Title: Just like the real thing
Pages: 54 - 61
Author: Robert Madge
JUST LIKE THE REALTHING
MODELMAKING IS WIDELY RESPECTED AS A FUNDAMENTAL DESIGN TECHNIQUE - IT'S DANGEROUS TO IGNORE IT. REPORT BY ROBERT MADGE
Christopher Wren relied on models to get his designs through committees. Cathedrals and other large buildings were modelled in the Middle Ages, usually by joiner-build wood models, to give continuity to long building projects and demonstrate the intended results. The first steam engines were designed with the aid of models. The Concorde project has used about 1300 models.
Today, modelmaking can have the status of a sort of three-dimensional 'drawing board' activity, demonstrating and confirming the designer's evolving ideas more accurately than could ever be achieved in two dimensions. Used in this manner, modelmaking has become indispensable to a growing number of industrial designers. Among engineering designers, modelmaking is frequently regarded as the only possible way to test performance in advance of expensive real-life production - and techniques for miniaturised performance testing can be startlingly accurate. But architects are different: like industrial designers and engineering designers, they use models for persuasive presentation of finished design work and, under mounting pressure from planning authorities and the public, their sophistication in this area (and sometimes the money spent on it) is considerable. But instances of architects working with models at the drawing-board stage are rare indeed. Most architects rely on the assumption that all the 3D implications can be seen from 2D plans alone. Their confidence is not always justified.
Among industrial design shops, design by-modelmaking is often obligatory. Kenneth Grange of Pentagram Design Partnership will seldom employ a design assistant who isn't capable of producing a reasonable model; and he usually has three full-time modelmakers also. He will commonly start with card or foam models to plan the disposition of components and get a feel for the aesthetics of the object. This will be followed by more detailed modelmaking, using wood, plastics or glass-fibre, repeatedly modified as the design develops.
Stephen Bartlett, of Butler, Isherwood & Bartlett, cannot imagine designing without the use of models. He likes to work in 3D as much as possible, and needs to sense the physical presence of an object in order to evolve the design. BIB normally uses card and perspex for its own design models, which at first may be very crude and for the designer's eyes only. But they stress the importance of colour in their design, introducing it right from the start.
David Carter, who has his own practice in Warwick, is emphatic about the need to model an object as early as possible in the design process. He has an intuitive reaction to 3D objects and his first model, preferably full-size, is likely to be roughly made of cardboard 'just to get an idea of how big it is'. The design progresses through a series of models, to test different aspects, to refine the appearance, or to suit technical requirements. The distinction between models, mock-ups and prototypes is blurred.
At Medway College of Design, which runs the only exclusive modelmaking course in Britain, John Gaylard teaches his students that models for industrial purposes are split into two types: those for explanation and those for demonstration. The explanatory models serve to determine a principle or sequence. For demonstration there is likely to be a series of 'sight models', starting as block models and progressing to quite detailed objects. A final model may be required for publicity purposes.
Designers have varying ideas about the degree to u hich their clients should be involved in the design as it progresses. BIB, for example, prefers to show only the finished model to a client for fear that the design will be misunderstood or that an unfinished aspect will receive undue attention - and the client will never be shown a model without the final colours. Ken Grange will tend to show either a highly-formalised, and therefore obviously unfinished, model or else the final presentation model. He does, however, find models an essential part of the conversation with clients, and he sometimes consciously modifies models so as not to give the wrong idea: for example, he once produced a model of a tv set for initial approval by a client, and he made it to a smaller scale so that it wasn't directly comparable with existing sets.
Modelmaking within the designer's offices has its own advantages: it is possible to alter the model rapidly, and the modelmaker can easily contribute to the interpretation of the design. Matthew Stewart of Schema Ltd actually designs as he sculpts his own models. He doesn't
(Caption page 54) Six stages in David Carter's design of a flower gatherer for Stanley Tools Inc. He starts by making a model which looks precisely right and progresses towards a working model
(Caption page 55) Models at different scales to develop and demonstrate the design for a Post Office container: 1. 1:12 scale model to evaluate concept of collapsibility; 2. 1:12-scale model to examine methods of opening door; 3. 1:6-scale model to demonstrate a particular production method and material; 4. full-size wooden mock-up to test functioning of the doors and to prove correct choice of heights; 5. full-size mock-up model to demonstrate final production appearance. A David Carter project.
believe that drawings alone can show the toolmaker exactly what he wants: typically, a thermo-plastics product may end up with too many straight lines and radii. Ogle Design, of Letchworth, combines design work with operating a model shop which also provides a professional modelmaking service for outsiders.
But nearly all designers turn to outside professionals if the construction of the model is particularly difficult or if a particularly slick presentation model is required. Laurie Bar, of Aeronautical & General Models, says that most designers don't progress beyond a rough cardboard model and it may turn out that their ideas are slightly rough also: they may not consider in sufficient detail the technical aspects of manufacture. He sees the professional modelmaker as a link-man between designers and manufacturers; he tells the designer if his ideas are practical and demonstrates how an object will be produced. His work is accurate enough to be used for pattern-making purposes. The skills which go into actually building a model lead AGM into projects such as the model spaceships for the film 2001, or the Maxwell House jar which expels instant coffee in the television ad.
It is quite common for industrial models to be used as a direct basis for tool-making, especially with plastics products. The accuracy with which models are constructed can obviate the need to employ a pattern-maker. Pantographic reduction from a larger-than-life model is even more accurate. Plastics do however represent a certain problem if an attempt is made to test the structural qualities by modelling. The behaviour of an object cut from solid plastics will differ from the moulded version.
Whenever models are tested for performance, compensation has to be made for scale. The more detailed the performance, the more complex and ingenious must the compensation be. Such techniques are highly developed in the aeronautical and ship-building industries. When hull design and vessel stability are tested in water tanks, the simulated waves move much faster than in the real-life situation, and the film of the model performing has to be viewed slowed down. Turbulence stimulators have to be included on the bow of the model; these are not scaled-down representations of the actual hull surface, but they accurately simulate the turbulence characteristics.
Results from wind-tunnel testing need to be adjusted to compensate for the scale of the model and the speed of the wind. Alternatively, a compressed-air tunnel can be used, to scale up the density of the air in the same ratio that the model is scaled down. Air viscosity is another variable which could be used to compensate for the scale reduction.
Architectural models are in another world altogether. Nearly all work done in this field is concerned with presentation only: for the purpose of planning appli
(Captions page 56) Top, a medium duty tractor-mounted welding head which used to be made by BOC.
Centre, a 'stage one' presentation model made by the designers BIB to demonstrate to the BOC management
a concept which would standardise what had previously been non-standardised components.
Bottom the final re-design now on the market.
This cooker is only 355mm high. A dozen models like this were made by Aeronautical & General Modelmakers Ltd, for use as sales aids.
(Caption page 57) A landscape model of part of the city of Sydney, used by the National Physical Laboratory for a wind-tunnel investigation into the effects of wind on the proposed Qantas centre.
A model of the British pavilion at Expo 70 also used for wind testing by the National Physical Laboratory at Teddington. The tests showed that the central piers could be unstable during erection, so holes were made in the piers to decrease wind resistance.
(Captions page 58) Top, an engineering model on a grand scale. This model built at the Hydraulics Research Station, Wallingford, represents the whole area of the Wash and its four main rivers. It tests water storage capacity.
Above, sea-keeping tests on a container ship model, by the National Physical Laboratory's ship division at Feltham. Models of ship hulls are usually about 6m long, commonly moulded in wax or cut from blocks of polyurethane
cations, 1:500-scale block models, which give an indication of height but are misleading about most else, are often all that is used. However, with the massive cost of hold-ups in mind, architects are discovering how convincing a good presentation model can be.
John Piper, whose firm was responsible for the Westminster parliamentary extension and Kensington barracks models, dislikes detailed models with surroundings represented in block-form only. He would prefer to model at a smaller scale to make the relationship between buildings clearer: overall detail, he says, undermines frivolous objection. On the other hand, when modelling a typical semi-detached house for Wimpey, John Piper did a very expensive 1:50 scale model, incorporating extras such as modelled tomatoes and lupine; the cost, he says, was justified by the number of times the house type was to be repeated.
Well built architectural models can be extremely beautiful objects, but they can be equally deceptive. They cannot be realistic, in the precise sense of the term, which leaves considerable discretion to the modelmaker. John Alcock, of the Hugh Watson Partnership, believes that colour has to be 'scaled' as well as size: it must be adjusted to be photographed just like the real thing.
Of course no self-respecting modelmaker would allow a model to leave the studio unless it looked good and would please the client, usually the architect. The Design Presentation Unit is frank enough to admit to modifying its models with this in mind by brightening up the colours, for example, if that makes the building look better; they leave ethics to the architect.
Many committees ask for models because they do not understand plans so well, but it also takes a degree of sophistication to appreciate models properly. It is almost safer to rely on photographs of models to judge their full-scale impact, because the scheme is then presented in a form in which real-life buildings are often viewed. The photographs must be taken from a low level and include some of the surroundings and street paraphernalia of an actual street scene. Modelscopes can be useful in this context; by conveying light up a thin tube, they can give the impression of seeing a model at its own scale. The quality of photographs taken this way is never very high, but it is more than made up for by the realism of the view. Systems are also being developed using video tape recorders.
If the presentation of a project is by photographs of a model, and not the 3D object itself, various short-cuts can be taken. A repetitive element of a building need only be modelled once and, by the use of mirrors or by montaging, the effect of the whole building can be created. Photographs showing only one side of a building require only window-deep 3D moulding. Few pictures ever show more than two sides at once.
(Captions page 59) Very sophisticated acoustic modelling, carried out by the BBC. This model of the main music studio in Maida Vale was made at a scale of l: 8, so frequencies had to be scaled up eight times, and absorption characteristics of surfaces had to be similarly scaled. The orchestra was modelled to be acoustically, rather than visually, accurate (expanded polystyrene, partly clad with felt). Since the noise absorption of air increases disproportionately with a rise in frequency, it had to be dried in compensation to three per cent relative humidity. Music recorded in a room with no reverberation was played eight times as fast within the model, re-recorded inside the model and re-played eight times as slow.
Below, a model of the Bristol United Press Building, to assist design development and layout of workspace and machinery, by Hugh Watson Partnership.
Below, an example of an intricate model, made for display in a museum: built to 1:12-scale by Peter Severn Lamb of Stratford-upon-Avon, this is an electrically-driven working model of an Allis-Chalmers Westinghouse Double Compound Generator, the type used for 50 years to power the New York subway. It was commissioned by the Smithsonian Institute in Washington
Modelmakers regularly discover faults in architects' designs: their work puts a building into 3D for the first time. A very small number of architects have recognised the implications of this, and so introduce models as an aid to basic design: when David Rock's team at Building Design Partnership produced their prize-winning Vienna scheme, one group of three architect/modellers did continual 3D representations, mainly in balsa wood, of each plan made by the other group of architects. But such examples are rare.
Although architects themselves are seldom modelmaking designers, their work is often assisted by modelled performance testing. Under artificial skies, internal daylight in project buildings can be evaluated, although the results are pretty crude. Shadow angles can be predicted by simulated sunlight tests - a method which determined the shape of the new London Stock Exchange, and which could usefully replace the arbitrary daylight standards of planning regulations. Models of building projects are now often tested in wind tunnels in an attempt to avoid some of the more horrendous mistakes of modern developments. None of these experiments require sophisticated models, and would usually include less detail than is needed to make aesthetic judgements. To test structural characteristics, a specially made model is needed: cut plastics sections can be used to show stress patterns.
The use of models to estimate the acoustic characteristics of concert halls and music studios has become increasingly sophisticated. Most concert halls are now modelled at 1:20-1:100 scale to analyse the reflection characteristics by 'impulse response testing' (this sort of experiment measures the 'objective' characteristics of sound in the hall, but doesn't simulate the appreciation of music played in the room).
Checking noise problems is also attempted but, in the view of Dr Delaney of the National Physical Laboratory, it is not often worthwhile to test models of specific building projects, such as housing flanking roads. He has been evolving design criteria for road-building based on noise tests conducted with a model built at 1:30 scale. The road surfaces are modelled by aluminium or sheet plastics, and the surrounding ground areas have to be carefully simulated (the reflection coefficient of the facades of houses has been found to be not particularly important). One noise source moving along the road is measured, and a computer is used to predict the effects with traffic flows. Dr Delaney has also had to overcome the problem of air absorption. His research has shown that roads in cuttings are best (as deep as possible, so long as the sides are not quite vertical), that flat-top buildings tend to work slightly better as noise shields, and even that garages set back in the gaps between suburban houses can make noise conditions worse at the rear of the buildings.
There is not much room for confusion when a formalised model is used for performance testing: its particular usefulness (and its limitations) can be understood by anyone. But the intentions of a model finished to presentation standards can be uncertain: it will mean different things to different audiences. This is especially true of architectural models, but any model may fall foul of the trap. People cannot rely on their eyes alone to judge a scale model. No scaled model simulates all the characteristics of a subject, nor suggests all the effects of the real object. Few would object to an exquisite model of Concorde in their living room; the full-size plane -passing 1000m overhead is something else.
(Caption page 60) Left a presentation model by John Piper. Below left, a quickly-made balsa model used by Edward Cullinan and Partners in an architectural submission.
(Captions page 61) Above, a highly-functional model of a petro-chemical works: nearly all projects such as oil refineries use this sort of model, built out of standard components to show the arrangements of the pipework and plant. Each element is labelled with a co-ordinate and reference to enable the designing engineer to work directly from the model. The contractors building the works also refer to the model, before construction is finished, it can be used as a tool for training the staff.
Left, John Piper model of the British Steel Corporation's works at Llanwern. For a modernisation programme, the model tested whether new equipment would fit within the old system. It also helped to develop phasing which was least disruptive, and was used as a basis for tenders and for re-training.
Below, an extremely detailed model of Burton-on-Trent, commissioned from John Piper to alleviate residents' fears about a new road scheme.