Why Steel Columns Read Out of Plumb When the Anchor Bolts Are Fine

Steel Erection · Site Quality & Tolerances

Every erection crew has had this morning. The setting-out was signed off, the anchor rods sit where the drawing says they should, and the first survey still comes back showing column C4 leaning 14 mm across 8 metres. The first thing everyone blames is the rods. In my experience the rods are usually fine, and the 14 mm came from somewhere else.

Here is where I would look, in the order I would look at it.

1. Start With the Tolerance, Not the Column

Before anyone argues about whether 14 mm matters, pin down which tolerance you are working to. The two codes most likely to be on your project do not agree with each other.

AISC 303 (Section 7.13.1.1) states the plumbness of an individual column shipping piece as 1/500 of the distance between working points. A working point is the actual centre of the member at each end of the piece. Not the edge of the flange you can see, the centre.

IS 12843:1989, Table 1, is tighter: out-of-plumbness at the column top is H/1000 or 25 mm, whichever is less, for columns up to 30 m. Where the column supports a crane, the cap drops to 20 mm.

On the same steel, with the same crew, one code gives twice the allowance of the other. I have watched a frame get accepted against one and rejected against the other a week later, and both engineers were right. Nobody had said which one governed until the survey was already on the table.

Measure it properly

Instrument, two directions at 90 degrees, taken at the working points. Sighting down a flange is not a plumbness check. You are measuring the flange, and the flange arrived with its own mill tolerance.

Settle it on the drawings, not on the survey report

The cheapest place to fix this argument is the drawing set. Before the first column goes up, ask for three things to be stated on the erection drawings: the governing code, the plumbness value, and how and where it will be measured. If your supplier prepares the erection drawings, this is a fair request to make at quotation stage — and how they answer it tells you something about the drawings you will get later.

For how base connections and anchor bolt detailing fit into the wider frame design, see our guide to portal frame spans, foundations and fabrication.

2. Grout Under the Base Plate

This is the one I see most often.

The sequence that causes trouble is familiar. The column goes up on levelling nuts, gets plumbed, and grout is poured under the plate the same afternoon because the concrete gang is on site and wants to close out. Now the column is fixed in whatever position it held at four o’clock. If the survey comes back the next morning showing movement, nothing can be adjusted without breaking the grout back out.

The quieter version is shrinkage. A site-mixed, cement-rich grout pulls back as it cures, and what it leaves behind cannot be seen from outside. The plate looks bedded. It is bedded on something no longer fully there.

3. The Bolt Cage Moved During the Pour

A rod cage that is dead right before the pour is not automatically dead right after it. Vibrators shift things. So does a concrete hose being dragged across the formwork, and so does anyone who steps on the template.

The templates are part of this too. A plywood template that has been through nine bases is not the shape it was on the first one.

Get the cage checked immediately before the pour, and again once the concrete is in but before it sets. A rod 6 mm out is a five-minute fix with a pipe wrench while the concrete is green. The same rod is a base plate modification once it has cured.

4. Levelling Nuts Doing a Job They Were Never Meant to Do

Levelling nuts are for levelling. On site they end up carrying the full weight of the column and a good part of the steel above it, sometimes for weeks.

The failure mode is not dramatic. A single nut under a plate thinner than it should be will slowly bed into the thread, and the column settles a millimetre or two at a time. Shims packed loosely around the rods do the same thing. By the time the frame is loaded the movement has already happened, and nobody traces it back to the nut.

Use a second nut to lock the first. Make sure the shim packs are steel, are tight, and sit close to the rods rather than out at the plate edge. And if a column is going to stand on its levelling arrangement for more than a few days, that is a question for the engineer, not a habit to fall into.

5. Tolerance Stacking Up From the Top of the Foundation

The foundation is built to a concrete tolerance and the steel is erected to a steel tolerance, and the two are not the same thing.

That difference has to be absorbed somewhere, and if it is not absorbed by a designed shim pack it gets absorbed by the column.

Add the rod projection, the flatness of the plate as it left the shop, and the mill tolerance on the column length, and a few millimetres here plus a few millimetres there stops being a few millimetres. Start 15 mm out because the slab was poured high, and the tolerance is gone before the column arrives. What the shop is actually held to before anything ships is its own subject — we have written a separate fabrication quality control overview.

6. The One That Gets Blamed on Everyone: The Sun

A column standing free in the sun is not the same shape at 2pm as it was at 7am. The face toward the sun runs hotter, expands more than the shaded face, and the member bows away from the heat. On a tall slender section that is several millimetres, and it reverses overnight.

I have seen a crew spend a full morning shimming a column that was never out of plumb, because every survey was taken in the afternoon and every check was chasing the same thermal bend.

7. What to Check Before Anyone Reaches for a Shim

In this order:

1. Is the grout cured? Correcting a column on green grout just moves it again.
2. Are the anchor rod nuts actually loose enough to allow movement? You cannot plumb a column that is already clamped.
3. Is the frame properly braced and guyed? This is not optional, and under AISC 303 (Section 7.10.3) it is the erector’s call.
4. Is it one column, or every column on the line leaning the same way? One column is a local problem. A whole line leaning is a setting-out, foundation or thermal problem, and shimming each one separately makes the frame worse.
5. Is the member bent rather than out of plumb? Straightness is a separate tolerance, H/1500 or 10 mm under IS 12843. Shimming will not fix a bowed shaft.

If shims are the answer, they should be steel, they should be tight, and the stack should not be a tower of thin plates. Once the plate is packed, the grout has to fill the whole void.

FAQs