Ties and anchors: testing, recording and what goes wrong

Ties are where a scaffold design meets a building nobody surveyed internally. The drawing specifies a pattern and a duty; the substrate decides whether that pattern is achievable, and the substrate is frequently not what anyone expected — render over blockwork, a cavity where solid was assumed, brick that crumbles under a drill.

The test is about the building, not the tie

Testing exists because the fixing is being made into a material of unknown strength. The tie assembly is a known quantity; the wall is not. That is why a result is specific to a location and a substrate rather than to the job, and why a pattern that tested fine on the south elevation tells you comparatively little about the north one if the construction differs.

What the required frequency, proportion and load are is set by current published guidance and by the designer, both of which are revised. They are the authority. What is worth understanding operationally is why the answer varies with the building rather than the scaffold.

The record is the part most often thin

A test result recorded as a pass, with no location, no substrate description and no value, cannot be used later — not to defend the structure, not to inform the next job on the same building, and not to explain why a tie failed. The useful record names where, into what, what was achieved, who tested and when.

It also has to reach the design decision. A test result that stays in a site folder while the structure diverges from its drawing is doing none of the work it could.

In practice: a pattern that had to change

A gang drilling for ties on a 1960s block finds the outer leaf is considerably weaker than assumed. The first three test results are marginal. The right response is not to proceed with more ties of the same type in the hope that the average holds; it is to stop and refer back, because the assumption underlying the whole tie pattern has just been shown to be wrong for this building.

What determines whether that happens is how quickly a gang can get an answer. Where referral takes a day, the pressure runs the other way — and marginal results get absorbed as normal variation rather than treated as information.

Common mistakes

• –Recording a pass without the location, substrate and value achieved
• –Treating results from one elevation as representative of a building with mixed construction
• –Continuing after marginal results rather than referring the pattern back
• –No record of who tested, with what equipment, and when
• –Results held on site and never reaching whoever owns the design
• –Not re-testing after the building is altered, or after a tie is moved

Different buildings, different problems

Substrates vary more than drawings assume. Modern cavity construction, older solid brick, render over blockwork, concrete frame with infill panels, and systems with insulation behind the outer skin all behave differently — and several of them look identical from the outside.

That is the argument for spreading tests across a structure rather than concentrating them where access happens to be easy. A result from one corner of a building with mixed construction describes that corner and very little else.

Getting this right

Results have to travel. A test recorded on site and never reaching whoever holds the design lets the tie pattern and the building diverge unseen, and the same record is what a later inspection needs to judge whether a tie deserves another look.

• –Spread tests across the structure, not where access is easiest
• –Record location, substrate, value achieved, who tested and when
• –Treat marginal results as information about the pattern, not as variation
• –Re-test after the building is altered or a tie is moved
• –Get results to whoever owns the design, not just into a site folder

Where this connects: when the building changes

A tie pattern is designed against a building as it was surveyed. Buildings change during the work a scaffold is there to enable — render removed, panels replaced, openings formed, substrate exposed or altered. Every one of those can affect what a tie is fixed into, and none of them generate a prompt to revisit the pattern.

Which makes this one of the clearest cases for the adaptation log covering the building rather than only the scaffold. A tie relocated because the wall behind it was removed is a change to the structure's basis, and it needs the same treatment as moving a standard.

• –Treat changes to the building as changes to the tie basis
• –Re-test after a tie is moved or the substrate is altered
• –Spread tests across the structure rather than where access is easy
• –Record location, substrate, value, tester and date
• –Get results back to whoever owns the design
• –Refer marginal results rather than averaging them away