CRGO lamination edge condition: shear, laser, and the impact on losses

1. Why edge condition quietly overrides your CRGO datasheet

You already trust the CRGO mill data: grade, thickness, core loss at 1.7 T / 50 Hz, polarization.

Then you cut it. Then the numbers change.

Cutting, joining, stress relief annealing, and stacking all modify the steel near the edges. Local hysteresis and eddy-current loss rise around the cut, so the real machine almost always shows higher iron loss than a model that assumes “ideal” material.

Two mechanisms matter for edge condition on CRGO-Laminierungen:

1. Mechanical damage and residual stress
   • Shear slitting / punching introduces a plastic zone: work hardening, residual stress, grain refinement and shear bands right at the edge.
   • Domain walls see this as a hostile region. They pin, jump, and dissipate more energy per cycle.
2. Electrical bridges
   • Burrs that pierce coating link laminations together.
   • Those bridges create extra eddy-current paths, both radial and along the stack.

In controlled tests, artificial burrs shorting many Lamellen have taken a small transformer core and almost doubled its total loss at high flux. That’s not a subtle tweak. It’s your no-load loss guarantee walking away.

So edge condition is less “finish detail” and more “hidden material grade upgrade/downgrade knob.”

2. Sheared CRGO lamination edges – what’s really happening

Most CRGO transformer cores are still produced from sheared or punched sheets, not fully laser-cut. For good reasons.

2.1 Microstructure around a sheared edge

Near a sheared edge, several zones show up under EBSD and nano-indentation: roll-over, burnished shear, fracture and burr. Each has different hardness and dislocation density compared to the bulk.

Rough picture for CRGO:

• 0–0.1 mm from edge – severe plastic deformation, ultrafine grains and shear bands, higher hardness.
• Out to ~0.3–0.5 mm – residual stress dominates, still harder than bulk.
• Past that – the steel gradually returns to “datasheet” behavior.

None of this shows in the mill’s core-loss certificate. It’s all added by your slitting and blanking line.

2.2 Burr height, coating damage, and interlaminar shorts

You’ll see three recurring numbers in specs and papers:

• “Keep burr height < 20% of sheet thickness” as a general process guideline for 0.3 mm CRGO, based on shear-slitting studies.
• “Max burr height 0.03 mm” in many electrical steel and magnet lamination specifications.

Once burrs get tall enough to bite through the inorganic coating, we move from “extra hysteresis” into “interlaminar short” territory. Models and experiments both show that these bridges can dramatically increase local eddy-current loss.

In a classic experiment with artificial burrs on distribution transformer cores, fully shorting groups of laminations:

• Total core loss at 1.8 T rose by almost 100%,
• Localized loss above 50 W/kg around the burrs was recorded.

Real cores rarely reach that worst case, but the direction is clear: burr height × burr continuity × coating damage = how much trouble you bought.

2.3 How much extra loss does shearing add, realistically?

It’s messy to quantify, but some patterns keep repeating:

• Cutting and punching alone can add on the order of 10–30% to iron loss versus models that ignore cutting damage.
• Slitting / shearing with tight clearance and sharp tools can keep the damage zone narrow, so the global impact stays in the lower end of that band.
• Poorly controlled shear with large, continuous burrs turns that into a different game: you’re no longer just degrading local permeability; you’re inserting extra loss components not in any EP standard.

So by the time a “M**H” CRGO sheet becomes an assembled transformer core, the original W/kg number is only a starting point. Edge condition decides how much of that advantage survives.

2.4 What purchasing can actually specify on sheared edges

If your lamination drawings only say “CRGO M0H, 0.23 mm, cut to size”, you’re funding experiments, not a process.

Typical contract-level points that bring edge condition under control:

• Max burr height
  • ≤ 0.02–0.03 mm on both edges, measured with a stylus gauge or microscope over defined length.
  • No continuous burr over more than, say, 20 mm without a gap.
• Cutting method and tooling
  • Coil slitting: specified clearance window and max slitting speed for each thickness.
  • Blank/punch: carbide dies for CRGO, defined re-sharpening interval.
• Coating integrity near edge
  • No visible flaking at edges after shearing.
  • Agreed sampling rate for cross-section inspection (e.g., etched microsection once per X tons).
• Stacking side control
  • Define which edge faces the flux in the core window and require burr to point away from high-flux regions, or be removed.

These are boring to negotiate, but much cheaper than a 6–8% no-load loss overrun discovered after tanking.

3. Laser-cut CRGO laminations – not always the hero, not always the villain

“Laser cutting = clean, burr-free edges, so losses should be lower.” That sounds nice. It’s only half true.

There are really two different uses of lasers on CRGO:

1. Laser cutting the laminations (shape)
2. Laser scribing / domain refinement (microscopic stress lines to reduce loss)

The physics and the outcome aren’t the same at all.

3.1 Laser cutting: heat-affected zone and degraded properties

Instead of a shear band, laser cutting gives you a wärmebeeinflusste Zone (HAZ):

• Local melting, re-solidification, and tempering
• Residual tensile stress, microstructural changes near the edge
• Coating damage or re-oxidation if parameters are wrong

Studies on electrical steels (mostly non-oriented, but the mechanisms carry over) consistently show:

• Increase in coercive field
• Decrease in effective permeability
• Higher specific iron loss near the cut.

In one recent experimental+simulation study, considering cutting damage in the model vs ignoring it led to iron losses roughly 30% higher once realistic cutting was included.

And detailed loss measurements on motors built from laser-cut laminations typically show higher magnetic losses than those using carefully punched sheets, if you hold material and geometry constant.

So laser edges are geometrically neat, but magnetically stressed.

3.2 Shear vs laser – which is worse for loss?

It depends where you stand on this triangle:

• Burr height / short-circuit risk (shear can be bad here)
• HAZ width and severity (laser takes this slot)
• Flux density and frequency in the application

Recent work on high-grade electrical steel shows:

• At 50 Hz and moderate flux (around 1.0 T), laser-cut samples often show higher ΔP than mechanically sheared samples.
• At higher flux densities (e.g., 1.5 T) and with very tightly optimised laser parameters, the ranking can invert for some steels.

You could say:

• Shear – more mechanical damage and burrs, but no HAZ.
• Laser – excellent geometric freedom, but thermal damage and often higher loss unless process is heavily tuned.

Für CRGO transformer cores running near 1.7 T at 50 Hz, the safest practical rule so far:

Prefer sheared / punched CRGO with strict burr control and proven core-loss performance. Use laser cutting for prototypes, specials, or when geometry forces you, but ask for data, not promises.

3.3 Laser scribing for domain-refined CRGO – different game

Now the confusing part: Laseranreißen is also a laser process, but with the opposite goal.

Instead of cutting edges, the laser writes shallow lines into the surface, deliberately introducing small stress regions to subdivide large domains. When parameters are in the sweet spot, domain-refined CRGO shows around 5–15% lower core loss than the same grade without scribing, in the 0.23–0.30 mm range.

Two important caveats for buyers:

• Domain refinement doesn’t magically cancel poor cutting. A beautifully scribed sheet can still have ugly HAZ or burr damage from later operations.
• Scribing is typically done at the mill or a specialized facility, vor your lamination supplier cuts the pieces.

So a reasonable spec stack is:

1. Ask for domain-refined CRGO if the loss budget is tight.
2. Still insist on burr and edge condition limits in your lamination PO.
3. Verify with actual core-loss tests on stacked laminations, not just Epstein strips from the coil.

4. Shear vs laser edge condition – quick comparison

Very approximate, intended as a design + purchasing guide, not a replacement for local testing.

5. Practical checklist for engineers and buyers

You’re finalizing a lamination PO or a transformer tender. What do you actually write?

5.1 Drawing / specification hints

Consider building clauses along these lines (adapt the numbers to your standards):

• Material and condition
  • “CRGO grade X (e.g., HiB), thickness 0.23 / 0.27 mm, domain-refined where available. Material certificates with Epstein W/kg at 1.7 T / 50 Hz attached.”
• Cutting process declaration
  • “Supplier shall declare cutting process for laminations (shear/punch/laser/EDM). Any process change requires written approval.”
• Grathöhenbegrenzung
  • “Max burr height 0.02 mm for t ≤ 0.27 mm; 0.03 mm for t > 0.27 mm. Burr height measured per ISO XXXX over at least 10 locations per coil or batch.”
• Coating and shorts
  • “No continuous edge burr longer than 20 mm that penetrates coating. Coating must appear intact when viewed at 50× magnification at edges.”
• Laser cutting control, if used
  • “For laser-cut CRGO laminations, supplier shall provide process window (power, speed, assist gas) and evidence of HAZ width < 0.1 mm with metallographic cross-sections.”
• Core loss verification
  • “Random sample cores assembled from production laminations shall meet specified no-load loss at 1.7 T / 50 Hz within +X%. Testing to IEC 60076 series at agreed tap.”

This turns “nice edge” into contractual reality instead of a vague promise.

5.2 Debugging: when measured no-load loss is too high

If a finished transformer shows 5–10% higher no-load loss than design:

1. Look at the edges before anything else
   • Simple stereoscope check for burr continuity and coating damage.
2. Check lamination source vs model
   • Was the model based on WEDM samples or punched/sheared ones?
   • If you used “ideal” material in FEA, add 10–30% allowance and compare again.
3. Take a ring-core or Epstein sample from finished laminations
   • Compare loss vs mill certificate; any big drift suggests cutting or annealing damage.
4. Check stacking and clamping
   • Over-clamping can force burrs together and increase the probability of interlaminar shorts.
5. Run a local heating or flux-mapping check if available
   • Hot spots often line up neatly with burr clusters or badly cut corners.

Not every overrun is an edge problem, but it’s often one of the cheapest things to fix next batch.

6. FAQ – quick answers for buyers and engineers