Beware: Don’t just bolt your beams!

Blog Post
Builders pointing at beams in the roof

The traditional approach used when two steel beams are provided to support a cavity wall has been to design the beams to each carry half the total load and then connect them with pipe spacers and bolts, even when the loading to each beam was not equal.

This view has been challenged by many structural engineers who say you should not rely on spacers and bolts to fix them together. This becomes more important where the beams are likely to carry different loads, a typical situation in a cavity wall where the inner leaf carries the floor joists, resulting in significantly different loads on the inner and outer beams.

It is common for structural engineers to specify I-beams, also known as Universal Columns (UC), for situations described above. The horizontal elements in an I-beam are called “flanges”, and the vertical element is called the “web”.

I-beams are excellent at resisting vertical loading and bending as these are parallel to the web. However, I-beams are generally weaker in torsion, which may lead to twisting or deflection in the beam, particularly when unevenly loaded – this could result in cracking of masonry above, and in severe cases this could lead to failure of the beams.

So, steel beams and their connection details should always be designed by a competent structural engineer. Depending on the individual situation and the loadings involved, the structural engineer may specify one of several options to connect beams together and reduce torsion, such as:


  1. using bolts and spacers at regular centres
  2. welding the beams together or providing a continuous welded plate over the top of both beams
  3. provide web stiffeners between the flanges of each beam

Ideally, onsite welding should be minimised where possible to reduce fire safety risks. If welding is unavoidable, it should be adequately planned and carried out in a safe manner by a competent person – and remember also to check if the client has a ‘hot works permit’ approval process in place before commencing.

Want to read more about traditional building techniques and structures?

Take a look at our articles from the "My dad told me about them!" series. Our most popular ones include:

You might also find this article useful: Beware of point loads on beams.


Please Note: Every care was taken to ensure the information was correct at the time of publication. Any written guidance provided does not replace the user’s professional judgement. It is the responsibility of the dutyholder or person carrying out the work to ensure compliance with relevant building regulations or applicable technical standards.

This article was updated on 1 August 2022


Avoid welding

Submitted 3 years 11 months ago

Last paragraph methodology correct in load design terms, but avoid welding on site where possible, to reduce site fire safety risks- design mech. fixings fit for purpose & to allow safe removal for a day when demolition is planned- CDM Regulation ethos & aid BCO's inspection, as welds are only as good as the welder/techniques etc. is my personal philosophy.


Submitted 3 years 11 months ago

I regularly specify double PFC's for existing & new cavity walls but I always specify steel spacers formed with angle one side of which is welded to 1 pfc and the other side is bolted to the other with a minimum of 2 no m16 bolts. The spacers are at a minimum of 750crs along the length of the beam including close to the support. In this way I consider that the two beams will act together and produce a more economical solution. There is unfortunately still a number of steel fabricators who think that a piece chs or rhs welded to 1 beam and a single through bolt is sufficient but this doesn't transfer the load from 1 beam to another. fabdetail has beenor small

Web diaphragms

Submitted 3 years 11 months ago

One solution which we have previously accepted from engineers, is to provide stiff web diaphragms between the beams (not simply spacers). These usually take the form of small sections of RSC's fixed tightly between the beam flanges and bolted through the beams.

Bolts and spacers

Submitted 3 years 11 months ago

Bolts and spacers will fix the beams together but do nothing to distribute load from one into the other. The idea presented as the traditional design method is flawed. It is more of a traditional observation that beams in pairs are often kept the same size for more practical reasons.

Unless beams are identical welding them together is not entirely reliable - best to get professional advice at that point.

No bolts

Submitted 3 years 11 months ago

Years ago they didnt even bolt them, the beams were just sat side by side on padstones, never had an issue , like most engineering these days its all well over the top.

No bolts

Submitted 2 years 1 month ago

Years ago they used much larger steel section. These days the beams are pared down in size, so some effective connection between them is needed.

UC sections

Submitted 3 years 11 months ago

Simple answer would be to use 203 UC sections (with a 250 wide plate tacked-welded on top if it's a cavity wall). Always OK for domestic work and never found deflection an issue.

UC section

Submitted 2 years 3 months ago

Indeed and good contractors know this and prefer them. Always design for bending and live load deflection. If you're struggling with dead load deflection then take that load out with wedges (25-50% deflection is the maximum required due to arching of brick and safety factors).

Beam depth

Submitted 3 years 11 months ago

I believe the beam depth should only be dictated by the span. I find 203's are not usable over 3m span.

Beam Depth

Submitted 2 years 3 months ago

This is an interesting exchange, but one that is largely unfathomable. 152UC's and 203 UC's are really practical ways of saving on headroom. As a previous writer says deflection isn't always as bad as you think because of arching etc. and the fact that in domestic uses the average Live Load is actually 0.25-0.5kN/m2. This is well know and well research by those of us working in historic conservation. Most deflection occurs in beams due to poor mobilising of the beam, Steel folding wedges or hardwood wedges can pre-load the beam and as a rule of thumb try to mobilise the beam by 0.25-0.5 x the assumed dead load deflection (using a vertical measuring stick as they use for fitting office ceilings). This is because of a) the 1.4 factor of safety and b) the likely arching of masonry in the floors above. I've never had any cracking to any of my projects in 35 years by making this advice, but its surprising how many engineers end up designing for deflection in domestic situations. I've recently been asked to peer review another engineers design of a 9.1m span picture frame which was governed by dead load deflection. This resulted in over £30k worth of steel. By pre-loading the steel with wedges and designing live load deflection and bending the saving was 45% in weight of steel if a 50mm deeper I beam had been used and 33% if a 50mm less deep UC was used. In office buildings this is why we design pre-cambered beams so we're not inefficiently designing for dead load deflection - which once its mobilised is static. So to all those designing long spans and tying themselves in knots as their beams are unnecesarily heavy, your issue isn't a deflection one its a poor engineering design one. Frankly if one of my younger engineers had proposed what I described above I'd be considering it a rookie design error. As for the person who said you can't use a 203UC for more than 3m, that is absolute nonsense, with web stiffeners and proper design approach (i.e not designing for dead load deflection), I've seen spans of 6m and beyond. Cracks are almost always caused by poor mobilisation on the steel beam by poor wedging or a complete lack of it prior to dry-packing. In summary if dead load deflection is your key design factor then you're not designing efficiently!!

Add new comment

This question is for testing whether or not you are a human visitor and to prevent automated spam submissions.

Sign up to the building bulletin newsletter

Over 48,000 construction professionals have already signed up for the LABC Building Bulletin.

Join them and receive useful tips, practical technical information and industry news by email once every 6 weeks.

Subscribe to the Building Bulletin