One or two of my readers have expressed dismay over my apparent abandonment of 3D printing. There have been several factors, including the fact that I do not have either the space or a layout to physically house the various models that have caught my interest. I have found that creating models in the computer is very satisfying, in its own right, and I am still learning how to tackle various design problems.

 

Another factor was that the Prusa printer I bought did not meet my expectations and initially produced prints that were inferior in some respects to those from the cheap Chinese printer that I originally chose as a learning tool. Fortunately, @MikeTrice was on the right path when he commented: “I changed from Prusa slicer to Cura which made a big difference”.

 

Recently my son needed a 3D print of an item that was not railway-related but provided me with a reason to have another attempt at optimising my Prusa printer. By using the ‘Cura’ slicer and tweaking many of the parameters closer to the values I had previously used successfully with my Geeetech printer, I managed to produce a good print. The main changes that I made to the default Prusa Mini settings were:

 

• Temperature 190°C (default 210°C)
• Retraction distance 6.5mm (default 3mm)
• Retraction speed 50mm/s (default 40mm/s)
• Linewidth 3mm (default 4mm)

 

So, re-motivated, and under some pressure from my son, I decided to try printing the ‘Twin Sisters’ engine, which I described in my previous post:

 

 

3D Model of the ‘Twin Sisters’

 

While I always keep in mind the possibility of 3D-printing my computer models, there were some dimensions of this model that needed to be modified to create ‘printable’ parts. Since I had kept the components of the model as separate ‘bodies’ in the ‘Fusion’ software, it was not too difficult to make the modifications needed and then to combine groups of bodies into ‘kit’ components, suitable for printing.

 

 

My ‘kit’ of parts ready for 3D printing

 

Some of these parts needed to be duplicated for my 3D-printed model. As shown above, I made the boilers as hollow shells with one end open, so that internal support structures were not needed. My other modifications were to increase the thicknesses of some rods and pipes, to ensure they were ‘printable’ and sufficiently robust.

 

Adding a Tender

 

I realised that, thus far, I have not included a tender, about which I had virtually no information, apart from the illustration by Isaac Shaw (1838), which I showed in my previous post.

 

From his illustration, it appears that the tender was very similar to that originally attached to ‘Rocket’, so I worked up a design based on an illustration from the Science Museum collection:

 

‘Rocket’ tender extracted from a drawing in the Science Museum collection

 

The general structure was simple enough to reproduce but the water barrel presented some interesting 3D design problems. My first step was to draw the profile of the bowed sides of the barrel. Then I used the ‘Rotate’ tool in the ‘Fusion’ software, to create the overall barrel shape.

 

To create the curved staves, which make up the body of the barrel, I used the ‘Sweep’ tool. For this, I drew a rectangle to represent the cross-section of a plank and then ‘swept’ this profile along a path corresponding with the curvature of the barrel. This produced a single stave from which I could create a ring around the periphery of the barrel by means of the ‘Circular Pattern’ tool. These steps are illustrated below:

 

 

Creating 3D model of a Wooden Water barrel

 

The other parts of the tender were made by my usual methods of extruding from the ‘Rocket’ drawing shown above. When I place the 3D tender model together with the locomotive, I feel it gives a better impression of the overall (small) size of this unusual engine.

 

My 3D model of Twin Sisters, with tender, rendered in ‘Fusion’

 

I think this 3D model compares very well with the illustration by Isaac Shaw (1838), shown in my previous post.

 

As usual, I divided the complete 3D model in the computer into components such that each part had a flat surface to form a firm base on the 3D printer bed. To make the barrel ‘printable’, I split it along its length, arranging the ‘split’ to correspond with a gap between a pair of staves in an attempt to minimise its visibility. I then grouped the parts for printing as shown below, laid out in ‘Cura’:

 

3D ‘Kit’ parts arranged by ‘Cura’ for printing

 

 

Printing

 

Once the design and layout work had been completed, all that had to be done was to convert the various files into ‘gcode’ by means of the ‘Cura’ software and then to load these files into the printer.

 

Locomotive

 

For the locomotive, I printed the larger parts individually and grouped the small parts together. I set the ‘Build Plate Adhesion’ parameter to ‘Brim’ for the small parts, to provide a larger area for the initial layer on the build plate.

 

After printing, the parts appeared as below:

 

Locomotive parts as printed on my ‘Prusa Mini’ printer

 

For the record, the print times for these parts were:

 

• Boiler (one): 17 min. X2
• Boiler top and chimneys: 43 min.
• Cylinders (both) + rods: 13 min.
• Steam Drum + Fire holes: 9 min.
• Side Planks, Motion Boards + Pipes: 13 min.
• Chassis: 12 min.
• Wheels (six): 30 min.

Total printing time: 2h 34m

 

Tender

 

For the relatively small and simple Tender, I decided to print all the parts together as a single ‘printer job’. This plan worked out well, as shown in the illustration below of the complete ‘kit’ before I removed it from the printer bed.

 

Tender ‘kit’ as printed on my ‘Prusa Mini’ printer

 

The printing time, as shown on the Prusa screen was 1h 45m for the entire job.

 

Assembly

 

Once I saw the parts on the printer bed, it brought home how small these early engines were, creating a very fiddly assembly job!

 

The main parts of the locomotive were not too difficult to assemble by using the gel type of superglue, which allows for a little leeway in positioning the parts before pressing them firmly together and then waiting for the joint to harden.

 

The tender proved more tricky mainly because, as usual, I hadn’t thought to provide any tabs or other registration marks to help bring everything into the correct alignment. I made silly mistakes too, like forgetting that the tender wheels had outside bearings whereas those on the engine were inside. As a result, the 1mm diameter axles that I had prepared were too short for the tender. I had also been too ‘realistic’ in modelling the stand for the water barrel, which proved impossibly fragile, so I resorted to a blob of ‘Bluetac’ to space the barrel from the floor.

 

Printing the water barrel did not go smoothly. For some reason that I’ve not identified yet, the Prusa printer created blobs of plastic in various places that seemed to be random, when I tried further prints. If anyone has experienced this, I’ll be grateful for any suggestions on how to cure it.

 

The blobs proved rather resistant to being removed, so I shall experiment with different orientations during printing.

 

My current model in its raw state off the printer, looks as below. I have not yet tackled items such as piston and connecting rods, for which I shall probably use brass rods.

 

My 3D printed model of ‘Twin Sisters’ with scale

 

 

My 3D printed model in a similar pose to the Isaac drawing

 

The prototype deserves recognition for its role in proving that coke could be used as fuel for a ‘smokeless’ locomotive engine. Later, ‘Rocket’ showed that the combination of a multi-tube boiler with a blast pipe could produce a similar result in a more conventional format. Nevertheless ‘Twin Sisters’ remained in service until after ‘Rocket’ had been demoted to ballasting duties.

 

Addendum

 

I thought it would be of interest to compare 'Twin Sisters' with its contemporary: 'Rocket.  I photographed my models together, against a backscene of Chat Moss - their natural habitat.

 

 

Mike

 

P.S.  Instead of printing the water barrel in two halves, I separated the barrel ends and orientated the parts as shown below on the printer bed:

 

 

This has produced a much cleaner print although there were two small areas of 'crud' that were easily removed with a finger nail.