CRGO lamination permeability and B–H curve: how to interpret the data

If you buy CRGO lamination stacks or sign off transformer core designs, you probably spend more time than you want looking at B–H curves and “µ” tables. The basics are clear. The tricky part is turning datasheet curves into purchasing decisions and practical margins.

This guide keeps the theory short and stays close to what actually changes when you choose one lamination stack over another.

1. Where the B–H curve on the CRGO datasheet really comes from

Most CRGO B–H and permeability numbers you see in lamination offers come from:

• Strip samples tested in an Epstein frame according to IEC / JIS / ASTM standards
• Longitudinal (rolling) direction only
• Stress-relief annealed before measurement
• One or two fixed flux densities, usually 1.5 T and/or 1.7 T, at 50 or 60 Hz

POSCO, JFE, Nippon Steel and others state exactly this in their catalogs: core loss and induction are measured after stress-relief annealing, mainly along the rolling direction, and typically quoted as W15/50 or W17/50 (loss at 1.5 T or 1.7 T, 50 Hz).

So the “smooth” B–H curve you see is:

• One-dimensional (no corners, no joints)
• Perfect grain alignment
• No punching burrs
• No clamping pressure, no tank stresses
• No gaps except the idealized Epstein joints

Suitable for comparing steels. Not the same as your stacked core.

2. CRGO permeability: material µ vs stack µ

Design tools usually talk about material relative permeability µr or initial permeability. Datasheets show either:

• µ at a given H (e.g., at 800 A/m)
• Or an “effective” µ between two points on the B–H curve

Hi-B grades can show µ values well above 30 000 in the rolling direction.

But what you actually build is a stack:

• Each sheet carries an insulating coating
• You have step-lap or mitred joints
• There are air gaps between packets
• Punching and bending add stress
• Stacking factor settles somewhere around 95–97 %, sometimes less if burr control is poor

That means the effective µ of the lamination stack is always lower than the material µ. How much lower depends on:

• Coating thickness and consistency
• Stack pressure
• Joint design (step-lap vs butt)
• Cutting and annealing practice
• Whether grain direction is respected at every limb and yoke

If you compare suppliers only on catalog µ, you are comparing something you will never actually see in operation.

3. How to actually read the B–H curve when you choose CRGO lamination stacks

Engineers know the B–H curve is just B versus H with hysteresis. The question here is: which parts of that curve should drive your lamination purchase?

Use this as a quick reading order.

3.1 Check the test point and notation first

• W15/50 = core loss at 1.5 T, 50 Hz
• W17/50 = core loss at 1.7 T, 50 Hz

If one supplier quotes W15/50 and another W17/50, or mixes 50 Hz and 60 Hz, you cannot compare their curves directly. Decide on one reference condition (often 1.5 T, 50 Hz for distribution transformers) and ask everyone to supply data for that point.

Also check:

• Whether values are “maximum guaranteed” or “typical”
• Whether the curve is before or after stress-relief anneal
• Whether it is longitudinal, transverse, or a mix of both directions

Without this, the prettiest B–H plot tells you very little.

3.2 Align the curve with your actual operating B

Most modern Hi-B CRGO grades operate around 1.7–1.9 T in the rolling direction, with core losses around 0.7–1.0 W/kg at 1.5 T, 50 Hz for thinner gauges (0.23–0.27 mm).

Your design might be at:

• 1.5–1.6 T for conservative distribution designs
• 1.7–1.8 T in more compact power transformers
• Local peaks higher at joints

When you look at a B–H curve:

1. Mark your nominal B on the curve.
2. See what H the material needs at that point.
3. Convert that H to magnetizing current and compare against your no-load current budget.

If your operating B sits on the very steep part of the curve, you are betting on tight manufacturing control. Some projects can accept that bet. Many utility specs cannot.

3.3 Watch the hysteresis loop width, not just single µ numbers

The area inside the B–H loop ties directly to hysteresis loss. Larger area, higher core loss at the same B and frequency.

Two steels can have similar µ at 1.7 T but very different loop shapes:

• Narrow loop: lower hysteresis loss, lower no-load loss
• Wide loop: more loss, more heating for the same flux

When you only see µ or a few loss numbers, ask the supplier for:

• A family of B–H loops at different peak B
• Or at least loss vs B curves, not just two fixed points

It is the shape that tells you about behavior during inrush, over-excitation and off-frequency operation, not one permeability figure.

4. Typical CRGO datasheet numbers vs stack reality

Here is a compact way to read the common numbers buyers and engineers debate over.

Table 1 – Reading CRGO B–H and permeability data for lamination stack decisions

*Ranges are indicative only and must be checked against the actual mill datasheet and your own design rules.

5. Why your measured B–H curve never matches the brochure

Even with the same steel, your measured magnetizing current or core loss will drift away from the “official” B–H curves. Main reasons:

1. Stress from punching and bending Grain-oriented steel is very sensitive to mechanical stress; punching, bending and even clamping change domain structures and reduce permeability.
2. Loss of domain refinement benefits Domain-refined CRGO shows lower loss and higher permeability, but repeated stress-relief annealing and rough handling can erase part of that benefit.
3. Anisotropy and grain direction errors Magnetic properties in the transverse direction are much worse than along rolling; turning a lamination the wrong way in a limb or yoke can spoil stack µ in that region.
4. Joint and gap design Step-lap joints reduce local saturation and loss, but only if lap step, overlap length and cutting tolerance are respected. Poor control opens the B–H loop locally and creates hot spots.
5. Coating and stacking factor Extra-thick coating or burrs lower stacking factor and introduce more effective air gap. That shifts your whole operating point to higher H for the same B.

If you never see vendor test reports on actual lamination stacks, only on bare steel, you are missing the most important part.

6. A simple workflow: engineer + purchasing read the same B–H data

You do not need a complex routine. A short checklist that both engineering and purchasing can use is usually enough.

Step 1 – Lock the reference condition

• Pick one reference: for example P1.5/50 and B–H curve up to 1.8 T at 50 Hz
• Ask every supplier to provide data in that exact form, with test method and standard noted

This removes half the confusion.

Step 2 – Plot your design point on each B–H curve

• Put your nominal B and over-excitation B (say 110–120 % voltage) on the vendor curve
• Note the corresponding H and estimate magnetizing current
• Flag any steel where your maximum B is already very close to saturation knee

Purchasing does not need to do the math; they just need a simple “OK / tight / risky” tag from the design team.

Step 3 – Compare loss curves across the operating band

Instead of only P1.5/50, ask for loss versus B up to your maximum flux. Then, for each candidate steel:

• Check loss at nominal B
• Check loss at over-excitation B
• Ask whether these values are “guaranteed maximum” or “typical”

Sometimes a steel with slightly higher datasheet loss at 1.5 T behaves better in the 1.6–1.7 T band where your core actually runs.

Step 4 – Ask for stack-level test results

For at least one reference core size, ask the lamination supplier to provide:

• No-load loss and magnetizing current at rated voltage
• Measured stacking factor
• Photos or drawings of the actual step-lap pattern and limb build

This tells you more about their punching, deburring and assembly than any isolated B–H curve.

Step 5 – Freeze a “data sheet for stacks”, not just for steel

Once you choose a vendor, capture in your internal spec:

• Grade and thickness
• Target P1.5/50 and P1.7/50 limits
• Minimum stacking factor
• Joint type and cutting tolerances
• Required stack-level loss and current for one or two reference designs

Then the purchasing team can run future RFQs against this spec without re-doing the magnetic homework each time.

7. FAQ: B–H curves, permeability and CRGO lamination stacks