A Great Western (livesteam) Rail Car
Mike Swain
Not so long ago I wrote an article about my making a simple steam rail car, the LNER Sentinel (Gazette Vol18, No12). Well, success went to my head and I thought it was about time for something a bit more complicated.
As so often happens, a friend, in jest, challenged me to make a working model of the Great Western No 93 steam rail motor recently restored for the Didcot Railway Centre. Eventually the planted seed began to grow.
Wanting to up the specification of a steam motor and now having a lathe in my workshop, I felt it was time to go for two double-acting oscillating cylinders with silicone O rings on the pistons and rod glands. This would make it self starting, unlike the single-cylinder Sentinel which needed a flick of the flywheel.
My initial design involved a long beam of 1 x ¼in brass with one cylinder on each side fore and aft and the drive take-off in the center, again using the belt from a video recorder.
It didn’t work. I got the porting wrong. All those drillings in the beam ended up causing a fight between the cylinder actions, both wanting to pull and push at the same time (I learn the hard way).
Out of that mistake came a much better design: the beam was shortened and the cylinders sat opposite each other with the drive take-off crammed inside one crank. I have since discovered that this arrangement has long been in existence and is called an Osmotor.
The unsuccessful long beam was made like that to accommodate the holding studs and springs for the cylinders. Now with a cylinder each side I had to find a way of pressing them to the beam; there’s not much room. One of the pictures shows how I overcame that problem.
Something else I have learned from this long beam mistake is that it takes a lot of energy to heat the mass of metal to working temperature and for a time condensation enters the cylinders instead of steam. Reducing the length of the beam helped to reduce that problem.
During the re-construction and testing of the full sized GWR rail motor No 93, the building group published a picture of the motor bogie complete without the body so, for the fun of it, I have modeled the scene. The good friend who suggested in jest that I do this build has a layout with little clearance past the signal boxes, and looking at the length of No 93 we thought a model of it wouldn’t get past on the curves. Help! What now?
His comprehensive literature on all things Great Western revealed Lot 1101, rail motor No 75, which was short enough to cope and I must say looks more sensible going round all our out of scale curves.
I used the same body side panel construction as with the Sentinel, a jig comprising of a piece of wood with cutouts where the doors would be and bits of 3mm ply where the windows are situated. Then separate strips of brass were arranged for the panels and with the window frame were laid on the jig. All this was soldered together with a small blow torch.
One feature I decided I must have was the mouldings round the windows. To achieve that I stripped 1mm lighting cable and with the earth wire I bent up little rectangles and soldered them in place.
It was a long and tedious job as each window was a different size but well worth it. I don’t construct with a lot of detail as it’s not my style nor is it compatible with steam running. Yes, I know there are those who can manage detail and excellent running and I have great admiration for them.
I expect I’m going to state the obvious to all others who build steam engines but it is something that I’m slowly learning, namely that steam economy is something serious to consider. The rate at which steam is used up depends on many varied things; first of all there is wastage, caused by: poor joints, poor sealing of glands and pistons. There is also the effect of the size of cylinders (volume) and the number of cylinders.
The Sentinel I built had one doubleacting cylinder and no O ring on the piston and that gave about 15 minutes run. The GWR rail car has two cylinders, O rings and gland seals but from that motor I only get 10-12 minutes run.
The painting and lining is my effort. Please don’t look too closely. The transfers once again were given to me by one of those much needed friends.
During its early life on the track prior to the body being fitted, the rubber-band drive (which is still going strong on the Sentinel) slipped on this loco mainly because it was of a lesser quality. All other recorder strip-downs only produced tiny notched belts that were too long; the upping of their design was no help to me.
I thought I would try the Delrin Chain and sprocket and that proved to be just the ticket. I also needed to fit an adjustable jockey wheel. Once I got the plastic sprocket pinned to stop slipping it remained trouble free for 20 hours of running; then chain stretch meant getting a new one.
Now, about a year and a half later, and as said having run for around 20 hours, (could that be 100 runs?) the rail car is back in The Works as a result of a poor design feature. The bearing, which I dare to call the main bearing, has failed dramatically; to see the surfaces you would think it was running on grit. What I had done was run a 5/16 in steel rod through the 1/4 in brass beam and hoped the forces either side wouldn’t cause too much problem. I was wrong.
I have had endless discussions with several friends who know far more than I about bearings and design for loading shafts but because of the constraints of space a compromise had to be reached. One of these friends put into my hands ball races small enough to get two into the 1/4 in beam. OK it’s not perfect but must be a lot better than the previous method. They are clamped in place and have an oil hole drilled above them. The assembly of cranks, sprocket and spacers on the mainshaft are all clamped up. I’ve done my best and at present it all works well.
Fill it up with Meths, water and oil, light
the wicks and sit back. Four to five minutes
later and it will gently move off; most
satisfying.
