'Canadian Pacific' in 7.25" gauge,
a new build Merchant Navy locomotive - Intro.
by Andrew Giffen and Martyn Dix.
The late Arron Homewood has a lot to answer for, and so does standing around at the Guildford gala in July 2023, chatting about a wish list for a new-build locomotive.
At the time, I had just brought my latest creation down to the gala, the 9 foot long South African 15F, to be delivered to its new owner Remo Bussola, and make its way back to Switzerland.
Murphy's law had the new loco drop a gudgeon pin, and move a return crank! - but it went home anyway and justified a very pleasant later remedial visit to the Winterthur area, (I wonder how many model engineers have packed a selection of tools and explained it all to security at Luton airport? - pic below: our test run at the stunning track in Staufen).
But those mishaps didn't prevent a very interesting conversation between Arron Homewood, Martyn Dix and myself, around the question… what next?
Now from my side, I had a leaning to the BR standard class 9F; yes ok, there are plenty about, but knowing the 15F was built pretty much at the upper limits of my mill and lathe capacity, I knew I could turn down 8" diameter castings.
A 7.25" gauge, 1:8 scale 9F has the same diameter wheels as the 15F, and is similar in many ways, so I was confident I could predict all those things that sensible buyers would be wanting to know.
As a professional builder, I find it is wise to avoid unknowns where possible - they make it hard to set expectations of time and cost, and expectations mismatch causes problems as ever.
Martyn was fairly keen on the 9F, and sensible, until Arron started talking about the Merchant Navy.
Of course, the modeller can choose between the original 'Spam Can' and the rebuilt version, but we more or less unanimously preferred the look of Jarvis' rebuild, not to mention the same advantages that justified the rebuild apply just as much to the model - amongst others improved access to all the valve gear, especially the outside obviously, and using conventional valve setups all helps to keep things a bit more predictable.
So suddenly, we were in the realm of not-so-sensible driving wheel diameters that my lathe could not handle, a crank axle, something I'd not tackled before, and another cylinder.
But wow, what a challenge, and it would push my model engineering up a level or two.
Also, as far as we know, there is only one other rebuilt version in 7.25" gauge out there, 'Clan Line', built by John Morgan and now owned by Andy Gelson, so it would be a rarity in the extreme.
So that gave some food for thought on my drive back up to Lincolnshire.
Further chats with Martyn started setting up our very effective modus operandi, which was to stand us in good stead when it came to lateral thinking to prepare for the unknowns this would mean.
As you can imagine, both of us needed to have the same mindset, of where on the scale between gold medal winner scale museum replica at the one extreme, and cheapest possible caricature on the other.
It was an easy one, in that we both naturally lean towards very accurate external appearance, although open to consider changes where they make sense from a practical point of view.
Whilst internally, freelance to favour robustness, speed of assembly and ease of maintenance, without going so far as to drop the middle cylinder as had been done before.
Needless to say, there are no modeller's drawings of this out there for 1:8 scale on 7.25" gauge.
There is, of course, Keith Wilson's 5" gauge "Ariel", of which numerous examples are around, but just as well known is that the design has a lot of errors - this web-site
shows 11 detailed pages of them.
As luck would have it, I got hold of a very complete set of SR/BR drawings; in fact, too complete, as they were over 4000 drawings of all the versions mixed in together of both Light Pacific versions and Merchant Navies, and streamlined and rebuilds of them all, but at least I found they were all there after a lot of sifting through a very interesting collection, and historical record.
Here a small sample of what I faced:
I wondered whether Keith Wilson was in possession of the same bewildering set of drawings, and that might explain how the "Ariel" design is a mix of MN and light pacific components.
After having a good look, I was able to whittle down over 4000 drawings to 208 for the locomotive, into 26 folders arranged by the sections, that I find logical when building a loco.
The sections are assemblies, such as the complete main frame, stretcher and motion bracket assembly, which are broken down into sub-assemblies and usually also sub-sub-assembles and so on, down to each component part.
At this stage there was no longer any doubt about using the Keith Wilson drawings, not only would I have to consider and correct all the known errors, but also scale everything up from 5" to 7.25" gauge.
So, there seemed to be no disadvantage in simply using the works drawings, not to mention the enjoyment of going to the source.
So, it was time to begin.
Martyn's investment was going to be considerable, and what we both needed was peace of mind that the parts we made and bought would fit, and the engine would run.
Every experienced builder knows how often things have to be remade, and when the project is seen as a hobby, that is OK, time is often not a factor, parts can be binned and remade trial and error fashion, and a loco can be built over 20 years.
It is a little different for me, although there is some leeway built in, my time has to be paid for, and mistakes are painfully reflected at the bottom line.
Industrial practice would be a prototype, but here I am building a one off anyway, and the prototype IS the production model, so thank goodness for 3D CAD.
The more effort you put into faithfully transposing everything into 3D CAD at this stage, the more the 3D model will serve as a valid prototype, and the more time and trouble it will save you once the metal arrives.
Ideally even bolt heads and nuts should be modelled and placed in position on the screen - why? you may ask.
CAD has one big downside - it will happily allow you without any warning to blindly copy a bolted bracket between the frames that a scaled-down fitter would be able to access by climbing in, and it is all too easy to draw that without realising that it will be difficult, if not impossible, to get the bolt there on the model or tighten it up.
So modelling in 3D is not just copying from the original drawing, there is a lot of simplifying and thought given to how to put things together in the workshop.
My original estimate for the loco build was 5 years, based on my experience of 4 years with the similar-sized, but much less complex 15F.
The first year of that 5 certainly could be taken up by non-stop drawing work, but I wanted to start having something to show for our efforts in metal after 6 months or so, so the drawing marathon would finish at the running chassis stage.
In effect, the main frames, all 3 cylinders, driving wheels and all the rods and motion running as a simulation on the screen at all cut-offs, in forward and reverse.
This little account would get very boring were I to show each stage of the drawing process, but after about 6 months, the simulator looked something like this:
So, with the main function of CAD being to allow you to make sure everything fits before making anything, the logical next level is to cut away the steam chest, and "manually" turn the driving wheels degree by degree to plot exactly what is happening in each steam chest.
So of course, I isolate each cylinder in turn and chart steam admission, expansion (port closing to steam) and exhaust moments for the revolution.
Were they all perfect from the start?
Nope, but close.
No explanations are needed for the huge advantage in being able to make the necessary changes to the interlinked dependencies of return crank/eccentric position, length and eccentric rod length, to achieve pin point accuracy.
I took my time at this stage, as this is the most critical part, and worth being ambitious about perfect valve events.
A video clip of the simulator in motion can be viewed by clicking here.
So, 6 months down the line, the computerless model engineer might well have had the main frames already complete, while we still had nothing in metal, but the many unfinished model Bullied pacific chassis gathering dust around the world are at least in part testimony to the importance of preparing the groundwork.
I can at least promise a lot more photos in the next instalments, as things start to come together.