Why USD/Ton Can Hide the Real Cost of a Transformer Core
Most transformer core RFQs are still decided on USD/ton.
That unit prices the one variable in core manufacturing that moves in the wrong direction when process control is weak: kilograms.
Core weight is not an input. It is an output of slitting, cutting and stacking control. In our own production at Chenfan Electric, finished weight is treated that way, and it is quoted that way.
Stacking factor changes what the buyer actually pays for
A drawing fixes effective cross-sectional area, which is geometric area multiplied by stacking factor. Run 96.5% instead of 97.5% and you need roughly 1% more stack to reach the same effective area. More laminations, more mass, more loss, and on a fixed USD/ton, a larger invoice. The buyer pays extra for the weaker process.
If the drawing instead fixes stack thickness, the penalty lands elsewhere. Effective area drops, operating flux density sits above design, and no-load loss and exciting current both rise faster than linearly.
Branch 1: Effective area fixed
Lower stacking factor means more stack, more weight and a larger invoice at the same USD/ton.
Branch 2: Stack thickness fixed
Effective area falls, operating flux density rises, and no-load performance deteriorates.
Neither branch is visible in a USD/ton comparison.
Now the loss arithmetic
A 12-ton three-phase step-lap core, 0.23 mm Hi-B, material loss near 0.85 W/kg at 1.7 T / 50 Hz. Ideal material loss is about 10,200 W. A building factor of 1.12 gives roughly 11,400 W. A building factor of 1.18 gives about 12,000 W. Same drawing, same grade, same mill certificate, roughly 600 W apart. That spread between suppliers is not exotic.
Utilities capitalize no-load loss at roughly USD 4 to 10 per watt. So those 600 W carry USD 2,400 to 6,000 of evaluated cost. A USD 80/ton advantage on 12 tons is USD 960.
FAT is where it stops being theoretical
IEC 60076-1 allows +15% on a component loss only while total losses stay inside +10%, and under a tender with loss penalties the real headroom is close to zero. A core landing 4% over guarantee does not fail a standard. It fails a contract.
Recovery is the expensive part: re-stacking, re-testing, holding a tank. A re-stacked core is also not the same core, because alignment, joint geometry and residual stress have all changed.
So evaluate quotes this way
- Price per finished core, using the supplier’s own quoted finished weight.
- Guaranteed finished-core watts at stated flux density, frequency, waveform and temperature.
- Repeatability across several sets, not one showcase test report.