Chenfan Electric
EN
Send drawing

ARTICLETechnical article

The same finished core can report a building factor of 1.15 or 1.24.

  • Chenfan Power
Back to technical library
CHENFAN ELECTRIC
Transformer Core Technical Note
Core Loss Engineering
Building Factor
The same finished core can report a building factor of 1.15 or 1.24.

Nothing in the core changed. Only the denominator did.

CHENFAN ELECTRIC
SAME FINISHED CORE One measured loss. Two denominators. The reported ratio changes although the physical core does not. BUILDING FACTOR BF = measured core loss reference loss SST REFERENCE Single sheet tester value BF ≈ 1.15 EPSTEIN-EQUIV. ≈ 0.925 × SST BF ≈ 1.24 0.925 Same core Same measured watts Different reference ≈ 8.1% change in reported BF
Same physical transformer core and the same measured core loss. Changing the denominator from SST to an Epstein-equivalent reference changes the reported building factor by about 8.1%.

For domain-refined GOES, IEC 60404-8-7:2020 uses the single sheet tester as reference and applies a 0.925 factor to convert SST loss at 1.7 T into an Epstein-equivalent value. Divide one measured core loss by each figure and BF shifts by about 8%.

SST-based BF
BF = measured core loss ÷ SST reference loss
Epstein-equivalent BF
BF = measured core loss ÷ (0.925 × SST reference loss)
That is why BF is one of the most useful loss figures in a transformer-core RFQ — and one of the easiest to misread.

Why it beats W/kg alone

Mill certificate W/kg is measured on samples magnetised along the rolling direction under controlled sinusoidal flux. Epstein strips are usually stress-relief annealed first, so cutting stress is removed before anything is measured.

A three-phase core does not work that way. At T-joints the flux rotates and leaves the rolling direction. At step-lap overlaps part of it crosses between laminations. Then add cut-edge stress, burr and clamping load.

CHENFAN ELECTRIC
WHY FINISHED-CORE LOSS RISES The joint is where laboratory simplicity disappears. Flux rotation Flux turns away from the rolling direction at the joint. Step-lap overlap Flux transfers between adjacent laminations. Cut-edge stress Shearing changes local magnetic behaviour near the edge. Burr Raised edges can damage insulation and add local eddy loss. Clamping pressure Mechanical stress can raise loss and exciting current.
The finished core adds magnetic rotation, inter-laminar flux transfer and mechanical effects that are not represented by sheet-loss W/kg alone.

Published measurements have shown local joint loss reaching about twice the nominal limb loss in some areas.

BF prices all of this in one number. W/kg prices none of it.

What BF will not tell you

  1. The reference. SST or Epstein-equivalent? Actual coil value or grade maximum? Each answer gives a different BF.
  2. The operating point. BF at 1.5 T and BF at 1.7 T are different numbers for the same core.
  3. The actual watts. Highly oriented Hi-B steel tends to show a higher BF because it loses more of its advantage when flux turns off-axis. A core with the higher BF can still have the lower total loss. Buy watts, not ratios.
  4. Exciting current. In one published comparison at constant overlap, moving from a mitred joint to a five-step step-lap reduced total loss by about 2–4.4%, while apparent power fell by roughly 31–37%. A loss-based BF barely registers that.
  5. Who caused it. BF mixes design and manufacturing. Joint share, window size, step-lap arrangement and overlap length all move it. Comparing BF across two different drawings says little about the two suppliers.
CHENFAN ELECTRIC
VALID SUPPLIER-TO-SUPPLIER COMPARISON Keep the basis fixed before comparing the ratio. Same drawing Identical geometry. Same GOES grade Same material basis. Same flux density Same excitation point. Same reference SST or Epstein defined. THEN COMPARE BF spread across consecutive cores Repeatability tells more about process control than one best result. Change the basis and BF stops being a supplier-to-supplier performance metric.
Drawing, steel grade, flux density and reference method must remain consistent before BF becomes a meaningful supplier comparison.

How to use it properly

  • Compare BF only on the same drawing, same grade and same flux density.
  • Ask each supplier to state the reference method and whether a conversion was applied.
  • Define whether the denominator uses actual measured coil loss or the grade maximum.
  • Request more than one excitation level.
  • Read BF together with actual watts and exciting current.
  • Look at the spread across consecutive cores, not the best single result.
Low average BF Useful, but not enough by itself.
Tight BF spread Stronger evidence of repeatable process control.
Actual watts Still the figure that determines real no-load loss.
A BF without its denominator is not a performance figure. If a core specification asks for one, it should also define exactly what the finished-core loss is divided by.
Technical references mentioned in this article
  • IEC 60404-8-7:2020 — grain-oriented electrical steel strip and sheet.
  • Materials, 2023, Vol. 16, Article 1648 — interpretation of GOES loss measurements.
  • Published three-phase transformer-core studies covering localized joint-loss contribution to building factor.
  • Cardiff University research on step-lap geometry, core loss and apparent power.
  • Published patent literature discussing Hi-B behaviour and building factor.
Chenfan Electric

Direct coordination

Start a conversation.

A drawing, a material requirement, or a project to discuss.

Send drawing