How to Size an IEC Circuit Breaker: 4 Checks Before You Buy

The sizing rule, stripped to its bones

Size for the design current, then prove the breaker can interrupt the fault current at that exact point in the system. Everything else is a refinement. Ib ≤ In ≤ Iz is not a suggestion — it's the minimum you need before the breaker is legal, and the only part you cannot argue your way out of with a building inspector.

Per IEC 60364-4-43 §433.1, three inequalities have to hold at the same time: the breaker rating In must be at least the design current Ib, must not exceed the cable's current-carrying capacity Iz, and the breaker's guaranteed tripping current I2 must stay under 1.45 × Iz. Get those right and the breaker protects the cable. Get them wrong and you've either built a nuisance-trip machine or a fire risk. That's the short version of how to size an IEC circuit breaker — the rest of this article is the three checks that turn those inequalities into a safe installation.

But that only covers overload. The other half of sizing is short-circuit: the breaker's rated ultimate breaking capacity Icu at the installation point must exceed the prospective fault current there. And the service breaking capacity Ics — the level it can interrupt and remain usable afterwards — matters more than most spec sheets let on. If Ics is only 50% of Icu (fairly common on lower-end molded case breakers), your breaker survives one bolted fault and then becomes decorative.

Then there's the trip curve. That's where the actual engineering lives, and where most site mistakes happen.

Why I check this on every emergency call

I'm a senior applications engineer at VoltShield. I've handled 600+ rush breaker replacements and panel upgrades in 11 years, with a good chunk of those being same-day turnarounds for automotive plants, cold-chain warehouses, and one very memorable data hall. That pace trains you to check four things before you touch anything else.

In March 2024, a client called at 6:40pm on a Friday. A stamped-metal line had been down for four hours. It looked like a fault — main ACB trip, no upstream alarm, no obvious smoking gun. What it actually was: a downstream 250A MCCB with a C-curve magnetic trip sitting right inside the inrush envelope of two 90kW VFDs starting within a 400ms window. Every third or fourth start, the breaker read that inrush as a fault. Nobody had done a selectivity study because nobody had needed one — until the line grew by a third and added those drives. Six hours of lost production ran them around $40k in throughput. The fix was a 400A frame with an electronic trip unit set to a 4×In short-time delay, plus a proper zone-selective interlock. The breaker wasn't the problem. The breaker was just the messenger.

We didn't have a formal selectivity review process for panels under 400A until that incident. Now it's on the first page of our site survey template. Should have been there from the start.

The four checks, in the order I run them

1. Design current — and it's not the nameplate

Ib is the design current of the circuit, not the sum of what the equipment stickers say. Add the diversity factor. A MCC feeding six 30kW pumps typically doesn't draw 150A continuous, even if the arithmetic says it does. IEC 60364-1 Annex A gives you the method; the short version is that a diverse load is lower than the arithmetic sum, and sizing to the arithmetic sum is the most common way people overspend on copper.

The other mistake I see: sizing to the next standard breaker frame. If Ib comes out at 62A, you do not round up to 80A "for headroom" unless you've also upsized the cable. Round In to the next rating, then round Iz up separately by choosing the cable. Those are independent choices, and treating them as one choice is how you end up with a breaker that's legal but a cable that's marginal.

2. Cable ampacity — the derating most people skip

Iz is not the number in the catalog. IEC 60364-5-52 gives the method; the installation conditions give the reality. A 4-core XLPE copper cable rated at 100A in free air is a completely different animal once it's in a tray with six other loaded cables, running through a 45°C plant room.

Typical correction factors I use most often:

• Ambient temperature, 40°C reference: ~0.91 at 45°C, ~0.82 at 50°C for XLPE. Check the manufacturer's table — it varies by insulation.
• Grouping, six circuits in one tray: ~0.75 to 0.80 of free-air rating. Some standards use 0.78 for six touching cables.
• Buried in a duct bank with thermal resistivity above 2.5 K·m/W: another 0.85 to 0.90 on top.

Stack those and a "100A" cable can honestly carry 65A. That shifts Ib ≤ In ≤ Iz by a full frame size. I've seen projects where the breaker was correctly sized for the calculated load but the cable had never been derated — the breaker was right, the installation was still unsafe, and nobody noticed until a thermal scan. Cable derating is where the quiet problems live.

3. Breaking capacity — and don't forget cascading

Icu is the headline number. The relevant number is the prospective fault current at the terminals of that specific breaker, which drops as you move downstream. A 50kA fault at the main switchboard can be 12kA at a distribution board three floors up, depending on cable impedance.

Cascading (back-up protection) lets a downstream breaker with a lower Icu sit behind an upstream breaker that will clear the fault first. It's permitted under IEC 60947-2, but only if the manufacturer has published a tested combination. If you're relying on cascading and you don't have the coordination table in hand, you don't have cascading — you have a hope. This is the part where the spec sheet alone will not save you.

4. Trip curve — where nuisance trips actually come from

Standard IEC 60898-1 curves: B trips at 3–5×In, C at 5–10×In, D at 10–20×In. Industrial IEC 60947-2 breakers add K (8–14×In) and Z (2–3×In) for specific loads. My working rules:

• Motors: D curve or a motor-rated MCCB, sized to locked-rotor current rather than FLA. A 45kW motor at 6.5× FLC inrush will walk straight through a C-curve.
• Transformers: first inrush peak typically runs 8–12× rated current for 10–100ms. If you put a standard MCB on a transformer, you're going to have a bad time unless it's D-curve or better.
• VFDs: high-frequency inrush and leakage. Type B curve, and check the VFD manual — many specify the maximum upstream breaker frame, not the minimum.
• Long cable runs: high fault loop impedance means lower fault current at the far end, which means the magnetic trip might not operate fast enough. TN system fault protection per IEC 60364-4-41 §411.4 requires Ia × Zs ≤ U0. Long runs fail that quietly, and the fix is usually an RCD, not a bigger breaker.

Selectivity between series breakers is where the numbers get honest. Current selectivity generally needs roughly a 2:1 ratio of ratings to be reliable — a 100A upstream breaker will usually ride through a fault on a 50A downstream, but not on a 63A one, on the magnetic trip alone. Time selectivity (adjustable short-time delay upstream) gets you further up the range. Energy selectivity, which is manufacturer-tested, is what actually works on modern electronic trip units.

And here's the most frustrating part of selectivity work: the coordination study says the breakers will discriminate, then a real fault trips three of them and you're standing in a dark plant room explaining to a plant manager why the math didn't match reality. You'd think tested coordination tables would prevent that. They prevent some of it. Cable impedance, arc resistance, and the actual fault path cover the rest.

Where arc flash changes the sizing decision

Arc flash doesn't change Ib ≤ In ≤ Iz. It changes what you're allowed to put in the room, and how fast the breaker needs to open.

Per IEEE 1584-2018, incident energy is calculated at a specific working distance for a specific equipment configuration and electrode gap. The result is in cal/cm², and that number drives the PPE category in NFPA 70E. There's a hard threshold at 1.2 cal/cm² — the arc flash boundary — below which you technically don't need arc-rated PPE for the task. Above it, you do.

The sizing levers that actually move incident energy:

• Faster clearing time. A breaker that opens in 3 cycles instead of 12 cuts incident energy roughly fourfold at the same fault level. This is the single biggest lever, and it's why zone-selective interlock and maintenance-mode settings exist.
• Current-limiting breakers. Above a certain fault current, a current-limiting MCCB opens before the first peak and drops the let-through energy.
• An arc flash mitigation circuit breaker with a maintenance switch. This is the option most clients don't ask about until after an incident. A VoltShield arc flash mitigation circuit breaker with a maintenance setting drops the instantaneous trip threshold during maintenance windows — when workers are actually in front of the equipment and the risk is highest — then returns to normal settings outside those windows. It's not a replacement for a calculated arc flash study. It's what you install after you've had one done.
• Remote racking and arc-rated enclosures. Sizing doesn't help you if the door is open at 480V.

One caveat that gets skipped constantly: IEEE 1584-2018 replaced the older 1584-2002 tables, and the incident energy numbers moved — sometimes higher, often significantly, especially on 480V systems with larger electrode gaps. If your arc flash labels are dated before 2018 and haven't been recalculated, they may be understating the hazard. That's not a breaker problem. It's a label problem, and it's the same size.

RCBO: protection layer, not a breaker substitute

An RCBO breaker (residual current breaker with overcurrent protection) combines an MCB and an RCD in one DIN-rail device. Per IEC 61009-1, it carries both the overcurrent rating and the residual current rating on the same label — something like "C16, 30mA, Type A." It does not add ampacity. It adds a second trip criterion: residual current to earth.

Where I specify them:

• Socket outlets up to 32A, per IEC 60364-4-41 §411.3.3 — 30mA is the common requirement in most national adoptions, TT systems included.
• Final circuits in wet or outdoor areas where touch voltage can't be met by disconnection time alone.
• Circuits where the earth fault loop is too long to guarantee Ia × Zs ≤ U0 within the required time. This is the most common legitimate industrial use, and it's usually cheaper than pulling bigger cable.

Where I push back:

Not every circuit wants an RCBO. Circuits feeding large VFD banks, welding equipment, or anything with substantial PE leakage current will nuisance-trip a 30mA device on startup. On those circuits, a Type AC RCBO is the wrong tool — the leakage isn't sinusoidal, and the device may either trip without a real fault or miss a genuinely DC-biased fault. Type F handles mixed-frequency residual current, and Type B handles smooth DC. Both cost more and require a supplier who will actually sell you one with short notice.

When that's the case, a VoltShield residual current circuit breaker on a separate RCD (not combined) at a higher threshold — 300mA for fire protection, or 100mA for equipment protection — plus properly rated Type B devices on the VFD feeders, is usually the honest answer. Less convenient than an all-in-one RCBO, more expensive up front. Also the configuration that doesn't trip the line on a Monday morning.

One more thing on three-phase industrial circuit breaker applications: RCBOs are a selectivity liability. A 30mA instantaneous device upstream of another 30mA instantaneous device has no time discrimination — both will trip. If you need selectivity, use time-delayed (Type S) devices, or separate RCDs with different thresholds, and verify the discrimination in writing. I have mixed feelings about telling clients to buy two devices when one could theoretically work — on one hand it's more SKUs to stock, on the other hand it's the difference between a tripped final circuit and a tripped production line.

When you should not size this yourself

I sell breakers, so take this with the appropriate discount: there are installations where calling a consultant is the right move, and buying a breaker off a datasheet is not.

The cases that trigger that recommendation from me:

• Prospective fault current is unknown or unverified at the installation point. Not "probably around 25kA." Measured, or calculated from the utility's declared fault level.
• Parallel cables or parallel breakers. Derating, load sharing, and selectivity math stop being intuitive.
• Generator-supplied systems with transfer switching, where fault current depends on the generator's subtransient reactance and drops dramatically in island mode. A breaker sized for grid fault level can be wildly over- or under-protected on generator.
• IT (unearthed) systems. Different fault path, different protection logic, residual current monitoring rather than residual current tripping.
• Any installation that needs an arc flash label under NFPA 70E. The label requires incident energy calculation, which requires a short-circuit study, which requires system data you probably don't have on a Friday afternoon.

If you're in one of those, my job is to get you the right breaker by Monday, not to pretend I did the study. That's a different job, and doing it wrong is how people get hurt.

For everything else — a standard IEC switchboard, known load, known cable, known fault level — the four checks above take about forty minutes and cover the vast majority of industrial low-voltage installations. Start with Ib ≤ In ≤ Iz. Prove the Icu. Match the curve. Check the arc flash. In that order. Then buy the breaker.