We work in Zone II at one end of the record and Zone V at the other — Bengaluru and Guwahati — and the difference is not that the Guwahati joints are wider versions of the Bengaluru ones. It is that a different calculation is in charge, and a joint sized by the thermal calculation in a Zone IV or V building is being asked to survive an event it was never selected for.
The zones, and what they mean numerically
IS 1893 (Part 1) divides the country into four seismic zones and gives each a zone factor Z, which is the design ground acceleration the structure is checked against:
| Zone | Zone factor Z | Description | Where we work in it |
|---|---|---|---|
| II | 0.10 | Low | Bengaluru, Hyderabad |
| III | 0.16 | Moderate | Kolkata, Chennai, Mumbai, Bhubaneswar |
| IV | 0.24 | Severe | Delhi NCR, Noida, Gurugram, Patna, Jammu |
| V | 0.36 | Very severe | Guwahati and Assam |
The step from Zone II to Zone V is a factor of 3.6 in design ground acceleration. That is not a refinement of the same detail; it is a different problem. Delhi at 0.24 sits at more than twice Bengaluru, which is the single clearest reason a system carried unchanged from a Bengaluru project to a Delhi one tends to be under-specified.
Why the seismic gap is so much wider than the thermal one
The purpose of a seismic separation is to stop two adjacent blocks colliding when they respond out of phase — pounding, which is a documented cause of structural damage in every major urban earthquake, and which does its worst where floor levels do not line up and one block's slab arrives at the other block's column mid-height.
IS 1893 derives the required separation from the calculated storey displacements of the two blocks, amplified by the response reduction factor R. The amplification is the part that surprises people. Design seismic forces are already reduced by R to account for ductility, so the displacements that come out of the analysis are correspondingly reduced; multiplying back by R recovers the displacement the building will actually experience when it responds inelastically. With R typically 3 for an ordinary moment frame and 5 for a special one, that is a large multiplier applied to the sum of two displacements.
Run it for a realistic case. Two 15-storey blocks in Zone IV, each with a calculated design displacement at roof level of the order of 40 mm, R = 5:
Set that against the thermal movement for the same building: a 45 m bay in Delhi swings around 20 mm. The seismic separation is an order of magnitude larger and governs outright. Once seismic is in play, the thermal calculation stops being the sizing case and becomes a check.
A 400 mm gap is not a wide expansion joint. It is a different product category, and it is why the SeismAtec line exists separately from the MegAtec one.
The detail everyone gets wrong: the gap grows with height
Storey displacement accumulates up a building. Two towers separated by a joint are nearly touching at the ground and furthest apart at the roof, so the separation the code requires at level 15 is substantially larger than the one required at level 2.
Structures are usually built with a constant separation, sized on the worst case at the top. That is correct and conservative — but it means the joint system at podium level is sitting in a gap far wider than the movement it will actually see there, while the one at roof level is at the limit of its capacity. Both need a system rated for the full gap, and only one of them is genuinely working hard. Specifying a single system for the whole run is the right call; assuming the movement demand is the same at every level is not, and it is where the anchorage detail at high level quietly gets under-designed.
What the cover has to do that a thermal joint does not
A thermal joint opens and closes. A seismic joint does four things at once, and a system that only does the first will be destroyed by the other three.
- Movement in three axes simultaneously. Opening and closing along the joint's width, lateral shear along its length, and vertical offset as the two blocks displace differently. Real ground motion delivers all three together, not in sequence.
- Return to position. After the event the cover has to come back, not sit displaced with a 300 mm hole beside it. This is what a centring mechanism is for, and it is the clearest functional difference between a seismic system and a wide thermal one.
- Stay captive. A cover plate that lifts out under vertical offset becomes a large piece of loose metal in an occupied building during an earthquake. The retention detail matters more than the plate.
- Keep its anchors. The frames are transferring real load into the slab edge at the moment the slab edge is least happy about it. Anchor type, embedment and edge distance are part of the design, not a site decision — and slab edges at joints are exactly where concrete quality tends to be worst.
This is what ASTM E-1399 cyclic movement testing exists to demonstrate. It puts a joint system through repeated movement cycles and reports the minimum and maximum widths it survives, so the rating on a datasheet is a measured result rather than a claim. When a specification names a movement class, that is what it is naming — and it is the clause a substituted product usually cannot satisfy.
Zone by zone, what we would actually do
Zone II — Bengaluru, Hyderabad
Seismic rarely governs outside specific tall-tower cases. The joint is a thermal component and the money is better spent on the waterproofing detail than on a seismic rating the building will not use. Specifying a Class III seismic system here is a common and expensive over-specification.
Zone III — Kolkata, Chennai, Mumbai
The transitional band, and the one that needs the most judgement. For low and mid-rise the thermal case usually still governs. For anything tall, or anywhere two blocks of different height and stiffness meet, the separation calculation needs doing before the joint is sized — it will sometimes come back larger than expected. Kolkata adds a second reason: deltaic soil amplifies ground motion and blocks founded differently settle differently, and both effects land on the joint line.
Zone IV — Delhi NCR, Patna, Jammu
Assume seismic governs at any separation between blocks until the calculation says otherwise. Delhi also carries the widest thermal swing in the country, so the two demands stack rather than substituting for each other. A system rated for cyclic movement in more than one direction should be the default, not the upgrade.
Zone V — Guwahati and Assam
The seismic case governs everything. Design ground motion is more than double Zone III, valley alluvium amplifies what arrives, and the separation figures are correspondingly large. On our Assam work the joint is specified as a seismic system from the first drawing — there is no version of the problem here where it is a thermal joint with an allowance added. Gaps in the 300 to 500 mm range are ordinary rather than exceptional, which is the top of what any joint system covers.
Where our systems sit
SeismAtec 450 is the seismic line — designed for movement of the order of half the gap width, across gaps up to 500 mm, with the multi-axis capability and the centring behaviour the four requirements above call for. MegAtec 340 covers trafficked floors where the demand is thermal with a seismic allowance, and self-centres. MegAtec 360 takes the wall and ceiling returns, which on a seismic joint are as important as the floor and are routinely left off the schedule until fit-out.
The full range and the zones we have built in are on the Guwahati and Delhi NCR pages — the two ends of the seismic problem in our record.