Building Kozo's New Shay - John Caldwell

Written January 2004

Why build a Shay? Very good question. I'm still asking myself that. After I started getting involved with model engineering I vowed I wouldn't build a locomotive. It didn't really sit well with me. I initially got involved in early 1997 when my family and I moved to the UK and I needed a hobby I could do in a small space. Until then, I had always worked in wood (furniture, boats, that sort of thing) and been scared of the precision needed when working metal, not to mention the sort of kit required. A lathe and milling machine seemed mandatory, and beyond my pocket up to that point. However, just before leaving New Zealand, I had occasion to get an item turned for the propeller of my boat and on asking around I was put in touch with a chap who owned a lathe and who was in fact a model engineer. I had a good chat with him in his workshop and thought, "This is the thing for me".

Once settled in the UK, I was busy looking for a lathe and eventually decided on a lathe / mill combination of Chinese manufacture then selling at a reasonable price. I must say that, although the finish is rough it has served me well. I still think that I work to my ability rather than the machine's capability. The disadvantage of a single machine is that I am forever breaking down a milling setup in order to switch to a turning setup. Also the milling head is too light - there is too much flexibility from the work, through the cross slide to the lathe bed, up the column to the mill head and down through the bearings to the milling cutter. The lathe however is quite robust. After a bunch of small projects, including a fair amount of tooling, I turned to making an 8 day long case clock. The mechanism is complete. I am just waiting for an opportunity to make the case. Back to that nasty, dusty, unstable, unpredictable wood stuff again!

In looking around for a project to follow the clock, I wanted something that was a complicated mechanical system which would take a good deal longer to build. That brief sort of pushed me towards a locomotive. I had a copy of Kozo's "Building the Shay" and the New Shay had started being serialised in Live Steam. The quirky look of a Shay appealed to my sense of humour too, besides meeting the "complicated mechanical system" requirement. Other appealing features were that the plans were very detailed and the design was in metric (NZ has been metric for at least 25 years. It is a unit system I much prefer, but let's not get into that argument). The New Shay had additional features such as brakes which were not present in the original design. A friend of mine had a Shay in 5" gauge that I studied closely whenever he had it at the track. More things pushing me to make a start. The first article appeared in July 1999. I started mine in March 2000. Rarely did the parts described in an issue take two months to make, so I rapidly caught up with the articles. In order to keep working at my pace rather than the pace of publication, I contacted Kozo via the editor (Joe Rice at the time) to see if he could supply me with the plans. Kozo has very kindly given me a private copy of them although I get the impression that is very unusual for him. Obviously this was for my own use and in no way conflicts with the magazine - indeed I still get and enjoy reading Live Steam.

What has happened along the way? Well, I have a nice collection of discarded (read faulty or badly made) parts. A friend of mine reckons he re-makes about one part in ten. I don't think my discard bin is quite that full. Or maybe it is that I try and recover from some of the mistakes. Here is an example.

Before I made the wheels, I was thinking, "This is going to be a tedious task". 8 rims, 8 hubs, 64 spokes, each spoke to a close length tolerance for silver soldering to the hub and rim. However, Kozo's method really works and it only took a few days to make all the wheel blanks. Then came the fun part of turning the blanks into finished wheels. Seven wheels finished. Eighth wheel in the lathe, turning the taper of the tyre, when I was interrupted by my young nephew. In explaining to him what I was doing I ended up taking off a millimetre too much. ARRRGGGGHHHH!!!! What to do? Make a new wheel? Are you joking, after the trauma of making all those spokes? Take a millimetre off each of the other wheels? Could perhaps get away with it, but it didn't seem right to mutilate seven perfectly good wheels. In the end I figured I could turn off the flange and shrink on a new tyre. Even though you know that, you would be hard pressed to pick out the odd wheel.

So I have formed three standards by which to work:

  • Know when to throw something away and start again because it is too bad to include - it would lower the standard of the project.
  • Know when you can retrieve a bad situation. The wheel example is one of these.
  • Know when you can get away with something being not quite right. Perfection may be a desirable goal but life is too short to be absolutely fanatical.

On a long project like this you need confidence that what already exists is of a good standard. If too many doubts creep in, the project won't get finished.

I cannot speak highly enough of Kozo's drawings, and his approach to explaining how to make a part. On most plans you are lucky to get a completely dimensioned side and end elevation of a part. On Kozo's drawings you get that (with a high degree of confidence that there are no drawing errors) plus a dimetric view to aid in understanding what the part looks like. Dimetric is like isometric, but the angles of each axis are 42 and 7 degrees, rather than being equal at 30 degrees. Then you get a sequence of machining stages to help you make the part. When a new kind of operation is introduced there is a complete run down on the method. For example, there is the generic silver soldering method for which the description starts on page 28 of September 1999 Live Steam. The drawings are completely self explanatory. I usually only need to read through the text once, and from then on work only from the drawings. I couldn't have had a better teacher. Even if you don't want to build one of Kozo's locomotives I think it is worth getting some of his books just as a technical resource.

I find most turning operations are interesting and enjoyable. Building the boiler was not so enjoyable. Working with large pieces of hot copper created a certain amount of tension in me. I borrowed a propane torch from a friend with a 45mm diameter nozzle and a great roar that was ample to get the boiler up to silver soldering temperature. Getting the boiler pressure tight proved to be quite an exercise. When I was drilling the holes for the side firebox stays, twelve stays on each side of the boiler, I was thinking, "Here come 48 potential leaks". Indeed, on the first test 4 or 5 leaks were apparent. Having sorted those out, more would emerge. Round and round the loop. There were times when I thought I would never get it pressure tight. Eventually however, I managed to maintain a water test to twice working pressure for half an hour and everything remained bone dry. What a relief.

Making the crankshaft was something I looked forward to, albeit with some trepidation. With three cylinders, a crank and two eccentrics for each cylinder, plus of course the main bearings, turning is required on ten centres. I wanted everything going in my favour on this one, so the first step was to strip down the lathe and get everything adjusted correctly. Kozo specifies free cutting steel or free cutting stainless steel. I could not get a decent finish on the first piece of stainless I tried, although it was supposed to be free cutting. I eventually found a source of free cutting mild steel, and that was a joy to work. A polished finish straight off the tool. It pays to make all the ancillary tooling before you start turning. Then it is just a matter of slowly and carefully following the instructions. Some amazing shapes appear during the process. There was immense satisfaction in completing the single most complicated part in the locomotive.

Where am I up to now? My family and I have been back in New Zealand since April 2002. My workshop contents, lathe included, are well travelled. From the UK, we moved to Seattle, back to the UK then home to New Zealand. The engine of the Shay is now complete and I am about to embark on the tank. With luck it should be on the track before the end of the year. Did I say which year?

Post Script - the Shay was first steamed in May 2004. It was then stripped for painting and completed August 2004. Some photos of this model appear in Kozo's book Building the New Shay which was published after the complete set of articles appeared in Live Steam.