This technical article is in English. Only the surrounding interface is localized.
Technical NoteCHENFAN ELECTRIC
0.20 mm vs 0.23 mm GOES: When Is the Thinner Gauge Actually Worth the Premium?
Sheet loss is only the starting point. Joint geometry, stack build, material grade, frequency and process control decide how much of the laboratory advantage survives in the finished transformer core.
1.7 T / 50 HzDomain-refined GOESThree-phase test coresFive-step-lap
Two domain-refined GOES sheets at 1.7 T, 50 Hz: 0.20 mm: 0.65 W/kg and 0.23 mm: 0.67 W/kg. In identical three-phase, five-step-lap test cores reported in JFE Steel patent US 11,495,378, the 0.23 mm steel finished at 0.82 W/kg, while the 0.20 mm steel finished at 0.92 W/kg.
Context matters. These are small laboratory cores and one reported data set. The point is not that one gauge always wins; it is that a mill-certificate advantage does not automatically become a finished-core advantage.
Why this matters: the thinner sheet can look better on the mill certificate yet fail to deliver the same ranking after cutting, stacking and joint assembly.
What the thinner gauge buys
Classical eddy-current loss scales with thickness squared (t²). For 3% Si electrical steel at 1.7 T and 50 Hz, the eddy-current component is about 0.17 W/kg at 0.23 mm. Moving to 0.20 mm cuts that component by roughly 24%, equivalent to about 0.04 W/kg. The rest of the total loss does not scale the same way.
The gauge effect is real, but it applies to only part of the total specific loss. Hysteresis, excess loss, local flux rotation and assembly effects can dominate the finished-core result.
The patent also reports a comparison at equal B8 and a 2 mm lap: 0.73 vs 0.69 W/kg on the sheet, and 0.92 vs 0.86 W/kg in the core. In that test point, the gauge advantage survived assembly.
Other levers can buy the same watts
Gauge is only one lever in the loss budget. Joint design and material grade can move the final result by a similar amount.
Lap length
0.90 W/kg
0.23 mm with 3 mm lap in the cited test
Shorter lap
0.96 W/kg
0.23 mm with 1 mm lap in the cited test
Thin gauge / short lap
0.91 W/kg
0.20 mm with 1 mm lap
Grade selection
0.70 W/kg
One European mill lists 0.20 mm and 0.23 mm grades at the same maximum
In the quoted data, increasing lap length on 0.23 mm material nearly matched the 0.20 mm / 1 mm-lap result.
Rotational loss is another hidden variable
The patent links opening-gap behavior to sheet performance under elliptical magnetization. That matters because standard mill certificates generally characterize one-dimensional magnetic loss, while local flux at joints and corners can rotate.
Where 0.20 mm tends to pay back
More likely to justify the premium
60 Hz systems: the eddy-current term is 44% larger than at 50 Hz, so the potential gauge saving is larger too.
High design flux density combined with aggressive no-load-loss capitalization.
A hard loss cap where the alternative is more steel, a larger core or more conductor.
Voltage distortion / harmonics: harmonic eddy-current loss rises approximately with frequency squared.
Often difficult to justify
Low no-load-loss capitalization.
A better 0.23 mm grade already reaches the target.
Joint optimization can achieve the same finished-core loss at lower total cost.
Process capability is not stable enough to exploit the thinner sheet consistently.
What changes in cutting and stacking
A thinner gauge is not only a material purchase. It changes the manufacturing workload and the sensitivity of the finished core to process variation.
More parts, lower stiffness and a larger burr-to-thickness ratio all increase the importance of flatness, alignment and handling discipline.
At equal finished stack height, 0.20 mm requires roughly 15% more laminations. That means more cutting, more handling and more joints to align. Bending stiffness drops by about a third because it scales with t³, so flatness and joint alignment need tighter control.
A 0.02 mm burr is already 10% of a 0.20 mm lamination thickness. Typical stacking factor can also be slightly lower; one catalog basis gives 95.0% vs 95.5%. If gross core section is held constant, the effective iron section changes and operating flux density can rise, consuming part of the theoretical loss gain.
Procurement implication: thin-gauge property stability is harder to hold. Do not evaluate only the guaranteed maximum. Ask for coil-to-coil spread, actual test distribution and the supplier’s process capability on the grade you are buying.
How I would evaluate 0.20 mm vs 0.23 mm
Price the watt, not the tonne. Calculate material premium plus extra processing cost per finished-core watt saved, then compare it with your no-load-loss capitalization.
Benchmark the 0.23 mm alternatives. Check whether a better grade, lower design B or a different lap geometry reaches the same target more economically.
Normalize the test basis. IEC 60404-8-7:2020 rates domain-refined grades by SST with a 0.925 factor, while conventional Hi-B grades may be specified using Epstein values. Do not compare unlike test bases as if they were identical.
Ask for finished-core evidence. The most useful comparison is a model-core or finished-core result on geometry close to your own design.
Built one transformer-core design in both gauges? The useful question is not the mill-certificate gap. It is how much of that gap survived cutting, stacking, joint assembly and the finished-core test.
Source note: Numerical examples in this article are based on the user-supplied reference to JFE Steel patent US 11,495,378 and the other catalog / standard references stated in the original draft. Small-core patent data should be treated as design evidence, not as a universal production guarantee.
