The old coil came out of the window. That does not prove a replacement will go back in safely. Dismantling may relax a winding, move a core packet or remove insulation. Survey the rebuilt assembly in its final defined condition before assessing the insertion path.
A fit assessment should follow the winding through three positions: as manufactured, during insertion and after final clamping. Each position can have a different limiting clearance. Reducing the problem to copper area divided by window area conceals these constraints.
Establish the retained geometry first
Survey window width and height at identified levels, core-limb size, packet offsets and the faces that locate the coil. Keep local deviations rather than replacing the window with one best-fit rectangle. The narrowest relevant passage may govern insertion, while a different location governs the final electrical clearance.
Define the winding dimensions in its stated condition. Is the diameter measured before drying, after drying or under final compression? Is the insulation build included? A dimension without that condition can be internally accurate and still unsuitable for the assembly chain.
| Dimension-chain term | Measurement condition | Decision affected |
|---|---|---|
| Core envelope | Final packet and clamp arrangement | Available coil passage |
| Coil inner envelope | After the specified manufacturing treatment | Core-to-winding interface |
| Outer winding envelope | Including leads and insulation projections | Window and adjacent-phase fit |
| Insertion movement | Planned orientation and handling allowance | Temporary minimum passage |
| Final spacers and barriers | Installed geometry with tolerances | Dielectric and cooling spaces |
| Axial compression | Defined final assembly load or condition | End clearances and alignment |
Calculate the clearance that actually matters
For an illustrative one-dimensional passage check, define C as available space in millimetres minus the combined coil, insulation and assembly envelopes in millimetres. State whether the difference is a diametral clearance or a one-side radial clearance before applying it to the drawing. The sign convention and measuring datum must be the same for every term. Evaluate C at each controlling location rather than averaging the readings.
In this assumed example, if the available passage is 300 mm and the complete moving envelope is 294 mm, the arithmetic difference is 6 mm. That number is not an acceptable installation allowance by itself. Measurement uncertainty, tilt, local projections and the qualified handling method still have to be considered. A worst-case sum is appropriate where terms can move together adversely; a statistical combination needs defensible independence assumptions.
Allocate the full insertion allowance
Assume measured passage W = 300 mm, moving envelope E = 294 mm and centring offset x = 2.0 mm toward one wall. The tight-side clearance is C = (W − E)/2 − x = 1.0 mm. Assume independent standard uncertainties uW = 0.20 mm, uE = 0.30 mm and ux = 0.20 mm, representing dimensional measurement rather than production tolerances. Propagation gives uC = √[(uW/2)² + (uE/2)² + ux²] = 0.269 mm [1]. With an assumed k = 2 and an assumed project rule C − 2uC ≥ 0.50 mm, the lower review value is 0.461 mm: reject this evidence despite nominal 1.0 mm clearance. Reducing the verified offset to 1.5 mm raises the review value to 0.961 mm. Confirm every controlling section, tilt and insertion state; this one-dimensional result cannot qualify the whole route.
Keep fit separate from design equivalence
A smaller coil may solve insertion while changing leakage impedance, conductor temperature or insulation stress. Moving a barrier can change a dielectric interface. Treat these as design changes. After rebuilding, compare the identified ratio, winding resistance and relevant winding-deformation evidence with appropriate baselines [2]; passing a dimensional screen alone cannot establish electrical equivalence.
References
- B. N. Taylor and C. E. Kuyatt, NIST Technical Note 1297, Guidelines for Evaluating and Expressing the Uncertainty of NIST Measurement Results, 1994, sections 5 and 6.
- U.S. Bureau of Reclamation, FIST Volume 3-31, Transformer Diagnostics, June 2003.