Why can't the circuit breaker interrupt this circuit?

Question

I have a 3-phase input in my home. There is 127 V between any of the phases and neutral, and 220 V between any pair of phases. The electrical appliances have 127 V, and the wirings are such that the air conditioners (that uses higher currents) of 3 rooms are connected to different phases.

I replaced one of them, and the new one needs 220V. I used the phase of the old one and another phase used for the lights and sockets of the same room (see picture).

It is working fine for some years, except that if I turn off the circuit breaker of one of the phases of the room (that used for lights), I have all kinds of unstabilities like blinking lights and beeping sounds from appliances. I need to turn off both S and T to properly turn off the energy of the room.

If only phase T is off, the sockets keep live. A voltimeter measures 127 V between live and neutral in the sockets. But an incandescent test lamp shines weakly as the voltage were smaller than that.

I suppose that somehow the phase S partially feeds T through the air conditioner wirings, but on the other hand, if the 2 phases were connected inside it, there would have a short circuit. There is a remote control for the air conditioner, and I don't turn it on during the experiment.

Answer 1

^{simulate this circuit – Schematic created using CircuitLab}

Figure 1. (a) The equivalent circuit. What you've got. (b) What you should have.

• When SW1 (your circuit breaker) is open you have created a potential divider between the S phase and neutral, N.
• The voltage on the "dead" T phase will depend on the ratio of the air-con and lights and sockets impedances (resistances).
• The T phase is no longer dead and the only protection is the impedance of the air-con unit.
• There is danger of electric shock if anyone is working on T after isolating that phase by SW1.

I suppose that somehow the phase S partially feeds T through the air conditioner wirings, ...

Correct.

... but on the other hand, if the 2 phases were connected inside it, there would have a short circuit.

No. A short circuit usually means a zero or very low resistance. Yours has the impedance / resistance of whatever is switched on in the air-con.

There is a remote control for the air conditioner, and I don't turn it on during the experiment.

Not relevant.

---

Your setup is not good. The air-con should be on a separate circuit to the others and should have a two-phase circuit breaker so that it can be isolated properly. Don't mess. Don't risk safety.

Answer 2

I suppose that somehow the phase S partially feeds T through the air conditioner wirings

That is most likely what is happening. The air conditioning unit should have a 2-pole breaker and should not be sharing a breaker with a line-neutral load.

if the 2 phases were connected inside it, there would have a short circuit.

No, there is a circuit that has a relatively low impedance compared to the lighting circuit.