Short answer: 30 mA is the sensitivity for protecting people; 100 mA is the sensitivity for protecting continuity of supply and equipment. A 30 mA RCD is the additional-protection device required on socket-outlets and outdoor final circuits in most IEC-based codes, because it limits the duration of a shock to a level a healthy adult can usually survive. A 100 mA RCD does not do that job at that current: it is used upstream, as a time-delayed or selective device, or on circuits where the risk is earth fault and fire rather than a person touching a live part. If a single RCD has to do both jobs, it cannot — you need two devices, in the right order.
The confusion comes from treating sensitivity as a quality grade: “more sensitive is better.” It is not a grade, it is a purpose. A 30 mA device on a long, leaky circuit will trip for reasons that have nothing to do with a fault; a 100 mA device on a socket circuit will not trip fast enough at the current that matters. This article sets out which device goes where, how to keep them from tripping each other, and what changes on 110–127 V and 60 Hz systems. We write as an independent trading company — we are not a manufacturer, and we are not an appointed distributor of any brand.
What IΔn actually measures
Residual current is the difference between the current going out on the line conductor and the current coming back on the neutral. In a healthy circuit that difference is zero; when insulation fails or a person touches a live part, some current finds another path back to earth and the difference appears. The RCD measures it with a toroidal core and knows nothing about overload, which is why an RCD never replaces a breaker and why the RCBO exists at all.
Two numbers matter on the label. IΔn is the rated residual operating current — the sensitivity. The trip time at IΔn is bounded, and at higher residual currents it gets shorter; for personnel protection the useful property is that at a few times IΔn the device opens in tens of milliseconds, not seconds. A 30 mA device therefore limits the let-go and ventricular-fibrillation risk window; a 100 mA device, even a fast one, is already past the current range where a shock is reliably survivable.
| IΔn | Typical purpose | Where it belongs | Watch out for |
|---|---|---|---|
| 10 mA | Extra sensitivity for wet or medical locations | Special locations, single circuits | Nuisance trips on any circuit with real capacitance to earth |
| 30 mA | Additional protection against electric shock | Final circuits: socket-outlets, outdoor equipment, bathrooms, portable tools | Accumulated leakage on long circuits; split the board rather than raise the setting |
| 100 mA | Earth-fault protection and discrimination step | Upstream of 30 mA groups; circuits without accessible live parts | Not a shock-protection device at this setting |
| 300 mA | Earth-fault / fire-hazard protection on distribution | Incoming or sub-main level | Must be selective in time, or it will take the whole board down |
| 500 mA | Large distribution, high leakage networks | Origin of large installations | Legal in some codes, insufficient in others — check the local rule |
IEC 60364-4-41 §411.3.3 is the clause most codes copy: additional protection by an RCD with IΔn not exceeding 30 mA is required for socket-outlets with a rated current up to 32 A, and for mobile equipment up to 32 A used outdoors. Note what the clause does not say: that 30 mA protects everything upstream, or that the device may be shared across an unlimited number of circuits.
The discrimination problem, and the 3:1 rule
Two RCDs in series with the same sensitivity have no idea which one should trip. The faster one wins, and on a shared neutral or common earth path that is often the incoming device — the whole board goes dark for a fault on one socket. Selectivity is created deliberately, in two ways: by sensitivity (a 100 mA upstream device is far less likely to see a small fault current than the 30 mA downstream one) and by time (a selective, “S”-type device delays intentionally so the downstream device clears first).
The working rule of thumb is a ratio of at least 3:1 in sensitivity between the upstream and downstream device, with the upstream one time-delayed. In practice that means 30 mA on the final circuits and 100 mA S-type upstream, or 30 mA downstream of a 300 mA S-type at the sub-main. Two 30 mA devices in series is not selectivity, it is a coin toss.
| Arrangement | Selective? | What happens on a 40 mA earth fault |
|---|---|---|
| 30 mA → 30 mA | No | Either device may trip; no predictable outcome |
| 30 mA downstream, 100 mA S-type upstream | Yes | Downstream sees 40 mA and trips; upstream holds |
| 30 mA downstream, 100 mA instantaneous upstream | Marginal | Downstream should win on speed, but impulse or surge currents can make it random |
| 30 mA downstream, 300 mA S-type upstream | Yes | Downstream trips, upstream remains closed as backup |
| Single 100 mA for everything | — | No shock protection on the final circuits at all |
The ratio is a rule of thumb, not a guarantee: leakage currents add vectorially and depend on cable length, filters, inverter capacitance and the number of electronic drivers on the circuit. The ratio sets the direction of the design; the manufacturer’s selectivity table sets the final answer.
Which device for which board position
The CHINT ranges map onto this logic fairly cleanly, and the differences matter when you are writing an order code rather than a blog post. Group the devices by what they contain: an RCCB has no overcurrent element, an RCBO has both, and the sensitivity options differ between the families.
| Device | What it is | Poles | IΔn options | Breaking capacity |
|---|---|---|---|---|
| NL1-63 | RCCB, IEC/EN 61008-1 (F and S variants per IEC/EN 62423) | 2P, 4P (manual also lists 1P+N, 3P+N) | 0.01 / 0.03 / 0.1 / 0.3 / 0.5 A | Inc 6 kA; 10 kA on the larger frames |
| NB1L | RCBO, IEC/EN 61009-1 | 1P+N, 2P, 3P+N, 4P | 30 / 100 / 300 mA | Icn 6 kA (some catalogue entries 6/10 kA), IΔm 6 kA |
| NB2LE | RCBO, IEC/EN 61009-1 | 1P+N | 0.03 / 0.1 / 0.3 A | Icn 6 kA, IΔm 3 kA, Uimp 4 kV |
| NXBLE-63 | RCBO, IEC/EN 61009-1 | 1P+N, 2P, 3P+N, 4P | 30 / 100 / 300 mA | 6 kA; NXBLE-63ZB adds self-test |
Two practical notes on ordering. First, the pole configuration is not cosmetic: a 1P+N device switches the neutral, a 2P device switches and protects both poles, and on a TT system or a 110–127 V two-wire supply that choice changes what is isolated during maintenance. Second, the 100 mA and 300 mA options are not “the same device with a different knob” — check whether the variant is the selective one, because an instantaneous 100 mA upstream of a 30 mA downstream does not give you the discrimination you drew on paper.
Type AC, A, F, S — and why the sensitivity is only half the story
Sensitivity tells you how much residual current trips the device. Type tells you what shape of residual current it can see at all. Type AC responds to sinusoidal alternating residual current only; Type A adds pulsating DC; Type F adds mixed-frequency residual current, which is what single-phase inverters and some variable-speed drives produce; Type B covers smooth DC, which is why EV chargers and PV arrays need a different conversation. The S suffix is not about waveform — it is the intentional time delay that makes selectivity possible.
Getting this wrong is common and expensive: a Type AC device on a circuit full of electronic drivers may fail to trip on a real DC-tainted fault. Sensitivity and type have to be specified together, and the exact ordering code decides which you receive — a certificate listing a family name does not tell you the variant. That is why we ask for the full code, not the series.
Leakage, long runs and nuisance tripping
A 30 mA RCD protects a limited amount of cable and equipment before its own standing leakage consumes the budget. Every metre of cable has capacitance to earth, every electronic driver has a filter with a small discharge current, and every inverter has capacitance on both sides. On a long circuit, or a circuit with thirty LED drivers, standing leakage can approach the trip threshold and the device trips with no fault present.
The wrong fix is to raise the sensitivity to 100 mA and call it solved — that removes the shock protection the circuit was built for. The right fixes, in order: split the circuit across more RCD groups; shorten the run or move the board closer; select equipment with lower earth leakage; and only then consider a different device on a circuit where shock protection is provided another way. On three-phase boards, keep the neutral with its own RCD group — borrowed neutrals create paths the RCD reads as faults.
110–127 V and 60 Hz sites
Lower voltage does not remove the shock risk: the current through a person falls somewhat, but body impedance falls too, and the 30 mA threshold is chosen for the physiology rather than for the system voltage. So 30 mA remains the personnel-protection setting on 120 V and 127 V systems. What changes is the rest of the board: devices must be rated for the system voltage, the neutral arrangement differs, and the local code may require the device at a different point in the installation.
Frequency matters less for the RCD itself, since it is a current-balance device, and more for what it coordinates with: leakage from filters rises with frequency, and 60 Hz systems often carry more electronic load per circuit. Specify 50/60 Hz rated devices and state the frequency on the enquiry — the ordering code may differ.
What to send with the enquiry
System voltage, frequency, phases and earthing arrangement (TN-S, TN-C-S, TT); the sensitivity and type you need, or the clause in the local code you are designing to; the number of outgoing circuits and whether the RCD is per-circuit or per-group; the rated current and breaking capacity at that board position; the ambient inside the enclosure; and the destination country, because the certificate requirement follows the market. With those six items an order code can be fixed rather than guessed, and the certificate quoted against that exact code.
FAQ
Can I use a 100 mA RCD instead of 30 mA to stop nuisance tripping?
Only if the circuit does not rely on that device for additional protection against electric shock. If it is a socket circuit or outdoor equipment, 30 mA is the requirement in IEC-based codes and raising it removes the protection. The correct fix for nuisance tripping is to split the circuits across more RCD groups or reduce standing leakage, not to desensitise the device.
Do I need a 30 mA RCD on a circuit protected by an RCBO?
The RCBO is the RCD and the breaker in one enclosure. If the RCBO is 30 mA, that circuit has its additional protection and needs nothing further. What it may still need is an upstream device for discrimination, so that a fault on this circuit does not take out the whole board.
Is 300 mA for fire protection or for people?
For fire and earth-fault protection on distribution. Several codes use a 300 mA setting at the origin or sub-main to detect leaking insulation before it becomes a fire, and it is paired with 30 mA devices downstream. It is not a personnel-protection setting.
Why does my 30 mA RCD trip when nothing is faulty?
Because standing leakage from cable capacitance, filters and electronic drivers is additive, and on a long or heavily loaded circuit it can approach the threshold. It can also be a borrowed neutral, a shared earth path, or moisture in an outdoor enclosure. Measure the leakage with a clamp meter before changing the device.
Does a 30 mA RCD protect against all electric shock?
No. It limits the duration and magnitude of earth-fault current through a person, which is what makes a shock survivable in most cases, but it does not protect against line-to-line contact, and it does not remove the need for insulation, barriers and correct earthing. It is additional protection, which is exactly what the code calls it.
Related: RCCB Type AC vs Type A · MCB vs RCCB vs RCBO · NB1L vs NB2LE vs NXBLE-63 · Type B RCD for EV chargers · How to size a distribution board · RCCB and RCBO range · FAQ · Request a quotation