Key takeaway
Preserve the chord, curve reference and quantity definition with a rail alignment value. Raw mid-chord offset and deviation from uniformity are different fields even when both use inches.
Do not compare a bare curve ordinate with a deviation limit
A rail geometry archive exports one column called alignment_inches. A later reviewer cannot tell whether it contains a raw mid-chord offset or a derived deviation from uniformity. Converting the values to millimeters would preserve that ambiguity rather than resolve it.
For the stated US scope, 49 CFR 213.55 distinguishes chord-specific alignment comparisons. FRA’s 2026 explanatory manual describes the curve reference and the difference between a measured mid-chord offset and deviation from uniformity. A curve has an ordinate even when it is uniform; treating that raw ordinate as the deviation changes the quantity being assessed.
The owner’s task is to preserve the measurement definition and calculation lineage. This article illustrates that task; it does not inspect a track, set an operating speed or authorize train movement.
Reconstruct two explicitly scoped fictional comparisons
The hypothetical example stipulates a Class 4 curve-body review under paragraph (a), excluding the separate qualified-cant-deficiency case in paragraph (b). It also stipulates the retained reference means below. Their supporting reference stations and selection method would still need to be inspectable in a real archive.
For the 31-foot chord, subtracting reference 4.000 inches from measured MCO 5.125 gives deviation 1.125 inches. For the 62-foot chord, 16.750 minus 16.000 gives 0.750 inches. Keep each measured ordinate, reference and derived result rather than overwriting the raw field.
Under the specifically selected Class 4 paragraph (a) curved-track columns, the isolated comparisons are 1.125 greater than 1.000 inches for 31 feet and 0.750 at or below 1.500 inches for 62 feet. They do not combine into an overall compliance or safe-track determination.
| Fictional chord and basis | Raw MCO | Retained reference mean | Derived deviation | Selected isolated comparison |
|---|---|---|---|---|
| 31 feet; Class 4(a) curve body | 5.125 in | 4.000 in | 1.125 in | 1.125 > 1.000 in |
| 62 feet; Class 4(a) curve body | 16.750 in | 16.000 in | 0.750 in | 0.750 ≤ 1.500 in |
Keep the reference stations attached to the calculation
The completed arithmetic does not establish that a real reference was selected correctly. FRA’s explanation addresses reference observations and the treatment of the curve body; transitions and other geometry can change the appropriate interpretation. Retain the locations, chord lengths, reference observations and stated calculation method used for the reported value.
If a packet supplies only the final deviation, a reviewer may be unable to reproduce its reference. If it supplies only raw MCO, the reference may not yet have been established. Export those states differently. Neither should be silently represented as a fully reconstructed result.
In the fictional example, the reference mean is an explicit premise. The artifact shows the subtraction and selected comparison that follow from it. It does not claim that reference stations were surveyed, that an instrument was calibrated or that an entire curve was evaluated.
Do not average away the chord distinction
The two rows are not interchangeable estimates of a single number. Their chord lengths define different spatial supports, and the selected regulatory table gives them different comparison columns. Averaging their deviations would erase the particular field result that the reviewer needs to examine.
FRA’s manual discusses using both chord lengths within its stated Classes 3–5 context, including the relevance of the shorter chord to shorter deviations. Preserve which chord produced each result. A pass-like result in one column is not a reason to delete or replace the other row.
Paragraph (b) of the rule addresses a separate qualified-cant-deficiency case. The example excludes it expressly rather than applying paragraph (a) as a universal template. The responsible technical reviewer must establish the actual track class, geometry and applicable provisions.
Identify the rule and the explanatory edition separately
The FRA listing records the 2026 manual as issued on 4 March 2026, with an effective date of 1 July 2026. Its release material explains that the guidance does not itself have the force of law. Preserve those dates and that role separately from the governing regulatory text.
The completed archive description now identifies raw MCO, chord length, reference basis, derived deviation and the selected comparison provision as separate fields. A later qualified reviewer can reproduce the stated arithmetic and assess the missing premises. A bare alignment number could not support that review.