Reverse-Dial Shaft Alignment Shim and Move

Because BOTH readings are rim readings and neither depends on a face being square. Two indicators sweep each shaft from the other, giving the relative position of the two shaft centerlines at two planes a known distance apart -- and two points define a line, so the misalignment is fully described without any face measurement at all. That is exactly why it is the method of choice on long couplings and spacer couplings. Everything after that is one straight line extrapolated to the feet. The single largest source of error is the direction convention: the distance to a foot must be measured in the SAME sense as the plane spacing, from plane A toward plane B, and a foot on the far side of plane A carries a negative distance. Readings of -0.014 in at A and +0.022 in at B ten inches apart give centerline offsets of -7.0 and +11.0 mils and a slope of 1.80 mils per inch, which projects to RAISING the front foot 3.8 mils and RAISING the rear foot 36.2 mils -- both the same direction, because over this foot spacing the line has already crossed zero before the front foot. That is worth stating plainly, because the moves being opposite is a real and different signature, and reading two positive numbers as opposite is how a crew makes an alignment worse instead of better. The projected centerline position at the coupling centre is reported so the result can be checked against a tolerance rather than trusted. This is the vertical solve on readings the user takes. It assumes both indicators swept with the shafts rotated together and that bracket sag has already been measured and subtracted -- on the long brackets reverse-dial invites, sag is easily larger than the misalignment being measured. It does not do the horizontal solve, which is the same arithmetic on the side readings and is corrected by jacking rather than shimming; it does not check the result against a tolerance, apply a thermal growth target, or verify soft foot, and every alignment number is meaningless until soft foot is zero. The machine manufacturer's alignment specification, the coupling manufacturer's data, and the plant's own precision maintenance procedure govern.

centerline offset at each plane = that plane's TIR / 2; slope = (offset B - offset A) / the plane spacing; move at a foot = offset A + slope x the distance from plane A to that foot, measured in the same sense as the plane spacing.

The reverse-dial alignment relations as standard millwright practice, by name. Both readings are rim readings, so no face measurement is involved and axial float does not affect the result. Bracket sag must be measured and subtracted. The machine manufacturer's alignment specification and the plant's precision maintenance procedure govern.

Similar-triangles arithmetic on two dial readings; no manufacturer table is reproduced.

Estimate. AHJ and licensed professional govern.

Field names used by the API: tir_a_in, tir_b_in, plane_spacing_in, front_foot_distance_in, rear_foot_distance_in, coupling_center_distance_in, offset_a_in, offset_b_in, slope_in_per_in, front_move_in, rear_move_in, coupling_center_offset_in

• Direction convention governs the answer a foot behind plane A carries a negative distancemillwright practice
• Bracket sag is not corrected here and reverse-dial invites long bracketsmillwright practice
• Vertical solve only the horizontal solve is the same arithmetic on the side readings, corrected by jackingmillwright practice