Jump to content
 

MikeOxon

Members
 View Profile  See their activity

  • Posts

    3,369

  • Joined

  • Last visited

 Content Type 

Blog Entries posted by MikeOxon

  1. MikeOxon

    Paddington Station 1840

    By MikeOxon,


    General
    Brunel’s great arched roof is to many people the epitome of Paddington Station but this was not built until 1854. The passengers who first travelled on the line from Paddington to Maidenhead, which opened on 4th June 1838, started their journey from a far less imposing structure – little more than a collection of wooden sheds.
     
    A London terminus for the GWR was needed in a hurry, after negotiations with the London & Birmingham Railway for a joint terminus at Euston broke down. With the line to Maidenhead almost ready, the GWR Directors desperately needed to start generating passenger revenue and, since authorisation for a route into Paddington was only agreed by Parliament on 3rd July 1837, there was no time for grand designs!
     
    A quick solution was to build the station offices into the arches of the new Bishop’s Road Bridge and provide simple wooden platforms to the West of this bridge. Goods facilities were established on the other side of the bridge, alongside the Forecourt from where passengers entered the booking office, under one of the arches. The great artist of early railways, J.C.Bourne, produced a lithograph of this façade of the station. I have annotated the locations of the passenger facilities on his illustration, as shown below:
     

    Paddington Station 1843 by J.C.Bourne (colourised Mike Flemming)
     
    I have not found any early illustrations of the layout of the platforms beyond the bridge but there are several early drawings, mostly in poor condition, that provide plans of the tracks and platforms, as well as details of the platform canopies. A selection of these drawings is available on-line, in the ‘Historical Engineering Collection’ of the Network Rail Corporate Archive (NRCA).
     
    3D Station Drawing
     
    I decided, as an experiment, to see whether I could use ‘Fusion 360’ to create an impression of the original station by using the techniques of extruding from drawings, just as I have done for my models of rolling stock.
     
    The NRCA drawings include sketches of various alternative proposals, from which I chose the plan view in NRCA161183 as a suitable base from which to create a 3D model. First of all, I had to digitally ‘clean’ the original drawing to create my working version, shown below.
     

    My annotated ‘working copy’ from NRCA161183
     
    There are many features of the track-work that seem strange to modern eyes. Note, in particular, the widespread use of wagon turntables and traversers for moving the small carriages and wagons of the time between tracks. Several tracks ended in carriage loading ‘shoots’ at the end of the central carriage road, between the Arrival and Departure platforms. This arrangement is shown in a lithograph of, Slough Station by J.C Bourne, of which I show an extract below:
     

    Slough Station (detail) by J.C. Bourne
     
    I also found a less detailed re-drawn plan of Paddington Station, which provided the all-important scale, in the book ‘Paddington Station - Its history and architecture’ by Steven Brindle, published by English Heritage 2013.
     
    With this additional information, I could import the NRCA plan into ‘Fusion 360’ as a canvas and use the ‘calibrate’ command to adjust it to the correct scale. I decided to work directly in ‘feet’, since these units are used throughout the NRCA elevation drawings.
     
    Bishop’s Road Bridge
     
    My first 3D extrusion in ‘Fusion 360’was from the plan view of Bishop’s Road Bridge (shaded pink in my annotated version, above). Initially, I extruded the rectangular ‘body’ to a height of 30 feet. I then compared my structure, marked with the locations of the various arches, as indicated on the plan, to check the proportions against the Bourne lithograph. (It is known that Bourne used a ‘camera obscura’ as a drawing-aid, so I was confident that his illustration is accurately proportioned.)
     

    Steps in creating my model of Bishop’s Road Bridge
     
    Once I had made sketches by tracing over the Bourne canvas, imported into ‘Fusion 360’, I could overlay these sketches onto the face of my rectangular Bridge ‘body’ and extrude the various arches.
     
    Platforms
     
    Details of the platform canopies are shown in drawing NRCA161326, of which my ‘cleaned up’ version is shown below. This drawing shows the end elevations of the canopies over both the Departure and Arrival platforms, together with details of the cast-iron support pillars.
     

    My ‘Working copy’ from NRCA161326
     
    As before, I imported this drawing into ‘Fusion 360’ and sketched the outlines of the roof trusses. I also created a model of a single pillar by drawing over the profile and then using the ‘revolve’ tool to create a cylindrical ‘body’, as shown below:


    Using ‘Revolve’ in ‘Fusion 360’ to create 3D-model of Pillar
     
    Once I had a single model of a pillar, it was simply a matter of using the ‘Pattern on Path’ tool in ‘Fusion 360’ to create the array of pillars shown on the NRCA plan of the platforms. Note that I have raised the platform surfaces and carriage road by 3 feet, above the level of the track bed.
     

    My 3D-model of the Arrivals Platform with Pillars and a single roof truss
     
    I was very pleased to find that the dimensions derived from the elevation drawings corresponded very closely to those derived from the plan view, indicating that my ‘calibration’ in ‘Fusion 360’ had been successful. After duplicating the trusses as required, again by means of the ‘Pattern on Path’ tool, I added roofs by extruding from the profile set by the trusses at one end of the structure.
     
    The end result of this stage of my modelling is shown below:
     

    My 3D-model of Paddington platforms viewed from the West
     
    Of course, the advantage of having a 3D-model in the computer is that I can choose to view it from whichever direction and in whatever detail I choose, for example:
     

    View across my model of Paddington Station from above the Forecourt.
     
    The above view demonstrates the sharp angle between the platforms and the approach road, by which passengers arrived at the station. Carriages could proceed through the central arch to reach the carriage dock set between the Arrival and Departure platforms. At that period, wealthy patrons of the railway loaded their carriages and their horses onto trains and, in some cases, chose to travel inside their own carriages, rather that the coaches provided by the railway.
     
    Adding other models
     
    Once the basic 3-D model was in place, I realised that it was perfectly easy to add some of my existing models into the scene. The ‘Insert Derive’ tool in ‘Fusion 360’ allows model data to be imported into a design from other folders already held in the computer. To demonstrate this procedure, I imported my existing models of Brunel’s Britzka and of a Carriage truck.
     
    I realised that these models had been dimensioned in millimetres, as 4mm-scale models. When first imported, they appeared at their small ‘model scale’. I could, however, select these model bodies and use the ‘Scale’ command in ‘Fusion 360’ to increase their dimensions by a factor of 76, so that they conformed to the overall building model. After re-scaling, I could use the ‘Move’ tool to position them as I wished within the Station, as shown in the example below:
     
     

    Loading my model of Brunel’s Britzka onto a Carriage Truck at Paddington Station
     
     
    Taking this idea a little further, I imported some other models, including my Horse Box, a horse, and the locomotive ‘Vulcan’, to create the following scene:
     

    Brunel’s carriage and horses being loaded for travel from Paddington Station
     
    Conclusion
     
    As I wrote at the outset, this is all experimental and I have had a lot of fun exploring the possibilities of scene modelling in ‘Fusion 360’. I realise that I am venturing into the territory of digital ‘Train Simulators’, which @Annie demonstrates in her thread can be very impressive and allow you to drive the trains as well.
     
    My own modelling has allowed me to bring to life some long-forgotten scenes and I intend to continue by re-creating some of the other buildings around the old Paddington Station. Apart from the offices within the Bishop’s Road bridge, there was also a remarkable ‘round-house’ engine shed, designed by Daniel Gooch, a Carriage shed and, on the other side of the bridge, the entire Goods Station, with sheds and offices.
     
    I think all this can keep me occupied and entertained for some time to come 🙂
     
    Mike
  2. MikeOxon
    Some time has elapsed since I first decided to paint the wagons red on my GWR 'North Leigh' railway but I still look out for any further information that may shed further light on when and where this colour was used.
     

     
    My 1st edition copy of 'Great Western Way'(GWW) stated: "it would seem to be about the end of the 1870s that wagons and vans first sported a standard pattern of painted lettering. It was about this time, too, that the change from red to grey as a body colour was apparently decided upon." Since that was written, however, opinions have been revised and many now think that red was still in use up to 1904. (the way things are going, I'm sure someone will eventually conclude that red continued until nationalisation, when it was renamed 'bauxite' )
     
    My old GWW also states that the earliest lettering took the form of the letters "G.W.R",...on the left hand end of the vehicle usually three planks up from the bottom of the body side" The text continues by stating that the lettering soon moved to the bottom plank, with the legend "To carry ... tons" above the initials. At the opposite end (RHS), the wagon number appeared, with the tare weight above. Then, in the early to mid 1880s, the pattern was reversed, with G.W.R at the right-hand end on the second plank up and the number and tare at the LHS. Legends for load and tare were now below the letters and numbers. Numbers also began to appear on the ends of wagons. GWW also reports that, in 1894, the use of cast plates for ownership and number information started to be applied and continued until at least 1905.
     
    With that received wisdom in mind and some knowledge of later opinions, I happened upon an interesting photo in Ian Pope and Paul Karau's book 'The Forest of Dean Branch - volume one' This photograph shows Cinderford ironworks and is believed to have been taken in the 1890s and certainly after 1880, because it shows a structure completed in that year.
     
    There is another photo, also taken in the Forest of Dean, in 1883, of a wagon at Coleford in what is described in GWW as "the earliest style of painting with lettering on the solebar".
     
    A selection of wagons appear in the foreground of the Cinderford photo, in a considerable mixture of liveries. For research purposes, I have copied two small sections of the photograph, to illustrate this point.
     

     
    In the upper row, five wagons can be seen, of which two appear to be of a darker shade than the others. Taking the two 'dark' wagons first: the one on the left has G.W.R low on the LHS with (presumably) load above, while the number and tare are to the RHS. The other dark wagon has the G.W.R above the load, to the left, and a number (level with the G.W.R) to the RHS, with some addition small lettering above and to the right of the number. The lettering on the 'lighter' wagons is less distinct but G.W.R always seems to be on the left side.
     
    In the lower row, the first wagon on the left is unclear but, next, is a dark-looking wagon, with G.W.R on the left and other lettering in the centre and RHS. Next is a lighter-looking wagon, with G.W.R at bottom left and (presumably) load above but what might be the tare seems to be below the number on RHS. The next wagon is not clear and the final wagon has G.W.R high on the left, with load below and number on RHS, though not on the visible end of this wagon.
     
    I have noted that there is a clear distinction between 'light' and 'dark' wagons and that the shades of these two types seem consistent, between the various members of each type.
     
    At this point, I decided to try some experimental archaeology.
     
    I set up a line of model wagons from my collection, some painted grey and some in my interpretation of GWR red. After photographing the group, I processed the photo using a Photoshop plug-in that allows a Black and White conversion, with an adjustable colour response. I chose a 'colour blind' response, typical of 19th century film technology, when film emulsions responded to blue, violet, and ultra-violet wavelengths but only very slightly to green and not at all to yellow and red.
     

     
    My experiment shows that the model wagons divide into 'lighter' and 'darker' shades, as in the Cinderford photo, with the red wagons being darker.
     
    If I extrapolate from my experiment to the Cinderford photo, it suggests that, at the time of the photograph (known to be after 1880 and probably around 1890), there were three red wagons in the yard, together with seven grey wagons. Both types, however, showed a mix of liveries, with G.W.R either above or below the weight information in both cases. If my guess about colours is correct, then it appears that both styles of painting were in use concurrently, both before and after the time of the change of lettering style!
     
    Could one have been predominant on the South Wales line from Gloucester, while the other was Swindon practice? I'm sure there are many knowledgeable people on the forums, who may have ideas on this matter.
     
    The books on the Forest of Dean railways are well worth tracking down for the superb illustrations of many industrial scenes from the 19th and early 20th centuries.
  3. MikeOxon

    A Matter of Colour

    By MikeOxon,


    General
    I’m currently pondering what colour I should use for the body of my Broad Gauge ‘Rob Roy’.  References to the colour originally used on Broad Engines declare it to have been ‘Holly Green’ but opinions differ on what shade that name represents.
     
    According to Christopher Awdry’s book: ‘Brunel’s Broad Gauge Railway’, the Boiler Cladding was ‘Holly Green (Dark Blue/Green)’ until 1881, after which ‘Chrome Green’ was adopted. (he quotes the Broad Gauge Society as the source of this information).  It interests me to observe that the ‘GWR Wolverhampton Green’ is also frequently described as a ‘blue/green’ hue, so might the original Broad Gauge colour have been more akin to this Wolverhampton colour than we now realise?
     
    I found some support for this view from an unexpected source. The ‘new’ Great Western railway franchise adopted a new corporate livery in 2015, which is claimed to have been based on the original Broad Gauge engine colour - see https://www.pentagram.com/work/great-western-railway-1/story for more information. I quote: "A bespoke paint has been created for train liveries based on the original ‘dark holly green’ used on the first GWR locomotives." What the sources for the chosen colour were is not revealed but it seems very unlikely that the present colour has anything to do with the original Victorian pigments. The ‘new’ GWR colour is defined in the Wikipedia UK Railways/Colours list as #0a493e, which can also de defined as RGB= 10, 73, 62 or HSL= 170°, 86%, 29%
     
    When I place this colour against the GWR Chrome Green colour described on the GWR Modelling website the result is somewhat startling. Were early GWR engines really this blue?
     

     
    One point to consider is the effect of ‘brightness’ on the perceived hue. As I explain on my website about colour perception , the human concept of colour can be divided into ‘colour’ and ‘brightness’ (or ‘luminosity’). If we maintain a constant ‘colour’ but vary the brightness, perception of the colour can vary considerably, as shown below:
     

     
    To my eyes, at least, the colour does appear more definitely ‘green’ as the luminosity decreases and all the reports of early engine colours agree that they were darker than the later chrome green.
     
    Another colour to consider is the modern interpretation of ‘Holly Green’, described as British Standard BS 4800 14 C 39 - Holly green / Hollybush / #435d50 Hex Colour Code. (RGB= 26, 36, 31). The ‘Encycolorpedia’ website also shows the effect of variations, as shown below, which confirms that darker colours tend to look more green.
     

     
    Finally, I went out into my garden, where there is a real Holly tree, and photographed some leaves, together with seasonal berries.
     

     
    These leaves have a greater green content than the various ‘official’ colours and I find their colour rather pleasing.  I thought I’d see what this colour looks like when applied (by Photoshop) to my (still unfinished) model:
     

     
    In the end, I shall probably look along the array of ‘rattle cans’ in the local motorists shop and make a personal choice, relying on the fact that no-one will be able to contradict me with any confidence.
     
    Mike
     
  4. MikeOxon

    GWR ‘Sir Daniel’

    By MikeOxon,


    general
    A comment on my recent post about modelling Rocket reminded me that my first scratch-built locomotive was an Armstrong 2-2-2 that I constructed 10 years ago and described in ‘Railway Modeller’, July 2014 , as ‘Simply Victorian’. I explained in that article that I was encouraged by a drawing of one of these engines in Russell’s ‘A Pictorial Record of Great Western Engines’ with the caption comment that: "The utter simplicity of these early engines can be seen." The idea of ‘simplicity’ appealed to me!
     
    Later, when contemplating additional engines for my ‘North Leigh’ layout, I considered other Armstrong types, including the earlier ‘Sir Daniel’ class. At that time, I wrote of my first model that “It really was simple to build - basically a brass tube over a brass plate, with a very simple 'chassis' to hold a set of wheels at the right distance apart! It was really just a wagon that can be pushed along by a motorised tender. I would recommend an early 2-2-2 as a good subject for a first attempt at locomotive scratch-building.”
     

    Principal Components of my 1st model
     
    Whereas the 'Queen / Sir Alexander' class that I chose to model were the last engines designed by Joseph Armstrong before his untimely death in 1877, the 'Sir Daniels' were his first standard-gauge engines built at Swindon, starting in 1866,
     
    Writing in 2015, I decided that “In summary, I can see that I could make a model of a 'Sir Daniel' by using exactly the same methods that I used for 'Queen'. At the moment, I feel tempted but concerned that the two would end up looking too similar! If I do tackle a 'Sir Dan', I shall have to choose a prototype with significant differences from my existing model”
     
    I felt at that time that my options were limited by what I could make using the traditional methods of brass-sheet construction. Since then however, my horizons have widened considerably following my adoption of 3D printing which greatly eases the problems associated with matters such as open splashers and other awkward features.
     

    Sir Daniel No.378 in ‘as-built’ condition
     
    One elephant remains in the room, however, and that is the fact that a ‘00’ gauge model really is a ‘narrow gauge’ model, within which true-to-scale boilers may not fit! In the case of my ‘Queen’ model that meant taking jewellers’ snips to the brass sheet I intended to roll into a boiler – fortunately, it still rolled successfully! It also meant that I had to adopt covered-in splashers to hide the gaping hole that I had created, so my model had to be of the later Dean rebuild of the type.
     
    Now, with 3D printing, I felt able to look back to the original design of 1866 for an engine that would be significantly ‘different’ from a rebuilt ‘Queen’ prototype of 30 years later. I still had to take into account the constraints of ‘00’ gauge, with a back-to-back dimension between the driving wheels of only 14.4 mm (3’ 7” at prototype size) against a boiler diameter of 4’ 2” (plus cladding).
     
    According to RCTS ‘Locomotives of the GWR’,Part 4, the main dimension of interest to a modeller were:
     
    ‘Sir Daniel’
    Cylinders. Diam. 17" Stroke 24" Boiler. Barrel 11’ 0" Diam. outs 4' 2" Pitch 6' 11". Firebox. Length outs 5' 4" Wheels. Carrying 4' 0". Driving 7' 0" Wheelbase 7' 8" + 8' 4", total 16' 0".  
    For comparison, the later ‘Queen’ class had shorter boilers but larger fireboxes and a longer wheelbase of 17’ 6”.
     
    Creating a Model.
     
    I started by considering the design of the boiler and the need to accommodate ‘narrow gauge’ driving wheels. My references were the photograph of ‘Sir Daniel’ as originally built, shown above, and a drawing by Jim Champ, which I imported as a canvas into ‘Fusion 360’
     
    Usually, I like to make the boiler from a brass tube and add a 3D printed cladding, as on the prototype. In this case, however, I had to accommodate the driving wheels, so I decided to 3D print the boiler itself and add weight by lead strips afterwards.
     
    After some thought, I decided on the following procedure:
     
    1. create a cylinder of length 44mm and diameter 18mm, to represent the external dimensions of the boiler plus its cladding
     
    2. create a pair of 28mm diameter driving wheels, with additional 0,5 mm flanges, and mount these either side of the boiler at the ‘00’ back-to-back separation of 14.4 mm
     
    3. Assemble these parts into their correct locations within Fusion 360 to create a visual impression of what modifications would be required to the boiler model, as shown below..
     

    Tackling the 00 gauge Boiler Problem
     
    I realised that the diameter of the cut-outs in the sides of the boiler would have to be wider than this, because it would simplify the design of the splashers, if I could also recess these into the sides of the boiler. I drew a circle of the required diameter in Fusion 360 and then used the ‘extrude’ tool in ‘cut’ mode to create recesses of the depth required to accommodate the driving wheels.
     

    Cutting recesses for driving wheels and splashers
     
    After dealing with that problem, the rest of the construction followed a familiar course. I extruded a pair of outside frames from the drawing canvas and set these at a ‘true scale’ distance apart. This means, of course, that the driving wheels are inset, relative to the prototype, but the overall arrangement of the locomotive is correct. If I wanted to run the engine on scale track, this would be possible, simply by extending the driving axles. At this stage, the frames are purely ‘cosmetic’ and I shall consider the design of the real chassis later.
     
    All the other parts were similarly extruded from the drawing or, for the dome and safety valve covers, I used the ‘revolve’ tool on a profile sketch.
     
    One thing I always have to bear in mind is the ‘printability’ of a 3D model created in Fusion 360. I aim to divide up the component parts, so that each one has a flat surface to lie on the printer bed, from which the rest of the structure can be ‘grown’, For example, I provide flat plinths on the boiler, so that the various boiler fittings can be printed from flat bases. The ‘exploded’ view below shows the collection of parts that can be printed individually.
     

    Components shown separately in Fusion 360
     
    As usual, I also like to do a ‘test fit’ of all the components by assembling them within Fusion 360 as a final check before printing – they can be seen against the ‘canvases’ from which the parts were extruded below:
     

    My model of ‘Sir Daniel’ assembled within Fusion 360
     
    The next stage will be to design a chassis, which will be based on an inside-frame for the wheels. As in the case of my existing ‘Queen’ model, I intend this one to be driven from a powered tender.
     
    Mike
  5. MikeOxon
    This post is a miscellany of ‘lessons learned’ on my journey to incorporate 3D printing into my railway modelling work-flow
     
    Removing window in-fill
     
    All the carriages I’ve printed so far have an amount of supportive in-fill within the various window apertures. This has proved surprisingly difficult to remove since, although the infill is very thin, it clings very tenaciously to the edges of the window opening. I tried several tools, including small cutting tweezers, wax-carving chisels, and needle files but, eventually, I had the greatest success with a blade from the Modelcraft Precision Saw Set (0.24mm)
     

    Clearing in-fill from window openings
     
    The Modelcraft set includes 6 very fine saw blades (0.24mm thickness). intended for intricate sawing and cutting in plastics, wood, and photo etched parts. These blades fit onto scalpel handles.
     
    Unlike scalpel blades which readily break, if used for sawing actions, these blades are flexible and can be used both to stab into the thin plastic film and then to work around all the edges of the window frame. In some cases, some of the infill was found to have folded back against the inside wall of the carriage and, in such cases, a small wax-carving chisel was a good removal tool
     
    Printer Bed Adjustment
     
    In my last post, I described problems that I had with adhesion of my models to the bed of my Geeetech E180 printer.  I am pleased to have discovered for myself how important the process of setting up the alignment of the bed is, to achieving successful results.  Now that I have attended to this detail, my prints have been very consistent and, while remaining firmly in position during printing, have been removable from the bed without ripping up the blue masking tape.
     
    The improvement in reliability has done a lot to raise my confidence in the 3D printing process as being one that can produce good results with very little user input being required, after the initial computer model has been drawn.
     
    Repeat Prints
     
    After an adhesion failure, which resulted in a distorted carriage body, as described in my previous post, I simply re-inserted the same gcode file into the carefully set-up printer and pressed the ‘print’ button. All went well and about 5 hours later, I had a non-distorted carriage, with no intervention being needed on my part.
     
    I was surprised to see that, when I placed the two carriage bodies side by side, many of the printing blemishes were in exactly the same places on both. I had thought that these were random errors due, perhaps, to backlash in the mechanisms or other mechanical tolerances but, instead, it looks as though they are systematic defects arising from the way the print head tries to follow the instructions provided by the ‘layer’ model from the slicer software.
     

    Comparison of two successive prints
     
    Curved Roofs
     
    I decided to do some more experiments, to see whether my 3D Printer could produce a reasonable curved roof for my carriages. My initial thought had been to use a rectangle of plasticard, moulded to shape by judicious application of a little heat, as described in a previous post , when I built other carriages.
     
    The possibility of using the 3D printer, however, was too strong for me to resist, so I had a go with the simplest structure I could devise. I sketched a pair of arcs, using the Fusion 360 software and then extruded the sketch by means of the ‘push-pull’ command, to form a 1mm thick single-curvature roof. I added some rectangles close to the edges of the under-side, to form a ‘base from which the roof could be ‘grown’. It was an experiment, since I was fully aware that the raised part along the centre of the roof would have no underlying material and would be dependent on support from the adjacent strips, as the printing progressed.
     

    Sketching the Roof Profile in Fusion 360 software
     
    In the event, I thought that the result was surprisingly good – while not perfectly smooth, the curved upper surface showed only very light grooving. On the other hand, because it started with all the edges lying on the printer bed, there were no alignment features to locate the roof into its correct position over the body.
     

    Printed Roof - Surface Finish
     
     
    Adding location lugs to a roof
     
    My next step was to add some vertical 'lugs' running downwards from near each end of the roof, placed so as to engage with the inside faces of the end walls of the carriage.
     
    This introduced the problem that the edges of the roof itself could no longer be in direct contact with the printer bed, so I selected the options in the Cura software to add support structures, where necessary. I used the default ‘zig-zag’ support style and allowed the print to take its course.
     

    Zig-zag Support Structure Underneath the Roof
     
     
    I was somewhat surprised to see how ‘dense’ the resulting support mesh appeared to be, at least at first glance. In fact, the ‘cellular’ structure only applied to the very first layer, in contact with the bed, and subsequent levels of the structure were a zig-zag pattern of very thin walls, rising throughout the structure. This support structure proved much easier to cut through with a knife-blade than I had expected, although it did leave lots of strips of thin material, on which I used a wax-carving chisel, to slice them away from the underside of the roof.
     

    Removing Support In-fill
     
     
     
    ‘Soldering’ with PLA filament
     
    In chiselling away these support structures, I inadvertently broke off one of the lugs that I had designed to hold the roof in position on the body. At this stage, I tried another technique that I had seen described on the web
     
    This is the somewhat surprising technique of ‘soldering’ the PLA plastic. By setting my soldering iron, fitted with a small conical tip, to a temperature of 200°C, it melts the PLA  in exactly the same way as does the print head itself.
     
    By running the hot tip of the iron along the damaged joint, the plastic readily melted and welded together the two sides of the joint, to form a strong bond. Larger repairs can be made by introducing additional PLA filament, in much the same manner as one would use a solder fillet!
     

    Using Soldering Iron for Seam Weld
     
    This made me realise that more complex structures could potentially be assembled from 3D printed parts by using a soldering-iron in much the same way as one would when assembling an etched-brass kit. I think this method of assembly could open up a whole new realm of possibilities.
     
    Finishing the Roof
     
    Although the surface finish of my 3D-printed roof was better than I had anticipated, it was far from perfect. To overcome the slight 'grooving' and achieve a smooth surface, I applied a film of self-adhesive vinyl to the 3D printed surface, which completely masked all the small ‘steps’ produced by the layered printing process. The final roof (sprayed with grey primer) is shown below, on my (unfinished) Composite carriage body and chassis.
     

    Vinyl-Covered Printed Roof
     
    The same methods could be used for more complex roof shapes and I plan to re-visit the clerestory roofs that I fabricated for some of my previous models. It should also be possible to add features, such as oil lamp housings, as part of the 3D print.
     
    Mike
  6. MikeOxon

    Creating a Pantechnicon

    By MikeOxon,


    General
    Even after a long career in research, I am still frequently surprised by the new information that emerges from small beginnings!
     
    After completing my Britzka model, I placed it on my BG carriage truck model for some photos. The Britzka was rather a tight fit between the truck rails, so I went back to Eddy’s Data Sheets from the Broad Gauge Society (BGS) to check the dimensions. It turned out that it is my carriage at fault and I shall revise my model to make it a little narrower (easy to do with a 3D printed design!). As I was looking at the BGS data sheets, however, I noticed that there was a very similar truck described as a 'Road Van Truck' (Data Sheet 105B).  I started to investigate further and so started a new line of research ...
     
    From the BGS data sheet, the ‘Road Van Truck’ looks similar to the carriage truck I have already modelled, except for a longer body and slightly increased wheelbase: body length increased from 15’ 4” to 18’ 6” and wheelbase from 8’ 6” to 9’ 0”. It was a straightforward task to increase these dimensions in my 3D model of the carriage truck, without having to re-design any of the components:
     

    3D-model of Road Van Truck – revised from Carriage Truck
     
    In describing this truck, Data Sheet 105B includes the information that “… of general interest are the notes in BGS Journal [Broadsheet] No.26, p10, which refer to the first recorded use of a Furniture Van Truck with Pantechnicon, 1847, transported by rail.”  Fortunately I have the complete set of digitised ‘Broadsheet’ issues, which are available from the Broad Gauge Society, so I could look up the article entitled 'Knee Brothers of Bristol'
     
    Knee Brothers, Bristol
     
    ‘Broadsheet’ No.26 contains an article about the firm of Knee Brothers, a removals firm in Bristol founded in 1839.  According to this article “In 1844 William Knee had a special mammoth heavy version of a horse drawn spring furniture van built to his own specifications at a wheelwrights in the St. Pauls district of Bristol. The felloes of its wheels were recorded as being 9" wide. This was the first pantechnicon to travel by rail.”   The first record of such rail transport was in 1847, when the furniture of a contractor for the construction of the Bristol and Exeter Railway was conveyed on the B&ER, using GWR rolling stock. There is an early advertisement by Knee brothers, which shows the Pantechnicon mounted on a GWR Road Van Truck, as shown below:
     
     

    Contemporary Illustration from Knee Brothers Advertisement
     
    According to the ‘Broadsheet’ article: “The Knee Brothers 'pantechnicon' had wagon sides and strouters (wooden supports to strengthen wagon sides) of blue-grey, similar to warship grey, with brown (possibly varnished wood(?) shafts and axles.  Springs were black whilst the wheels were red (similar to wheels of later G.W.R. horse drawn road vehicles). Its tarpaulin(?) covering was dark charcoal grey with side and front owner signs of cream background with black sign writing. An italic "No. 1" was inscribed on the wagon side just inside the front strouter.  Strouters on this vehicle sides, from just behind the region over the front wheels (i.e. about two-thirds along the wagon side from its rear), were slightly tilted from the true vertical towards the vehicle front for remaining length of the wagon side, whereas the other strouters appeared to be truly vertical.  This practice seemed to have been commonplace with techniques of road wagon construction in the nineteenth century.”
     
    It is also states that “The Great Western Railway broad gauge van truck on which the 'pantechnicon' was pictured was in all chocolate brown livery except for its wheels and springs which were black. Lettering "G.W.R. No. 66" was also in black” This description does not conform with other GWR livery references, which state that all parts of goods vehicles were brown, including wheels and undergear. The lettering was probably incised into the solebars, as described in Great Western Way (1st.ed).  Whether it was coloured black seems questionable
     
    Creating a Model
     
    Apart from there being no dimensions, this description and illustration above seemed adequate for me to attempt making a model. The length of the Road Van Truck has been given as 18’ 6” and, in the above illustration, the ends of the pantechnicon are around one foot from the ends of the truck, so I have guesstimated the length of the pantechnicon as 16’ 6”. I used the perspective tools in Photoshop to ‘square up’ the side of the body to estimate other dimensions on this basis. I have described this method in more detail in an earlier post.
     
     

    Estimating dimensions after perspective adjustment in Photoshop
     
    I also had to estimate the width of the body and decided, somewhat arbitrarily, to make the width across the panelled sides 4’9”, widening to 6’ 0” above the large rear wheels. This allowed sufficient clearance for wheels with 9” felloes (i.e. rim width) to fit within the side-rails of the Road Van Truck.
     
    Using ‘Fusion 360’ modelling software, I extruded the van body from a rectangle measuring 66 mm x 19 mm to a height of 14 mm. I then extruded the upper part of the sides to an overall width of 24 mm. I used the ‘Shell’ tool to hollow out the interior of the vehicle to leave an outer wall width of 1 mm.
     
    For the side panelling, I drew a pair of rectangles of appropriate size, one above the other near the rear end of the side. I then used the ‘Pattern’ tool to create an array of two rows of 19 panels along the length of each side. I see this ability to create repetitive features like these panels as a great advantage of 3D modelling over other construction methods.  In accordance with the description given above, I modified the front six panels so that the uprights between the panels were inclined at 5 degrees from vertical.  At this point, the 3D extruded body appeared as shown below:
     

    My first-stage extrusion of Pantechnicon body
     
    My next step was to add the front board above the body and a series of hoops to the same profile along the length, to the tail-board. Then I designed the curved strouters, which help to support the overhanging part of the upper sides. I also created wheels of 9” width and placed them in appropriate positions, so that I could check the clearances within my Road Van Truck.  The hoops were created as individual bodies, so that they can be printed separately, lying flat on the printer bed.  The ‘test assembly’ of these various parts within ‘Fusion 360’ is shown below:
     
     

    Test fit of my pantechnicon body on road van truck
     
    Designing the Chassis
     
    It’s not possible to glean much about the chassis from the available illustration but there appear to be separate under-frames for the front and rear axles, providing mounts for semi-elliptic springs above the axles.
     
    The front underframe is fitted to a swivel mount below the floor of the vehicle body and also carries attachments for two pairs of shafts, to attach the horses. In the illustration, these shafts are shown swung upwards against the front board of the vehicle. There is no indication of a drivers seat, although this would have been placed near the top of the front board.  By drawing on my experience from designing the Britzka carriage, I sketched out an underframe comprising a rectangular frame to carry the springs for the rear axle. From this, two longerons reach forward to carry the fixed ‘fifth wheel ‘, which provides a bearing for the pivoting fore-carriage.  The fore-carriage is a separate component carrying the frame for the springs over the front pivoting axle, plus mountings for the shafts, for two horses.
     
     

    My design for the pantechnicon chassis
     
    For the shafts, I turned to Janet Russell’s book: “GWR Horse Power”, which contains dimensioned drawings of various types.  I used one of her drawings as a ‘canvas’ in ‘Fusion 360’, from which I extruded two sets of shafts to attach to the front of my chassis.. I arranged these shafts to fit onto a cross-shaft, so that they could be raised against the front panel of the body, as shown in the illustration above, or lowered for attachment to a pair of horses. Once I had brought all the parts together in ‘Fusion 360’, the 3D design looked as below:
     
     

    My 3D design of the complete Pantechnicon undergear
     
    I could now place the body over the chassis in ‘Fusion 360’, to assess the overall assembly before printing the various components.
     
    Printing
     
    As usual, after all the design effort, the printing came as something of an anti-climax. Once again, I was pleasantly surprised by how cleanly my printer produced even the smallest and most delicate parts. I laid them all out for a photo, just as I took them from the printer bed (although one hoop somehow disappeared to the carpet monster and I had to reprint a second set)
     

     
     
    For the record, the print times (estimated by the Cura software) were:
     
    body 67 min wheels (set of four) 20 min underframe 10 min fore carriage 7 min front board 7 min axles (pair) 5 min hoops (set of four) 4 min shafts 3 min  
    The assembly was very straightforward. Only the front board, hoops, and axles needed to be fixed into place I fixed a ‘perch pin’ into the centre of the ‘fifth wheel’ underneath the body and the fore-carriage was then slid onto this pin. Similarly, the wheels and shafts were slid onto lengths of 0,7 mm wire, which provided the pivot points.
     
    I painted the body light grey, as described in the ‘Broadsheet’ article, while the axles and shafts are brown, and the wheel faces red.
     
    I produced the lettering using ‘Cooper Std’ and ‘Garamond’ fonts, which are reasonably good matches to the prototype illustration, shown below.
     
     

    My layout of the front panel text
     
    After laying out the text and the Bristol coat of arms in Photoshop, I printed the front and side panels on ink-jet printable self-adhesive vinyl, which I stuck onto the 3D-printed model.  The end result, without a tilt cover, is shown below:
     
     

    My model pantechnicon (less tilt cover)
     
     

    Rear View, showing Tailgate
     
    Canvas Tilt Cover
     
    For the cover, I first wrapped the body of the van in cling film, stretched over the hoops and shaped by hand. I then wrapped a length of wet plaster bandage over the cling film and smoothed it down into place. Once I was content with the general appearance, I placed the model over a radiator to dry.
     
     
     
    My model wrapped in cling film and plaster bandage
     
    Once the plaster had dried, I applied a rather thick layer of black acrylic paint over the bandage, working it into the texture of the surface to make as opaque a coating as possible. Once this had dried, I trimmed off the excess plaster, below the top of the body and took the portrait below:
     
     

    My model with Tilt Cover added.
     
    As I indicated at the outset, the thought of building a pantechnicon was not in my mind until I followed the train of thought that started from the BGS Data Sheets!  It has proved to be an interesting model to create and I think the use of 3D printing provided a simpler course than ‘traditional’ modelling methods. All the parts were extruded from 2D drawings so, if you can create a drawing for a ‘Silhouette’ cutter, it can be turned into a 3D model! In fact, some of my first 3D-printed models were created in this way from drawings that I had previously used for Silhouette-cut carriage sides.
     
    I notice that there is an earlier thread on this site about modelling a 4 mm scale pantechnicon from a kit. I suspect that it was the ‘Gem’ kit, which is no longer available.
     
    Mike
     
  7. MikeOxon

    William Dean's Express Tank

    By MikeOxon,


    general
    Being somewhat of a back-water in railway history, the Witney branch via North Leigh became home to many unusual vehicles, rarely seen elsewhere. I have already shown the standard-gauge Tilt Wagon ('Hat Box') but a particular claim to fame for this line was that it became the 'stamping ground' for William Dean's experimental 4-2-4 express tank engine, shown below entering North Leigh station.
     
     
     
     
     
     
     
     
     
     
     

    William Dean's experimental 4-2-4T No.9 at North Leigh
     
    My first encounter with this locomotive was many years ago, through the pages of John Gibson's “GW Locomotive Design - A Critical Appreciation”, which portrayed it as such an embarrassing failure that all traces of it had to be expunged from the record, with an order going out that 'this thing never existed'.
     
    If that were true, Dean would hardly have provided an article for 'The Engineer, Sept.24th 1886', with a detailed drawing of the unusual outside Stephenson valve gear. Dean explained that this arrangement was intended to get rid of the excessive cramping up of the valve gear, necessary when large cylinders have the valve chests placed between them.
     

    Outside Valve Gear on Re-built GWR No.9 (from 'The Engineer')
     
    In fact, Dean had arrived in office knowing that the Broad Gauge was nearing its end and a new type of locomotive would soon be needed, to replace the Gooch singles. Amongst the ideas to be tested was the possibility of using express tank engines, as designed successfully by Pearson, for the Bristol & Exeter railway. E L Ahrons, who arrived at Swindon in 1885, described the original configuration of the engine as having a 30' overall wheelbase, including two carrying bogies of wheelbases 7' 3" and 5' 6" respectively. He also stated that the length of the platform was 36' 5-5/8".
     
    Very much later (around 1940), E W Twining produced an outline drawing, purporting to show the layout of the tank engine. A comparison of his sketch with the detailed drawing of the valve gear in 'The Engineer' reveals that it is simply not possible to fit this gear between the front bogie and the driving wheel, if the longer bogie is placed at the front! In my opinion, Twining's chassis seems to bear a very considerable resemblance to the Lehigh Valley inspection saloon, still in existence at that time, even down to the bicycle-style 'mudguards' over the front bogie wheels.
     
    My own solution places the shorter bogie at the front and, with this configuration, a well-balanced design results. The front bogie fits neatly under the smokebox, leaving room for the outside valve gear, while the rear 7' 3" bogie fits immediately behind the firebox, within the overall 30' wheelbase. Apart from the reversed bogies, the dimensions conform to Ahrons' description.
     
    The construction of my model followed similar lines to those I have described previously, when I built my 'Queen'-class engine.  I built the engine in two parts: a rectangular chassis, carrying the driving wheels plus two bogies, and a box structure for the upper body. The boiler is represented by a half round section soldered into position between the two side tanks.  The steps in my construction are shown below:
     

     
    I have fitted an extended cab roof, similar to those used on other GWR tank engines in the same period, and a rear spectacle plate, appropriate for high-speed reverse running. Other features include cutaways in the tanks, to allow access to the motion, and tanks that extend a short distance in front of the smokebox, both as mentioned by Ahrons. The long (11' 6") boiler, in two rings, was designed to provide similar steaming capacity to the large-diameter boilers of the Gooch singles.
     
    My model wears the new livery, introduced in 1881, with Indian red frames and vermilion lining (though at that time the wheels were still green). With this presentation, I believe that the engine is worthy of the description given by David Joy, when he visited Swindon around 1882 "I saw all about a mighty 'single' tank engine ... I saw drawings and all, and she looked a beauty".
     

    My Interpretation of GWR 4-2-4T No.9
     
    The real significance of this engine was that, while it showed that the 4-2-4 express tank engine concept was not appropriate for the standard gauge, it addressed an urgent need to re-visit valve-gear design for more powerful engines. The eventual outcome was, of course, the famous Dean Singles, with the same length of boiler and direct-driven valves, mounted below the cylinders. The family relationship can be seen in the following extremely rare photo of these two engines, side-by-side at North Leigh:
     

    Dean 'Single' alongside its 4-2-4T pre-decessor at North Leigh
     
    My model currently only has 'cosmetic' outside valve gear, cut from plasticard with my Silhouette cutter. It is powered by a Tenshodo WB28.7 SPUD, contained within the rear bogie, in the same way that my 4-2-2 uses a similar one in its front bogie. The model negotiates my small-radius curves with ease and does not share a tendency for de-railing, which apparently afflicted the prototype!
     
    Update:  I have created a 3D computer model of this engine - see https://www.rmweb.co.uk/blogs/entry/26347-william-dean’s-express-tank-revisited/
     
    Mike
     
     
    references:
    Great Western Locomotive Design, John C Gibson 1984
    The British Steam Railway Locomotive 1825 - 1925, E L Ahrons, 1927
    Swindon Steam, L A Summers, 2013
  8. MikeOxon
    Much of what we now accept as standard railway practice was actually developed towards the end of the 19th century, so that it is easy to forget that there was a long period of evolution, before the 'norms' with which we are so familiar became established. Whereas we have plenty of photographs to guide our perceptions of the later years of the century, our 'picture' of the earlier years remains much hazier, so it is easy to lose sight of the many major changes that occurred.
     
    For example, my modelling of some early coaches demonstrated to me how it was not until the 1870s that the design of railway coaches began to shake free of their stage coach roots. There was a dramatic change in both the scale and the construction methods of railway vehicles in the latter half of the century. Two of my (unfinished) standard-gauge models illustrate the point. On the left is an early 2nd-class coach, probably built during the 1850s, while on the right is the type of coach that was built in the new carriage shop at Swindon, to diagram S5 in 1875
     
    .
     
    Now that I have started looking at the Broad Gauge era of the GWR, linked to a specific event in November 1868 - the accident near Bullo Pill - I have started to look more closely at developments around that period. Research is an endless task and it is easy to become overwhelmed by all the facts that start to emerge, as one digs more deeply, but one has to start somewhere so, here is a look at the state of GWR rolling stock in 1868.
     
    The photographs of the accident at Bullo Pill clearly show the telegraph wires running alongside the railway. It seems almost inconceivable from our standpoint that these wires were not used to control the movements of trains. Responsibility for the running of the trains rested almost entirely on the driver, just as it had rested with coachmen in earlier times. In 1868, an express train could be running at 50 mph or more, with virtually no knowledge as to whether the line ahead was clear. According to the accident report: "The rules of the company require the signalmen to warn the driver of a passenger train when there is, after dark, a goods, cattle, or mineral train, less than 20 minutes in front of such passenger train." Since the interval at Lydney was 22 minutes, no warning had been given.
     

     
    In addition, the express train had no means of stopping quickly, if something untoward occurred, because there were no continuous brakes in 1868 and only a few carriages (including the Mail Coach) had hand-wheel brakes, operated by a Guard, who could respond to a whistle signal from the engine. Sanders' Automatic Vacuum Brake was trialled on a standard gauge GWR train in 1876 and only adopted more generally in 1878, long after my chosen period
     
    Early carriage brakes were often of the 'clasp' type, with large wooden shoes acting on the rims of the wheels. A typical arrangement, taken from a drawing of an early brake van, is shown below. The layout of the various levers varied but the principle remained the same.
     

    What does all this mean for the design of my model of a Broad Gauge Mail Coach, which was one of the three coaches on that ill-fated mail train? It turns out that my model will be rather different from my initial pre-conceived ideas, which were formed largely on impressions given by later re-builds of such vehicles. Much of the information in the following paragraphs is taken from MacDermot's 'History of the GWR', Vol.2.
     
    We all 'know', of course, that GWR coaches were painted 'chocolate and cream'. Except that they weren't!  GWR coaches were painted brown all over until some time after October 1864, when the Directors decreed that the tops of the carriages should be painted white. The cream appearance only developed as several coats of varnish were applied. Furthermore, the Broad Gauge had already been in decline for many years by then, with no new carriage stock having appeared since 1863, and the existing stock steadily became increasingly shabby and dilapidated.
     
    Although the Broad Gauge 'hung on' in the West Country until 1892, a major conversion had already occurred, much earlier. In May 1872 all the South Wales lines (plus other lines West of Gloucester) were converted to standard gauge. With this major upheaval already pending in November 1868, it seems very unlikely that much attention would have been given to updating any of the Broad Gauge stock running in that area, at that time.
     
    So, the colour and the brakes of my model should be more appropriate to a date in the 1850s, when these coaches were built, than to 1868. What other details might differ from my expectations?
     
    Early coaches were built with iron spoked wheels. These wheels were still fitted to almost all Broad Gauge coaches and vans as late as 1874, although Mansell wooden-centred wheels had been tried as early as 1866 and adopted for all new stock in 1868. My coach, therefore, requires spoked wheels.
     
    The long wooden foot-boards, widely associated with GWR coaches, were only introduced in 1876, so my coach should have a small iron step at each doorway, just like those fitted on stage coaches in earlier times.
     
    Less surprisingly, the axle boxes should be of the grease pattern, as oil boxes were a much later innovation, dating from 1886. Gas lighting had arrived a little earlier, in 1882, but, as with most such changes, it was many years before they were fitted in all stock. (I believe some oil lamps were still around in the 1930s)
     
    Even the 'communication cord', initially slung along the eaves of the coaches, was only introduced on the Broad Gauge in 1869, after initially running only from the leading Guard's van to the tender. Throughout the 1860s, luggage rails were fitted to the roofs of many Broad Gauge coaches but it is not clear whether this applied to the Mail Coaches. One photograph (P.W.Pilcher, 1885) suggests that luggage may have been carried on the lower section of the roof.
     

     
     
    With all this additional information, I am relieved that I have not yet started building the under-frame of my Mail Coach. I could easily have added many later details, without realising, and bought completely the wrong type of wheels. Some aspects will now be simpler, since I have already encountered the difficulties of modelling later brake gear, with all its rods and safety straps to prevent bits falling onto the track!
     

    sketch of planned model
     
    Mike
     
    Link to Part 4
     
    As an annexe to this post, I have placed the various dates mentioned above on a 'timeline', to help visualise when the various innovations were made.
     

     
  9. MikeOxon

    Building 'Tantalus' - Part Three

    By MikeOxon,


    General
    At the end of the previous Part, I wrote that I intended to turn my attention to the chassis. Before that, I had broken off from work on the chassis, to give some thought to the problems posed by the curved splashers over the driving wheels.
     
    Now that I have proven the technique of ‘hybrid’ construction, combining traditional etched brass model-making with 3D-printing, I’ve decided to pursue similar methods for the chassis. For wheels, I recovered a spare set from a ‘Mainline’ Dean Goods that I converted into a 2-4-0 ‘Stella’ a few years ago. My conversion was described in the April 2013 issue of ‘Railway Modeller’, These wheels have larger flanges than are desirable, so I shall probably replace them at a later stage but they are useful for setting the clearances for the splashers.
     
    Rather than relying on measurements, I prefer to base my models on ‘templates’ that are derived from the actual components that I am using. In this case, I laid the chassis, with wheels attached, in my flat-bed scanner and made a scale image as the basis for my drawings
     

    Scanned image of mainframe and wheels
     
    I then ‘pasted’ this image as a background layer in my drawing program and drew the splasher outlines to fit around these wheels. I transferred my 2D drawing into ‘Fusion 360’, where I used the ‘push/pull’ tools to extrude the tops of the splashers so that they would ‘frame’ the wheels.
     
    The resulting 3D model only took a few minutes to print, so it was easy to run off a test-print and check the fit before proceeding any further. This proved worth while, since my first print was a little too shallow to leave clearance for the valance to fit in front of the wheels.
     
    While considering potential refinements to my initial print, I realised that some other features could be added to the footplate, including the tops of the springs, which appear from behind the splashers, and the sand-boxes, which are placed alongside the smoke-box. After adding these additional features, my 3D model appeared as shown below.
     

    3D model of splashers with springs and sand-box
     
    I extended the backplane of the print slightly below the top of the chassis, to provide a ‘tab’, so that the splasher assembly could easily be glued to the back of the mainframe.
     
    One very welcome feature of 3D printing is that, in order to produce the splashers for the opposite side of the engine, it is only necessary to click on the ‘mirror’ command and the appropriate part is created automatically.
     
    Once both sides had been printed, I glued them to the sandwich mainframes and inserted the boiler assembly between the frames, as shown below:
     

    Hybrid brass / 3D-printed model of ‘Tantalus’
     
    It was now time to turn back to brass construction and I cut out the etched valances provided in the BGS kit and glued them to the outer edges of the splashers to give the result shown below:
     

    ‘Tantalus’ after adding etched-brass valance
     
    There are still many details to be added but I think my model has reached the stage where it allows the rather ‘gawky’ appearance of these early Broad Gauge engines to be appreciated.  This might seem to be a short post but most of the work was in the planning - the actual printing and assembly only took a few minutes
     
  10. MikeOxon

    There's Nothing New ...

    By MikeOxon,

    More than three years ago, I wrote about some experiments I did to add to the debate about GWR red wagon livery. I showed photos of wagons at Cinderford Iron Works, taken around 1890, which had marked differences in (monochrome) image tone.
     

     
    I thought, at the time, that I was doing some ‘original’ research but recently I have been looking through back issues of the 'Broad Gauge Society' magazine ‘Broadsheet’.  There I found that this ‘well known’ photograph was discussed at length, as part of a series of articles on wagon liveries, as long ago as 2001 (Broadsheet No.45).
     
    Brian Arman wrote “This photograph is the one frequently claimed [my italics] to be an example of red and grey wagons intermixed. I presume that the wagons supposed to be grey appear to be a darker shade, as in the fourth wagon from the left on the near siding, and the second and fourth vehicles in the back siding. However, following John’s photographic searches we may confidently surmise [my italics] that these vehicles are in all probability simply newly painted as opposed to ‘re-liveried’. The positioning of the lettering varies from wagon to wagon, and adds an element of further interest to the photograph. In fact the whole scene is full of historical interest ...”
     
    His conclusions are exactly the opposite of mine, since I thought the darker shade was caused because the photographic emulsions of the time were colour-blind to red!
     

     
    We are still left with the conundrum that the different styles of lettering appear on wagons of both shades, on the Cinderford photograph, which seems to mitigate against the idea that the dark shade is simply the result of more recent painting. Surely, lettering styles changed at specific dates – or did they replace the lettering in the ‘old’ positions when re-painting (or relivery-ing)?
     
    So, although I had discovered nothing new, the photograph still presents a puzzle and may indeed show both red and grey wagons – but which is which?
     
    Mike
  11. MikeOxon
    I feel it is time for an update on progress with my model of GWR No.184. After a rather static period, when I did little actual modelling but thought quite a lot about the details of this engine, I have at last been cutting metal.
     
    In fact, apart from the fact that it is now made of brass, I have made few changes to my initial paper template. My method of construction followed that used on my first scratch-built model (GWR 'Queen'-class - described in Railway Modeller July 2014). In summary, I cut out the components from scale paper drawings and stuck these (using a glue stick) to 10 thou (0.25 mm) brass sheet.
     
    One simplification that I made, following experience with 'Queen' was not to make a double-fold in the main chassis in order to mount the splasher fronts, but instead to provide tabs on the splashers 'module', to fold under the chassis sides. The main components that make up the body look like this:
     

     
     
    As before, I had little difficulty with rolling the boiler, although I suspect that it helped to have a paper layer stuck to the outside of the brass sheet while rolling. This seemed to help with the smoothness of the curvature and also protected the surface of the brass from scratching. Since the prototype boiler had a diameter of 4' 2", I calculated the scale circumference (3.142 x diameter) and then cut out a rectangle of brass, to roll into a cylinder. I made this tube the full length of firebox + boiler + smokebox, to provide a rigid structure and then added the raised firebox and smokebox as simple wrappers. The stages in assembly are shown below:
     
     

     
    I used a diamond slitting disk on my mini-drill to open up the apertures for the wheels in the floor plate. For the photograph, I've just rested a chimney and safety-valve cover in place, to make it look a bit like a locomotive (!) and have not bought a dome yet - probably a 517-type. The next step will be to fit 'cosmetic' outside frames with embossed rivet detail and then use thin (5 thou - 0.125 mm) brass sheet, to form the curved running plates between the outside frames and the splasher fronts.
     
    Anyone who read my last post on this engine might remember that I had planned to begin with the chassis but, in the end, I decided to repeat the part that was already familiar, having done a very similar body for 'Queen'. I may produce a simple rolling chassis that can be pushed along by my 2,500 gal. motorised tender. I find that tackling a job in easy stages helps to keep up motivation, when I see the outline of an engine start to appear, and allows me to consider different options as I start to add the details.
     
    So far, the outlay on brass sheet has been minimal but costs start to rise once all the detailing components are added, especially wheels, gearbox, and motor. At this stage, I feel that there is a reasonable chance of producing a model that will have at least a passing resemblance to its prototype.
     

     
    GWR No.184 in 'Wolverhampton' livery
     
    As an aside, I have mentioned before that many interesting old books can be downloaded from the Internet Archive website. I recently found that Ahrons' classic 'The British Steam Railway Locomotive 1825 - 1925' can be downloaded and, although I have a facsimile copy, it is useful to be able to copy drawings directly from the digital version. Another interesting find was Sekon's 'The Evolution of the Steam Locomotive 1803 - 1898', which has lots of fascinating detail about many very early engines, and many anecdotes about events during their trials and early use.
     
    Mike
     
    Continue to next part
  12. MikeOxon

    Cautionary Tales

    By MikeOxon,


    General
    Accuracy of Drawings
     
    In an earlier post, I wrote: “I used the same method that I described in my previous post to extrude the saddle tank from a drawing – this time a pencil sketch by F.J.Roche, reproduced in the ‘Broadsheet’ article. This drawing was useful for the front elevation but I feel the drawing in Mike Sharman’s compilation by the Oakwood Press is more dependable for the side elevation.”
     
    Some recent correspondence within the Broad Gauge Society (BGS) e-group suggested that the drawings in the Mike Sharman compilation may not in fact be that accurate, I quote: “The Sharman's book drawings were transcribed from originals published in the Loco Magazine. The one in question here* was published in 1903. The transcribers varied greatly. The originals are believed to be accurate but ... this particular transcription was one of the least accurate.”
    * This quote refers to a drawing of a Dean 2-4-0 convertible of the 3501 class, not my engine, but it sowed seeds of doubt in my mind.
     
    On looking more closely, I noticed for example, that the spokes of the bogie wheels on ‘Aurora’ were not placed accurately on the drawing, which showed up clearly when I designed my own wheel, using ‘Fusion 360’s Pattern command to produce nine equally spaced spokes.


    3D-printed wheel laid over Drawing
     
    It’s a small point but a warning not to believe the correctness of all the details on the drawing. I have now measured the wheel base and other key dimensions on my 3D-printed model and have been relieved to find that they are all accurate. I shall pay more attention to the accuracy of any drawings I use as a basis for extruding models in the future.
     
    Front Elevation
     
    Looking through my own small collection of drawings, I found three showing the front elevation of one of these 4-4-0ST engines. Two of these, by F.J.Roche and by Ian Beattie are said to be of ‘Lance’ (both drawings are from the BGS magazine ‘Broadsheet’ no.17), whereas the one by Alan Prior (in his book ‘19th Century Railway Drawings') is of ‘Corsair’. Putting these three alongside one another shows that there are many significant differences between them:
     
     

    Front elevation drawings compared
     
    Faced with discrepancies such as these, I turned to photographs and, in particular, the one of ‘Aurora’ that I showed in an earlier post The tank seen in this photo appears to have the more rounded profile shown in the Roche drawing above, although the sand pipes are not so arched. Of course, those early engines often showed a iot of individual variation, so the dictum to work from photographs as far as possible is very sound but can be difficult to apply, when photos were so far and few between.
     
    Following ‘rules’ may not be best
     
    Another interesting point appeared when I started to create the trial prints of the wheels that I designed for my model of ‘Aurora’
     
    I made the initial prints by laying the inside of the wheel flat on my printer bed. This meant that the widest part of the wheel, with the flange, lay on the bed so that there were no overhangs as the printing progressed, which is the ‘preferred’ method.
     
    In practice, the flanges came out thinner than expected and were damaged when I removed them from the printer. This may be a result of using a printer with an unheated bed but it was another case where disobeying the rule book yielded a better result!
     
    My later prints were made with the outer face of the wheel on the bed, so that the flanges actually overhung the main part of the wheel, as it printed. Nevertheless, the flanges printed cleanly and, by printing in this orientation, I could include sleeves extending from the backs of the wheels, to guide the pin-point axles and ensure the correct back-to-back measurement between the wheels.


    3D-printed wheel-sets with integral axle sleeves
     
    Each sleeve contains a clearance hole for the 2mm axle, while the wheel itself is an interference fit onto the axle. I assembled each wheel-set by dropping a pin-point axle into the sleeve and then tapping it gently home into the wheel with a light hammer, as shown below. Once one wheel had been attached, I turned the part over and tapped the axle into the other wheel, until the two spacing sleeves meet at the centre line.
     
     

    Assembling a wheel-set
     
    Once I had produced a set of wheels, I could place them under my model, so that it began to look like a real engine!  It has a purposeful look, well matched to its task of hauling important passenger trains over the South Devon banks. The famous Gooch singles may have stolen the limelight but it was these tank engines that maintained services across the more difficult routes of the South Western peninsular – they must be lauded for that capability
     
     

    ‘Aurora’ on her Wheels
     
    A few years ago, I could not have contemplated making a model like this and I am pleased that, during its design and construction, I have gained an appreciation of the prototype’s remarkable qualities. No longer shall I call it an ‘ugly duckling’
     
    As usual, there’s a lot of finishing still to be done. One day, I must get together my collection of unfinished Broad Gauge locomotives and have a session of handrail fixing,-plus all those additional fittings and polished brass-work that make make them into beautiful swans.
     
    Mike
  13. MikeOxon

    A Missing Link?

    By MikeOxon,


    General
    In my last few posts, I’ve been delving into the almost lost world of the early days of the GWR broad gauge. I notice that my previous post aroused little comment so, perhaps, I have moved rather too far from what most people think of as ‘railway modelling’ - but I do like using models as a way of improving our understanding of these early engines.  I do appreciate the various 'likes' that many of you have given me.
     
    Before I move back into more familiar territory, there is one more piece of history to record, regarding an engine that may not have existed at all – at least not in the form in which it was described.
     
    Francis Whishaw on the GWR
     
    One of the earliest books to survey the British railway scene was Francis Whishaw’s “The Railways of Great Britain and Ireland” of which the 2nd edition, with additional plates, was published 1842.
     
    He wrote “THE Great Western Railway is by far the most gigantic work of the kind, not only in Great Britain, not only in Europe, but, we venture to say, in the whole world. Mr. Brunel, not satisfied with the beaten track pursued by those who had gone before him, determined on carrying out this important work on entirely new principles; and, notwithstanding the numerous adversaries he has had to contend with from every quarter, has thus far been eminently successful in his favourite project, and will, no doubt, ere another summer shall have passed away, pronounce this mighty work to be completed throughout.”
     
    I should add, though, that Whishaw did go on to criticise the great increase in the costs incurred, over the original Parliamentary estimate … something with which we are all too familiar!
     
    Concerning locomotives, Whishaw left us with a puzzle, when he wrote :“we have ... classified these machines according to their magnitude, and the gauge of way to which they severally belong. Thus the six-wheel engines for the seven-feet gauge, as those on the Great Western Railway, belong to class A ; ...An engine belonging to class A is shewn in elevation in Plate 1, the frontispiece. and sections and details of an engine belonging to this class are exhibited in Plates 7, 8, and 9.”
     
    Whereas Plate 1 shows a locomotive of the ‘Firefly’ class, the engine shown in Plates 7, 8, and 9 does not correspond to any engine known to have worked on the GWR!:
     

    Whishaw Plate 7
     

    Whishaw Plate 8
     
    Whishaw does not specifically state that this is a GWR engine, although that assumption has been made by many later writers and is reasonable, considering that no other British railway used a 7 foot gauge..  The image has also sometimes been referred to as being a member of the ‘Firefly’ class, which it superficially resembles, at first glance.
     
    I have also noticed that there are many similarities between these drawings and the example of Stephenson’s ‘Patentee’ engine, illustrated in Tredgold’s ‘The Steam Engine, Vol II’, Plate LXXXXIX (99), published 1838. To demonstrate this, I overlaid the two drawings to the same scale, as shown below:
     

    Comparison between Whishaw’s Plate 7 and Stephenson’s ‘Patentee’
     
    Creating a 3D Model
     
    I decided to create a 3D model based on Whishaw Plates 7 and 8, shown above. As usual, I imported the two images as ‘canvases’ into ‘Fusion 360’. I used the various sketching tools in this software to create drawings, which I could then extrude to create the main components of the engine. I also copied details of the frames and axle guides from the drawing of the ‘Patentee’ engine in Tredgold’s book..
     

    Examples of Components extruded from Whishaw plates
     
    Once I had extruded all the necessary components, including use of the ‘revolve tool’ for the chimney, safety valve, and manhole cover, I assembled them together, within the Fusion 360 software, and created a ‘rendered’ 3D image:
     

    My 3D model created from Plates in Whishaw.
     
    Comparing Models
     
    When viewing the images of the Whishaw engine in isolation, it is not difficult to conclude that this image has similarities with the GWR ‘Firefly’ class. That idea is quickly disposed of, however, when I bring together the above model with the model of ‘Argus’, a member of the Firefly-class that I created in 2021.. Note that both these models were created to the scales indicated on the original drawings.
     

    My 3D models of the Whishaw engine and Firefly-class ‘Argus’ brought together
     
    Once the two models are placed together, the profound differences between their overall dimensions is immediately apparent! Notice too, the much lighter frames and smaller driving wheels of Whishaw’s engine.
     
    Similar Engines
     
    There was a Stephenson 2-2-2 of the ‘Patentee’ type, named ‘Harvey Combe’, which was built n 1835 and was used by Messrs. Cubitt, the contractors, during construction of the London and Birmingham Railway near Berkhampstead. The scene was captured by the artist J.C.Bourne, as shown above in my header image:
     
    According to an article in ‘The Engineer’ by J.G.H. Warren, dated 24th Sep.,1926, Nicholas Wood, in a “Report to the Directors of the Great Western Railway,” December 10th, 1838, gave results of experiments on the ‘Harvey Combe’ of the London and Birmingham Railway, to compare with contemporary experiments with the North Star of the Great Western Railway.
     
    I have not read Wood’s report in full but it suggests the possibility, at least, that similar engines might have been used by contractors engaged in construction of the GWR. It does seem to me that an engine similar to that shown in Whishaw would have been far more suitable for that task than the large-wheeled engines specified by Brunel, as the initial engines for the GWR.
     
    Perhaps, then, the engine shown in Whishaw’s book represents a ‘missing link’ between the engines built by Stephenson for the London and Birmingham railway and an engine or engines used during the construction of the broad gauge GWR?
     
    Mike
  14. MikeOxon
    I thought it would be easy to modify my U29 model, to represent the body of a diagram G13 Family Saloon but various gremlins struck with a vengeance!
     
    Omit reading the ‘Gremlins’ sections, if you just want to follow the progress of the model itself.
     
    Gremlins
     
    I have now learned the necessity for taking great care when selecting entities to modify in Fusion 360.
     
    For example, when selecting parts of a drawing in, say, the X-Z plane, it is easy to include parts of a drawing that are in the Y-Z plane, by mistake.  So, unintended changes are made to the ‘other’ drawing as well.  I think this explains how I managed to lower the height of the clerestory, without noticing until I had done a lot more design work.  Although I could ‘go back’ to the point where the problem started, I still had to re-do all the work on the window apertures and partitions, to fit the corrected body shell.
     
    I have also found some difficulties when importing old DXF drawings into Fusion 360.  Some panels, which had seemed to be ‘closed’ in other software, could not be extruded.  I eventually spotted some very tiny gaps, which I managed to fill, although I am still not sure exactly which incantations are needed to make the ‘extend’ command work on drawings imported into Fusion 360.  I had almost given up, when the command suddenly started working, for no reason that I could determine!  
     
    Having modified the drawing, I tried to save it and this brought up a message that “This file type requires cloud translation, which may take a few minutes.”.  For some reason, the ‘translated’ DXF file was unreadable in other programs, such as Studio, Autosketch, or Irfan View.  In the end, I decided to go back to Autosketch and repair the file there, so that it now works correctly in both Fusion 360 and other programs.  
     
    I mention these things, to show how learning 3D CAD programs can be very frustrating.
     
    Printing the G13 body and clerestory
     
    Once I had modified my carriage design to represent diagram G13, I printed the body and roof, as previously described for my U29 carriage.  
     

     
    This time, however, I wanted to investigate how I might use the 3D printed body as a basis for ‘finishing’ with parts produced on my Silhouette Portrait cutter.  For this reason, I omitted the inset panel details, which I had included in my U29 model, and left the outer surfaces smooth.  I did, however, include other inset items, such as window apertures and drop-light frames.
     
    3D-printed Carriages
     
    There’s nothing really new in this – I’ve simply substituted a robust 3D-printed body shell for the relatively frail laminated plasticard that I had used before.  I am impressed by the toughness of the PLA plastic.  It’s supposed to be bio-degradable, however, so perhaps it will crumble away in a few years time – I’ll leave it to my Grandchildren to find out!
     
    My planned method was to use my Silhouette Portrait machine, to cut out the body framing from 10thou (0.25 mm) plasticard, and my HP Inkjet Printer to print overlays for the body sides.  The ‘registration’ facility of the Silhouette Cutter could then open up the window apertures in these overlays, in registration with the printed body sides.
     

    The Three Printers used for This Project

    For most of my earlier carriage building, I printed coloured sides onto HP Photo Paper, although I did carry out some experiments into printing directly onto plasticard (described in posts in JCL’s Silhouette Cutter thread)  I found some of my old ‘printed-on-hairspray’ samples in a drawer and they still look fairly good – better than I remember achieving at the time.
     
    The actual printing and cutting procedures were exactly the same as I had used previously, when building my earlier versions of these carriages.  To summarise:
     
    “Open the colour image in Studio, copy it and paste it into the cutting diagram. Use the 'Arrangement' command to sent the image to the back and then re-size and align the image with the cutting marks as accurately as possible on the screen (use a magnified image and make sure that 'Snap to Grid' is 'off'). Once you are content with the alignment, turn 'on' the Registration Marks in Studio and Print the image on your colour printer, using the highest print-quality settings. The print-out will include the alignment marks, which will be read by the Silhouette cutter so that the various apertures will be aligned with the colour image.”
     
    I already had all the necessary drawings in DXF files and the colour images as TIFF files, so I could simply open these in Silhouette Studio.  The only differences from my previous method were that I printed the colour overlays onto printable vinyl and cut the frames out of 10thou plasticard, to be applied on top of the vinyl sheets.
     
    Finishing the Model
     
    The first task in finishing the detailing of the 3-D printed body was to cut out the frames, using 10thou (0.25 mm) plasticard.  After cutting out, which the Silhouette cutter handled very well, I painted the frames black and hung them on hooks to dry.  If I were making a lot of these, I would try to source some black plasticard.
     

    Silhouette-cut Frames
     
    Next I printed the body sides from the Silhouette Studio software, using my HP ink-jet printer.  I found, after the first test, that the colour setting of my printer needed to be re-adjusted for printing on vinyl sheet, since what should have been ‘chocolate’ came out more like ‘Indian red’.  After a couple of test prints, I achieved a more acceptable colour for the lower body sides.
     
    I had found before that the vinyl does not stick well to the body shell after spraying with grey primer so, while painting around all the window opening in Indian Red,  I decided to continue painting the whole side with enamel, hoping it would provide a better surface for attaching the vinyl.
     
    It was quite tricky to get the vinyl sheet into good alignment with the window apertures, since there is no ‘slide’ and the adhesive grabs very quickly.  As it happened, the first attempt went well but it does need considerable care!  The clerestory was particularly tricky because I had to tuck the top edge under the roof overhang, for which I used cocktail sticks to push the vinyl layer into place before tamping down over the rest of the side.  It was good to have painted the entire sides first, as any very slight mis-alignment round the windows was not apparent.
     
    Finally, I had to decide how to attach the outside frames, which are rather ‘floppy’, since they are just a skeleton of narrow beadings.  I did not want to damage the appearance of the printed vinyl sides by applying excess adhesive.  I decided to go back to the can of hair-spray that I bought some years ago for experiments on printing directly onto plasticard.  It acts as a mild ‘hold’ adhesive while leaving no visible traces on the surface and I was pleased to find that, in general, it worked well, although adhesion in a few places was poor.  I felt, however, that once the main part of the frame was correctly positioned, everything could be made firmer by applying an overall coat of varnish.  This was another very tricky part of the assembly and the alignment on my first attempt was not perfect.  The ‘First’ lettering on the doors is not quite centred within the frames.  I shall continue to experiment, if I try further assemblies of this type.
     
    Conclusions
     
    Overall, this method of construction worked well.  I haven’t really broken any new ground, since I have previously added new printed sides and framing to one of the well-known Ratio carriage kits.  The only new aspect in the current model was in using all-plastic materials, rather than photo-paper and card.
    Perhaps I would have felt a greater degree of satisfaction if I had built a new type of carriage, rather than repeating one I had made before.  There is no doubt that the use of a 3D-printed body has resulted in a much stronger overall structure.  I’m sure this could be achieved for Silhouette-cut sides by laminating more layers, although that approach brings risks from warping and mis-alignment.
     
    A photograph, which had previously cruelly revealed the defects in my earlier carriages, does show that the new overall structure, especially the clerestory roof, stands up well to close scrutiny.
     

    G13 Family Saloon – final details to be added
     
    For this photo, I placed the body on the Cleminson chassis that I built several years ago for my original Family Saloon.  I have nostalgic feelings for the older carriage, which now has a well-worn look but is noticeably out-of-true in several places.  The really huge advantage of the 3D-printed body lies in the clerestory roof, which I never managed to fabricate well.
     
    From my point of view, it was all part of a learning exercise, helping me to feel more confident in tackling more complex shapes, where the 3D-printing method will come into its own.  I shall now switch my attention to such structures.
     
     
    Blanche Wilcote is pleased that the new carriage is much less draughty than the old one, which had various gaps along the seams!  So that is some reward for my efforts
     
    Mike
     
  15. MikeOxon

    What big wheels you’ve got!

    By MikeOxon,


    General
    I reached something of an impasse at the end of my previous entry in this blog. Lots of problems to be addressed but no clear routes to solutions.
     
    Possibly the most striking feature of the ‘Waverley’ class 4-4-0s was their wholly-exposed coupled driving wheels. I therefore felt strongly that this was an aspect that my model had to capture. Unfortunately, there seems to be no commercial source of 24 spoke, 7 foot-diameter driving wheels and, while I fantasised a little about building my own, I quickly realised that it was beyond my skills and the tools that I have available.
     
    Then, looking at my Tri-ang ‘Lord of the Isles’ model, I suddenly realised that its driving wheels had the requisite 24 spokes! For once, however, Tri-ang had not made their driving wheels grossly under-size but they did have massive tyres and flanges, which offered scope for turning them down to somewhere near 7 feet. So, the next step was to buy some spare wheels (Hornby X275/X276) for only £1.90 a pair.
     
    Not having a lathe, I mounted the wheels onto a spindle and fitted them in the chuck of my stand-mounted Dremel drill. I then used a variety of tools, including a diamond slitting disk fitted in another mini-drill, and an assortment of files, to reduce the size of the massive flanges. Even with a minimal flange, the overall diameter (including flanges) was still 30 mm, so I decided to remove the flanges altogether from the leading pair of drivers. With a long, rigid wheelbase, I felt that this would give the locomotive some chance of negotiating curves.
     

     
     
    Another important reason for reducing the overall wheel diameter is that the coupled wheels are mounted very close together, so any excess size increases the minimum possible wheelbase, which I felt would be even more obvious than over-size wheels. After completely removing one set of flanges, I found that I could use a wheelbase of 31 mm, which I was prepared to accept as close enough in scale to the prototype’s 7’ 5”.
     
    Nevertheless, this compromise meant that I had to redesign the inside frames, so as to provide extra clearance between the driving wheels. Since I had drawn the original frames with the Silhouette Studio software, it was easy to modify the details, without having to do a complete re-design. I feel that the revised frames do not ‘look’ appreciably different from the original frames shown in my previous post.
     

     
     
    I ‘tacked’ the frames temporarily to the boiler assembly by means of Uhu adhesive and then checked that all the wheels could be fitted, without fouling each other. Now my model was beginning to look like a Broad Gauge locomotive.
     

     
     
    I assessed the appearance from different angles and believe that it has captured much of the ‘character’ of the prototype, even though it is not a strictly accurate model
     

     
     
    Finally, I set the part-finished model head-to-head with my model of the ‘Rover’ class 4-2-2. This comparison revealed that I need to make further adjustments to the boiler pitch and the frame height but, since nothing is fixed as yet, this will be easy to rectify.
     

     
     
    Of course, I’m not completely out of the woods, yet. There is still the matter of those bicycle-like splashers, with their brass facings.
     
    Even so, I am very pleased to have reached the current stage, which at one time seemed an impossible task for me. I find it rather inspiring to see this mid-19th century locomotive coming to reality on my work bench.
     
    Mike
  16. MikeOxon
    I've read that some people like blog posts to be fairly substantial, so I'll include a resumé, in this post, of the story so far, before showing photos of my model of GWR No. 184. There are more details in the earlier posts in this series.
     
    Background
     
    I chose this prototype since it was one of the first standard-gauge engines to run on the GWR, following the absorption of the Oxford, Worcester, and Wolverhampton Railway (otherwise known as the 'Old Worse and Worse'). No.184 was one of six engines built by E.B.Wilson in 1853 and this particular engine survived almost until the end of the 19th century, appearing in several photographs that document its re-builds at Wolverhampton Works in 1871 and 1893 It is also the subject of a sketch by C M Doncaster, which shows it at the head of a train of 6-wheelers approaching Reading Station. I decided that I would like to emulate this scene on my model railway.
     
    Method of Construction
     
    The first step was to create some drawings of the engine from a mixture of both photographs and drawings of similar engine types. I used these drawings to construct paper templates, so that I could check clearances, etc. for an 00-gauge model. I then cut out the major components from brass sheet and rolled the boiler by hand. The following photographs illustrate the major stages in the construction, together with dates at which each stage was achieved.
     

     
    Because I was working mainly from paper templates, I made very few dimensioned drawings and fitted together the various parts by 'trial and error'. I often cut out a part from paper, initially, and then, after trimming the paper to provide a close fit, I replicated the component in brass. The major components of the engine body fitted together as shown below. Although I was concerned, in the early stages, that the structure felt rather weak, it all became pretty rigid, once the various sheets were soldered together. I was pleased to find that very little filler was needed, since the boiler and firebox fitted closely around the wheel arches. I used some book-binders' glue (which appears to be a tough form of PVA), to fill any remaining gaps. This proved very effective in providing a resilient seal, rather like bath caulk.
     

     
    Chassis
     
    I built the chassis from a pair of brass strips, soldered together while drilling the holes for the axles and frame spacers. For more details,see part 6 of this series. Initially, the model will be powered by its tender but I have made sure that there is sufficient space for a motor to be fitted inside the boiler.
     
    To achieve this space, I cut away the lower half of the boiler, where it is hidden behind the splashers and the firebox. Top and underside views of the body are shown below:
     

     
    Detailing
     
    I made boiler bands from 5 thou brass sheet, cut into strips and stuck onto the boiler barrel using super glue. I also added rivet detail to the outside frames by using the brass detailing strips supplied by 'Mainly Trains'. These were also fixed with super glue.
     
    I made sandboxes from small wooden blocks, wrapped in 2 thou brass foil, with a separate brass 'lid'. The springs are white metal, mounted above the footplate on nickel-silver wire uprights. Handrails are, similarly, nickel-silver wire (0.5mm). Outside cranks and coupling rods are from 'Alan Gibson'.
     
    Since the main dimensions and appearance of the prototype changed very little throughout its re-builds, I keep in mind the possibility of back-dating my model to an earlier time period. I found that 'RT models' supply 4mm scale brass dome and safety valve cover in the E.B.Wilson style but, for the present, I have installed an 'Alan Gibson' '517-type' brass dome.
     

     
    Painting
     
    I decided to paint my model in a representation of GWR 'Wolverhampton 'livery, which I understand to have been based on the former OW&W livery. The Wolverhampton style of painting differed from Swindon, in that all items above the footplate, including splasher fronts and sand-boxes, were the same colour as the boiler, while the frames below the footplate and the wheels were brown.
     
    My first step in painting, after a thorough clean-up of all the brass-work, was to apply an overall spray of grey primer. Then, for the boiler colour, I chose to use 'Rustoleum Painter's Touch' dark green enamel, which looked a reasonable match to descriptions of the 'blue-green' colour used at Wolverhampton This is a water-soluble paint,with a rather 'plastic' feel that dries to a 'wet look' finish. I found that it brushed more smoothly when thinned with a little water I am fortunate to have inherited some 'Winsor & Newton' series 3A sable brushes, which have superb paint-holding capacity and super-fine tips. They really make hand painting a pleasure! I think the finish will need some matt varnish in due course, since even 19th-century cleaners would find it had to maintain such a gloss!
     
    For the brown frames and wheels, I used 'Winsor & Newton' Burnt Sienna artists' acrylic colour. Again, I thinned the paint from the tube, to produce a smooth brushing consistency with excellent covering power. If any of the 'painting gurus' on this site happen to read this, I should be interested to know why artists' acrylic is rarely mentioned for model painting, since I found it very pleasant to use.
     
    For the black areas (footplate, smokebox, etc.), I used Humbrol black acrylic, which I find dries to a good-looking 'oily' sheen.
     
    Lining and Lettering
    .
    I prepared artwork for the lining and number plate by the methods I have described previously in this blog.
     
    Wolverhampton lining was black and white, so I used white inkjet transfer paper. My copy of 'Great Western Way' (1st.ed.) gives few details, except to state that the white edges of the lining were broader than the equivalent orange edging used at Swindon. Another factor to take into account is that the printed black centre lines tend to spread slightly into the unpainted white edges. I made a few test prints to determine appropriate widths that seemed to match the visual impression seen in old photographs.
     
    I scanned one side of the model to provide 'actual size' templates of the cab sides and sand boxes, on which I 'drew' the lining on the computer screen, as shown below:
     

     
    For the brass fillet between the boiler and smokebox, I cut a sliver of 1 thou brass shim and attached it to the boiler with PVA glue. Cutting this very thin material produced a natural curl, which I exploited to help hold it firmly against the boiler.
     
    The 'Current' Model
     
    I cannot describe the model as 'finished', since there are still many additions and improvements that I wish to make, but it has reached the stage where photographs can give a reasonable impression of what I have been trying to achieve.
     

    My model of GWR No.184 attached to 2500 gal Dean tender (powered)
     
    As I mentioned above, the model is not yet 'self-powered' but can be 'driven' by means of the 2500 gal Dean tender that I described in a previous post.
     
    When seen in isolation, it's not easy to appreciate its small size and rather 'antique' feel, so I have taken a couple of photos to set it in context.
     
    One of my aims, at the outset of this project, was to represent the scene drawn by C M Doncaster. Here is my interpretation, built around some of the models I have built of GWR 6-wheel coaches:
     

     
     
    and here it is alongside Armstrong's 'Queen' class express 2-2-2, the first of which was designed 20 years after No.184 was built:
     

    GWR No.184 alongside 'Queen' class at North Leigh
     
    Conclusion
     
    This was the third model that I have completely 'scratch-built' from brass sheet; the others were the 'Queen'-class 2-2-2 and my interpretation of Dean's 4-2-4T, No.9.
     
    It proved a trickier build than the others, mainly because of the 4-coupled chassis but, also, the raised firebox and the need to leave space inside the structure to add a motor at a later date. My 'Queen' was only ever intended to be tender-powered, while the 4-2-4T is powered by a 'Tenshodo' SPUD in the rear bogie.
     
    Construction took longer than I expected because of many non modelling-related interruptions and I was beginning to get a little impatient towards the end. There are quite a few mistakes that could be corrected in a re-build, especially the shape of the top of the firebox. I've decided, however, to set engine building on one side, for a while, so that I can concentrate on some scenic work, which is badly needed on my layout, if it is ever to represent the sort of 19th-century scene that I have in mind.
     
    If anyone else feels moved to have a go at scratch-building a 19th-century type of locomotive, I would recommend starting with a 2-2-2, since these are very straightforward, with none of the alignment problems associated with coupled engines. I feel that, if you can build a wagon, then a 2-2-2 engine is not that different! Also, I have found that brass sheet is an easy material to use – in some ways easier than plastic card, because it can be folded and rolled into curved shapes. Once you've cut your teeth on a simple prototype, you can follow whatever course catches your imagination.
     
    I also realise how much I have learned, and absorbed, from various books, and should mention, in particular, Iain Rice's “Etched Loco Construction”, which helped me over many hurdles.
     
    Mike
  17. MikeOxon
    A couple of members on the forums have indicated that they would like to know more details of how I create my own transfers for lining and lettering pre-grouping models.
     
    I have already written a little in my earlier post "Cheats Lining & Lettering" and I must also admit that, because I model GWR prototypes, much of what is needed can actually be bought from the trade! I am one of those "scratch-builders" that likes to do things "because I can"
     
    I use a Windows PC, together with Adobe Photoshop Elements (PSE) software, to produce the 'artwork' and print onto white Inkjet Water-Slide Decal Paper from http://www.craftycomputerpaper.co.uk/ , using a Hewlett-Packard Deskjet 6980 printer. Of course, other equipment and materials could be used but these are what I am familiar with and will describe in detail. I use the 'white' decal paper, rather than 'clear', because printers are designed to print on a white background and the inks tend to be too transparent to show up against a dark background. In addition, there is no white ink, so white areas are formed by letting the white background show through.
     
    I shall begin by describing a rather basic operation:- that of creating boiler bands.
     
    To do this, we have to set up the computer and associated printer to work to scale and then use these tools to produce the various coloured lines which, together, make up the band. For GWR locomotives, during the period 1881 - 1906, the standard boiler bands consisted of a 1½" wide black band, flanked by two 1/8" orange-chrome lines.
     
    Since 1/8" is the smallest dimension in the lining that I am designing, I base the settings in PSE around this figure. I feel that a 3 pixel (px) wide line is the minimum for reliable and even printing, so my first calculations assume that 3 pixels will equate to 1/8" in the prototype. Hence, if 1/8" = 3 px then 1" will need 24 px and 1 foot will need 288 px. I model in 4 mm scale (1/76), so these 288 px will be required to represent a distance of 4 mm on the model, or 72 px = 1mm. Hence, I need to set up a new page in PSE to a resolution of 720 px/cm. Click on 'New' and enter the dimensions shown below into the box that appears (I use a rather old version of PSE, so your input box may look different), These dimension will provide 10 cm lengths of lining - adjust the width of the drawing area if you need more or less.
     

    New Page Set-up in PSE
     
    When producing transfers, I aim to surround the patterned areas with the underlying body colour, to make for a smooth transition to adjacent painted areas. Thus, I flood-fill the new page with the basic green colour of the boiler. When I open the 'pre-1928' green from http://www.gwr.org.uk/liverieslococolour.html , the PSE colour picker indicates the colour as R,G,B = (1,46,3), so I fill with this, as the background colour.
     
    Now use the 'Line' tool to draw a horizontal black line, with the width set to 36 px (= 1½") - hold down the 'Shift' key to constrain the tool. 'Simplify' this line, then draw an orange-chrome line of width 3px, close to the first line. For 'orange-chrome', I used R,G,B = (255,128,64). Zoom in to 'Actual Size' and use the 'Move' tool to place the orange line immediately adjacent to the black line. Repeat for the orange line on the opposite side of the band, and then 'Flatten' the image, in the 'Layers' menu. If you want to know more about moving and manipulating layers in PSE, you might like to read my article at http://home.btconnect.com/mike.flemming/layers.htm
     

    Placing line with the 'Move' tool in PSE
     
    If you want more lengths of this lining, select the complete band then copy and paste duplicate versions on your transfer sheet. Use 'Print Preview' to see how the complete transfer will appear on the page and, preferably, move the image towards the top edge of the sheet, so that the rest of the sheet can be used for other images. If you want to save the image for later use, save it as a TIFF file, rather than JPEG, to avoid any artefacts, which will spoil the crisp edges of the lines.
     

    Print Preview screen in PSE
     
    It's now necessary to set up your printer for high-quality printing at maximum dots per inch (DPI). On my HP printer, I select 'Presentation printing' under the 'Printing Shortcuts' tab and 'Maximum dpi' under the 'Paper/Quality' tab. This results in 4800 x 1200 dpi printing. Other printers will have different set-up procedures, so experimentation may be needed to get the best results.
     
    I place a single sheet of decal paper in the feeder tray and, after printing, leave about 15 minutes for the ink to dry thoroughly. I use a guillotine to cut off the printed area - the rest of the sheet can be saved for further use. Next, it is vital to spray or brush the whole surface of the decal with waterproof varnish - I use Humbrol clear gloss, as I find the matt finish gives a slightly 'milky' look and softens the detail. Some printer inks (e.g. Epson) are claimed to be waterproof but I have not tried these and probably wouldn't trust them without varnish!
     
    I cut out the bands individually from the transfer sheet, leaving a green edge to blend with the boiler paintwork, and immerse each one in clean (demin or distilled) water, with a couple of drops of washing-up liquid added as a wetting agent. The decals will curl up at first but should flatten after a minute or so - this can be helped with a paintbrush. I also 'paint' the body of the model with the same water and then use a paintbrush to slide the transfer off the backing paper onto the model, and tamp it down into place with the brush. Job done
     
    Actually, you may find that 'true scale' lining is too 'subtle' in 4 mm scale (or smaller), so I have found that a slightly wider orange line can be desirable. Trial and error may be needed to meet your personal taste. The example shown below is my GWR 'Stella', converted from a Mainline Dean Goods. It was my first attempt and I may have over-done the orange a little, at least for photographic purposes - it looks quite good (to me), in practice.
     

    GWR 'Stella' with 'home-made' Lining and Lettering
     
     
    Next time, I will write about panels, corners,and adding logos, etc.
     
    Mike
  18. MikeOxon
    In Part One , I wrote that “this engine had several very unusual features” and, in regard to building a model, “I had to start somewhere and, with so many peculiarities, it was hard to choose. As a ‘gentle introduction’, I decided to start with the two bogies.”
     
    I intend to continue, as far as possible, to follow a line of ‘least resistance’ but before going any further, I collected as much potentially useful information , photos, and drawings as I could.

     
    In his book ‘The British Steam Locomotive 1825 – 1925’ p.106, Ahrons noted that “no description of them, beyond the meagre details in Colburn's " Locomotive Engineering," page 73, has ever been given, and the following account of their constructional details may therefore be of interest.”
     
     
    He continued with quite an extensive article, providing many key dimensions and descriptions of the rubber suspension system. This suspension anticipated Alex Moulton’s work for the Austin Mini by more than a century! Ahron’s also referenced an article in ‘The Engineer’ supplement, 16 Dec 1910 , which provided various sectional views of the engine from which I could make a start.
     
     
    Using these drawings, I started by creating some of those major components for which I had already gained experience in modelling other engines.
     
    Ahrons gave the boiler dimensions as 10ft. 9in. long by 4ft. 0½ in. inside diameter, so I created a tube of length 43 mm, inside dia. 16 mm, with 1 mm wall thickness, which provided a good match to an ‘The Engineer’ drawing. I then added firebox and smokebox by tracing over the drawing and extruding as required, to create solid ‘bodies’ in ‘Fusion 360’. The results were as shown below:
     

    3D model of Boiler Assy. Referenced from an illustration in ‘The Engineer’.
     
    The next ‘familiar’ item on the agenda was a pair of driving wheels. although these are a little larger than usual, at 36 mm diameter, and flangeless. The stages of my usual method are shown below:
     

    My steps in 3D modelling the Driving Wheels
     
    It always surprises me how rapidly something resembling an engine emerges, especially once the platform, chimney and safety valve housing have been added.  The chimney and safety valve cover were created by my usual method of tracing the profile and then using the ‘Revolve’ tool to create the cylindrical ‘bodies’.
     
    The platform was a simple rectangular extrusion from the plan drawing, with the exception that small ‘humps’ had to be raised over each of the bogie wheels. Those early designers did nothing to make life easier for the workmen having to fabricate these shapes by hand!
     

    First impression of my evolving model
     
    That has completed most of the straightforward parts of this engine and it is now time to start tackling its (many) peculiarities!
     
    Outside ‘Frame’
     
    According to Ahrons “The inside bearings were only 5in. long, and therefore additional outside bearings, 9in. long, were provided, the hornblocks of which were riveted to the triangular queen truss " frame," shown outside the driving wheel”.
     
    This ‘frame’ was my next subject for modelling. I created the truss and the outside rim of the splasher as a single ‘body’ in ‘Fusion 360’ by extruding from a drawing, as shown below:
     

    My sketch of the outside frame supporting the driving axle
     
    Most of the sketch was made by using the ‘three point arc’ tool in ‘Fusion 360’ and then I extruded the area coloured blue to form the frame. Next, I extruded the rectangle representing the axlebox, to house the outside bearing. I used another useful tool – the ‘offset’ tool – to create an outer rim around the frame and then ‘pushed’ the annulus (coloured yellow) backwards, to surround the driving wheel as the splasher.
     
    Yoke and Boiler Support Plates
     
    Now it was time to address two more peculiarities – the yoke across the top of the boiler and the curvaceous plates that supported the boiler from the platform. I created both these features by sketching over one of the front-elevation illustrations from ‘The Engineer’.   Again, the most useful sketching tool was the ‘three point arc’ and I then extruded the outlined area to form plates of 0.5 mm thickness, as shown below:
     
    There were two identical ‘yoke’ plates in tandem (blue in my sketch) straddling the top of the boiler, with pivoted brackets between them, to carry the vertical suspension rods above the riving wheels.  Two curved support plates (green) were placed, one on either side of the boiler, immediately ahead of the driving wheel splashers.
     

    Tracing the shapes of the Curved Plates around the Boiler
     
    I then move and rotate the parts I have made and align them against my reference 'canvas'.  I have hidden the rest of the model for clarity:
     

     
    After producing models of these plates, I moved on to the suspension units for the driving axle. These were illustrated and described by Ahrons.  I created my models of the suspension units by sketching the profile over the drawing from ‘The Engineer’ and then using the ‘Revolve’ tool in ‘Fusion 360’ to create the cylindrical rods and the brass pots that contain the india rubber ‘springs, as shown below:
     
     

    My extrusion of a suspension arm alongside ‘The Engineer’ Drawing
     
    Motion Plate
     
    The motion plate was, again, unusual in that its outer edges were shaped to follow similar curves to those of the boiler support plates. For details of their appearance, I had to turn to another drawing, shown in the Broad Gauge Society (BGS) journal ‘Broadsheet’ No.49 (Spring 2003).
     
    The original is one of those ‘split’ drawings, with halves of two different cross sections shown together. To help my visualisation of the engine, I 'mirrored' each half to provide two separate complete cross sections, although I cannot guarantee that all the details on the two sides were perfectly matched.
     

    Two cross-sections created from drawings in BGS ‘Broadsheet’ No.49
     
    In my drawing ‘B’, I have shaded the split motion plate, which has square apertures for the connecting rods to pass through. I copied the outlines of these motion plates, as for the other boiler support plates, and extruded them to 0.5 mm thickness.  The valves were placed between the cylinders and the valve rods are shown with supporting slide bars. There are many other unusual features in these drawings that I shall return to later, including the well tank suspended below the boiler on brackets from the motion plate.
     
    Cylinders
     
    The cylinders themselves were too long to fit within the very short smokebox, so they extended for about 1 foot forwards of the smokebox. They were connected across the width of the engine by a box, which enclosed most of the front end of the valve gear, although there were two tail rods protruding from the front face of this box. A photograph of the front end of No.42 also shows a small steam cock on the centre of his cover
     
    I constructed this box and the cylinder ends as a separate ‘body’, which I then fixed to the front of the smokebox.
     

    My model of the cylinder ‘front end’
     
    That has completed all the main components needed for a ‘top-side’ view of the engine. I have added buffer beams, copied from a different Broad Gauge engine, and extruded the outer sides of the coke bunker. Although simple in external appearance, this bunker has many unusual internal features but I shall come to these later.
     
    In the meantime, my 3D model in ‘Fusion 360’ now looks as shown below:
     

     
    I think this model is beginning to capture something of the ‘presence’ that the original engines must had. I still have a lot of work to do on the underpinnings – especially the attachment points for the bogies and the well tanks but right now I feel it’s time to pause for the Easter break!
     
    Mike
     
     
  19. MikeOxon

    Aeolus Valve Gear

    By MikeOxon,


    General
    Interpreting the Valve Motion
     
    At the end of my previous post, I commented on the surprising layout of the valve gear, as shown on the Lane sketches. The Works drawing of the engine, as originally designed, shows a more usual arrangement, with the weigh-bar placed under the boiler behind the smokebox.
     
    Following further research, however, I have re-interpreted the drawings of Aeolus by E.T. Lane and have, therefore, revised this post on 16th April.
     
    According to the recently published book by Brian Arman: ‘BG Engines – Part 3’, several engines built during the 1840s were fitted with a regulator that was placed immediately above the cylinders. Drawings of, for example, the ‘Prince’ class show that the regulator was operated by levers from a shaft across the front of the smokebox, by means of a control on the right-hand side of the footplate.
     
    I had not been happy with my previous thought that these components might have represented a ‘throw back’ to much earlier type of valve gear as used on Stephenson’s ‘Planet’ series of engines. It appears that Gooch started fitting his fixed link expansion gear from 1843, so I now suggest that the rebuilt Aeolus may have been one of the first engines to have been fitted with this gear. Far from being a ‘throw back’, this engine may have a been a test-bed for the new arrangement, which was fitted to Great Western and other engines shortly afterwards.
     
    This change of interpretation does not actually affect the design of my model very much, it is simply that the parts that I have modelled serve a different purpose from what I initially had thought. The new interpretation is much more convincing than my earlier thoughts.
     
    Initially, I produced every piece of the regulator mechanism as a separate ‘body’, so that I could move them all around independently to what seemed to be the most appropriate positions. The plausible outcome looked as below:
     

    Potential Layout of regulator mechanism in ‘Fusion 360’
     
    The next requirement is to provide supports for all these gubbins and, on this matter, I find the available sketches and drawings very confusing.
     
    Lane’s sketches, which were clearly used as the basis of the much later G.F.Bird drawings, show what appears to be a rather massive support structure carrying a shaped boss to hold the transverse shaft, presumably in the centre-line of the locomotive. There is no clear indication of how the outer ends of the shaft were supported. As an interim measure, I decided to add support brackets extending forward from the smokebox, as shown in drawings of  ‘Prince’, but there is no indication of such supports in the Lane sketches. There also seems to be a pillar supporting horizontal bars but with no indication of where these were placed laterally. Was there perhaps a curved rail around the front of the engine?
     

    20th Century drawing by G.F. Bird, derived from Lane’s sketches
     
     
    Another puzzle is provided by those rather splendid curved side-brackets shown on the original Works drawing of ‘Aeolus’. It is not clear whether these were retained when the smaller wheels and consequently lower outside frames, were fitted during re-building. It is also not clear where they appeared along the length of the engine, although the photograph of ‘Vulcan’ suggests they were at the front of the smokebox and rear of the firebox, with intermediate straight supports to the sides of the boiler. Bird seems to have assumed this in his drawing but it not clear on Lane’s sketches.
     
    Adding Surface Details
     
    One of the pleasures of using 3D modelling tools is the ability to add rows of rivets, either in regular arrays or following a defined path, by using a few keystrokes and the ‘pattern’ commands.
     
    For the frames, I drew one rivet and then used the rectangular pattern tool to create a 2 x 36 array for all the rivets on one side and then the Copy command to replicate them on the other side. For the smokebox front, I drew one rivet and then told them to follow a path around the edge of the smokebox – job done! With these additions, my 3D model in ‘Fusion 360’ looks like this:
     

    Assembly of parts within ‘Fusion 360’
     
    Preparation for Printing
     
    As I have pointed out before, there is a difference between a 3D model and a printable 3D model.
     
    My approach is to break the complete model down into several parts that can be printed individually and then assembled as a ‘kit’. In deciding how to separate the parts, I try to ensure that each part has a flat surface which can be laid on my FDM printer bed. An advantage of this approach is that the print times for individual parts can be quite short so, if some re-design is necessary, it can be done without having to re-print the entire model.
     
    Smokebox Front
     
    I felt that the most demanding task for my printer would be to reproduce the details on the smokebox front, including the supports for the valve gear. For this reason, I did a test print of this part first, since the overall success of my model depends on how well this region can be represented. After printing, my first trial looked as below (Note that I shall print the weigh bars and levers separately.):
     

    My first test 3D print of the Smokebox Front
     
    There were a few stray strands of filament that had to be carefully removed with fine tweezers but overall, I was very pleasantly surprised to see how well the details had been rendered, including the supports for the valve gear and the protruding piston rods, etc.
     
    Further Components
     
    Encouraged by this first trial (which took only 10 mins to print), I continued to select and print the various other components.
     
    I have learned to keep the various parts in the form of individual ‘bodies’ within ‘Fusion 360’, so that I keep as many options as possible open for printing. In this case, the front of the smokebox, shown above, is a ‘body’ in its own right, separate from the rest of the smokebox. The back of this part is completely flat, to lie on the printer bed, while the front has some very fine details, including the piston rods and eccentric rods, protruding from the front surface.
     
    Most of the other parts are simpler and very similar to the equivalent parts on models that I have constructed before. For example, the boiler, smokebox body, and firebox appeared as below, immediately after printing:
     

    My 3D Printed Boiler Components on Printer Bed
     
    I am especially pleased with the way the combined dome/safety valve cover has turned out. I printed this in two parts: the fluted barrel and the curvaceous top cover. I provided each part with dowels and mating sockets, so that they plugged together on top of the firebox. After fitting together, the printed parts look like this:
     

    My 3D-printed Dome/Safety Valve Cover on Firebox
     
    I am reminded of Brunel’s statement in a letter to the engineer T. E. Harrison on 5th March 1838. Forgive the sexist remarks but I am quoting verbatim: “Lastly let me call your attention to the appearance - we have a splendid engine of Stephenson's, it would be a beautiful ornament in the most elegant drawing room and we have another of Quaker-like simplicity carried even to shabbyness but very possibly as good as engine, but the difference in the care bestowed by the engine man, the favour in which it is held by others and even oneself, not to mention the public, is striking. A plain young lady however amiable is apt to be neglected. Now your engine is capable of being made very handsome, and it ought to be so.” [MacDermot, History of the GWR]
     
    I do think this dome is very handsome and has printed very well!
     
    Now, I have all the parts needed to complete the boiler assembly. The major parts all plug together, while the smokebox front is glued to the flat front-face of the smokebox itself. I have not yet attached the weigh-bar and rocking levers, to actuate the valve rods, which can be seen protruding from the front face:
     

    My Model ‘Aeolus’ Boiler
     
    This is a good place to pause before embarking on the running gear, which will support these completed parts. That will be my next post.
     
    Mike
     
     
     
  20. MikeOxon

    A Year in Retrospect

    By MikeOxon,


    General
    Last year, as 2020 drew towards its close and we prepared for the holiday season, I showed a collection of my North Leigh engines ‘on shed’. I think that, at that time, we were all hoping that the difficulties caused by Covid would soon be over. Sadly, as another year draws to its close, we are still in a period of uncertainty, waiting to learn what sort of threat the latest variant may pose.
     
    Whatever else has happened, I have found plenty of time to practise my modelling abilities with my 3D-printer and its associated software. I have reached a point from where I can look back on my earlier efforts and think that I learned quite a lot over the course of the year! Tasks that I found rather difficult are now just routine matters although, as with all complex software, I’m sure that I still have a great deal to learn. After all, I’ve been using Photoshop for processing photo images for very many years and am still learning new tricks!!
     
    In particular, I have developed my method of designing parts by extruding from drawings and then making a test assembly in the software, to ensure that everything will fit together correctly.


    Overview of my 3D-modelling Method
     
    I feel that I’ve had a creative year, which started by my producing models of a train from the early days of the GWR, based on a lithograph by J.C.Bourne. These models included a Fire-Fly class engine, together with a train including a Passenger Luggage Box, Open 2nd-class Carriage, Posting Carriage, Horse Box, and Carriage Truck.


    My Model of Fire-Fly class ‘Argus’
     
    As any regular readers of my blog will know, I started my Broad Gauge modelling by creating models of the engines and stock that were recorded in the official report on the accident which took place near Bullo Pill in 1868.
     
    I then branched out into building various other models, until the wheel unexpectedly turned full-circle. I set out to model one of the Gooch ‘bogie-class’ engines, which I had thought of as South Devon engines, only to find that several worked the Forest of Dean line from Bullo Pill!  I also discovered that some members of the Sir Watkin class were allocated to Bull Pill for a few years, before gauge conversion took place there in 1872.


    My model of Sir Watkin class 0-6-0T
     
    For the lining on the side tanks, I returned to a method I described in detail a few years ago.  The main difference now is that I used my Silhouette cutter to create the shapes from self-coloured adhesive vinyl, which I teased into position with cocktail sticks before pressing down firmly.
     
    Bullo Pill & The Forest of Dean Branch
     
    Bullo Pill lies on the railway from Gloucester to South Wales, at a point where the line runs between the Western bank of the River Severn and the Forest of Dean. At the beginning of the 19th century, a tramway was built to bring coal and iron ore from the mines and quarries in the Forest down to a small dock at Bullo Pill. The tramway passed through what was then the longest tramway tunnel in the world at Haie Hill. In the 1850s,  Brunel was tasked with converting the tramway into a Broad Gauge railway, which involved widening of the Haie Hill tunnel and the construction of additional tunnels further up the line. It was at that time that my wife’s Gt-Gt-Grandfather arrived in the village of Soudley, at the upper end of the Haie Hill tunnel, where the family lived near the Iron Works.
     


    Soudley Iron Works – Haie Hill Tunnel entrance is behind the dark tree on LHS
     
    Several of the sons from this family, including my wife’s Gt-Grandfather, started their GWR careers at Bullo Pill in the 1860s, as I have described in an earlier post. Bullo Pill Junction connected the Forest of Dean Railway to the GWR main line, with additional branches into the dock area. Tipping machinery was installed at the dock, to transfer coal and iron ore from railway wagons into barges bound for South Wales and across the River Severn.
     
     

    Bullo Pill Dock
     
    There was quite a complex network of railway tracks around the dock and to loading jetties on the banks of the River Severn, as shown in the section of Ordnance Survey 25” map, below:
     

    Map of Bullo Pill Dock
     
    Th site of the Mail Train accident, which started my interest in Broad Gauge modelling, was about one mile (1.6 km) South of here, on the main line than can be seen on the LHS of the map, running North-South.
     
    End of Year Overview
     
    My modelling activities this year have added several items of stock which are appropriate for operations on the Forest of Dean Branch, down to this dock.
     
    I thought that, as an ‘End of Year’ overview, I would show my Broad Gauge models in a ‘Bullo Pill’ context. Collecting them together has revealed just how much detailing remains to be completed.  Please don’t look too closely!  I even had to have an extra 3D-printing session, to produce some additional roofs, wheels, and other small parts, which had been ‘shared’ between different vehicles!
     
    After taking my photo, I spent some time with Photoshop, blending in a ‘Bullo Pill dock’ back-scene. It’s given me some ideas for creating a diorama – something to think about for next year.
     

    My BG Model Collection, set in a Bullo Pill context
     
    May I conclude by thanking my readers for all the encouragement you have given me over the year and send you my best wishes for the coming Holiday season.
     
    Mike
  21. MikeOxon

    Easy-Peasy Carriage Build

    By MikeOxon,


    General
    Easy-Peasy Carriage Build or ‘How to build a carriage with no drawing, no measuring, and little time’
     
    One of the vehicles I wanted to add to my collection for use with my ‘Firefly’-class locomotive was the early type of 6-wheel ‘open’ 2nd-class carriage. There is a full-size replica at Didcot Railway Centre, as shown below:
     

    Didcot Railway Centre – Replica ‘Fire Fly’ and train
     
    At first glance, those panelled sides might look to be a modeller’s nightmare but 3D-printing makes it extremely simple!
     
    There are drawings of this carriage, by Eddy Lane in the appropriate Data Sheet from the Broad Gauge Society, so my first step was simply to import a copy of his drawing, as a ‘Canvas’, into ‘Fusion 360’. The drawing shows the overall length as 27’ 2½“ (i.e. 108.8 mm in 4 mm/ft scale).  Again, ‘Fusion 360’ makes it very easy to scale the ‘Canvas’ by means of a ‘Calibrate’ command, which simply requires the length of one reference line to be stated.
     
    Now that I have learned about some of the tools that ‘Fusion 360’ has in its armoury, I have realised that there are many aids to creating arrays of similar features, such as the regular panels on this carriage. In fact, I only needed to create four types of rectangle: ( i ) the outline of the complete side, ( ii ) a ‘window’ opening with rounded corners, ( iii ) the upper side panel, and ( iv ) the lower side panel. I simply created one each of the required types, in registration with the ‘Canvas’ . After that, I could use the ‘Pattern’ tool to replicate as many identical rectangles as were needed to complete the entire carriage side! The same method can be used both for the ends and for the internal partitions.
     
    Thus, a few simple steps were all that was necessary to turn a published drawing into a three dimensional model. The overall procedure is illustrated below:
     

    Stages in converting a published drawing into a 3D model, using 'Fusion 360'
     
    I now had three ‘printable’ bodies that I transferred to my ‘Cura’ software, to produce a ‘sliced’ model suitable for my printer. Since these individual components only take a few minutes each, to print, I laid out all five partitions together as a batch, for which the total print time was only 45 minutes.  I also printed the sides and ends in pairs, with similarly short print times.
     

    Set of Printed Partitions created as a single job
     
    To assemble the parts, I fixed the ends and all the partitions to one side, using super-glue, starting from one end and adding the partitions one at a time. I then weighted the assembly until the glue had cured.
     

    Steps in assembling the components parts of the model
     
    After completing this side by adding the final end panel, I applied super-glue to the outer edges of all the partitions and the two ends. I then offered up the remaining side, pressing it into place until the glue ‘caught’. I laid the model on its side and weighted the upper surface until the glue had cured.
     
    To show the end result of less than a day’s work, I added a chassis and a roof from one of the carriages I had already built:
     

    My ‘easy-peasy’ carriage
     
    Of course, there are lots of finishing touches to add. The purpose of this short post was to show how a seemingly complicated design can be created quickly and easily, by making use of the tools available in 3D design software.
     
    Mike
  22. MikeOxon

    Into a New Dimension

    By MikeOxon,


    general
    Five years ago, I came across JCL's splendid thread about the Silhouette cutters at an appropriate time for dropping heavy hints before Christmas. This resulted in my acquiring a 'Silhouette Portrait' machine and loads of ideas for making my own coaches and buildings.
     
    It proved very fortuitous since, only a few months before, I had decided to return to railway modelling by re-furbishing my old small layout and turning back the clock to the 19th century period. The cutter was a great source of inspiration, since it allowed me to explore a wide range of carriages from the Dean period and earlier. I greatly enjoyed constructing models of long-forgotten prototypes.
     
    Time has moved on and I have been feeling for a while that I should consider machines capable of working in three dimensions, rather than just cutting out flat sheets. Until recently, the affordable 3D printers all seemed rather ‘geeky’ and needed a rather daunting amount of ‘setting up’. In addition, they were not the sort of thing that could easily fit into the domestic environment, where I do my modelling!
     
    Then, while browsing on the web, shortly before the recent Christmas, I spotted a very neat-looking mini-printer called the Geeetech E180 at a very reasonable price (<£200). It had received several favourable reviews and was noted as being especially suitable for use by children. That fact re-assured me that it should not prove too difficult to set up and operate! I wasn’t looking for anything particularly sophisticated but wanted to ‘dip a toe’ in the water and explore the possibilities for making various small parts and fittings for the ‘odd-ball’ locomotives and other vehicles that I enjoy creating. The overall build volume of 130 x 130 x 130 mm seemed adequate for my purposes.
     
    Once the festivities were over, I got around to setting up my new machine and was very favourably impressed by its compact size and attractive appearance. There was no doubt that it would sit easily on my work-desk but could be lifted out of the way when not required. I also discovered that the dust cover for a Kenwood Chef mixer was a perfect fit, to keep the machine dust free while in storage.
     

     
    When it came to switching on, I realised that the supplied ‘manual’ actually contained very little information and that the operating instructions were cryptic, to say the least. Even the software has to be loaded from a ‘user forum’ and it was not at all obvious where to begin. At first, I downloaded an old version of the software by mistake, but eventually managed to track down the (apparently current) version of ‘EasyPrint3D v.1.2.6’. This can load an STL-format 3D model and slice it into layers, for laying down by the printer. The ‘layer’ model can be saved onto a mini-SD card, which fits into a slot on the printer. (It’s also possible to connect directly to the printer though a USB lead)
     
    Initially, the ‘slicer’ would not start on my Windows 7, 64-bit machine. The EasyPrint software uses a version of the ‘Cura’ engine that needs the 32-bit version of vcomp140.dll to be installed in the SysWOW64 subdirectory of the Windows directory. (on my machine, it wasn’t present). The main EasyPrint.exe file, in the Program Files (x32) folder, also needs to be run in ‘Adminstrator’ mode, so that the preferences can be set for the E180 printer.
     
    An SD card was supplied with the printer but with no information about what it actually contained. The file was called bitonga8.gcode and I have subsequently discovered that it creates a largish pot. (I think bitonga is Chinese for pen-holder) Not quite what I wanted but it provided somewhere to start.
     
     

    There is a real need for a simple guide to starting out with a printer like this but it would be presumptuous of me to attempt to write up my findings in detail, before I have gained a lot more experience. The following are a summary of my findings, so far.
     
    Setting up the Printer
     
    The first task is to level the print bed and I found some 'YouTube' videos, which demonstrated that their authors didn’t really understand the machine either!
     
    I worked out that one starts with the central point on the ‘levelling’ display (called position 5) and uses the up/down buttons on the touch screen, to adjust the height of the printer head until it just grabs a sheet of paper laid on the bed. It was not immediately obvious that the height can be varied in either 0.5 mm or 0.05 mm steps, to achieve the required result. After that, the print head can be moved to the four corners of the bed in turn and small screws (initially hidden under the cutting mat) have to be turned until the paper is ‘grabbed’ by the same amount as at the centre. By going around the four corners a couple of times, I found it quite easy to get an even ‘feel’ at all these pre-set locations.
     
    The next task is to feed the filament to the print head. There is a small lever under the feed mechanism (not mentioned in instructions) that releases the grip and allows the thread to be pushed by hand along a clear PTFE tube to the ‘hot end’, which carries the actual extruder. Some cryptic symbols on the display allow the filament to be ‘motored’ forwards or back very slowly.
     
    Starting a Print
     
    Now, with the filament in position and the SD card (containing the model file) inserted, a press of the print button allows the file to be selected, when fans start whirring and the print head moves to the start position. Almost at once, filament starts to be laid down on the bed and, very slowly, layers begin to build. The display shows an estimated time to completion and, for the sample file, this was around 23 hours! I watched for a while and then, after the first few layers, the model detached itself from the bed. Obviously, still quite a lot to be learned!
     

     
    One of the child-friendly features of this printer is there there is no heated bed and the hot end has a maximum temperature of 200°C, although all the easily touchable parts remain cool. This limits the type of filament that can be used to PLA but does mean that there are no unpleasant fumes emitted while the machine is working.
     
    Choosing a Model
     
    I decided to look on the web for some STL models that were of greater interest to me and also substantially smaller than the example provided on the SD card. I found several railway-related models on the Cults website (https://cults3d.com/en/collections/stl-file-train) and chose a free version of Thomas the Tank Engine for my next trial.
     
    I loaded the thomas_body.STL file into EasyPrint, rotated and centred the model within the 3D box (set up for the E180 printer – which is not the default, as I quickly discovered). Then I pressed the ‘slicer’ option to generate the .GCO file, which I saved onto the mini-SD card. (I have an adapter, to use this card with my regular SD card reader/writer)
     

     
    Completing a Print
     
    Once the SD card was in the printer, I pressed the ‘print’ button on the touch screen and the machine whirred into action. It estimated that the job would take about 4 hours. All started well but then after the first couple of layers, the model again came loose on the bed, so I stopped the printing. I decided to apply a little Pritt Stick glue to the bed, to improve adhesion, and started again. This time all went well!
     

     
    Because I was impatient to see how the printing would progress, I used the controls to accelerate the print speed to 150% and then sat back to watch progress. After about an hour, the model was building up nicely and I invited my wife to come and watch the process in action. Unfortunately, I then managed to knock the power lead out of the printer, which promptly shut down.
     
    One of the claimed features of this printer is its ability to recover from a power failure. So, after restoring the power, I hit the ‘resume’ button. Something went wrong, because the top of the model became displaced from the lower part by about 3 mm along the length axis. It may have been ‘finger trouble’ on the touch screen but I need to investigate further and make sure that all the plugs are firmly attached in future.
     
    Anyway, it got here in the end, with a somewhat mis-shapen model The surface finish seemed quite rough, especially on the curved surfaces and the chimney came out a rather odd shape (far from round) but I don’t know if this was due to shortcomings in the model or in the printer. There were also several stray lengths of fine filament, bridging different parts of the model. I was surprised by how ‘solid’ the model was, with a completely filled body of honeycomb structure.
     

     
    Next Steps
     
    I am hoping that experience will enable me to make improvements but I am satisfied that the machine does work, albeit with some teething troubles. There are plenty of selections for varying parameters such as hot-end temperature and speed, so I will try more experiments. I think this machine should be capable of making small parts of complex shapes quite quickly and easily.
     
    The greatest difficulty will lie in creating the 3D models themselves on the computer but I shall take this in easy stages – one step at a time. I am planning to try out the Fusion 360 software, which is currently being offered free for non-commercial use.
     
    EDIT (13th Jan): As I mention in the comments below, this post records my initial 'warts and all' experience with my printer. I have been doing a lot of reading since writing the above and am somewhat amazed that I managed to produce anything at all, in view of my slender understanding of the processes involved. In particular, I am becoming aware of the roles played by the Cura 'slicer' software, which has an enormous influence on the final result. Already, I am finding out how to achieve a far better result
  23. MikeOxon

    One Step on a Broad Road

    By MikeOxon,


    General
    After spending a long time reading and thinking about Brunel's broad gauge railway, I knew I had to make a start somewhere. The thought of plunging straight into a layout was proving too daunting so, I decided to buy a display case of the type sold by Antics models This case has a plain wooden base, measuring about 330mm x 80mm, onto which I could build a short length of broad-gauge (BG) track.
     
    I wanted to build my track base using materials and techniques that mirror the original construction devised by Brunel. In searching for materials, I found that Cornwall Model Boats supply a wide range of materials and fittings, many of which are potentially useful to railway modellers. For the 'baulks' of my planned track, I bought lengths of 5mm x 2mm mahogany strip, while the transoms are made from 1.5mm x 2mm strips.
     
    One feature of the BG trackbed is the use of pinewood packing underneath the running rails themselves. After some thought, I decided to simulate the appearance of this packing, by sticking narrow strips of 2mm squared graph paper on the top faces of the baulks. The rails themselves would then be glued over these strips. The appearance of one of my baulks, built up in this way is shown below:
     

     
    I drew out a scale template, with the baulks in their correct relative positions and with the locations for the transoms marked at 8' (32mm in 4mm scale) intervals. Because the wooden strips were slightly warped, I used a straight-edge to hold them in position, exactly over the template lines, while I glued them down with PVA adhesive. At this stage, the actual running rails were not fitted. The 'bridge section' rails, to scale dimensions, were obtained from the Broad Gauge Society (BGS). I glued down one length of running rail, again using a straight-edge to ensure it ran down the centre-line of the baulk. I then used a roller gauge from the BGS to fit the other rail in position on the opposite baulk, at the correct gauge (28.08mm) for 4mm-scale track.
     

     
    To complete the 'wood-work', I finally added the transoms, gluing each into position over the reference lines marked on my template.
     
    The next task was to add ballast, which I first spread dry into the rectangular openings between the transoms, smoothing down the dry material by hand (finger tips). On the basis of photos from the Bullo Pill accident site (shown in a previous post), I chose a 'medium' ballast in dark brown, to match the ironstone colour typical of the Dean Forest area. After laying the dry ballast, I fixed it down by adding a dilute solution of PVA glue from a dropper. I use about 3-parts water to one of PVA and add one drop of washing-up liquid to the mixture, in order to make it flow freely. This last step is important as, without it, the solution tends to stand in beads on top of the ballast.
     

     
     
    Once the ballast is in place, one really gets a good impression of the 'different' appearance of broad-gauge trackwork,
     

     
     
    Finally, I printed a simple back-scene. I set my printer to 'draft' mode, which produces a low-saturation image that naturally recedes into the background, when photographed. I placed my old 'Gooch single', originally built from a 'K's Milestones' kit, many years ago, and, for the first time in her life, she stood on some track, in a pose reminiscent of 'official' Swindon photos of the period.
     

     
     
    My methods were fiddly to carry out and probably not suitable for a more extensive working layout - and then there is the issue of points! One thought I had, to simplify the process, is that a laser cutter could be used to make wooden frames in the equivalent of 30' sections, onto which the rails could be mounted. Frames for pointwork could be made in a similar manner. Perhaps an idea for someone to take up?
     

     
     
    Mike
  24. MikeOxon

    A Distraction - Tilt Wagon

    By MikeOxon,


    general
    In my previous post, I made the self-fulfilling prophesy that I would be distracted by the forum thread on GWR standard gauge 'tilt' wagons, started by drduncan. Initially there was some discussion as to whether the photo shown was, in fact, of a Broad Gauge wagon but the dimensions (especially the height) seemed sufficiently different to indicate that the vehicle under discussion was indeed Standard Gauge.
     
    Something 'clicked' for me and I decided that I had to add one to my stock, so I began to prepare simple drawings by scaling the photograph, using the assumption that the wheelbase was 9' 9", as in the BG versions. The result, produced in Autosketch by tracing over the photograph, looked like this:
     

     
    Because the original was of metal construction, I wanted to do the same with my model but I also decided to try a new way of marking out my 10 thou (0.25 mm) brass sheet, making use of my Silhouette Portrait cutter. Previously, my method has been to cut out sections of drawings on paper and stick these to brass sheet, using a glue pen. I then simply cut out the parts by following the printed lines with jewellers' snips.
     
    This time, I decided to use a diamond scriber in the pen holder of the Silhouette cutter, to mark out the outlines of the components directly onto brass sheet. As well as the outlines, this method also enabled me to scribe details, such as planking and guidelines for attaching surface details.
     

     
     
    In addition, I realised that if I drew the outline of my rectangular brass sheet on a sheet of paper and also added the Silhouette registration marks, then I could scribe both sides of the brass sheet, in registration. To do this, I lightly taped the brass to the paper, aligned with my outline drawing, and then scribed the detail. I then turned the brass over and scribed the other side with the appropriate designs - remembering to flip the Silhouette image to correspond with the way I turned the sheet over.
     

     
     
     

    Scribing Brass Sheet with a Silhouette Cutter
     
    After cutting out the individual components, I assembled the basic shape of the Tilt Wagon, as shown below. I have a set of socket spanners in a wide range of diameters which provide useful 'jigs' for setting the curvature of the end bonnets. I used super-glue to fix the bonnets inside the folded wagon sides and then inserted the curved ends into the bonnets. I prefer using super-glue to solder where there are lots of different small parts to be fitted together, as it avoids earlier joints melting while new ones are being made. Inevitably, some glue extrudes from the sides of the joints and I use a small stainless steel chisel, intended for wax carving, to remove this excess while it is still at a 'cheesy' consistency.
     
    A prominent feature of the prototype is the extensive use of rivets! I decided to 'cheat' and use the rivet strips that are currently available from 'Mainly Trains'. I realise that this means the rivet heads are on a raised 'plinth' but I find that the near-perfect alignment is preferable, at normal viewing distances, to my attempts at embossing even lines! Additional details are the angle-iron stiffeners along the tops of the sides and on the side doors (1mm brass angle), and the wooden cross-bar at the top of the doors (plasticard). I still have to add the rails above the sides and between the tops of the end bonnets.
     

     
     
    There remains the little matter of a chassis! I find that the GWR W-irons from MJT are still listed as "temporarily out of stock", as they have been all year! In addition, I see that 'Mainly Trains' have a notice on their website that "After 35 Years of trading we are beginning the process of winding down." It looks rather ominous for the future supply of many very useful detailing components!
     
    Perhaps I shall have to turn to completely scratch-building the chassis as well....
     
    In the meantime, my 'work in progress' looks like this, making an interesting comparison with a round-ended 3-planker from David Geen.
     

     
     
    Continue to next part
     
    Mike
  25. MikeOxon
    I meant to mention, in my previous entry, how I made the curved plasticard roofs for my early GWR coaches. I have read about wrapping plasticard sheet around an empty wine bottle, filled with boiling water, in order to 'set' the curve. Somehow, I'm always uneasy about pouring boiling water into glass bottles, so looked for an alternative - beer cans came to mind but these seemed of rather too small a diameter for my coach roofs. After searching around the kitchen (strange, alien place), I found a stainless-steel coffee jug that seemed just about the right size. As shown below, I taped the rectangle of 20 thou (0.5 mm) plasticard, for the roof, to the side of the jug, using broad strips of masking tape:
     

     
    I was pleased to find, after the water in the jug had cooled, that the plasticard had acquired just the right curvature and sat neatly on top of my coach sides. Only time will tell if the new shape is permanent.
     

     
    In building these coaches, I have realised that there was a revolution in the construction of railway carriages during the late 1860s, as their stage-coach origins were finally left behind. The new coaches of the 1870s were on an altogether more massive scale, with much more robust framing and iron solebars. I have taken a couple of photos to illustrate these changes:
     

     
    The train on the left is composed of the Dean type coaches (mainly Ratio kits), typical of the late 19th century, whereas on the right is a mix of earlier designs. I like the undulating roof line created by the juxtaposition of low, almost flat roofed stock, with the more impressive clerestory roof stock, much used by the GWR.
     

     
    I regret that the coaches are not yet finished (pace Mikkel). When I have finished enjoying contemplating the various styles that they represent, I shall get down to all those fiddly details
     
    Mike
×
×
  • Create New...