When the anchor bolts do not line up: what gets measured first

The base plate comes off the truck and down onto the pier, and three of the four rods will take a nut. The fourth stands far enough out that the plate will not sit flat over it, and it is not going to be argued into place with a pry bar. Note S2-009, field, hospital wing addition, south column line, read at 6:40 in the morning before the crane was called. Four 3/4 in diameter anchor rods, 12 in of projection above the pier, an 18 in by 18 in bolt pattern on the setting drawing, a W14x90 column on a 24 in square base plate 1 1/4 in thick with 1 in holes.

What gets measured before anything gets touched

The order never changes, and it runs from the paper to the concrete. Setting drawing first, open on the tailgate: bolt pattern 18 in by 18 in, group centered 6 in off the column line, rods 12 in above the top of the pier, plumb held to 1:40. Then the instrument, a 25 ft steel tape, not a laser. A laser has nothing to bounce off a round rod at close range; a tape hooks the rod shank and reads the same number for two people.

Then the readings, in a fixed sequence: both directions across the group, both diagonals, the group center back to the column line, and the plumb of the worst rod, read with a torpedo level and a tape held at the top. Only then does anybody look at the plate, because measuring it first is how a crew talks itself into a fix before it knows what moved.

What the tape read on this pier

North-south across the group: 18 1/16 in against 18 in on the drawing. East-west: 18 1/32 in. Both inside anything the drawing would argue about.

The diagonals were the first thing that looked wrong. 25 7/16 in one way, 25 5/8 in the other, unequal by 3/16 in. Equal diagonals on a rectangle mean square, so this group sat out of square by a little under half a degree, which moves the far corner about 1/8 in.

The group centerline was the story. Tape hooked on two rods and pulled back to the column line, the drawing says 6 in. It read 6 3/4 in, twice. The base plate carries 1 in holes for 3/4 in rods, so each bolt has 1/8 in of play in any direction. Three quarters of an inch is six times the slack.

How do you tell a group that shifted from a group that is out of square?

They are different problems with different fixes.

A group that shifted moved as a unit. The pattern still measures 18 in by 18 in, the diagonals still match, and every rod is wrong in the same direction by about the same amount. That is a layout problem, and it usually traces back to the formwork or to a grid line set from the wrong benchmark.

A group that is out of square has rods that are each close to right and collectively rotated. The pattern dimensions read true, the diagonals disagree, and the error grows as you walk away from the corner the setter worked from.

This pier had both, the shift doing the damage. The plumb check found 5/16 in of lean at the top of the fourth rod over 12 in of projection, 1:38 against the 1:40 the drawing carried. Those are this job's numbers, not a rule for all concrete everywhere.

The options, in the order this shop works through them

One rule sets the order: you can always take more material out of a plate, and you can never put it back. Each option closes the door behind it.

Re-read and re-document first. Recheck the dimension against the setting drawing and against the foundation drawing, because a plate built to a superseded revision looks exactly like a misplaced bolt group. Here the two agreed, so that door was shut.

Second, the plate hole. Opening a 1 in hole to 1 3/8 in on the two worst rods is a mag drill and twenty minutes in the field. Slotting it and covering it with a plate washer is the same idea with more take-up, detailed in AISC Design Guide 1. Neither is a crew call: the hole is bearing area and the net section around it carries the column.

Past that come chipped concrete, bent rods, cut rods and adhesive anchors, the ladder Modern Steel Construction has been writing about for years. All of them are a design change with a paper trail, not a field decision.

What is the one thing nobody does on this shop's jobs?

The rod does not move.

Everything else in the assembly can be measured, re-cut, re-drilled or replaced, and this shop has done all of those. The rod is the one part already cast into cured concrete, and it was designed as a straight bar with the load running down its axis into the embedment. Bend it and the load path has a kink in it, at the point where the bar is also taking the largest moment from the column above. Heat it and the steel is no longer the steel on the certificate behind it.

The plate moves, or the hole changes, or the engineer changes the design. The rod stays where the concrete put it.

What this note does not cover

It does not cover the concrete. Embedment depth, the elevation of the top of the pier, concrete strength that morning, the grout under the plate: none of it is in this note, and any of it can move a base plate as surely as a bolt can.

It does not cover the rod material. These rods carried an F1554 stamp on top; the certificate behind that stamp, and the embedment design that goes with it, live in a different folder.

It does not cover the frame. Bolted connections above the base plate belong to the RCSC Specification for Structural Joints Using High-Strength Bolts. Anchor rods are not part of that document; they get snugged, not pretensioned.

It does not say what the fix is, or who pays. The AISC Code of Standard Practice, Section 7.5, is where responsibility for setting anchor rods sits, and that conversation goes easier with a dated tape reading in hand.

The gesture that costs nothing

If you are the general contractor, the owner's rep or the erector on the next one, walk to the pier before the column is on the hook and read the bolt group yourself: both directions, both diagonals, the group center to the grid line, the plumb of the worst rod. Write the numbers down with the date, the time, the instrument and the names of the two people who read it.

That is fifteen minutes on a morning when nothing is in the air yet. The difference between a fix that costs a plate and a fix that costs a pier is usually decided by who measured first, and a reading with no date on it is worth almost nothing six weeks later.