News and Press

Designing a glass spiral staircase

15 April 2008, The Structural Engineer

Bob Barton of Barton Engineers Ltd describes
the challenge of designing a glass spiral
staircase at Hyde Park Square, London

Glass staircase

This project is the latest in a series of collaborations between
Cantifix of London Ltd and Barton Engineers Ltd. An existing
house, in Hyde Park Square near Marble Arch, had been remodelled and contained a large double height living space. The architect and owners wanted to keep the space as light and open as possible, and with Cantifix Ltd, we had already created a roof using glass beams, as well as a glass floor and balustrades around the edge of a mezzanine kitchen area. The final act would be to link the upper and lower spaces with a staircase.

Having already used transparency to create such a light and airy
space, the architect and client were both very keen on using the same principles for the staircase. Space requirements had made a spiral stair inevitable so, with the team, we considered various ways of building the stair. One obvious structural form would be to use a steel central post and perimeter curved stringer plate. Such a structural arrangement is common and lends itself to the use of a number of different tread materials, including timber and even acrylic, as well as glass. There was a clear desire to innovate, having done so much in the rest of the building, and so conversations about the staircase form continued, as the steel ribbon spiral was rejected by all of us.

The concept of the Georgian stone staircase was discussed, a form we have used a number of times in stone, concrete, and even steel and timber. Ideas of using glass treads and risers in combination to distribute gravity forces like stone treads were discussed. However the architect was uncomfortable with the use of ‘solid’ risers, feeling rightly that the stair would appear too solid and massive for the space.

Eventually we proposed the idea of ‘stacking’ plates of glass and
then clamping them in position. The same principle has been used in
concrete staircases for many years. Identical precast concrete treads are stacked one on top of another, and clamped in place using a single vertical steel bar through the central spine. The principle seemed easy to implement and would allow the use of open risers. However, as always, swapping materials in a structural form usually generates design problems. One tends to forget how well structural form and material weave together in design.

The basic scheme was agreed and comprised a series of identical
glass treads and circular glass disks, stacked together in an alternating sequence, and displaced radially around a common axis, like hands on a clock or blades on a Swiss Army Knife, to form the spiral of the staircase. Common sense tells us that the weight of the treads and a person standing on them will tip the stack over. In structural engineering language, eccentric forces will induce bending in the central spine that cannot be resisted because the joints between plates have no tensile resistance, and the staircase is not heavy enough to act as a gravity structure. The next stage of the scheme was to introduce a steel rod through the centre of the spine to resist tensile forces. However, tension would still exist in the extremity of the spine cross section and gaps
would still open up between treads as the spine flexes under bending forces. This is analogous to the ‘cracked section’ in reinforced
concrete theory.

We then considered the idea of post tensioning the steel bar, and so pre-compressing the glass column. If the post tensioning force is
high enough this would put the whole of the spine cross section into compression, even when full design bending forces are applied by
walking on the treads. This post tensioning increases the effective stiffness of the spine by preventing the dry joints between the glass
plates opening up under tension, and is identical to the mechanism that prevents concrete from cracking in the tensile zone of post
tensioned slabs and beam structures. The post tension used in the steel bar was determined to be 5t.

Because we were concerned about the architectural impact of restraining the staircase spine at the top, we decided to form the staircase as a vertical cantilever from its base. The poor condition of the existing ground slab made a new foundation inevitable, and a mass
concrete footing on the dense gravel beneath the lower ground floor seemed a sensible option. Stability checks on the glass column
showed the self weight of 600kg plus the 50kN post tensioning well within the spines’ buckling capacity, assuming the full effective
stiffness of the cross section is mobilised.

Normally in floor design we use laminated annealed glass. The combined strength of the laminates resists long term loading, and each
individual glass lamination is capable of resisting short term applied loads should the other laminate crack. The sheets are designed to act in unison, rather than compositely in the conventional sense, because we have found from our own experiments that the normal resin materials used in laminating
glass creep very significantly over a short period of time. The tread widths were profiled to suit the architectural scheme and to ensure that we had enough effective width of tread where the highest bending stresses occur at the tread-to-spine junction.

The layout of the staircase meant that the treads formed part of the spine, incorporating the circular spine profile and a hole for the
steel bar. Because of this the laminating resin would be compressed by the post tensioning, and would result in massive loss of pre-stress force due to creep. Our initial solution to this problem was to propose laminating a steel plate into the soft resin zone, and so making the spine section of the tread effectively incompressible. However, it proved impossible to control the lamination thickness accurately, and the amount of resin around the steel
plates. Prototypes were made and load tested, and all failed due to large very local distortions in the resin interlayer as the spine was compressed and the resin immediately adjacent in the tread was not. The tread manufacturers had not been able to control the amount of resin between the steel plate and the glass, resulting in a thicker interlayer with the steel plate ‘floating’ within it, and still having a significant amount of squashable resin within the spine.

We gave some thought to alternatives for the steel plate and then proposed a modified acrylate adhesive. This material is much stiffer than resin, is clear with no coloured tint, can be cold poured into narrow gaps, and will cure in anaerobic conditions under ultraviolet light of particular wavelengths. It is however relatively expensive, certainly orders of magnitude more expensive than normal cold pour resins. Our calculations were backed up by further prototype load tests, and the decision was made to complete the tread manufacture using modified acrylate as thelaminating material.

The staircase was assembled on site by hand, the heaviest component being the glass treads weighing just under 20kg each. A simple framework of softwood timber was constructed as the staircase was built, to act as a falsework support for the projecting treads. Care was required to keep the mating surfaces of glass clean and so prevent local stresses occurring during post tensioning. Then, once everything was in place and
complete, a specialist post tensioning company was brought in to carry out this task.

Immediately after completion we became concerned that the staircase was nowhere near as stiff as predicted. Climbing more that two or three treads up the staircase, the whole assembly started to oscillate at low frequency. Even a light push at mid height resulted in a similar low frequency wobble. The staircase was immediately re-propped and we rushed back to recheck our stability calculations. The only explanation for the observed behaviour
was that the spine was unstable; the combined weight and post tensioning force would have to be close to its buckling capacity. This meant that the spine was more flexible than predicted, and this could only be the case if the Young’s modulus of the material was incorrect or the post tension force was lost or had not been high enough.

We then asked Sandbergs LLP to check the force in the central rod and they found it to be 40% of that specified. Further enquiries revealed that apparently the subcontractor had been reluctant to apply the full force because he was frightened of breaking the glass! Once the full force was applied the staircase performed as originally predicted, and we could all breathe again.

To comply with the 100mm sphere rule for openings within the stair, intermediate toughened glass plates were also incorporated into the stack. These closed the open riser gaps between tread plates. They also provided a tension anchor for the toughened glass balustrade plates, which were boned to the edge of the treads and intermediate plates using a similar modified acrylate adhesive. An additional architectural feature was a string of LEDs threaded through the hole in the central spine, to provide a ‘body colour’ light to the spine that then glowed in a series of different colours.

Overall the staircase has proved to be a great success. The architect and client are both very happy. Both Cantifix Ltd and ourselves have gained a great deal from the collaboration, and we are currently developing a series of glass staircases using many of the lessons learned. The value of prototyping and testing construction methods has been reinforced. We would like to express our thanks to Will and Charlie Sharman of Cantifix of
London Ltd for their skill, patience, and good humour throughout the sometimes difficult process. We would also like to thank the staff of Sandberg LLP for their expertise and integrity in the testing of the components and the final assembly. And lastly, of course, we must thank our client, who had the vision and determination to want the staircase.

Credits
Architect: Michael Mallinson Architects
Structural Engineer: Barton Engineers Ltd
Glass Design and Build Contractor: Cantifix of London Ltd
Testing Services: Sandberg LLP

Barton Engineers on Institution of Structural Engineers Task Group / January 2012

Autodesk Q&A / November 2011

Top 200 Consultants / Building Magazine October 2011

In The Footsteps of the Great / RIBA Journal October 2011

Hanging Bicycle Velodrome on Architecture Today cover / July 2011

Barton Engineers are appointed to The National Gallery / June 2011

Chilver Hall at Cranfield University wins at LABC Building Excellence Awards

Top 250 Consultants and Top 100 Engineers / Building Magazine October 2010

Structural Analysis of Leutschenbach School / AT June 2010

Top 250 Consultants and Top 100 Engineers / Building Magazine October 2009

Hind House Feature & Cover / AT June 09

Barton Engineers Newsletter Summer '09