Project 24 - Battledown Flyover #2 (design approach)

With only photographs of the bridge to hand the challenge is to accurately determine dimensions of it. First stop was the satellite image on Google maps and using the scale measure provided I estimated the length of a side to be 150 feet, that's 600 mm in 4 mm scale and since the sides are skewed the model is going to be huge. Next, I needed to work out its height and radius of the curved top.

My photo of the north side was opened in Photoshop and scaled by reference to the brick pillars and a standard brick size and as a check measured the length for comparison with the previously determined 600 mm. It was 601.5 mm - good enough. 

The arc radius was estimated by overlaying a circle and enlarging it until it mirrored the arc. This gave an arc radius of 1600 mm.

I now needed to consider how I was going to fabricate it since it is too big for my 3D printer! I found two natural break points which resulted in splitting it into three equal parts. The two ends are nearly identical in construction so, I only need to design one end and print twice per side. The center section is of a different form.

I use the FreeCad Application 'Part' tool to design models. This utilises three dimensional geometric elements (blocks, spheres, cylinders etc.) to construct a model. I soon found a limitation in trying to create a 1600 mm radius part and had to learn the 'Sketch' feature of the Application to create it in two dimensions like a normal  drawing. Then it had to be extruded using the 'Part Design' tool into a three dimensional object that I could integrate with other parts using my normal approach. Oh dear, have I lost you with this esoteric discourse. Just ignore that and lets move on.

Here is the completed end design. The strut that seems unconnected is a separate piece that is fitted when the end is glued to the centre span. What is missing is the inside face of this side. I decided to make this a separate piece (easily created by mirroring the design of this side), with a view to painting the inside faces for full paint coverage of hidden areas. However, I reasoned it would not matter. I am glad I made a separate piece though, as you will learn.

I noticed in photographs that there is a difference between the two ends in that one end has one upright with double thickness sides made from rivetted plates. To achieve this in the model it was simply a case of creating two designs of the upright swapping them around and exporting two different .stl files. A .stl file is needed by the 3D printer Lychee Slicer Application for preparing the design for print and outputting machine readable code.

And this is what the Lychee did to it. All those orange bits are supports necessary for the printing process. I used its automated facility to both orientate the part and determine where it will place supports. However, I am learning to supplement the result manually, placing supports in places where I think more are needed since relying solely on the automated facility can result in deformations in areas of the model.

The Application tells me that this part will take 6hr 43 mins and 9 seconds to print! Time to print is dictated by the overall height of the part above the print bed. Subsequently, with growing experience, I now set the orientation manually. For essentially flat parts like this I set X & Y to 10 degrees each and this gives a lower print time without degradation. I noticed that the side connected to supports does not sustain the same quality of fine relief detail as the non support side. To overcome this we should always apply supports on the side with least detail - and this is a good reason why I made the inside face a separate piece since it has the same high level of detail as the outside face.

There is a lot of waste in this process. There is probably more plastic in the throwaway supports than in the model! The quantity may have reduced if I had inclined it more but, that would take much longer to print.

Rivet counters can have a field day spotting the difference in rivet numbers between the prototype and model (right). For others, be amazed at the level of detail 3D resin printing offers.

To Part 3.

To Part 1.

Read about the SLA (stereolithography) resin 3D printer I use here.

Project 24 - Battledown Flyover #1 (preamble)

Having successfully completed my first Resin 3D print project I was considering what to create next. The thing about Resin 3D printing is its ability to produce miniscule, crisp details. This drives us to go the extra mile in model design. In particular, as far as railways are concerned - rivets. Yes, I have become a rivet counter! 

Within walking distance of my home is Battledown Flyover - The Gateway to the West. It carries the up line from Southampton whilst the Salisbury up & down lines pass underneath, at a slew.

Battledown Flyover by Sandy B, CC BY-SA 2.0,
via Wikimedia Commons

If I modelled this scene where could I put it? My railway room is overflowing with layouts. 

My Misterton model railway layout in its current form has existed since 2016, having evolved from an earlier incarnation dating from the 1980s with the station building and goods shed created in the 1970s! It is based on 1960s Crewkerne station in Somerset, England. It being a main line station with goods handling facilities. 

I realised that from an operational point of view it mainly gets used to run trains round and round for the enjoyment of grandchildren. Shunting and Goods Yard operations are as rare as hens teeth. This might be the time to replace it with the Battledown Flyover scene. It would be exciting to see a train running over the flyover whilst another passes underneath.

I am a bit hesitant abandoning a station scene but, I do have a small branch line layout that maybe could be placed somehow on the other side of the oval. I am not sure at this stage whether a new layout will come to fruition. But, a flyover model diorama is certain, if only to make use of my 3D printer and satisfy my creative soul.

I researched the web and books, gathering images of the flyover and visited it several times to photograph elements of it. When I saw the beast close up I was taken aback by its complex construction and use of riveted plate girders. So many rivets! Its construction is compounded further by the struts not all being the same design. Some have different form (why?) and width meaning each need to be tweaked or designed independently for the model. I'll answer that question in a future posting.

To Part 2.

Read about the SLA (stereolithography) resin 3D printer I use here.

Ballast Cleaning Train - Part 19 (SLA Resin 3d Print Version)

 This is the final part for this project.

Matisa 3B5 Ballast Cleaning Machine (late 50s early 60s BR era)

Aerial view of the machine and its electric generator wagon

How they were coupled for travelling.
In operation the generator was coupled to the other
end of the machine and an electrical cable connected between them.

To the beginning of this SLA Resin Printed 3D model series.

Read about the SLA (stereolithography) resin 3D printer here.

Read about the FDM 3D printed version and its train here.

Movie below features the FDM 3D printed model and its train.

Ballast Cleaning Train - Part 18 (SLA Resin 3d Print Version)

Photo shows the cabins for both wagons and two other bits for the generator wagon after a spraying with black gloss car paint.

Why black for wagons that are to be finished in 1960s departmental yellow? Black is a common base for figurine painting as it enhances shadows caused by creases in clothing when overpainted to let the black through in places. The same principal is applied here except there is so little relief in these parts that shadows are minimal. Furthermore, I found that it changed the hue of the final yellow coating requiring multiple layers to recover. I will not use black for the remaining parts of this build. However, it was useful in simulating the rubber surround of the window panes that were revealed when the masking tape was removed.

It was really difficult to emulate the yellow colour of the prototype. The Railmatch 2304 I chose was far too yellow. My first approach was to tone it down with a white wash.  This made it too pale so I followed up with a brown wash that brought it as close to prototype as I could achieve.

There are seven components that are resin printed, including the chassis with all its levers, rods, axle boxes, leaf springs etc. printed as one piece. The railings are galvanised wire, the couplings Hornby, the wheelsets proprietary, the lights are ends of a biro ink tube with polystyrene gell glue infilled and the vacumn pipes are wire with thinner wire wound around. These could have been integrated with the chassis print.

Here is a comparison between the FDM printed model (background) and resin printed model (foreground).  For the uninitiated there does not seem to be much difference. But there is!

First, we may notice the colour change. The FDM paint finish is completely wrong having a green tinge. Next is a difference in equipment. I believe the resin model is more representative of the picture in the book and an early period photo I used for reference. The equipment on the other is based on a later version that was photographed in a yard awaiting scrapping.

Other notable differences where the resin printed model wins over FDM:

• No visible print layers
• Finer details e.g. axle boxes and leaf springs fully formed - as good as injection moulded. (For the FDM model I butchered a proprietary wagon and implanted its injection moulded axle box and leaf springs to achieve detail.)
• Sharper corners and edges
• Fewer parts

All the parts for the ballast cleaning machine have been printed and painted. Assembly next, which will be featured in the next blog posting.

To Part 19.

To the beginning.

Read about the SLA (stereolithography) resin 3D printer that I used here.

Read about the FDM 3D printed model here.

Ballast Cleaning Train - Part 17 (SLA Resin 3d Print Version)

Spent some time viewing tutorials on weathering to learn a method for the two chassis.

First stage was grey primer even though one recommendation was not to bother because the plastic base will take acrylic paints and the more layers of paint that are applied the more the definition of fine details will be lost.

However, the next layer was not acrylic. It was gloss black car paint that I had to hand and this needs a primer. The finish gave the models an ex works pristine appearance.

Before applying weathering a spray of matt varnish is necessary to dull the gloss and give good adhesion for the weathering layers. The varnish I purchased was Pebeo Auxiliary Matt Varnish which, due to a mix up, was not the one I intended to buy. I intended to buy a rattle can but this one is for brush application and has a white appearance in the bottle. At first I was not at all sure if it would be suitable. Anyway I mixed 50:50 with water and sprayed a test piece. It quickly dried giving the required matt finish and also suppressed the black intensity to a dark grey which is also what I wanted.

The first weathering layer was a wet on wet application of watery mid grey paint. What this means is water is brushed over the model and then paint applied which runs into crevices for highlighting and gives a mottled effect on flat areas.

The light grey areas in the photo are where masking tape has been removed. Those areas are for gluing the superstructure in place. 

The next layer is various coloured, scraped pastel sticks (powders) brushed onto the chassis sides. This final effect was quite pleasing until I sprayed with an artists fixer compound to seal the powders. It glossed the surface and suppressed the powder colours! I remedied with more powder covering and left it like that. Not quite as good as the pre fixer coating.

The Generator Wagon has footboards painted the same colour as the superstructure. The closest match to that seen in prototype photographs that I found was Railmatch 2304 Early Warning Yellow. However, once applied it looked too bright so a wash of watered down white was brushed over. It is still not right. But, who is to say old photographs accurately portray colour anyway.

To Part 18.

To the beginning.

Read about the SLA (stereolithography) resin 3D printer here.

Read about the FDM model here.