A large part of the appeal of 19th century railways, for me, is the appearance of the panelled coaches. Perhaps there is something lodged deep in 'folk-memory' from the days of finding warmth and comfort around the fire in simple timber-framed buildings. Whatever the reasons, wood frames and panelled in-fill stir emotions that plain steel sheets fail to do. The GWR even used papier-maché for panels, which doesn't seem that far removed from wattle and daub.
Carriage Shop at Swindon Steam Museum
I started with the Ratio GWR 4-wheelers, before developing the urge to have something 'different'. The first step was to build some alternative versions, by using the etched sides from Shire Scenes. Somehow, though, while locomotives are definitely best if made from metal, model coaches take more naturally to softer materials. My discovery of the Silhouette computer-controlled cutter, through JCL's thread on this website, made me feel that it might be possible to realise my aim at reasonable cost.
There's lots of helpful advice in the thread at http://www.rmweb.co.uk/community/index.php?/topic/79025-a-guide-to-using-the-silhouette-cameo-cutter/ , where the principle of building up a coach side from a series of laminations is described. The outer layer is the most complex as this carries all the frame details, which are overlaid onto the base layers.
Once I had decided to start this project, the first step was to select a suitable prototype. There is plenty of information on 20th century designs but details become much sketchier as we move back into the 19th century. Books, such as Russell’s 'Great Western Coaches' Part 1, contain a selection of photos but there are many gaps. A particularly useful source is the website at http://www.penrhos.me.uk/index.shtml , which has extensive information about GWR Short (i.e. 4- and 6-wheel) Coaches. As a result of searching these sources, I made a shortlist of potential subjects - all 6-wheel composite coaches - and chose diagram U29 as a particularly attractive subject. This diagram is quite similar to U27 and U28, which have been the subjects for K's and IKB kits, but has the interesting feature of the transitional GWR 'up and down' window line, where the windows alongside the doors were taller than those in the doors themselves. I made a coloured illustration of this coach in my previous post in this blog.
Having chosen a subject, the next step was to prepare the artwork for the cutter. Since I had a basic drawing of the side elevation, my first thought was to use the 'trace' function in the Silhouette Studio software, to make a vector drawing. I soon found that it would need a much 'crisper' image than I had, to achieve a reasonable trace, and, even then, it would be difficult to replicate the curvature at the panel corners, without introducing unwanted curvature into the straight sections, because of the way in which the trace function operates. For many simpler scenery objects, however, I think this method has a lot of potential.
So, I decided that, for my coach sides, I would have to produce my own artwork, from scratch. It was clear that this would involve a substantial investment of time and effort and I felt reluctant to commit this to the Silhouette Studio software, because it uses a proprietary format, which could not be used for any other purpose in the future.
As it happens, I have an old copy of 'AutoSketch' (ver.5.03), which uses 'industry-standard' drawing formats but, until now, I had only used it for rather simple rectangular components. The need to create more complex shapes, and to arrange them in the correct positions on the coach sides, was clearly going to need a much better understanding of how this CAD software operates.
The complexity arises largely from the provision, within the software, to specify every detail of the forms ('entities') that it can create. For example, in the case of the curved corners of the panels, the radii of the corners can all be specified separately and there are also various tools for aligning the different entities. I soon found that the 'help' screens were very good at explaining what the various drawing tools can do but were rather poor at explaining how the various operations could be achieved. As a result, I went through a rather lengthy process of 'trial and error' - initially, mainly 'error'. Fortunately, I found a basic manual on the web at http://boeingconsult.com/tafe/dwg/Skf-Manual.pdf
This explained the process of 'trimming' entities (using the Edit | Trim menu), which turned out to be the key to constructing complex shapes. It still took me some time to discover how to adjust the trim tools to my specific requirements, which is done through the Edit Toolbar.
Another useful feature of the software is that it can import bitmap images and place these on a background layer, which can then be used as a template for accurate drawing and placement of the entities. Again, it took me some time to understand how to control the behaviour of the layers, since all are normally visible but only one is active at a time. The trick is to copy the background image to an 'editable' layer and then convert this to a 'background' layer, which is then 'immune' to the selection and editing tools.
For the record, I shall now explain the method that I eventually devised, though this is still work in progress, so I may revise my ideas as I gain more experience. I am setting this out for my own use and in the hope that it may prove useful to some of my readers. It turned out that Diagram U29 is somewhat 'the design from hell' , since every panel seems to be unique! It certainly gave me plenty of practice in using the drawing tools.
I like to set up the drawing area in AutoSketch so that it resembles a sheet of graph paper, with major intervals at 10mm spacing over a 1mm grid. I set the 'snap' distance at 0.25mm, which represents 3/4inch in 4mm:1ft scale. The 'Drawing Properties' screen looks like this:
The first stage in drawing my coach was to import an image, to use as a template. I found a drawing of U29 on the web at http://www.penrhos.me.uk/IDGuide.shtml (which Richard Spratt has given permission for me to reproduce here) and pasted this into Photoshop as a TIFF image (not JPEG, since this blurs the line outlines). I then opened the image in AutoSketch, made a new layer, and pasted the image into this layer. I then converted the layer to a 'Background'. This is done by using the 'Graphic Properties' screen, shown below:
I then drew 'over' the background image, using the 'Rectangle' tool to create the panel outlines. The rounded corners are produced by using the 'Trim' tools in the 'Edit' menu. This is a rather complex procedure in AutoSketch. It's necessary to select the 'Round' option for the trim tool, then edit the tool properties for the desired radius. I used 1mm radius on the larger panels and 0.5mm on the smaller. Each rectangle has to be 'Exploded' ('Edit' menu) and then, with the 'Round' option, first click on one side of the corner and then on the other, which produces the rounded effect. It's a bit tedious, when there are lots of panels to do, but goes fairly quickly once into the swing of things! For windows and other constant-sized objects, the shaping only needs to be done once and then the object can be copied to all the locations where it is required.
When all the panels had been drawn, I selected the whole drawing and moved it above the background image on the drawing area, as shown below. In order to transfer the drawing to the Silhouette Studio software, it must be saved in AutoCAD R12/LT2 DXF format. I chose the wrong format the first time, and was dismayed to see a highly distorted image!
I found that it was necessary to re-scale the image again, to the correct dimensions, in 'Studio' and then I made my first cut! For this first trial, I used plain white card and accepted the settings recommended in the software, except for setting Speed to '1' (slowest). I also checked the 'Double Cut' box. All the settings are shown on the screen-shot, below.
I thought that the cutting proceeded quite quickly and the 'chads' lifted off the mat easily, using the point of a No.11 scalpel. The end result is shown below. I had used rather cheap card, which tended to delaminate, hence some ragged edges, but the cutter seemed to have followed the outline very well around the 1mm radius curves, though a little less consistently round the 0.5mm. The photograph, as always, exaggerates the defects but I am pleased with this 'first try'.
Next step is to do the other laminations needed to create a complete coach side