03.07 Cladding tolerances

Categories: Buildability

Introduction
The achievable accuracy of erection of a cladding system depends in part on the accuracy of manufacture of the cladding components in the factory.  It also depends on the accuracy of erection of the frame and the ability to adjust brackets and fixings to accommodate tolerances.

The required accuracy of erection of the cladding will depend on the location, visibility, purpose and required performance of the building.  The cladding is the skin and the clothes of the building and both appearance and sealing are dependent on the accuracy of the finsished wall.  This is described in greater detail in Section 03.06.
 

As-built tolerances
The acceptable as-built tolerances for different forms of cladding are given in the table below.  These are the tolerances allowed by the Standards referenced in the table.  A lower accuracy of construction may be acceptable and appropriate on some low grade forms of construction with the agreement of the client.  In other cases a higher standard than those given may be required.  In any case every specialist contractor will have their own standards to which they work and these may be in excess of the minimum values set out below.
 

Notes

1)  Manufacturer’s value
2)  Measured value (BS 5606)
3)  Tolerances for cast/punched openings within one concrete panel
4)  Tolerances for formed openings between concrete units
 

Geometrical form
As more components and more complex geometries are incorporated into the building, their inter-relationships become increasingly complex, with many implications for buildability:

• What manufacturing and erection deviations in shape (e.g. agreement tolerance of diagonals) are reasonable and permissible?
• What setting-out and checking procedures can be used?
• What means of stabilising the construction are required (e.g. due to bending of long members oriented out-of-vertical)?

The complexity of the cladding in terms of shape and number of components has implications for the required accuracy of components.
 

Single curvature
A barrel vault roof or curved vertical wall will take the form of a faceted curve or will contain curved glass.  With a faceted curve the effects of the facets become more pronounced as the radius of curvature is reduced or the panes are increased in size.  However faceted glazing is easier to install than curved glazing.

When setting out and fixing faceted glazing around a curve it is easy to overlook the need for each faceted pane to lie in its correct plane.  Setting out of vertical curved walls is done by drawing a curve or setting off from a base chord.  However, the fit of the glass depends on the spacing of the mullions which must also be checked as the frame is installed.

When installing glazing to barrel vault roofs the shape of the supporting steelwork and relative positioning of consecutive frames is important to the finished line and appearance of the vault.

The use of curved panes of glass can give rise to problems of tolerance and appearance.  The process of bending glass involves heating and shaping the glass at temperature.  This may induce deviations in terms of radius of curvature, and hence chord length, that are far greater than the deviations normally associated with glass.  Secondly a warp may be induced in the supposedly cylindrical shape of the bent glass.  This adds three critical dimensions to the pane; top chord, bottom chord, and plane (no warp).
 

Double curvature and more complex geometries
Complex multi-faceted walls and roofs may be neither plane or cylindrical.  These will involve the use of non rectangular glazing as may happen with some simpler shapes of roof, for instance hipped roofs and other intersecting planes.

The use of CAD packages can easily produce complex geometries in which no piece of glass even has a right corner.  When such systems are used to produce drawings of complex geometries it is important that the glass cutting procedures are of commensurate accuracy.

A greater problem is the setting out and fixing of glazing frames in three dimensions.  It is essential that members are cut to length accurately for in these types of construction a member of the wrong length can often be made to fit.  This of course is at the expense of distortion of the frame which only becomes evident when the glass will not fit.  Setting out should be by offset and not by measuring angles.  With the more complex geometries trial panels or bays are often erected before commencement of work on site to prove the buildability of the cladding.
 

Manufacturing tolerances
The accuracy of manufacture of cladding components will depend on;

• Materials used,
• Manufacturing process,
• Size of component/assembly.

The manufacturing process is closely linked to the materials used and there are inherent limits to accuracy associated with each material.  However, better accuracy can be achieved in most cases by using better manufacturing equipment and by better workmanship.  For instance a metal bar may be sawn in one of the following ways and accuracy increases going down the list:

• Hand sawn at site
• Hand sawn at factory
• Machine sawn
• Machine sawn to an end stop
• CNC machined

In all cases accuracy will also depend on checking of the sawn length and calibration of CNC machinery.

Some manufacturing process are inherently inacurate.  These include:

• Cutting plates and making panels or units that have no right corner,
• Bending of sheets and panel,
• Bending of framing members.

Production of curved glass or aluminium sheets is difficult to achieve with great accuracy.  It is difficult to control the final radius of the component and if the flat sheet is not fed squarely into the rolls then a warped panel is produced rather than a cylindrical one.

The greatest problem of tolerance is the production of perfectly flat or cylindrical surfaces.  The eye readily sees any such imperfections and judges the quality of the cladding accordingly.  Lack of flatness is often associated with thinner materials that are less stable and with thermally induced bowing associated with toughened glass and with some welded panels.  Flatness should be measured in terms of deviation from plane within a given length.  Bow refers to deviation from plane in the length of the member.  Ripple is quoted as xx mm in yy mm and is usually taken over 150mm or 300mm.  To check whether a curved component is truly cylindrical a circular tempate or a simple three point gauge may be used instead of a straight edge.  To check that there is no warp a curved edge of the panel is measured for flatness by standing it on a plane surface.

The table below gives accuracy of manufactured components as required by the standards referenced.  These are minimum standards and many manufacturers will be able to produce to greater accuracy.  However, accuracy often comes at a cost and components of an appropriate accuracy should be specified.
 
 

Notes
1)  Applies to Aluminium Alloy I;
2)  d_max denotes lateral curvature;
3)  ‘Normal tolerances’ are stated;
4)  Manufacturer’s value for maximum depth of roller wave of toughened glass;
5)  Class A tolerances are stated;
6)  For a section within a circumscribing circle of diameter > 100mm;
7)  For a section within a circumscribing circle of diameter > 80mm and length £ 8m;
 

Achieving fit
It is a basic requirement for achieving fit that there is sufficient space for the various components. This can be achieved by using overlapping components so that variations in size can be accommodated by changing the amount of overlap but for most elements this is not possible and specified sizes must be chosen to avoid overlap or conflict. For example if a window frame is to be installed in an opening the maximum size of the window should not be greater than the minimum size of the opening.
This process involves a number of steps.
 

Identifying critical dimensions
A critical dimension is defined as one in which normal permissible deviations may subsequently prevent the fit of related components. Achieving fit requires the critical dimensions to be identified. The likely deviations associated with these dimensions can then be evaluated and allowed for in the design.

Difficulties of assembly are related to the degree of dimensional control required, that is, the number of critical dimensions. Reducing the number of critical dimensions will therefore reduce problems of lack of fit. For example a structure in which the external cladding fits within the frame of the building is more likely to give problems of lack of fit than one in which the cladding is fixed outside the frame with the only contact at fixing points.

In the example of the window above the critical dimensions are the height, width and squareness of the opening and frame.
 

Evaluation of induced deviations
The values of induced deviations will, in many cases, follow a normal statistical distribution around the mean size or position. Randomly occurring induced deviations may be treated according to statistical principles, so that account can be taken of the relative probability of small and large values, and may be expressed in terms of the standard deviation, as a measure of variability.
Values may also exhibit a bias reflecting systematic, rather than random, variability, such as a fixed deviation due to maladjustment of measuring instruments. Systematically occurring deviations apply to batches or groups of components or measurements and must generally be treated as definite, recurring values.

BS 5606 provides guidance on the accuracy that can be achieved in masonry, concrete, steel and timber members and structures based on measured and estimated survey data.  This reflects the standards of construction/erection and manufacture achieved by industry in 1979 and 1990 respectively. For example, BS5606 indicates that the edge of a suspended concrete floor slab may vary from its intended position by +/-15 mm in plan and +/-25mm in level. These values have a probability of 1 in 22 of being exceeded and to reduce this probability to 1 in 80 the deviation would be increased to +/-19mm and +/-31mm respectively. For a probability of 1 in 370 the tolerances would be 22.5mm and 37.5mm.

Standards covering the accuracy of cladding components are given above.
 

Joint width
The interface between components constitutes a joint and in most cases it will be necessary to allow a space for the joint. The design joint width will depend on both the width necessary to construct the joint satisfactorily and the width necessary to allow the joint to function as intended. For example the joint must have sufficient width to be able to accommodate inherent deviations. One of the main functions of the joint is to compensate for the tolerances in the adjacent components and it will be necessary to determine the range of widths over which the joint will be satisfactory.
 

Assessment of specified dimensions
To ensure that components fit, the specified values for the critical dimensions must take account of both induced deviations and the width of the joint.

One approach would be to specify maximum and minimum values rather than mean values. In the example of the window, the size of the opening would be specified as a minimum value and the size of the frame would have maximum values specified. To ensure fit these values would have to differ by the minimum width of the joint and the overall variation in permissible joint width would have to equal or exceed the combined tolerance on the opening and frame.

Thus if the width of the opening is W with a tolerance of +2w and -0, the width of the frame is F +0/-2f and the width of the joint is J +/-j then:

W  ³ F  +  ( J - j )

    and

f  +  w  £ j

If it is not possible to satisfy these conditions simultaneously it will be necessary to modify the design by for example requiring tighter tolerances, using a joint with greater range of acceptable width or increasing the number of joints.
In practice the method given above will be conservative as it allows for the possibility of the extreme deviations occurring in both critical dimensions simultaneously. Due to the statistical distribution of the deviations this is unlikely to occur. To give a more economical result the overall tolerance on a dimension made up from a series of components each with an associated tolerance is given by the square root of the sum of the squares of the individual tolerances. Thus in the above example the required range of joint width is determined from

( f ² + w ² ) ^0.5 £  j

It should also be recognised that the statistical distribution of deviations means that there will be a small chance of the actual deviation being greater than the stated tolerance. Values of tolerances are often quoted with a chance of approximately 1 in 22 of being exceeded. This may give an unacceptable risk of lack of fit and it may therefore be necessary to allow for greater tolerances, which will have a lower risk of being exceeded.
 

Buffer zones
In some situations the risk of lack of fit can be reduced by incorporating a buffer zone in addition to the allowance for normal tolerances. The buffer zone then provides a contingency for cases where the specified tolerance is exceeded. Buffer zones are most commonly used between the building frame and the cladding.
 

Design of fixings
Fixings must also be designed to accommodate tolerances. For example if a curtain wall is fixed to the edges of the floor slabs, the brackets must be able to adjust to accommodate the tolerances on the construction of the floors described above. The required adjustment can be determined in the same way as the range in joint width.
Adjustment provision can be incorporated by one or more of the following features:

• Oversize or slotted holes;
• Site-drilling or welding after positioning of components;
• Shims, packing pieces or washers (low friction for movement provision);
• Sliding/serrated connections or adjustable bolts.

Where the adjustment is provided in steps as in the case of serrated surfaces it is important to ensure that the size of the steps is smaller than the tolerance on the erected cladding.
 

Construction procedures
Even where the cladding has been designed to take account of tolerances, the erected cladding will only be satisfactory if the erection of the cladding is carried out in accordance with good practice.

• A fundamental requirement is to ensure that all setting out is carried out from defined datums using equipment of the required accuracy. Setting out fixing locations from the edge of the floor slab rather than reference grid lines is unlikely to lead to a satisfactory result.
• The maximum number of shims/packing pieces should be specified.  Excessive shimming, which is usually the result of a structure being out of tolerance laterally, can lead to inadequate penetration of fixing bolts and to both a reduction in pull-out strength and an increase in the bending stresses in the anchors/bolts.  It may also prompt the substitution of longer bolts, again inducing unforeseen bending loads.
• The final position of a bolt in a slotted hole may need to be specified so as not to restrict post-installation movement.
• Panel-to-panel butt joints must be kept within the practical (installation) and physical (working range) of the joint seal.
• Minimum lengths of overlap must be maintained for weathertightness.
• Lateral adjustment (of masonry or concrete cladding panels) should not lead to a loss of bearing of the seating.