A Novel Steam Motor Development (continued)

By Ray Rolt

A DETAILED ASSESSMENT

Having described this in broad outline, we will now look at its concept and construction in greater depth. For many years, I have been constructing small live steam locomotives in a small workshop consisting of a bench, vice and vertical drill. This was made possible by buying in finished components such as cylinders or steam motors and boilers.Thanks to modern CAD, a variety of configurations can be designed for the main fabrication unit, to suit individual requirements, and used for the production of laser cut components. These can then be fabricated and machined to receive the cylinder units bought ‘off the shelf’, and the other moving components. The configurations can be as multiple ‘in line’, ‘V’ or ‘radial’units, and thanks to the slotted piston with gudgeon pin, conventional ‘

When I decided to build a large passenger hauling live steam locomotive, I was confronted with two problems. These were the cost and extra space required to install a machine shop, with lathe, milling machines, etc. and the time scale required to learn how to operate them competently and then make all the components needed. Being retired, I have a more limited time to make this locomotive if I am going to have a reasonable period of operating enjoyment when it is finished.

One obvious solution is to build one of the excellent kit locomotives now available, with the fully machined manufactured components only requiring simple assembly, delivered and paid for in instalments to spread the cost. The main disadvantages, apart from the cost, are being limited to what is available and the lack of personal creative input that I have enjoyed when building my small locomotives.

RATIONALISATION

Here we will look at the main components that make up the finished Locomotive. These are:-

a) The Rolling Chassis, which consists of the complete main frames, including frame stretchers and buffer beams (and dragbox if a tender locomotive) and complete wheel sets, including axle boxes, coupling rods and spring gear.

This can be made in a simple workshop as all the main frames, stretchers and buffer beams can be marked out and cut by hand, including the normal cut outs for the axle boxes and drilled using the vertical drill. Assembly can be done using nuts and bolts, riveting or welding. In the case of welding there is the risk of distortion. An alternative method is to prepare drawings and have the frames, etc., laser cut by one of the firms offering this service. Items like buffers and drawgear are readily available.

Complete finished wheel sets will have to be purchased, complete with ball race bearings. These are used in conjunction with swing arms or compensating beams to eliminate the machining work needed to fit normal horn guides and axle boxes. It is unlikely that the bearings will need to be replaced in normal use. The size of the frame cut outs may have to be bigger to accommodate this arrangement. Firm rubber suspension is required to limit the movement of the swing arms/compensating beams, as they move in an arc and not in parallel like normal axle boxes. This is because the wheel centres will vary in relation to each other requiring increased bearing clearance in the coupling rods to prevent binding. This also applies using normal axle boxes to some extent.

The coupling/connecting rods can be marked out, drilled and shaped by hand. Any fluting can also be done by hand, but milling out by a friend willing to do it will save a lot of time. Thick walled brass or bronze tube, cut to length and drilled and reamed to suit the crank pins, can be used for the bearings with the holes in the rods a light press fit, allowing them to be inserted using the vice, or a loose fit for securing with ’Loctite’of suitable grade.

b) The Boiler, which can be scratch built to a suitable published design for the locomotive being modeled, built using flanging blocks/ready cut components/complete set of flanged and rolled material supplied by the trade (if prototype selected to suit what is available), or professionally built (preferably to a published design where suitable flanging blocks may be available and design preparation can be avoided). This can be done in a simple workshop, with the silver soldering done in the open or by a friend/professional.

c) The Power Unit, based on the ‘Rolbro’ Motor. In a conventional model, this is the component which involves the most machining complexity and accuracy, the main item being the cylinders. By using the ‘Rolbro’ Motor, this can be eliminated, only requiring some form of drive to a coupled axle. Alternatively, the cylinder/piston assembly can be used in a direct drive form as suggested under Research and Development.

d) Finishing and Superstructure, this involves the final assembly of the locomotive, including all the pipe work and fittings (all the valves, gauges, injectors, pumps and safety valves, etc. can be supplied by the Trade), fabrication of the smoke box, ashpan, boiler cladding and general superstructure/bodywork. All this can be done in a simple workshop. To bend up sheet I use two lengths of angle iron of suitable length bolted together at each end with a bolt and wing nut to form vice jaws, screwed to a carpenters ‘horse’, a ‘Workmate’ would be a good alternative.

The above breakdown into units equally applies to a traction engine or ‘Overtype’ steam lorry, with two cylinder/piston assemblies housed in the ‘cylinder block’ driving onto two overhung disc cranks set at 180 degrees giving the equivalent of a single double acting cylinder, using ‘balanced’ vertical slide valves at the front in place of the ‘uniflow’ exhaust. A pinion on the crankshaft would drive the flywheel/driveshaft. Obviously some liberties have been taken and it would not be self starting, but it would enable something to be built in a shorter time in a simple workshop.

Similarly, the complete ‘motor’ could be used in an ‘Undertype’ steam lorry or steam boat.

THE ‘ROLBRO’ MOTOR

LUBRICATION

Back in the late 20s, a firm called ‘Bowman’ put on a demonstration at a national show in London of their ‘0’ Gauge live steam locomotives by running one continuously for the duration of the show, during the opening hours, in which time it ran quite a number of miles! Though of a crude basic design, it was very robust with a large over scale boiler giving a running time of over 40 minutes, when it was refilled with meths and water and lubricated ready for another run. The secret of its success was the lubrication!

It had two long stroke single acting oscillating cylinders driving a single pair of driving wheels with the cranks set at 180 degrees, being of 4-2-2-0 wheel arrangement. The pistons were plain plug type but with a felt plug extension at the outer end. This was well lubricated at the end of each run and was the only form of lubrication to the cylinder. By maintaining a continuous oil film on the cylinder wall, it made the piston steam tight and well lubricated.

This demonstration has always impressed me and when my thoughts turned to designing this steam motor, I decided to use the idea in a more sophisticated form. The fact that the cylinders are single acting allowed this to be done. As the oil reservoir is accommodated within the piston sleeve, I decided to use a lamp wick to retain the oil. To minimize the risk of fraying, the depth of the groove is slightly deeper than the ‘dry’ thickness of the wick, which will swell up with the oil and give a continuous film of oil on the cylinder wall. As the wick is woven to form its width, there are no cut edges to fray. This will be cut to length to suit and stitched in place!

With the piston on B.D.C., the wick is liberally oiled via the top oiling point on the cylinder, which can be in the form of a cup with spring loaded cap or a sight glass form of reservoir. By using the wick, the oil is retained and will be concentrated in the lower half, where the wear will be greatest, with oil being ‘wicked’ up to the top to maintain a continuous film of oil to the cylinder wall. The slide valve will be lubricated with steam oil in the normal way.

One problem is that the depth of the oil reservoir is governed by the ‘O’ ring size, but this could be accommodated by turning down the piston further for the length of the sleeve, leaving a step up for this wide enough to allow the ring to roll the required amount.

THE CYLINDER

With conventional cylinders, these are normally in the form of a casting (alternatively they are sometimes prefabricated) which are bored out, the flanged ends machined and covers turned up to suit, passage ways machined between the ends of the bore to the port face of the piston valve or slide valve, which have to be machined along with the valve chest. The piston and rod are then turned up and the gland box in the rear cover machined. This has to be repeated for the number of cylinders required.

Having seen what can be produced automatically on a CNC lathe at Andy Webb’s Workshop, I decided that the whole cylinder had to be made in a single operation to make full use of this facility. This would allow quantity production of consistent quality at a very competitive unit cost. How could this be achieved?

My first car was a 1953 Jowett ‘Javelin’, which had a flat four engine with wet liners. Why not use a thick walled liner as a cylinder secured by the cylinder head, I thought? It would mean that it was a single acting cylinder, but this would mean that there was no packing gland to make and maintain.

The result of this has already been covered in my initial description. Having now had the opportunity for further thought on the design, I propose to make an amendment that would have an advantage. With the original arrangement, to remove the ‘cylinder’ it would require the removal of the valve chest/cylinder head. By using a separate clamp plate, with the flange recess machined into it, bolted to the cylinder head plate, removal of the bolts would allow the complete cylinder/ piston assembly to be removed. By doing this, the machined clamp plate can be part of the cylinder- piston assembly. This means that the cylinder head plate now only needs to be drilled to suit the holes .

By producing the cylinder in this way, the ‘uniflow’ exhaust and oiling hole will require a separate machining operation. This can be done in the home workshop using a ‘V’ block and clamp and the vertical drill, and would also allow the use of a balanced slide valve instead. Care will be required in finishing the bore after drilling. Alternatively, it can be supplied ready machined as an optional extra.

It will be noted that the ‘uniflow’ exhaust is only on half of the cylinder circumference. This is to ensure the structural integrity of the cylinder and allow it to be rotated to direct the exhaust in any direction. In the ‘motor’, this would be handed to discharge into a central connection. Because of this, the oiling point needs to be drilled to suit. Assuming that a side outlet is being used and the holes are pre-drilled, they can be drilled straight through and tapped, with a plug screwed into the bottom hole of each cylinder.

The simplest way to form the exhaust opening would be to cut a slot using a thin cutting disc in the milling machine, which could not be done in a basic workshop, but this would damage the wick and ’O’ ring, hence the drilled holes to form bridging across the gap.

The exhaust timing can be advanced by forming a chamfer to the piston crown.

On my “Javelin”, the bottom flange of the wet liner had a copper ‘Wills’ ring as a seal, which raised the top flange above the aluminium cylinder block. When the head nuts were tightened, the ring was compressed and prevented the coolant from contaminating the sump oil. I decided to use a similar seal for the ‘motor’ cylinder, in the form of a simple soft annealed copper ring made from thick wire.

This is used on “Braunton”, a Bulleid light pacific that I have worked on as a volunteer for over eleven years, helping to restore it from a ‘Barry wreck’ to full working order on the West Somerset Railway. Here they are used as a seal to the piston rod gland box end covers in the form of two concentric rings housed in half round grooves in the cover flange. The annealed copper wire was pressed into the groove and a scarved joint formed, which was filled with soft solder and filed down flush.

On the ‘motor’, this can be made in a similar fashion in the chamfer formed on the cylinder flange. When the clamp bolts are tightened in a diagonal sequence, the ring will be squeezed into the recess until the cylinder head is in contact with the flange face to give a secure fixing. With the modification now suggested to enable the cylinder to be removed from the head plate, a new ‘ring’ will have to be made each time this is done. An alternative could be to use a suitable ‘O’ ring which would be reusable, but this may not give such a positive mechanical fixing for the cylinder.

When used in the ‘motor’, I believe that this cylinder design will prove satisfactory as the forces on the pistons are linear, thanks to the use of the ‘scotch crank’. It will be interesting to see if this is still the case with the use of direct drive using a connecting rod, with the variable angular forces on the gudgeon pin and longer stroke! Worth a try I think.

THE VALVE CHEST

This will be fabricated in brass as a complete unit with an external flange for bolting to the cylinder head plate and will extend up to accommodate the drive from the valve rod. The drive to the slide valve plate will be via an eccentric on a cross shaft within the valve chest and extended at one end, via an ‘O’ ring seal, with an arm driven by the valve rod. The eccentric will be drilled and tapped on one end and secured by a lock screw, to allow for setting the valve. Effectively a form of ‘bell crank’.

The sides of the valve chest will be thick enough for drilling and tapping for the fixing bolts for the valve chest cover, which will have a light bronze spring strip on the inside to hold the valve on its seat.

The fixing flange for the valve chest will be a brass sheet forming a ‘floor’ onto which a brass/bronze valve seat will be soldered, with the single port drilled and filed to connect with the port formed in the ‘cylinder head’ plate.

THE VALVE GEAR

As already indicated, this will be Hackworth Valve Gear. Many years ago, I had seen the articles by L.B.S.C. in the ‘Model Engineer’ on building “Mona”, a 3 1/2” gauge 0-6-2T, and bought a reprint of these at a TME auction night a few years ago.

The use of disc cranks on the driving axle as an eccentric drive for the Hackworth gear was an interesting idea, though I had reservations about the valve events with such a short drive rod, due to the boiler restrictions, and the imbalance of the crank discs. Not having seen an example in operation, on making enquiries in the Club I found two members that had owned or operated examples who said that they ran very well. My valve gear will be built generally as the article in terms of the dimensions and arrangement.

The late K. N. Harris considered that good valve events could be obtained, in his articles on radial valve gears, but axle movement affected this. This is not a problem in the case of the ‘Rolbro’ motor. By fortuitous coincidence, an article has just been published in ‘Engineering in Miniature’, by Simon Bowditch, about improving this gear on the ‘Sweet Pea’ 5” gauge locomotive. Having improved events, he then goes on to advocate the use of ‘Woolf’ valve gear to make further improvements.

This is a derivative of the “Hackworth” with an offset drive pin on the drive (eccentric) rod for the valve rod and an unequal pivot arm introduced into this rod to give greater length of valve travel. This gear was developed in America and extensively used by their traction engine manufacturers and has improved valve events.

When developing the ’Rolbro’ design, I proposed to introduce a ‘bell crank’ in the valve rod to drive the vertical slide valve, in the form already described.

An eminent C.M.E., Churchward I believe, has been quoted as saying that it is easy to get steam into the cylinder but more difficult to get it out again! This is an inherent problem with slide and piston valves in getting good admission events and a free exhaust release. The ideal solution is to control inlet and exhaust separately which can only be achieved by poppet valves, as used in the ‘Duke of Gloucester’, or other equivalent, or ‘Uniflow’ exhaust as proposed in this instance. This means that a simple slide valve is used, just to control the steam admission.

Effectively, the valve gear could now be ‘Woolf’ gear with the offset pivot, and “bell crank” arrangement as the pivot arm. However, the restriction of the eccentric would not allow the longer valve travel, which is not needed, as stated above. The setting of the valve would be done visually with the chest cover off.