Replacing a failed expansion joint, in the right order

India built a great deal between the late 1990s and the mid 2000s. Those buildings are now twenty to twenty-five years old, and the movement joints in them are reaching the end of a service life nobody planned for. We are called to more of these every year: a mall concourse that rattles, a hospital corridor where the cover has lifted, a podium that has started leaking into the car park below.

The people specifying these replacements are in an unusual position. Unlike the original specifier, they have first-hand evidence of what the cheaper option cost. That makes them the best-informed buyers in this market — provided the replacement is approached in the right order.

The order almost everyone uses

Remove the old joint. Measure the hole. Order something that fits it. Install.

Every step of that is reasonable in isolation and the sequence as a whole is wrong, because it lets the damaged existing condition determine the specification. The new joint inherits the old joint's assumptions, including the ones that failed.

If a joint failed, something about it was wrong. Replacing it like-for-like is a bet that the only thing wrong was its age.

The order we use: survey, movement, substrate, system

1. Survey — before anything is removed

Measure the actual gap, at three or more points along each continuous run, and record the ambient temperature while you do it. Both halves matter.

The gap will not match the original drawing, and it will not be uniform. More importantly, a gap measured on a 43 °C afternoon in Delhi is at its most closed — take that reading as the design width and the replacement will be too narrow every January. The temperature turns a single measurement into a usable one.

Survey before demolition, not after. Once the old frames are out, the rebate edges are broken and the honest evidence of the original condition is gone.

2. Movement — work out what it should have been

Do the calculation as though the building were new: bay length, exposure, the local temperature range, and the seismic separation if the blocks are structurally independent. The method is in calculating the gap width.

Then compare that answer with what was installed. One of three things is true, and each leads somewhere different:

• The original was correctly sized and simply wore out. The least common of the three, and the only one where like-for-like is defensible.
• The original was under-specified. Movement exceeded what the system was rated for, so it was being destroyed a little at a time from the day it went in. Replacing it with the same thing reproduces the failure on the same schedule.
• The joint was never the problem. The movement was fine and the failure was installation or substrate. Changing the product fixes nothing.

3. Substrate — where remedial jobs actually fail

This is the step that gets skipped, and it is the one that decides the outcome.

The slab edge that will carry the new frames has spent twenty years being the most abused concrete in the building. It has held anchors that worked loose under cyclic load, been drilled and re-drilled, taken impact from whatever crossed the joint, and in a leaking joint it has been wet for years. Spalling, delamination and carbonation at a slab edge are the norm, not the exception.

A new joint anchored into that will fail faster than the one it replaced, because the new fixings are going into concrete that is worse than the original ever was. Reinstating the rebate — cutting back to sound material and rebuilding with a high-strength repair mortar — is usually the largest single item in a remedial package, and the one most often left out of the budget.

Two related things to check while the joint is open, because there will not be another chance:

• Is there a membrane in the gap at all? A great deal of older work has none. If the joint is over occupied space or on a podium, the replacement is the only realistic opportunity to add one, and it means working below the visible cover rather than at it. See where the joint meets the waterproofing.
• Does the joint cross a fire compartment line? If it does and there is nothing in it, that is a compartmentation gap that has been open for two decades.

4. System — and only now

With a surveyed gap, a calculated movement and a known substrate condition, the selection is straightforward and defensible. Choose against the movement you calculated, not the gap you inherited, and size to the surveyed width rather than the drawn one.

What actually fails, from what we are called to look at

Worth stating plainly, because it changes where the remedial money should go. In this market the failure is almost never the extrusion — aluminium in a building interior outlives everything around it. What we find instead:

1. Fixings worked loose. Cyclic load over years, in a slab edge that was marginal to begin with. This is the single commonest cause, and it presents as a rattle long before it presents as a failure — a rattling joint is a warning, not a nuisance.
2. Membrane termination. Not the middle of the membrane; the lap where it was supposed to bond to the structural waterproofing, or an upstand that was never formed.
3. The finish, not the joint. Tiles or stone tented alongside the joint because the finish had no movement joint of its own over the structural one. That is a flooring failure that looks like a joint failure — see why tiled and stone floors tent.
4. A later trade. On commercial buildings, a tenant fit-out that tiled over, boxed in or drilled through a working joint. Common enough that it is worth asking what changed in the area before the symptoms started.

Doing it in an occupied building

Almost every replacement is in a live building, which constrains the method as much as the specification does.

Work comes in short possessions, often at night, and each length has to be handed back clean and trafficable at the end of the shift. That rules out anything needing a long cure or a wide exclusion zone, and it means the run is planned in sections with a safe temporary condition at each end. On the Chennai Egmore concourse the station never closed at all — see how that one was delivered.

Two practical points that save more time than they cost:

• Do not remove more than can be reinstated in the same possession. An open structural gap in an occupied building is a safety issue and a water risk, and the weather does not consult the programme.
• Expect the gap to be worse than the survey. Removing the old frames reveals the true edge condition. Build a contingency for rebate repair into the programme rather than discovering it at 2 a.m.

What to send us

A remedial enquiry is usually diagnosable without a site visit if it arrives with four things:

1. Photographs — along the joint line, and close up at a failure point.
2. The gap measured at three points on one continuous run, with the ambient temperature at the time.
3. What is underneath — occupied space, a basement, external, a podium over a car park.
4. The age of the building, and anything that changed in the area before the symptoms started.

From those we can usually say which of the four failure modes above you are looking at, and whether this is a joint replacement, a substrate repair with a new joint on top, or a flooring problem that a new joint will not fix.