0

I saw this question about measuring AC current without splitting wires. I understand why this is difficult and inherently imprecise.

Is it, however, easy to just detect whether an appliance is on or or off without splitting hot from neutral by the presence of current? For example something like 0.3A at 220V (~70W)?

I've heard about sensors in the past that have 2 hall sensors angled at 90 degrees to measure current (link), but here I want something more simple I think.

DennisFrett
  • 101
  • 2

2 Answers2

1

The problem is the same as in the linked question except that you want to detect current above a certain threshold rather than accurately measure it and display a reading. As you can see from my answer to the previous question, measuring the current in a two-core cable containing the feed and return wires is difficult and requires fairly precise alignment of the cable.

I've heard about sensors in the past that have 2 Hall sensors angled at 90 degrees to measure current ...

A link or reference would be nice if you can find it. Then we can discuss it.

enter image description here

Figure 1. The https://moderndevice.com/product/current-sensor/ device.

The linked article explains:

It theory you shouldn’t be able to sense a current from two AC carrying wires in close proximity. This is because the magnetic fields from the two, 180 degree out-of-phase, power conductors should cancel one another. If you’ve ever used an AC clamp (current) meter, you know that it is important to put only one conductor inside the clamp. Putting both conductors inside the clamp meter will result in a zero, or very low reading as the fields cancel.

The trick the Modern Device Current Sensor uses is getting the two separate hall elements as close as possible to a cables separate conductors, one sensor closer to one wire and the other sensor closer to the other (out of phase) conductor.

The sensors are mounted in anti-parallel (180°), not at 90°.

There is little benefit in this arrangement as the feed and return currents will be the same so you might as well clamp the PCB to the cable, rotate until you get the strongest signal (from either live or neutral) and then run some sort of calibration. Putting some hot-melt or other adhesive between the cable and PCB prior to calibration might be a good idea.

Transistor
  • 168,990
  • 12
  • 186
  • 385
  • I added the product in my original question. – DennisFrett Aug 15 '21 at 17:11
  • See the update. – Transistor Aug 15 '21 at 18:29
  • Or just use 3 sensors. But yes, if you're close up on a cable, one EMF should be detectably stronger than the other (unless some madman made the wires coaxial lol). – Harper - Reinstate Monica Aug 16 '21 at 00:02
  • @Harper, these are Hall effect sensors. They sense current, not EMF. I don't know why one would use three. – Transistor Aug 16 '21 at 06:06
  • And what kind of circuit would be needed to do this? A single hall sensor together with an opamp to amplify the signal? Or would a single hall sensor by itself be enough? – DennisFrett Aug 16 '21 at 09:30
  • The data provided on the link says that the output signal can go straight to the microcontroller's ADC. The signal is large enough that you shouldn't require further amplification. I see now that the PCB does have two Hall chips on it. The [schematic](https://moderndevice.com/wp-content/uploads/2012/11/MD_CurrentSensorSchematic.pdf) and PCB close-up shows that the board can be populated with SMD chips or 3-pin chips. There is a filter and peak detector with adjustable gain on the PCB. – Transistor Aug 16 '21 at 13:21
  • because 3 would eliminate having to rotate the sensors against the cable to find the strongest signal. Any orientation would give you a good differential signal between at least 2 of the sensors, assuming the load was not itself 3-phase. I was referring to ElectroMagnetic Fields. – Harper - Reinstate Monica Aug 16 '21 at 18:47
1

The simple answer is no.

The problem is that, assuming that all return current is in the neutral wire, the net current through a loop will be zero. The two currents, flowing in opposite directions, produce magnetic fields which overall cancel each other. As a matter of fact, GFCI mechanisms work by looking at the magnetic field from a pair of wires, and if it is NOT zero (since some of the current is flowing through ground rather than neutral) tripping a circuit breaker.

As you've discovered, it is possible to detect current, essentially by measuring the field nearer one conductor than the other - but this isn't going to be much help in rolling your own without a much experience.

WhatRoughBeast
  • 59,978
  • 2
  • 37
  • 97
  • The PCB in the OP's link has two Hall sensors in close enough proximity to the 2-core cable that it can get enough of a difference in field strength between the near and far wires to make a reading. – Transistor Aug 16 '21 at 13:24