Why a Transformer Trips the Breaker When Nothing Is Wrong
Switch on a transformer at the wrong instant of the mains waveform and it will draw a current spike ten to twenty times its rated value, decaying over a few cycles. The breaker trips, the fuse blows, or the contactor welds — and the transformer is blamed for a fault that is actually a physics problem.
Inrush current is a saturation phenomenon, and it can be calculated, predicted and mitigated. This guide explains the mechanism, gives a working calculation method, and lists the countermeasures that actually work in production.
The Mechanism
When a transformer is de-energised, the core retains remanent flux — typically 20–40 % of Bsat for grain-oriented steel, higher for some amorphous grades, lower for gapped cores. At the moment of switch-on, the flux must satisfy Faraday’s law:
B(t) = Brem + (1/N·Ae) · ∫v(t)dt
If the switch closes at voltage zero crossing, the integral of voltage over the first half cycle is maximum — the flux swings from +Brem to +Brem + 2Bnominal. If that sum exceeds Bsat, the core saturates, the magnetising inductance collapses, and the primary behaves almost like a short circuit limited only by the winding resistance and the source impedance.
Worst case: switching at the zero crossing with maximum remanence in the same direction. That case is typically 2–3× the nominal peak flux, which is why transformers are designed at 75–80 % of Bsat and still saturate on inrush.
Estimating the Peak
A practical estimate for the first-cycle peak, ignoring decay:
Ipeak ≈ (2 · Vrms · √2) / (ω · Lair) when the core is saturated
where Lair is the air-core inductance of the winding — the inductance with the core effectively removed. Because Lair is small, the resulting current is large. For a 10 kVA, 230 V transformer with Lair of 1 mH, the first-peak estimate is over 100 A.
For a faster field estimate: inrush peak is commonly 8–20× rated current, decaying with a time constant of 2–10 cycles; the first cycle carries the most energy.
| Core type | Typical inrush (multiple of rated) | Notes |
|---|---|---|
| Grain-oriented silicon steel, ungapped | 10–15× | High remanence, low Lair |
| Amorphous metal | 15–25× | Higher remanence and lower flux density limit |
| Gapped steel core | 5–8× | Gap reduces remanence and raises Lair |
| Ferrite HF transformer with active pre-charge | <3× | Soft-start through the converter, not the mains |
Mitigation That Works
1. Point-on-wave (controlled) switching
Close the contactor at the voltage peak rather than the zero crossing, so the flux integral over the first half cycle is near zero. This is the most effective method: it can reduce inrush to 1.5–2× rated current, but it requires knowledge of the remanent flux, which is not directly measurable in a cheap device. Practical implementations either assume a remanence polarity from the last switch-off, or accept a partial reduction.
2. Pre-charge resistor or NTC
A resistor in series for the first 100–300 ms limits the peak and is then bypassed by a contactor. Simple, reliable, adds cost and a failure mode (welded bypass contactor). For transformers above ~30 kVA this is the industry default.
3. Design changes
- Reduce design flux density to 70–75 % of Bsat: costs 5–10 % more core material, cuts inrush meaningfully.
- Add a small air gap to the core: lowers remanence and raises Lair; the gap also reduces permeability, so turns must increase. Common for transformers exposed to DC bias.
- Increase Lair by using a taller winding window or fewer layers — a modest but free improvement.
- Choose a core with lower remanence: gapped ferrite or powder cores have far lower remanence than grain-oriented steel.
4. Protection coordination
- Use a breaker with a magnetic trip setting above the predicted inrush, or a D-curve device.
- Fuses must be selected with time-current curves that tolerate 10–20× rated for 10 ms.
- Do not use an RC snubber to “fix” inrush — it does not change the flux integral.
What Does Not Work
- Soft-start on the load side does nothing for the transformer’s own magnetising inrush.
- Adding a capacitor in parallel can create a resonant transient that is worse than the original problem.
- Relying on the breaker’s thermal element is a guarantee of nuisance trips: the magnetic element responds in the same few milliseconds as the inrush peak.
Measurement
Verify with a current probe and an oscilloscope triggered on the switch-on event, capturing at least 10 cycles. Record the peak, the decay time constant and the switch-on phase angle. Repeat with the voltage at 0°, 90° and 180° relative to the previous switch-off — the variation between those cases is the real design window, and it is often a factor of five.
Cores Selected to Control Inrush
Inrush behaviour is a core property as much as a winding property. Shaanxi SiRui Technology supplies gapped ferrite and powder cores with controlled remanence for designs that must tolerate DC bias or frequent switching, plus nanocrystalline cores for low no-load loss where inrush is managed by point-on-wave switching. If inrush is a constraint in your design, tell us the switching scheme and we will propose a core-and-gap combination that meets it.
Related reading: Air gap design guide · UPS and energy-storage transformer design · Low-frequency transformer design
About Shaanxi SiRui Technology Co., Ltd.
Gapped and ungapped magnetic cores, isolation transformers and custom wound assemblies for power distribution and energy storage.
Siri Yuan, Sales Manager · Email: sales@sirui-tech.com · Web: www.sirui-tech.com
No. 2555, Chanba Avenue, Chanba Ecological Zone, Xi’an, Shaanxi Province