Electrical Puzzle

[TheChair]

A few weeks ago my immersion heater took to occasional cutting my power. Turn off the heater and restore the circuit breaker, wait a while, turn the heater on and all was fine.
Then the situation got worse, in that, as soon as the heater was turned on, the house power cut.
Today I replaced the element and thermostat, and the problem is still there. Further investigation revealed the following:
The heater is connected to a control unit where it goes through a 20A switch, and thence to a 25A overcurrent switch (excuse the terminology, I'm no expert). From there it goes to a 40A switch in the main control unit. In the main control unit is an RCD, and it's this that triggers as son as the heater is connected - even if the contacts are open.
Heater element measures 14 Ohms across the terminals, which seems about right for a 4KW device. There is no measurement between either of the element terminals and earth.
I'm completely stumped.

[The Builder]

If you've replaced both heater and element. then the fault must lie in the switching/wiring.

An RCD will operate on EITHER a leak from LIVE to earth or a leak from NEUTRAL to earth.
That's why it'll operate even when the connection to LIVE is open.

I would try an insulation resistance reading ... testing live to earth, and neutral to earth.
You'll need an electrician's tester such as a "Megger"

[youngmafbog]

Yes, the RCD measures current out and current in, get a difference and it trips.

RCD's are primarily to protect you, the mcbs are to protect the circuit.

You will have a leak to earth somewhere but it may wel not be measurable as it only takes a small leak to trip it (to protect you in case you are the leak to earth). I have a Fluke but a Megger or similar will do. The problem is my Fluke is about £700 so get a Sparky in to test it.

[Ryzen]

Try disconnecting the Line and Neutral connections that feed the immersion. If the problem persists then the control is at fault.
An unrelated issue would be the protection at the distribution board i.e. 40a is too high for a potential 16.5a load..what size cable feeds the immersion circuit.

[satprof]

Sounds to me as if the RCD has found a problem or is itself faulty. With lots of water about, immersion heaters, especially oldish ones might easily have developed such a fault. Even if the immersion is switched ‘off’, that says little unless double-pole switches are used.

With the dangers inherent in water and electricity, I’d recommend getting an electrician to investigate ASAP.

[Hymie]

Having replaced a number of immersion heaters, I can advise that the normal failure route is a build up of lime-scale on the heating element, resulting in reduced heat dissipation to the surrounding water. The excessive temperature of the heating element eventually causes the metal wall of the heating element to split – resulting in water reaching the mains electrical parts of the heater. The electrical current flowing through the water to earth, then causes the RCD to trip.

To maximise the life of the immersion heater, install the longest length the tank will accommodate – this will result in the lowest power dissipation per unit length of the heating element.

[TheChair]

Thanks for all the suggestions. In case anyone is interested, we finally found the cause: nearby is a disused solar tank, the live of which had been disconnected by someone in the past, but the earth and neutral were still attached to the same busses as the immersion heater. The connection box at that end had leaked during a recent rainstorm.