The long walk to find "Boscastle".
A new CAD designed rebuilt S.R. West Country Pacific No. 34039 in 3½ inch gauge
An article by Martyn
My interest in railways and locomotives, like most of us on this site, started at a young age with a Hornby HO gauge electric train set.
My favourite locos were always the British Pacifics, such as Flying Scotsman, Mallard, and the Duchesses.
Why this should be for someone from Australia I'm not exactly sure, but certainly these locos looked a lot better than the very functional Australian / American / European steam locos.
This love of steam was reinforced by several trips on Victoria's famous Puffing Billy and to the Newport Railway Museum.
This love of things mechanical led onto a lifelong passion for building things whether it was motorcycles, houses, machines, businesses or ideas; and at university it led to bachelors and masters degrees in mechanical engineering.
My working career in engineering was primarily focused on small hi-tech export orientated design and manufacturing companies, where I held a range of sales and marketing and product development roles.
Most relevantly I worked with Peter Farley on the formation of Farley Manufacturing, which designed and built state of the art CNC controlled profile cutting machines using primarily electric arc plasma but also oxy-acetylene, water jet and lasers.
The most innovative machine we designed at the time (mid 1980's) was the "Fabricator" which combined a high-power plasma with a high-power drill and automatic tool changer.
This machine allowed very large and thick plates to be drilled and profile-cut in one pass, and our first customer was British Steel.
Another Farley CNC profile cutting machine, a Trident, was used to cut the frame plates of A1 class 'Tornado' in 1994.
See the photo of Dorothy Mather, widow of Arthur Peppercorn, initiating the profile cutting of the frames for the locomotive Tornado at British Steel in Leeds on 13th July 1994.
When travelling to the UK working for Farley, I became exposed to the scale live steam scene, and decided that I would try to collect a set of the British Pacifics in 3½ inch gauge.
I chose 3½ inch gauge for several reasons, the principal ones being that SWMBO allowed them to be displayed in the house, and I could lift the locos by myself in and out of the car.
5 and 7¼ inch gauges just seemed unmanageable when dealing with Pacifics.
That said I enjoy driving larger scale locos and I have a friend in Western Australia who has a 12 inch gauge, 2/7 scale railway with a pair of Western Australian Government Railway W class 4-8-2s and a pair of South Australian Railways 400 Class 4-8-2+2-8-4 Garratt locomotives.
The power reserves of locomotives of this scale make them very easy to drive, with SWMBO taking control on a number of occasions.
Because there are many 3½ inch gauge designs for the LNER, LMS and BR Pacifics, now, after some 25 years of patiently searching the usual UK commercial dealers' websites and quickly buying when they became available, the collection is essentially complete with all the Pacifics included, except the GWR Great Bear.
All the Pacifics are finished except, as will be described below, the rebuilt West Country currently under design and construction.
Unfortunately, although machine shop experience was an integral part of my engineering degree, and despite possessing a reasonably well-equipped workshop myself, I have not had the time nor skill to build any of the Pacifics and have purchased well engineered versions from all the familiar British dealers.
A very accomplished model engineer based in Adelaide, South Australia has been employed for a number of years to bring these locos up to operating condition.
I had thought that the Aussie boiler inspectors would be difficult with regard to approving the use of these foreign boilers, but with careful calculation of each boiler's design strengths and comparison with the Aussie code, followed by careful inspection and testing, we've been able to get all except one of them approved.
A new design
Because there are no published designs for the SR Pacifics in 3½ inch gauge, these have proved to be the most difficult to find.
Much surfing of the internet looking for these SR Pacifics led me to this, 'Bulleid's Locomotives' website, and what a resource it has proved to be!
Monitoring the dealers' websites for more than a decade led finally to securing an as-built Merchant Navy which was scratch-designed and built by Len Gillet, described elsewhere on this website.
But I never had any luck with a rebuilt, either Merchant Navy or West Country class, so began considering the need to design and build one from scratch.
Reflecting on Keith Wilson's design, and all the difficulties with it described in the many model engineer's stories here, I decided that I would design, from scratch, from some 750+ original SR and BR drawings, two new West Country locomotives in 3D CAD; first a rebuilt version and subsequently an as-built (original).
The aim was always to leverage the opportunities presented by designing in 3D CAD - i.e. using the CAD component design files to drive modern manufacturing techniques; in particular 3D printing directly in metal, 3D printing of plastic patterns for casting larger components, and laser profile cutting of tab and slot designed simple brackets.
Given the very significant amount of work involved in the design and production engineering, I always imagined that I would in one way or another make the design or finished components available to other model engineers to enable them to build high quality locomotives more efficiently.
So, development of the new rebuilt design began in May 2015.
All the 3D CAD design (utilising IronCAD software) production engineering and prototype locomotive construction work is being undertaken by Phil from Newcastle, NSW.
Phil gained his Production / Manufacturing Engineering diploma from Swinburne University in Melbourne, and has nearly 40 years of professional mechanical and production engineering experience including extensive experience with 3D CAD systems.
Phil has been a model engineer for nearly 50 years, having built a pair of 5 inch gauge Simplexes and modified/rebuilt several larger locos.
The specifications for both rebuilt and as-built West Countries features the following in desired aspects in the design :
• Three cylinders with Walschaerts valve gear for the rebuilt loco, and Stephenson's valve gear for the as-built (as per Len Gillet's as-built Merchant Navy described elsewhere).
• 100 lb/sq. in. AMBSC/CE approved boiler with tapered barrel, Belpaire fire box, combustion chamber with water tubes and three element radiant superheater.
• A Steam injector, eccentric driven mechanical and double acting emergency hand pumps.
• Drop down (or alternatively, rocking) fire grate and ash pan for cleaning and emergency fire disposal.
• Easily removable combined i.e. single piece ⅔ split smoke box + door + ring, to allow ease of access to the blast nozzle, blower, internal pipework and superheater.
• For as-built locomotives, an easily removable heavily insulated single piece boiler casing
• Large, easily-removable cab roof section for ease of driving.
• Remote lubrication system comprising readily accessible pots and fixed plumbing to all internal plain bearings including axle boxes, cranks, eccentrics and slide bars.
Advancing 3D printing
As a means of speeding up the construction of the locomotive, many of the complex brackets, stretchers and other components will utilize 3D printing directly in metal.
So far several of these brackets have been trial 3D printed, first in an engineering plastic to check their design and dimensions, and then in metal as the finished part.
Currently the main metals which are readily available 3D printed in the laser-melting process (SLM), which starts with the metal in a powdered form, are alloys of titanium, Inconel or stainless steel.
We're still working on the trade-off between cost, machinability and finish quality for the two stainless steels, 316L (yield stress 460MPa vs mild steel 250-350) is not as hard/stiff as 15-5PH (950MPa).
This will make for very strong and light components, but unfortunately the stainless components are more difficult but not impossible to machine once manufactured, hence the trials in plastic first.
That said it is possible to machine stainless steel utilizing the usual cutting parameters for stainless, so as to avoid work hardening, i.e. a good sharp tool, good coolant supply, sufficiently high feed rate to always be cutting metal and low speed.
The first 3D prints in stainless steel of the motion brackets have been very successful with regard to quality, finish and accuracy.
The design of the brackets and stretchers calls for these to be attached to the frames by means of bolts and nuts, so as to avoid the need to tap the printed stainless steel.
The resultant 3D design is very detailed, so parts can be made at any scale from 3½ to 7¼ inch gauge, possibly even 10¼.
The attached figure shows the Inside Motion Bracket and Reversing Screw Bracket printed in both 3½ and 7¼ inch gauge scale.
The limitation on scale is currently a function of the size and economics of 3D printing in metal.
Currently the cost of printing the various brackets is between A$200-$400 each, which is expensive when the number of these components for a whole locomotive is considered.
However, I'm assuming that the cost of 3D printing will fall over the coming years following a typical technology cost curve.
The cost is much more dependent upon footprint size in the build chamber and the total laser 'tool path' distance, rather than on raw material cost.
The attached diagram shows the reduction in the cost of personal computers of 95% over the 15 years from 1998 to 2013.
Similar cost reductions have occurred in photovoltaics, DNA sequencing and many other technology products.
On this basis we can be reasonably confidently predict that 3D printed components will be ⅓ the current cost in 5 year's time.
Bronze Cylinders
Other major components will be cast in bronze from 3D printed patterns, e.g. the cylinders - see the photos of the raw outside and centre cylinder castings.
Based on the BR drawing, the centre cylinder in particular is very complex, and all the full size steam passages are included in the design, and therefore the casting.
Together, the design of the inside and outside cylinder raw castings (the latter are designed so as to not be "handed" before machining) took approximately 700 hours.
Although designed with prototypical BFB outer/back-side alternating 'teardrop' recesses, the wheel patterns have been initially designed without the undercut "pocket" that exist in the full-size design, so as to facilitate casting them in steel.
The driving wheels have also been designed with cast-in counterweights.
Finally, water jet or laser profile cutting will be used for flat parts, e.g. frames; with tab and slot design for suitable slightly more complex parts, e.g. bogie bracket stretcher and front stretcher (in front of leading coupled wheels).
To date the valve gear has simply been scaled from full size.
Before assembly of the valve gear begins, a simulation of the valve motion will be undertaken with one of the appropriate computer programs.
A professionally built copper boiler, which very closely matches a scale version of the full-size West Country boiler, was purchased from a UK dealer some time ago, and used to create the current simple 3D CAD model.
This purchased boiler has been inspected & tested here in Australia, and will be utilized in the build of the first production prototype locomotive.
In due course this simple model will be developed to create a 100 lb/sq. in. operating pressure boiler to meet AMBSC/CE regulatory requirements.
So far, Phil's time has been primarily focused on the design, and working with suppliers to ensure ease of production of the various components utilizing the various manufacturing techniques described.
This has turned out to be something of an iterative process with the design being modified subsequent to the production of the prototype components.
All through the design process Phil has been concentrating on three aims; the manufacturability of the parts, the buildability of the locomotive with these parts, and the functionality/usability of the finished locomotive.
A good example of this process was the design development of the reversing gear gearbox, see photos.
This involved sourcing suitable gears and designing the housing to both replicate the full-size gearbox, whilst accommodating the readily available components.
This gearbox has not yet been completed to the stage of a working prototype.
To date Phil has not spent much time assembling the chassis, which is only at a very early stage in order to confirm assembly technique and bracketry alignments.
Before moving onto the tender Phil will make a priority of assembling the chassis to an advanced stage, and sufficient to run on air.
As of end July 2019, Phil has done 5,000 hours of work on the design and production engineering.
With the design of the chassis complete (includes approximately significant 300 parts), Phil has turned his attention to production engineering including sourcing third party suppliers, refining the manufacturing techniques, modifying the CAD design accordingly and locomotive construction; which in turn has led to refinement of the CAD design so as to improve manufacturability.
Approximately 20-25% of Phil's time so far has been spent on this process.
A best guess for the time required to complete the rebuilt West Country design, production engineering and construction is between 2,000-3,000 hours, suggesting a total time of 7,000-8,000 hours.
Completing the as-built West Country variant will require an estimated further 3,000 hours.
If the model engineering fraternity indicate their interest in purchasing sets of rebuilt West Country parts, then the best guess is that first release of 3D printed and water jet cut parts for them will be in 2021.
Interested model engineers would be able to purchase as many or as few of the finished components as they desire, and then be required to machine all the usual small parts, commission or build a boiler, purchase the usual third party manufactured fittings and finally assemble, test and paint the whole loco.
Guesstimated time to build a high-quality loco with the contemplated supply of parts is 1,500-2,500 hours or approximately ⅓ of the time of a conventional high-quality build.
Through exploiting the power of 3D CAD and printing techniques, we're confident that our design of Bulleid's rebuilt West Country will fill the remaining gap in the list of British Pacific designs, and hope to soon make parts available for model engineers to build detailed and true to scale locomotives in a range of gauges.