Figure 1. The prototype spring and axlebox of a McConnell tender (photo courtesy Harry Jack)

I soon discovered that the usual suppliers had nothing that remotely resembled this. Not even a spring casting, still less the combination of spring, hangers and axlebox. I was building two tenders so I needed a dozen of them. When I started to think about building them from scratch was when I began to lose the will to live. The first one would be a challenge to get it looking right, the second one would have to look just like the first … by the time I got to the third or fourth, even gardening would start to be more attractive.

The usual solution would be to make a master and cast sufficient of them from it. Some years ago I experimented with casting white metal in rubber moulds, but decided that to achieve acceptable and consistent results one must do it as the professionals do, vacuum degassing the rubber mould while it sets to remove all air bubbles and centrifugally casting to ensure the white metal fills the mould completely. Which all started to get a bit complicated. And on top of that, let's be honest, I loathe working with white metal. It is dimensionally unstable, the surface finish is poor if the mould is less than top quality, and above all, I cannot solder it. I expect you can. Lots of people tell me they can and some of them have explained, usually at length when it is near closing time, how they do it. But it does not work for me. No, if it has to be cast, I want a ‘proper’ metal like brass.

I leave casting brass to the experts with the right equipment. Unfortunately, those same people, at least the ones I talked to, were not to be convinced that there was an unmet demand for models of McConnell tenders and that the entire gauge O community would come beating on their doors for them. My paltry order of a dozen was not going to be cheap.

This is where ideas about 3D printing started. I started reading up on the subject and first I had to get past a whole new vocabulary of words like ‘makerspace’ and ‘thingiverse.’ I needed to move beyond people who were all making the same strange looking boat model and some rather derivative fantasy figures. I also learned that desktop printers don't always produce good results and often need some careful setting up and experiment. And they work with plastic or resin materials. I prefer not to, but I did not completely write-off these materials at that stage. Industrial 3D printers work with a lot of other materials including steel, copper and brass, but they operate at high temperatures, use high power lasers and cost somewhere north of more than you can afford, sonny. Not something to run in the spare bedroom. But there are companies operating on-line who will take your uploaded solid model and turn it into a component in one of many materials, including the aforesaid metals.

That was worth investigating. I created a solid model of the combined spring and axlebox in Fusion 360 and checked it over carefully against the published limits of what can be made in different materials. A few links and pieces had to be carefully thickened up to allow for this. Those spring leaves were problematic, they were just too thin for any 3D printing process that I could find. Reluctantly, I reduced the number of leaves so that the thickness of the remaining leaves could be increased. (I have been accused of being a rivet counter. Now I am a spring leaf counter). To do a trial print, I chose Shapeways. There are lots of others to choose from, but Shapeways have no minimum order and get generally good reviews. I ordered one print in brass, and just for comparison, one in a plastic material that was about half the price. I uploaded my file, paid the bill, and sat back and waited for my components to arrive.

Figure 2. Trial printing in brass. The light from this angle catches the layer marks

On first inspection, the way that they were laid down in layers was evident in both materials. The brass component had layering that looks quite pronounced on the spring itself, showing up as vertical lines in Fig. 2, but not so much elsewhere. However, those lines were more or less apparent depending on where the light was coming from. They also became a lot less evident when it was painted, but at the same time the divisions between the spring leaves also diminished. The plastic component was a lot worse, the lines between the spring leaves being hardly visible (Fig. 3), and that was because the plastic could not resolve the spring leaves properly, not because the dividing lines were too thin and shallow. The trial prints were an interesting and useful experiment, and if nothing else stopped me handing over a wodge of cash for a desktop printer that would only work with plastic.

Figure 3. Trial printing in plastic. The detail reproduction is poor, but that is because this model is pushing the limits of what can be achieved with this material

Feeling sufficiently encouraged about the brass material, I tweaked the solid model to make the divisions between the leaves broader and deeper and to increase the size of the rivet heads slightly. I sent off another order for a dozen in brass. In truth I was really chuffed when they arrived. Not only do they look the part, but dimensionally they are absolutely accurate and went in place with no filing or fettling (which, in my experience, rarely happens with castings). Fig. 4 shows the Mark 2 version exactly as received with no hand finishing at all. Fig. 5 is the tender with a full set of fittings.

Figure 4. Final printing in brass. The detail of the spring leaves, links and rivet heads is much improved

There are fewer spring leaves than there should be, but I think I pushed the detail as hard as the printing process would allow. That is the only real compromise and I am very happy with the result. I know the question you are all asking: how much do they cost? All in, about £12 each.

Anyone still reading? You might as well stay with it, we are near the end now. In fact, that price is not dissimilar to what a casting would cost for this project. Where castings win are in numbers, if you buy a hundred or a thousand the unit price drops considerably. With 3D printing, there is little or no discount for quantity and so for very short runs it is competitive. The price is largely based on the mass of material used and the type of material. I would think very carefully about using it for scales larger than ours, or even for large O gauge fittings. Brass is actually one of the cheapest metals for printing. I had more expensive quotes for copper (it is a more expensive material to start with) and steel (it melts at a higher temperature and requires more power). The plastic trial part was about half the cost of the brass one, but the quality just was not good enough for what I wanted and could not be improved without simplifying the detail considerably.

I know there are many people out there who will consider this is far too expensive, but, as the saying goes, you get what you pay for. If you want a high resolution of detail and dimensional accuracy, it will cost more. I was pleased with the result and 3D printing in brass is a process that I will use again in the future. Selectively: if there is a good quality casting available from one of our suppliers of course I will use it, but if there isn't, 3D printing is something I will consider carefully.

Thanks are due to Harry Jack who provided the prototype photograph, and Warren Haywood who painted the test piece for me.

Figure 5. The tender frame with fittings attached. I was very pleased with the result