As I wrote in the previous post, I found plenty of inspiration from my research into Broad Gauge cattle wagons. It gave me the impetus to move straight on with the design of a 3D-printed model.
Whereas I created my 3D printed carriages by extruding the carriage from one end, I felt that the sides of these wagons might be better printed while lying flat on the printer bed. This would mean assembling the wagon from separate parts for the sides and ends, mounted on a rectangular base, rather like assembling a plastic kit. The reasons for taking this approach are that there are several different layers, representing frames, planks and strapping, plus several large apertures, both horizontal and vertical, which would need a lot of additional support if printed vertically.
The first step in building a model is to prepare a suitable drawing. In this case, there wasn’t a lot of written information or specifications but a line of wagons appears in the background of a locomotive photo. This photo shows a little more than half of one side of a cattle wagon, although many details are obscured by the foreground. There is more detail of the central section of another wagon, which shows details of hinges, door fastenings, and strapping. I have annotated two sections of the photos below to show various features:
These sections were the sources for the ‘composite’ image that I created by photo-editing these fragments.
I decided to accept the key dimensions of 16’ 6” length and 7’ height for the body, plus a 9’ 9” wheelbase, as described in the Broad Gauge Society magazine ‘Broadsheet’ No.52. Taking these proportions, I set about ‘matching’ the sketches of the closed wagon, shown in the same article, to the partial views in the photograph. For this, I imported the composite image into ‘Autosketch’ and traced over the main outlines to obtain a proportionate sketch.
‘Autosketch’ screen, showing lines traced over the photograph.
I then scaled the ‘not to scale’ drawings in the ‘Broadsheet’ article, to see how well the locations of the various features were in agreement between the two drawings. Apart from the wider door posts in the photograph, the two drawings match up well, in terms of the main features, so I felt confident to go ahead with preparing a 3D-print from my own drawing.
My sketch overlaid on the ‘Broadsheet’ sketch (re-scaled)
For some reason, I have been having difficulty with transferring my recent DXF-format drawings into ‘Fusion 360’, whereas older drawings (of standard-gauge carriages) went well**. Fortunately, the current drawings were relatively simple and I was able to work through the various ‘open ends’, where lines should have met, and match the ‘nodes’ by hand.
I then worked through the various panels, using the ‘push-pull’ tool to recess them behind the main frame components by suitable amounts. As always, I found it tricky to make the various precise ‘incantations’ required for ‘Fusion 360’ carry out my requirements in exactly the right order and format but, after several false attempts, I arranged everything to my satisfaction. I think part of the problem arises because the software works ‘on-line’, so there are often short delays between initiating an action and it being completed. It can be a little like trying to hold a trans-Atlantic phone conversation when there’s a delay on the line.
I started on another learning curve when creating the sloping slats above the window opening. This involved using the ‘chamfer’ tool, for which an edge has to be selected and then two distances along the faces that meet at the edge are specified. Following my usual approach, I applied trial and error until I got the effect that I wanted.
There are 8 slats in all and the first 7 went smoothly enough but no.8 simply refused to work properly! After a lot of head-scratching, I found it was another problem in the drawing, where two edges had failed to ‘fuse’ together correctly and, as a result, the chamfer kept being applied in the wrong place.
Sadly, after all the attempts I had made to understand the problem, something went wrong with the locations of the bars across the window openings and some of them ended up protruding from the back surface of the model, which was supposed to lie flat on the printer bed. I have no idea how this happened but I had to undo several steps to unravel the mystery and then re-do my slats.
I mention this to illustrate some of the mishaps that can so easily occur when creating 3D models. It is definitely not an activity for anyone lacking patience! When frustration sets in too badly, I find it best to stop and come back later, when the mind has cleared and it is easier to assess the causes of a problem. A good feature of ‘Fusion 360’ is that it maintains a ‘time line’ of actions across the bottom of the screen, which makes re-tracing one’s steps fairly easy.
My 3D drawing displayed on ‘Fusion 360’ screen.
At this stage, I decided to do a test print and was pleasantly surprised by the result. It’s a pity that some of those details that involve so much effort are almost invisible when produced in 4 mm scale but it’s nice to know they are there. Encouraged by this success, I decided to add some more details, such as bolt heads on the door straps and lines between the planks (shown above). I was concerned not to over-do the plank edges and ended up with them being barely visible. Again, a reminder that details are easily ‘lost’ in the final print.
The final result looks good, although I have ‘cheated’ a little with the photo to change the garish blue colour of the filament I use to a more ‘sympathetic’ brown. The detail, however, is exactly as it came straight out of the printer.
My 3D-printed side (re-coloured)
Next steps are to print the ends and then to decide how to assemble a complete model.
* EDIT - After another look at my images of the real wagon, I have realised that the support on each side of the doors appear to be angle irons, rather than the broad wooden door-posts I depicted in my model. I shall make an appropriate modification to my 3D drawings for a re-print.
** EDIT - I've now realised that I had become lax in setting 'snaps' in Autosketch - hence the small gaps.
Edited by MikeOxon