Monday, August 24, 2009

How I made the swing bridge

NEW- A video showing the construction of the bridge

Despite there being no prototypical need for a swing bridge, I decided that the plank bridge across the entrance to the sheds would benefit from being detailed. A quick calculation showed that the ten foot length of the plank more or less coincided with the overall scale length of the swing bridge on the Southwold Railway. (See also p10 of the SR Trust Nov 2007 Newsletter).

Having read about using uPVC fascia trim for the construction of girder bridges on the G Scale Mad discussion forum, I went off to my local EuroCell store and purchased four 5m lengths of 25 x 23 cm angle trim and a bottle of superglue for less than £20.00.

Using a scale drawing of the first Southwold Railway swing bridge, I worked out the dimensions of the model in 16mm / foot scale.

Although the main members of the original bridge were either 15" or 6" wide, I decided to slightly over engineer my model to ensure it was robust enough. For the wider members, I trimmed the angle down to 20 x 20cm and 12.5 x 12.5cm for the narrow members. This meant that the 12.5cm off-cuts could be used as tie-bars.

A carpenters' marking gauge was used to score lines down the length of the angle trim.

A craft-knife was then used to cut through the rest of the trim. I found the knife had a tendency to wander if too much pressure was applied, particularly with the first two or three cuts.

The bottom member was marked up. The first two verticals either side of the middle are 8' (128mm in 16mm scale) from the centre, the remainder are 10' (160mm) apart. The final two members at each end are 2' 6" (40mm) apart.

The verticals were then cut to the following lengths:
  1. (Middle) 9' 9" (156mm in 16mm scale)
  2. 13' (208mm) with a slot cut at 9' 9" (see below)
  3. 9' 3" (148mm)
  4. 8' 6" (136mm)
  5. 7' 6" (120mm)
  6. 6' 3" (100mm)
  7. 4' 9" (76mm)
  8. 3' 3" (52mm)
  9. 2' 9" (44mm)
Using a tri-square, the verticals were superglued to the horizontal:

It is possible to buy an accelerator spray from Eurocell which is designed to set the superglue virtually instantaneously. However, I used clothes pegs to hold the pieces together while the glue set. I'd suggest using disposable latex gloves when using superglue. If you're anything like me you'll find it impossible to keep the glue off your fingers and having the painful job of separating them from whatever they've become attached to (including each other!).

For the central verticals, I glued two pieces of angle back-to-back to make a T-section. This was for symmetry. On the original, the majority of members were T-section or channel, but I decided to compromise for ease of construction.

For the longer uprights either side of the middle beam, slots were cut in one of the sides with a needle file, similar slots were cut in the upper horizontal beam to enable these to slot together:

Working outwards from the middle, I glued the uprights to the upper beam. I soon realised that I needed to anchors the lower horizontal to something substantial to ensure it was not pulled out of alignment. Fortunately, I had an 8 foot fence post to hand and screwed the lower horizontal to this. The two sides were thus completed:

Rather than modelling the ends as in the original, I simply angled the upper beam down to the lower beam and glued them together.

To compensate for the vagaries of my model making, the lengths of each diagonal was measured:

and cut:

and the angled ends trimmed by eye:

with a pair of scissors:

before being superglued in place:

I figured that the diagonals which are 'uphill' when moving away from the centre would be under compression and hence would need to be constructed from angle, whereas the diagonals which are 'downhill' from the centre would be under tension and hence would be from plate (ie the 12.5mm offcuts). This would greatly simplify the construction as the second diagonal could simple be threaded under the first diagonal.

The same approach to measuring:



and fixing:

was adopted for the second set of diagonals.

Once the glue had set, the sides were removed from the fence post and a representation of rivet detail was added using blobs of superglue gel. This proved to be a somewhat haphazard method as the consistency of the gel varies and hence the flow of glue is difficult to control. Consequently, some rivets are more regular than others:

Those with more time and patience would probably want to use a more reliable method.

Painting was completed via two 500ml aerosols of Halfords grey primer:

The pre-existing plank bridge was somewhat basic and needed a little modification to accommodate the new structure.

Firstly, the deck needed widening to the scale 12' 3" (18.4mm) of the original. This was achieved by screwing a few lengths of treated fencing battens to each side.

The bridge sides were then screwed to these. A couple of pieces of angle were glued across the uprights to link the sides together:

A square hole was cut into the base of a large paint-pot of roughly the right diameter (19' - ie 30.4mm in 16mm scale) for the central caisson. This was slotted over the existing centre post and the base buried into the ground.

Four legs, recycled from a plastic shelving unit were cut to fit beneath the ends of the swing bridge. Discs were cut out from a redundant piece of pine shelving to fit inside the legs.

The discs were screwed to the base of the bridge and with three screws the legs were attached to the discs.

Cross members made from spare pieces of uPVC strip were attached to the legs and then the caisson and struts were sprayed matt black.

The bridge was then tested for clearances:

I have made a few compromises which might upset the rivet-counters. The deck of the bridge should be an open lattice with longitudinal timbers beneath the rails. I would imagine that this method of construction would happily support the track and traffic, but given the location of my bridge I've gone for something more substantial. As indicated above, the diagonals ought to be 6" (8mm) wide whereas I have gone for 12.5mm (10") and the main members ought to be T-section. I've brought the upper and lower horizontal members together by completing the arc, rather than truncating them.

At present, the bridge still swings from one end rather than the middle (see also How did I bridge the gaps?) - maybe one day, I will devise a mechanism for more prototypical central swinging. However, I think the model captures the elegance of the original structure sufficiently well and certainly provides a focus for the line.

No comments: