Geyser tripping the 30 mA RCCB after 15 minutes
A customer's bathroom has a 15-litre, 2 kW, 230 V storage geyser fed from a 16 A MCB with a 30 mA RCCB at the main board. For the last week the RCCB trips, but only after the geyser has been heating for 10-15 minutes; it never trips when the geyser is switched off at the wall switch. The previous electrician replaced the RCCB, then the thermostat; nothing changed. At the geyser terminals you measure 228 V. The element reads 26.5 ohm end to end. With the geyser isolated and its wiring disconnected, a 500 V insulation tester shows 0.6 Mohm between an element terminal and the metal tank. The tank-to-earth path, measured from the earth pin of the plug to the earth bar, is 1.2 ohm, and the geyser shares its earth rod with the kitchen. The family has a function in two days and the owner asks you to 'just disconnect the earth wire from the socket, it will stop tripping' so the geyser can be used meanwhile.
1. Interpret the three readings (26.5 ohm element, 0.6 Mohm insulation, 1.2 ohm earth continuity) and explain why the RCCB trips only after 10-15 minutes, and why replacing the RCCB and the thermostat changed nothing.
Read the worked answer
Element, 26.5 ohm: P = V squared / R = 52900 / 26.5 = about 2.0 kW, and I = 8.7 A. The element is continuous and correctly rated, so the fault is not an open element.
Insulation, 0.6 Mohm at 500 V: this fails the usual acceptance value of at least 1 Mohm for a Class I heating appliance and points to moisture inside the magnesium-oxide filler or a crack in the sheath. Cold leakage is only about 230 / 600000 = 0.38 mA, which is why the RCCB holds when the unit is cold.
Why the delay: as the element heats, moisture migrates and the filler's resistance collapses, and a hairline crack opens with thermal expansion. Leakage to the tank and earth climbs past the 15-30 mA band in which a 30 mA RCCB must trip. Switch off, the element cools, resistance recovers, and the RCCB resets - exactly the reported pattern.
Why the parts swap failed: the RCCB was doing its job, not malfunctioning. The thermostat is a temperature switch in the live line; it cannot stop leakage from the element sheath to the earthed tank.
1.2 ohm earth path: higher than a sound protective conductor (should be well under 1 ohm). Not the cause of tripping, but a loose or corroded joint that must be corrected, since it raises touch voltage during a real fault.
2. Respond to the request to disconnect the earth wire, and set out the repair and verification plan you would follow before returning the geyser.
Read the worked answer
The request must be refused, politely and in writing on the job card. The tank is metal, the bathroom is wet, and the user stands on a wet floor with low body resistance. With the earth removed the tank sits at line potential through the leaking sheath: even 30 mA through the chest is enough to cause ventricular fibrillation, and there is no low-impedance path to divert current and operate the protective device. Disconnecting the protective conductor also defeats the appliance's Class I design and is a breach of the safety code, so the shop carries the liability. Offer instead the repair today, or leaving the geyser switched off at the isolator.
Repair and verification plan
- Isolate, close the water inlet, drain, remove the element and inspect the tank for scale and corrosion.
- Fit a genuine 2 kW / 230 V element with a new gasket; do not reuse the old gasket. Clean the earth stud and tank flange.
- Refill with water completely before energising - never dry-fire a new element.
- Insulation test at 500 V, element terminal to tank: at least 1 Mohm; a new element normally shows tens of Mohm.
- Earth continuity plug pin to earth bar: under 1 ohm after cleaning; recommend a second rod if it stays high.
- Run 20-30 minutes with the RCCB in circuit, then re-test hot - it must still hold at least 1 Mohm and must not trip.
- Check the thermostat and thermal cut-out cut in and out correctly, and record all readings on the job card.
3. The geyser circuit uses 1.5 mm2 copper in conduit on a 16 A MCB. Is this coordination acceptable for an 8.7 A load? What would you change, and where does the 30 mA RCCB fit in the requirement?
Read the worked answer
Coordination check: 1.5 mm2 copper in conduit carries roughly 11-13 A at 40 degrees C ambient. A 16 A MCB will pass 1.13 times its rating (about 18 A) indefinitely and will not protect the cable in a sustained overload, so the arrangement is not correctly coordinated even though the steady load, 8.7 A, is below the cable rating. The load itself does not overload the cable in normal use; the problem is that the protective device cannot back up the cable.
Two acceptable fixes:
- Keep 1.5 mm2 and fit a 10 A MCB. A resistive 8.7 A load is fine, and the device holds 11.3 A without tripping; inrush is negligible for a heater.
- Preferred: rewire the run in 2.5 mm2 and keep a 16 A MCB, giving margin for a future 3 kW element and for derating in a hot conduit.
RCCB: a 30 mA RCCB is the standard additional protection for bathroom circuits and socket outlets, and it is what saved this family. But note the logic clearly for the customer - the RCCB detected the element's deteriorating insulation. It does not replace earthing, it complements it: the earth keeps the metal at touch-safe potential, the RCCB disconnects the supply. Neither is a substitute for repairing the element; the RCCB will keep tripping until the element is replaced.