miniature steam engine.
I suppose that I have seen twenty radio-controlled, steam-driven boats at work other than my own. Performance has varied widely from boats that get blown down wind when going flat out to quite fast boats. Most run for a short time, often less than 10 minutes and for most of that time are at full speed. Sometimes boats are put into reverse to ring the changes but my impression is that most just want to enjoy watching their creation or possession performing on the water going as fast as it can before it slows down by itself. People do not seem to take any interest in the boat or in the wake that the boat makes and see no special attraction in running on a calm surface. The big thing is to have real steam plant and to see steam coming from the chimney to prove it.
Starting procedures seem to be protracted and this appears to be due to a want of understanding during the fitting out stage. How else can one account for a boat with its exhaust directly in the base of the funnel and the condensate from the engine running back into the flue? There seems to be self-imposed rules that inhibit the design of steam-boats. Often they go back 50 years or more. Why am I accused of cheating when I put hot water into my boliers?
The duty of the engine.
The first requirement of a miniature marine steam engine is that it must not lock up with water trapped in its cylinders. In a model boat the engine must drive the boat at an acceptable speed but, in order to satisfy its owner, it may be required to do other things. It will certainly be required to drive the boat in reverse but this does not necessarily require the engine to be reversible because reversing may be done using gears like those on the back gear of a lathe or a Kitchen rudder. It must therefore be self-starting. The engine may be required to reverse “instantly” as it would be used in steering competitions. It may also be required to run slowly with some confidence that it will not stop.
It may also required to do any or all of these things using steam economically.
In order to design an engine we need to be able to give some order of priority to these various requirements. I will look at each in turn.
Some designers use very shallow grooves in the piston and rely on oil to fill these grooves and form a seal. This might conceivably permit a leakage but not I think if the piston fit is good. Others use a deeper groove and fill it with packing. The packing is not designed to expand outwards under pressure and might well leak steam and therefore water as well. O-rings are fitted to some engines. When these are used the ring works by being pushed out by the live steam to stretch the O-ring and produce a seal on the cylinder walls and the groove in the piston head. The seal is so good that neither steam nor water can leak past. I use O-rings that have been cut radially to make them behave like a piston ring as used in ic engines and locos and push out against the cylinder walls under live steam pressure but leave a very small leakage path for water to flow through if necessary. This prevents excessive water pressure. However there is another solution that was used many years ago on locomotives.
When I was a railway apprentice for a time I worked in the brass shop and we made “D” type slide valves for the pannier tank engines that were used for shunting. These engines might easily stand for 30 minutes without moving giving enough time for steam to condense in the cylinders. I recall quite clearly being surprised to find that what seemed to me were quite deep grooves were cut into the exhaust side of the valve face like those shown in figure 23. No one offered a reason for them. Ours was a production shop. However they make sense now even if 50 years has elapsed. These grooves prevent water being trapped in the cylinder after cooling for 30 minutes but the relatively small size of the grooves offers a resistance to flow of steam so that the pressures do not suddenly collapse as if the valve had no exhaust side overlap. No doubt a lot of cutting and trying went into finding out the best size for these grooves. It seems to me now that the grooves should extend across the face until the uncut width of the face is equal to the width of the steam port. Then water cannot be trapped but its discharge from the cylinder will be retarded because of the small cross section of the grooves. As a result the engine will run slowly until it becomes hot enough to stop condensing steam.
We can use these grooves in a miniature engine and indeed just grooves in the lower exhaust overlap made my troublesome first engine into a powerful runner that would throttle and reverse. I made the grooves as small as would get rid of the water and during warm up I normally keep the engine turning by hand and then let it run for a short time to clear the water before launching.
This seems to leave us free to use the pressure volume diagrams to design an engine but we ought really to look at the consequences for the diagram of introducing these leakage paths.
We are in a position to propose a likely diagram for our steam engine. The first thing that we can settle is the point of cut-off. If the engine has two cylinders it must be at about 60% of the stroke. (For a three-cylinder engine it would be at about 40% of the stroke.) Figures 21 and 22 show that the point of opening to steam could be before the dead centre by say 10°. This gives enough information to start to draw a diagram.
I drew the points of admission and cut-off and they correspond closely to a lead of 45° ie the eccentric leading the crank by 135°. These two points rather fix themselves but the positions of release to exhaust and the closure of exhaust are dependent on the choice of exhaust side overlap and this can be chosen.
In figure 24 I have shown four different exhaust overlaps starting with no overlap and drawn the all four diagrams in blue. It is apparent that the use of no overlap gives the least expansion of the steam to point C and the least loss of useful work to compression from point D. However, for the greatest exhaust overlap, the work gained by the later cut-off is offset by the increased work lost to compression after exhaust valve closure. It presents a difficult design choice.
Now we must take the use of leakage grooves into account. Once the decision to use expansion in the cylinders is taken, compression at the end of the exhaust stroke is inevitable whatever exhaust overlap is used. The attempted compression of condensate in the cylinders will occur primarily during at start-up and the grooves are needed to avoid this. But they also provide a leakage path for steam and this will affect the diagram. I have constructed figure 25 where the lead is 45° and drawn a diagram for no exhaust overlap ABCD and the largest overlap shown in figure 23 ABC¢D¢ There is clearly not much difference between the areas enclosed by these two diagrams. On the figure I have added lines representing the effect of having leakage grooves. They are the new curves shown dotted red and dotted blue. The effect of the grooves is to reduce the work gained by expansion but also to reduce the loss of work to compression. So, provided that the grooves are as small as is practical the net effect is quite small. Of course small grooves give a dotted line near to the original and big grooves will take the dotted line a long way from the original.
In a real engine the opening and closing of the valves is not instantaneous and the effect on the diagram is to round off all the corners.
So it is possible to have expansive use of steam and to avoid problems of condensate in the cylinders. The valves, whether Dee valves or piston valves are not difficult to make and permit some experiment with exhaust overlap and the depth of grooves.
If you are happy with a brief run and prefer the easy life then make a two-cylinder engine with no overlaps and Stephenson’s gear with the cranks at 90° and off you go. It will be self-starting and reversing.
If you fancy your chances as a model engineer and are looking for an engine that starts and reverses and runs economically then there are decisions to be taken.
The first must be the starting. If a two-cylinder engine is to be self-starting steam must be admitted to one or other of the cylinders throughout the 360° of each revolution. As the engine could stop on one of the four dead centres the steam must be admitted for more than 360° and the lowest figure for admission must be about 100° from the appropriate dead centre. This limits the work that can be obtained by allowing the steam to expand in the cylinder.
For increased economy one might consider designing a three-cylinder engine when the need for self-starting can be satisfied with admission for only 70° of crank rotation but the problem of valve gear becomes difficult.
Reversing is a different sort of problem in that it involves choosing a mechanism.
I have said that the Stephenson’s link motion is not very suitable for speed control of a miniature marine engine and its best application is to use the motion for finding the best running position both forward and reverse and using the gear to switch eccentrics. Then the speed can be controlled using a throttling valve. For manufacturers this leads to an expensive engine and they seek alternative, cheaper ways of switching between eccentrics.
The two ways that I know are the slipping eccentric and reversing by gearing.
The slipping eccentric is usually used on two- or three-cylinder engines where the eccentrics are on a secondary shaft driven by gears from the crankshaft. The driving gear on the eccentric shaft is one part of a dog clutch that drives the shaft and the clutch can be in either of two positions depending on the direction of rotation. One position leads by 135° in one direction and the other leads by 135° in the other direction. In itself this is admirably simple but it becomes less attractive when the switching has to be done by radio control in a smallish boat. The part of the clutch on the eccentric shaft is fitted with a gear and when the engine is stationary a rack is moved across the gear and, as it does so it drops down to engage with gear to move it and then rises to disengage. This switches the eccentric shaft. It is not certain to work as sometimes the rack fails to engage. From the control point of view the sequence is, stop engine, operate reversing switch on Tx, open throttle to see which way the engine runs. If it has reversed, OK, but if not, the process has to be repeated with the extra step of reverting to the original switch position first. A better option that was used by Cheddar models is shown in figure 26. The two large gears continually mesh with both the gear on the crankshaft and the one on the eccentric shaft and themselves. Moving the idling pair bodily round the two shafts advances the eccentric shaft. I do not know the timing of the engine and this may not work for an engine with overlaps.
Judging by the variety of mechanisms for in-line and off-centre gear-operated reversing systems people just enjoy the exercising of ingenuity involved.
In figure 27 I have shown the reversing gear used on the Myford lathe in its two operating positions “a” and “b”. The reversing lever carries two intermediate gears that
continually mesh together and one meshes with the lead screw. In position “a” the drive is through one intermediate gear and in “b” it is through both gears. This essentially simple mechanism can be the basis of a reversing mechanism for a miniature engine.
Figure 28 shows it in use on a three-cylinder engine. Gears on the crank shaft and the eccentric shaft are connected through an idler wheel to drive the engine in one direction. Then the drive is taken back from the gear on the eccentric shaft via the reversing gears to drive the free output shaft that is co-axial with the crank shaft.
It throws up one restriction inherent in these mechanisms, the size of the gears is restricted by the spacing of the crank shaft and the eccentric shaft. The gears will be heavily loaded.
A respected engine designer I told me that a model steam engine did not need a flywheel because the propeller can act as a flywheel. I wondered why I spent so much college time studying the matching of flywheels to engines. It was quite complicated and obviously necessary.
The speed of a reciprocating engine varies during each revolution. This is easily seen for a four-stroke IC engine. During the working stroke the piston acts on the flywheel to increase its speed, then, during exhaust and induction the speed falls a little before the big drop in speed whilst the flywheel provides the energy to compress the air and fuel ready for the next cycle to begin. Throughout, the engine will be giving up energy to the load that it is driving. The speed will be lowest at the point of ignition and highest at the point of release to exhaust. The difference will depend on the moment of inertia of the flywheel and on the mean speed of rotation. Big flywheels reduce the fluctuation in speed during the cycle. This in turn reduces the fluctuating load on the gears, if any are used, and on the universal joints and generally improves the running and the life of the transmission. Propellers are not, and never can be, flywheels to any advantage. Nor can paddle wheels.
However miniature marine engines have to be mounted in boats and a large flywheel requires the engine to be mounted high enough to clear the bottom of the boat if it is glass or the keel if it is wooden. As we want the engine to be low this gives an incentive to keep the diameter of the flywheel down. Then it has to be longer and, as the weight near to the axis is least useful, it should be counter-bored to remove as much of this non-contributory weight as possible. Keep in mind that the crank webs make some contribution to the flywheel effect.
NB. Universal joints must be used in pairs, they simply do not work singly as any look at the geometry will show.
In my view it is essential to have a lathe and a universal milling machine and not to be too mean to buy new milling cutters. For preference the lathe should have a digital scale on the saddle. The lathe is used for turning. It is a great help if the milling machine has digital scales on both horizontal axes of the table to get rid of the effects of backlash in the screws and stops. It is also a great help, if not in fact essential, to have a dividing head and a rotary table for the milling machine.
Given the right machinery engines can be made without the need for castings by use of stock bar and silver soldering.
If you have no self-imposed rules there is no need to turn the crankshaft from the solid. It is difficult to do and necessarily requires split big ends on the connecting rods and caps on the main bearings. A crankshaft built up from crank-webs, short axles with appropriate flats and setscrews make a perfectly satisfactory crankshaft. The crank webs will contribute to the flywheel.
Engine frames can be fabricated and can be soldered with plumber’s solder.
If you need gears they can be cut quite easily using profiled milling cutters. You do need to understand how to work out the addenda and dedenda from circular pitch etc. but it is easy really. People are frightened by the mystique of gears. You also need a horizontal milling machine and a dividing head. My paddle steamer would not have been possible without this equipment as I had to change the gear ratio once and the width of the gears as well.
Ordinary stock brass bar can be used for cylinders and stainless steel for the pistons. The engine is unlikely ever to wear out and even if it did it will still run.
I see designs for miniature engines where the unit of length is 1/32² and perhaps 1/64². This won’t do, we need 1/1000² accuracy for the valves at least.
Engineering runs on sizes that are in geometric progression and not on arithmetic progression. When metrication took place this was not understood or worse ignored by NPL and we have metric sized stock bar in arithmetic progression. We were better off with Imperial sizes than the unsuitable metric sizes that emerged. BA screws are metric threads anyway converted to Imperial and have the great advantage that the diameters are in geometric progression.