THEORY & PRACTICE OF MODULAR COORDINATION

Alan. J. Brookes1

ABSTRACT

In this paper Professor Brookes will review the programme of introduction of dimensional coordina-tion and modular coordination in UK starting from the Hertfordshire Schools of 1955. He will de-scribe the various attempts by government and other bodies to force the mandatory use of dimensionalcoordination and his own involvement with the PSA Method of Building programme in 1970’s and thebuilding of a full size test mock-up using five different types of structural frames at Building ResearchEstablishment, Garston.

He will describe his own PhD thesis on tolerances and jointing and mainly the general conclusionthat 5% of building costs arise from remedial work caused by inaccuracy of construction lying outsidethe tolerances allowed for in design.

He concludes that in recent years the situation has not necessarily improved. As a claddingconsultant he is often called to advise on jointing failure arising from gaskets and sealants beingoutside their permissible joint sizes due to inaccuracy in the structural framework.

In his own work at Singapore Arts Centre and Federation Square Melbourne, he was well aware thatallowance for 3-dimensional tolerances must be provided at all critical interfaces and that all membersof the building team including subcontractors must be aware of and agree the tolerances inconstruction related to critical dimensions and grid lines before starting manufacture and assembly.

KEY WORDS

Modular, Claddings, Tolerance

INTRODUCTION OF MODULARCOORDINATION IN THE UK

Following the second world war various types ofprefabricated houses were developed in UK and inthe schools building programme HertfordshireCounty Council (1955) led the way in establishinga flexible means of construction based on modularassembly (Russell 1981).

A key part of the Hertfordshire philosophy wasthe use of a planning grid enabling the plan tochange size and shape at any point of the grid. Inan article signed “the County Architects Dept” inthe Architect and Building News in 1948 (Hert-fordshire Architectural Department 1948) theadvantages to the grid are simply set out aftermaking an analogy between the movement of achess piece and the movement of components onthe planning grid.

At the same time in USA Charles and RayEames were designing their own house in Santa

Monica using standard windows and doorsstraight from the manufacturers cataloguemounted on to a cross braced steel frame(MacCallum 1959). This house became the icon forsystem building and went on to influence the UKdesigners of the consortia building programmessuch as SCOLA and CLASP.

Product Development

327

Figure 1: Eames House. Santa Monica. USA.

1 Prof. Dr., Technical University Delft, P.O. Box 5043 2600 GA Delft, The Netherlands, T +31 (0) 15 27 84297,F +31 (0) 15 27 84178, a.j.brookes@bk.tudelft.nl

The Eames house and others engendered enor-mous enthusiasm amongst architects that anindustrialised vernacular was not only possible,but at hand. In UK the Modular Society wasfounded in 1953 by Mark Hartland Thomas whichsought to bring about the idea of a dimensionallyrelated building industry vernacular through aseries of ground rules the most crucial one beingthe use of the basic module of 4 inches or 100 mmfor all component sizing. This was not particularlya new idea. Albert Farwell Bemis had proposed asimilar three dimension dimensional modulebased upon the 4-inch dimension in USA in 1936(Bemis 1936).

Many of the protagonists of the modular idea,including the Prime Minister Harold McMillansaw it as a way of drawing Britain and the rest ofEurope closer together. Driven by the zeal ofThomas and others the Modular Society becameexceedingly influential in making the case formodular coordination through articles and publi-cations including the Modular Quarterly and theModular Primer. The latter written by Corker andDiprose and published in the Architects Journalfor 1 Aug 1962 became the handbook for a wholegeneration of architects intent on bringing some

rationalisation into the British building indus-try(Corker and Diprose 1963).

1961 also saw the publication in English ofKonrad Wachsmann’s. The Turning Point ofBuilding which contains one of the clearest state-ments of the case for mass production and indus-trialised buildings using the ideas of modules andstandardisation (Wachsmann 1961).

The process of metrication began in Britain in1965 in the building industry. The programmeallowed ten years to 1975 for the changeover andmade necessary the rewriting of Britain standardsand related documents. This gave an opportunityfor modular ideas to be incorporated and tobecome mandatory. Several Government Depart-ments were involved in introducing in their ownlarge building programmes the principles ofDimensional Coordination (Martin 1965).

A determined effort to establish the viability ofthe performance standard approach for compo-nent procurement linked to an open system ofcomponents using dimensional co-ordination wasmade by a group set up in 1968 at the Ministry ofPublic Buildings and Works London led by ColinPain and known as Method of Building Group.

Pain went to visit Ralph Iredale in USA in 1968.Iredale had previously worked for CLASP andNENK systems and then emigrated to Californiato work with Ezra Ehrenkrantz on the CalifornianSchools Programme SCSD.

In turn Ehrenkrantz had studied withWachsmann and thus the link towards componentbuilding using ideas of modular construction,space frames, lightweight steel framing etc wereformed. Later Norman Foster and Richard Rogerswent on to study at Yale in the Master Programmeof the same School of ideas.

I joined Method of Building Group in 1970 toprepare documentation on design discipline anddesign requirements that by 1971 had becomemandatory for the MOPBW building programme.A similar approach was also adopted by the

Proceedings IGLC-13, July 2005, Sydney, Australia

328 Theory & practice of modular coordination

Figure.3: Adaptability of components(Konrad Wachsmann)

Figure 2: Movement of a chess piece demonstratingadaptability

Department of Education & Science and the Min-istry of Health.

During this time although the advantages ofprocurement using performance specificationswere becoming clear there was a nagging doubtabout the limitations of conventions for fixing andjointing within the dimensional framework partic-ularly as the strict rules of dimensional co-ordina-tion concerned the question of components“keeping station within their modular (basic)space”. A full size mock-up was tested at theBuilding Research Station in Garston UK to checkprinciples of jointing using five different types ofstructural frameworks.

In 1967, J.F. Eden put the problem moredirectly to the dimensional co-ordinators when hecited the development of the engineering industrywhose success in mass production architects weretrying to emulate (Eden 1967). He pointed out thatthe accuracy and interchangeability of parts inengineering relied upon the definition early in itsdevelopment of sets of standard for joints.

Thus the way in which the parts fitted togetheris very much more crucial that the parts to bethemselves dimensional co-ordinated. Later hisadvice was taken by Bruce Martin in his book‘Joints in Building’ (Martin 1977) but at theheyday of dimensions co-ordination in 1968 suchadvice was hard to swallow.

The main obstacle was that many weatherproofjoints need overlap or rebates thus the rule that“every component should not exceed its ownbasic space” could not always apply.

Further the effect of tolerance if allowed withinthe basic space can lead to uncomfortably widejoints at the edges of the component.

Early studies of modular theory had made someattempts to point out the implications of manufac-turing erection and movement tolerances in com-ponent sizing but had not put any serious values ofthis allowance. Studies by Pietro N. Maggi in Italyhad identified problems of tolerances in buildingas early as 1961 (Maggi 1961). This led to a pro-posal to use statistics as a method of combininglikely errors in an article in Modular Quarterly bySefton Jenkins and Mak in 1963. However thisproposal came too late for the statistical conceptto be included in the British Standard (BS3626:1963). “Recommendations for a System ofTolerances and Fits for Building”.

Unfortunately early thinking was pre occupiedby a misunderstanding of terms. This is illustratedby a paper on definitions included within ModularQuarterly Summer 1958 in which “deviations” areconfirmed as ‘the allowance to take account oftolerances’ rather than the other way around. Sec-ondly there occurred not only in the Modular soci-ety but by others in the field an over simplificationthat each component must keep station within theco-ordinating lines. Even though the workingcommittee of the Modular Society showed in itsreport Winter 1959 the column face deviatingover the modular plate this diagram was declaredincorrect at the main discussion meeting in March1960. The extent of the joint size variation result-ing from this over simplification was either notappreciated or ignored.

BUILDING RESEARCHESTABLISHMENT

Following widespread dissatisfaction with BS3626 work was started at the Building Research

Product Development

Alan. J. Brookes 329

Figure 4: Dimensional control lines on the Method ofBuilding Rig showing components within their own

‘basic’ space.

Figure.5: Tolerances in Construction meant thatcomponents sometimes lay outside their own ‘basic’

space.

Establishment mainly by Harry Harrison and RonBonshor to determine more sensible methods ofwork sizing and determination of tolerances. Thisled finally to the production of a new British Stan-dard DD 22 in 1972. “Recommendation for theco-ordination of dimensions in building toler-ances and fits for building”.

Even so there were still limitations to the meth-ods as proposed.

The first was that permissible deviations forvarious methods of construction were not avail-able for insertion in the formulae suggested byDD22. The British Standards had published in1969 the provisional document on accuracy inbuilding (PD 6440) (14), which has a sectiongiving permissible deviations for erected buildingelements. In order to check the validity of thesevalues an accuracy surveys programme was initi-ated by BSI in December 1972. The programmewas devised by BSI with the assistance of BREand called for the checking of about 50 000 indi-vidual items of some 320 building projects ofvarious types.

Building contractors were asked to execute theactual measuring and to feed back data forms con-taining the required measurements of the erecteditems. Sixteen different data forms were devisedto cover the different types of items (distancebetween walls, levels of floor, straightness ofwalls etc.). This survey eventually provided arealistic Code of Practice of standards of accuracythat can reasonable be achieved for differentforms of structure. The designer was then able toinclude the deviations from this new Code into theDD 22 formulae.

The second main difficulty with the applicationof DD 22 was that the methods tends to produce asingle size peculiar to each application of a com-ponent. For this reason BRE (Sefton Jenkins andMak 1963) had produced graphical aids for Man-ufacturers and Designers whereby a range ofsizes, rather than a single size, are given depend-ing upon the conditions of use of the component(Bonshor 1972). These graphical aids are basedon an idea by John Ritter to use computers forfinding the optimum sizes of components usingthe DD 22 formula. The important merit of theseaids is that they enable choices—and their conse-quences—to be seen rather to dictate particularaction.

As such, they offered guidance on suitabilityfor particular uses and suggested the steps to betaken to achieve satisfactory use.

PHD THESIS AT LIVERPOOLUNIVERSITY

Thus by the time I started my PhD thesis at Liver-pool University in 1973 the stage was set. Mymain tasks were to study:1) To what extent is remedial work necessary as a

result of variability in construction lying out-side tolerances in design?

2) Where do the major difficulties lie?—An in-vestigation of issues arising from the case stud-ies

3) What are the costs of these difficulties?4) Why doesn’t the Building Industry avoid these

problems through the use of design aids andStandards?—A review of existing Standardsand designers attitudes.

The research (Brookes 1976) encompassed thestudy of thirteen building projects then under con-struction using precast concrete panels with insituconcrete frames. Measurements were taken of thevariability achieved on site and the factory and theresultant range of joint sizes between compo-nents. In all cases the variability (accuracy ofassembly) exceeded the tolerances specified.Joint sizes were always larger or smaller thandesigned.

The case studies pointed to three main conclu-sions:1) As the numbers of interrelated critical dimen-

sions increases the chance of misfit increasesand thus the critical points in assembly shouldbe kept to a minimum.

2) Three dimensional tolerance requirementshave the effect of increasing the number of crit-ical dimensions.

3) Misunderstandings between designers, manu-facturers and constructors occur when the per-missible deviations on critical dimensions arenot given or made clear from the drawings or inspecification. The use of assembly drawingsshowing cladding components related to thestructural framework with all ancillary fixingand jointing methods and with their tolerancesclearly stated would reduce this misunder-standing. It was estimated that the cost of re-medial and rectification work resulting fromaccuracy of assembly lying outside the toler-ances of the design was 5% of the cost of build-ing for the critical activities in the buildingprocess such as cladding.

At that time there were many ideas for furtherstudy of this problem. For example CIB W49 sub-group W6Z ‘Economical Tolerance’ was formedto analysis faults arising from inaccuracy and sev-eral congresses were held to study tolerances anddimensional control. Standards were producedand many learned papers were presented on this

Proceedings IGLC-13, July 2005, Sydney, Australia

330 Theory & practice of modular coordination

subject. As my career progressed first as an Archi-tect and later as a Cladding Consultant it becameclear that over and over again both the designteam and the contractor did not communicate toeach other the three-dimensional tolerances ofcritical moments of the assembly.

I will now review two projects where I haverecently been involved where the need to allowtolerance in the fixings was critical to the perfor-mance of the assembly. The first is Singapore ArtsCentre (D.P. Architects) the second is FederationSquare Melbourne (architect Bates & Smart).

THE SINGAPORE ARTS CENTRE

The Singapore Arts Centre roof was originallydesigned as a monolithic shell by MichaelWilford. I worked with Atelier One London todevelop a dynamic form of triangular glass panelswith aluminium sunshades. Here we were consul-tants to the client assisting the German subcon-tractor Mero to devise a system capable ofmeeting the client’s requirements not only forappearance of the performance requirements forweathering and wind loading. The concept origi-nally developed by the engineerings office AtelierOne was a sculpted non-linear form of the concerthall and lyric theatres with an articulated surfaceusing a space frame grid with glazed infill and alu-minium shading devices. The modelling of sur-face geometry using Micro station creates a meshof equal-length elements to standardize their man-ufacture.

Working closely with the architect MWP/DPAa generic solution of a double layered jointingsystem at the edges of the triangulated glasspanels with their moveable sunshades.The main problem was the large potential numberof variants in the aluminium sunshades and theneed to respond to the complex geometry whist atthe same time providing assurance that the build-ing would not leak particularly if the Prime Minis-

ter was sitting the audience at the time of a heavySingapore rain pour.

The system eventually developed and con-structed by Mero Germany consists of a series ofnodes and chords with an ingenious net ofe.p.d.m. drained gaskets to form the jointsbetween the triangular shaped panels. All thisneeded careful prediction of the forces at eachnode condition could only have been done usingCAD prediction methods. Even the forming andshaping of the aluminium sunshades was possibleby the use of such techniques.

FEDERATION SQUARE MELBOURNE

At Federation Square Melbourne again workingwith Atelier One we were struggling with translat-ing the architects (Bates and Smart) idea for usingan apparent random arrangement of stone, zincand glass panels based on fractal geometry, wedecided on a three point pinwheel to fix the panelsrather like changing the wheel on a car.

This simple concept allowed us to translate theincredible complex structure for this multi formedfacade into a complex but efficient claddingsystem fixed by Permastellisa Australia.

Product Development

Alan. J. Brookes 331

Figure 6: Early model of Singapore Arts Centre.

Figure 7: Mock up of aluminium sunshades and theirfixing to the glazing system.

Figure 8: Apparent random fractal geometry based on apinwheel used at Federation Square.

Recently in Britain there has been a re-interest in aprogramme of industrialized housing led by aGovernment Initiative (John Prescott) for lowcost single family dwellings.

Unfortunately it is not likely that architects andcontractors involved in this programme will lookback to the lessons of the 1960’s and 1970’s inBritain. Each generation has to develop its ownwheel.

REFERENCES

Bonshor R.B. (1972). “Component Standardisa-tion—Graphical Aids for Manufacturers andDesigner”. BRE IN 23 /72, March. Later pub-lished as a Current paper.

Brookes A.J. (1976) Implication of change inbuilding Technology PhD Thesis Liverpool1976.

Bemis A.F. (1936) The Evolving House VolumeIII Rational Design. The Technology PressMIT Cambridge. Mass.

Corker E. and Diprose A. (1963) The modularPrimer. The Modular Society London.

Eden J.F. (1967) “Metrology and the Module inArchitectural Design” March 1967 p 148 alsopublished as BRE current Paper March 1968.

Hertfordshire Architectural Dept. (1948). ‘Plan-ning of new Schools’. Architect and BuildingNews 16 Jan 1948 pp 41–51.

Mac Callum I. (1959) Architecture USA.Reinhold, New York and Architectural PressLondon

Maggi P.N. (1961) “The Problem of tolerances inbuilding”. Associazione Italiana Prefabrica-zione. Quaderna 1 E 2. Milan.

Martin B. (1965) The co-ordination of Dimen-sions for Building. RIBA London.

Martin .B. (1977) Joints in Building. GeorgeGodwin, London.

Rusell B. (1981) Building Systems Industrialis-ation and Architecture. John Wiley.

Sefton Jenkins R.A. and Mak W.K. (1963) “Tol-erance in Building” Modular QuaterlyAutumn 1963.

Wachsmann K. (1961) The Turning Point ofBuilding: structure and design. Reinhold Pub-lishing Corporation New York

Proceedings IGLC-13, July 2005, Sydney, Australia

332 Theory & practice of modular coordination

Figure 9: Complex Structure at Federation SquareMelbourne (Architects Bates and Smart).