Notes on Construction - Thalia
Generally ship models are either supplied by a manufacturer in kit form and everything required for construction including step by step instructions is supplied in a box or as a ‘scratch’ model where the builder has to work from a copy of the original drawings. In this case the drawing of the ship’s lines was at a scale of 1:25 and consequently this was the scale chosen for the model and fixed the overall length at approximately 1.60 metres.
The model of the ‘Dampfschiff Thalia’ is a scratch model and consequently the “How to build it?” and “What materials to build it from?” questions had to be resolved at the very beginning of the project. Material choice has also to take account of weight and stability apart from cost and availability.
Basic Decisions
The two fundamental decisions were to construct the frames and keel from 4mm plywood and the planking from 10mm wide by 2mm thick wooden strips. The planks are only available 1 metre long and that introduced a requirement for joints somewhere along the length of the hull. To avoid developing a plane of potential weakness in the model, joints in adjacent planks need to be offset as far as this is practical. Also, for reasons of strength, the frames at the joint positions were doubled in thickness to 8mm in order to provide an increased gluing area for the planks at the joint. Another point that was considered was that the joints needed to be positioned at a location where the plank was under least stress due to twisting or bending. Therefore the thickened frames (H and L) are at the ends of the central section of the model where the cross section is essentially constant. An unavoidable result of these decisions was the expectation that some wastage of material was inevitable.
Plank 1 lies between Frames A and L. 83 cm long with 17cm off cut.
Plank 2 lies between Frames A and H. 53 cm long with 47 cm off cut.
(As will be seen later in these notes the concern about wastage was largely mitigated).
Frame Design
The drawing included cross sections for 14 out of the twenty frames. One side of the drawing gave the half frames forward of the midpoint and the other side, the half frames from the mid point to the stern. This drawing was scanned and recorded as two separate half images split on the vertical centreline. These images were then flipped over and also recorded. Each of these was printed 14 times and the appropriate pairs glued together so that there was a complete cross section for each of the fourteen frames shown. Half frame drawings then had to be drafted to scale for positions E, G, I, K, M and O from the information on the lines drawing so that the skeleton of the model had frames at 7 _ cm spacing throughout. To ensure accuracy the beam of the boat was checked before the half frame drawings were glued together
Planking Trial
It was considered that a planking trial would be beneficial and a small number of planks were ordered from a model shop in England as no local stockist could be found. A scanned copy of the boat elevation (Frames A to D) was glued to a board and half frames (A to D) cut and glued to the board in the correct locations. It is customary to insert the edge of the first plank into a groove (rabbet) cut in the side of the keel. A rabbet was cut although it proved very difficult to shape it accurately as the angle the plank made with the keel varied along its length.
Experience derived from this trial led to four decisions:-
Planking would be curtailed at Frame ‘A’ and the bow forward of Frame ‘A’ would be carved from solid wood as it proved to be very difficult to shape the planks and achieve a strong joint where the plank met the stem.
A rabbet would not be cut in the keel because of the difficulty of cutting it accurately and the risk of weakening the keel by reducing its thickness by at least half. It was decided to glue a plank to either side of the keel as this provided a 2mm deep rabbet either side without any weakening. The resulting overall keel thickness is probably out of scale but this sacrifice was considered worthwhile.
Preparing the Planks
To suit the shape of the bottom of the hull – vertical at the bow and horizontal at about one third length - planks have to be soaked, twisted through about ninety degrees and allowed to dry. With the small number of planks required for the test piece the ends of the planks were held in G clamps that were rotated relative to one another and propped up on the bench. Because there are a much larger number of planks in the model a simple jig was made to hold the soaked planks at the required angle of twist. Normally four planks were cut approximately to length before soaking in cold water for about six to eight hours.
The edge length of each frame ‘A’ to ‘G’ and ‘L’ to ‘T’ varies and is greatest near the middle of the model and at its least at the bow and (to a lesser extent) at the stern. Subject to certain practical limitations it is customary to have the same number of planks at the bow and stern as there are in the middle of the model. This requires the plank to be tapered. A short length of untwisted plank can be held in a vice and with care it can be planed with normal woodworking tools. When the plank is circa 80 cm long it cannot be held in a vice and when it has also been pre-twisted it requires some special arrangement to enable the taper to be planed whilst keeping the plank edge at 90 degrees to the plank face along the length of the taper. Because of the extreme flexibility of the plank, pushing a plane along the plank edge was not favoured as the plank could buckle and snap.
I decided that it was better to hold the plank at one end and pull the wood across the plane blade because when the plank is in tension it cannot buckle and snap. The photo shows a standard shoulder plane mounted on a board and also the ‘gate’ screwed to the board to hold the plank flat as it passes the blade. This ensured that a twisted plank was always passing the cutting blade at right angles. This arrangement satisfied the basic requirement for tapering and if any minor adjustments were required to the width of the plank to fit it more accurately to the previously placed plank this was achieved by using a small ‘bull nose’ plane.
Clamping
During the planking trial the difficulties of clamping the planks at the frame positions due to the changing width of the structure and the variation in angle from frame to frame was recognised. Clearly this difficulty would be exacerbated when dealing with a plank circa 80 cm long passing across 12 frames. A whole series of options were considered and rejected as follows:-
G cramps. Conventional metal cramps would have the jaw opening necessary to span the model at any position but these do not work on surfaces that are inclined to each other. The weight was also a draw back as it was considered that these would impose an unacceptable load on the structure.
Modern screw and friction-grip cramps. There are many varieties of these cramps made of metal, plastic, carbon fibre etc. available on the market. The basic problem of only cramping on parallel surfaces was not resolved and in addition the unit cost was too high to be attractive.
Spring clamps. There are many varieties available but these were rejected because of the limited jaw opening and the possibility that the pressure exerted by the clamp spring could damage the plank.
Recognising that a properly shaped and twisted plank would require only sufficient pressure to hold it in place whilst the glue hardened led, to the consideration of a device clamped to the frame so that a wedge could be used to supply the clamping force between the plank and the frame edge.
The initial design used a bolt and butterfly nut to clamp the device to the frame. A wedge could then be used between the plank and the ‘L’ shaped end of the device. It was effective only where the frame was big enough to have a section cut from the centre to lighten the structure and when there was sufficient finger access to enable the wing nut to be tightened.
From this design a right angled bracket evolved. It was made from scraps of 4mm and 8mm ply from the frames and off-cuts of planking from the planking trial. To improve grip strips of sandpaper were glued to the inner surfaces of the planking. The bracket was held in place with a small spring clamp.
This last design was effective and used throughout construction except when gluing the final two planks as then the gap was too narrow and an external clamp (see later) had to be devised.
Home made ‘G’ cramps were also used. Four of these were assembled from scrap plywood and stock nuts and bolts. The objective was to produce a clamp that was strong but very light in weight and these were often used to hold the free end of a plank; at Frame ‘L’ for example where the clamping force is nominally perpendicular to the width of the plank.
Building Board
The literature on model building advocates the use of a keel clamp which allows the model to be positioned at the optimum angle for ease of work. This model is too big for a keel clamp but it is essential to have a stable support for the model that allows it to be turned end for end or even removed safely from the bench to allow other work to be done. A plank of planed softwood 200 * 20 * 2 centimetres thick was purchased as a building board. A centreline was established using a string line stretched between measured centre marks at each end of the plank to avoid any curvature in the plank being transferred to the model.
To support the keel 5 pieces of wood strip (4 pieces 2*1 cm and 1 piece 4*1 cm) were carefully positioned and screwed to the building board after a 4mm wide slot had been cut to half depth in each strip. Brackets were made to support the bow and stern vertically on the centreline. The illustration shows the status before the bow support was fixed.
On the ship builder’s drawing the frames were identified as 0, _ , 1 _ and so on but during construction this was amended to A, B, C etc and A1, B1, C1 etc for ease of reference in the ‘daily’ construction record.
Constructing the skeleton of the model.
The paper patterns for the three keel sections and 20 frames were carefully glued to sheets of 4mm ply and cut out using a powered fret saw.