Switching an AC circuit whilst also drawing power?

Hi all

I've seen online various devices that offer capacitive touch light switches, (for example: Touch Panel Light Switch). These are capable of taking power from the live wire of an AC circuit whilst also switching it, by allowing a small amount of power to bleed through past the switch - but not enough to illuminate the bulb as the voltage is so low.
In switches where there is a neutral connection available, this is pretty straightforward; the problem arises where that's not the case and only the live connection is available.

There's a transformer on the control circuit to lower the voltage to a suitable level for logic circuits, and this is connected to either side of the relay. Simplified, it's like this:

I'd like to be able to do the same so that I can do something similar with an Arduino as part of a control circuit, but I can't seem to find any guidance anywhere on how to do this. Looking for some advice please that will allow me to proceed without having to power the Arduino circuit from a completely separate source...

Note: I'm aware of the hazards of working with mains electricity, and the need for proper isolation etc.

You could use a rechargeable battery in the control circuit to power the electronics and the relay for the times when the light is on.
During the off times the control circuit would have to recharge the battery.
The size of the battery would be dependant on the the maximum on time of the light.

There has to be a current path to neutral, whether its a bulb filament, fluorescent transformer, or just a resistor. Note that in your drawing the relay contacts would create a short circuit across the control circuit and the relay would drop back out. There has to be enough resistance between relay and load to drop enough voltage to operate the control circuit. Maybe just a small inductor.

The touch switches you refer to are dimmers in the time-domain - there is always
some portion of the AC waveform during which they are off, allowing enough power
to be bled off. They are also likely to use triacs which need much less operating power
than a relay.

Thank you for the responses. I've since managed to get my hands on one of these commercial devices, and so can offer some clarifications.

Firstly, the circuit I've drawn above is incredibly simplified, and there is not a "resistance free" connection between bulb and relay as I have implied.
Secondly, it is indeed a mechanical relay that's performing the switching.

mauried has provided a very smart answer that I definitely hadn't thought of - the only problem is that due to space constraints that's not possible.

I believe that by introducing a small resistance immediately after the relay, I would be able to maintain enough voltage to power the control circuit. (As per klubfingers's comment above).
Time to get to the calculations...!

You could use earth as a return conductor, if it's available at the switch. In the US, it's quite common for an electronic light switch to connect between hot and earth, or ground, as we call it. My guess is that the same is true in the UK, where I think that you are.

If you try that, be aware that the local authorities may take a dim view of such an installation without appropriate regulatory approvals. If a residual current device, or ground fault interrupter, is in the upstream supply, it will see the earth current as a low-level fault. If you have other devices on the same circuit with noticeable earth leakage currents, it may cause the RCD to trip, or to behave as if it were overly sensitive.

In other words, doing this may be illegal where you are, and it might cause you operational problems, too.

[Edit: add this] US listing agencies limit the current that a light switch can deliver to earth to 0.5 mA. Not much, and, in the US, enough to allow the switch to use no more than 60 mW. Your locale may have similar rules. If you choose this route, be sure to identify and follow those rules.

As for whether it's a bad idea from a technical standpoint, I don't see a practical reason not to like it. If the device short circuits, it connects hot and earth. That's a bad thing, to be sure, but it's exactly the situation that the earthing conductor was made for - to carry a lot of current, and thus force the upstream breaker to trip, when the hot conductor contacts something that's earthed.

It's worth noting that recent editions of the most popular US electrical code, called the National Electrical Code, require that either neutral be installed, or provisions be made to easily install neutral, at switch locations for just this purpose. Nobody really likes the idea of intentionally putting current on the grounding conductor, and the code is trying to make that necessity disappear.

Firstly, the circuit I've drawn above is incredibly simplified, and there is not a "resistance free" connection between bulb and relay as I have implied.
Secondly, it is indeed a mechanical relay that's performing the switching.

Simplified or not, the AC voltage connection on the right is shown on the "dead" side of the load. It should moved to the right of the load where it says "AC neutral".
What does the circle with the cross in it represent ? Is that the touch switch ? If so,
it is the "load" and the control circuit ac line on the right should be to the right of the load.

tmd3:
You could use earth as a return conductor, if it's available at the switch. In the US, it's quite common for an electronic light switch to connect between hot and earth, or ground, as we call it. My guess is that the same is true in the UK, where I think that you are.

You are quite correct, I am indeed in the UK.
The problem with this idea is that RCD fuse panels are standard here, and so by feeding off via the earth connection once this is installed into a few switches will cause a trip unfortunately.

raschemmel:
Simplified or not, the AC voltage connection on the right is shown on the "dead" side of the load. It should moved to the right of the load where it says "AC neutral".
What does the circle with the cross in it represent ? Is that the touch switch ? If so,
it is the "load" and the control circuit ac line on the right should be to the right of the load.

The circle with a cross in it is a lightbulb; the touch sensor forms part of the simplified box I've proposed as a control circuit. I've redrawn the circuit here which allows more visibility.

Fundamentally I need to allow a connection at all times via the lower AC Live route, to ensure that there is a continual circuit through to the neutral side. By ensuring it's of a low enough current, the light will not be illuminated.
The transformer requires a minimum of 90V, although this could be ditched and replaced with a non-isolated method of conversion as the unit is sealed (only interaction is via capacitive touch).

I've got the DC logic and touch control side of things completed, I just need to continue to understand what I'm doing with the "AC bypass" part of the circuit. By doing something simple with an inductor as is just suggested above on a test bench, the logic circuit is powered successfully.

Simplified view of test (transformer circuit is incomplete, missing caps etc. but conveys the point)

Unfortunately using the only inductor I have to hand that's suitable (1μH) results in too low a voltage across the transformer when the relay is closed.
I also need to test with a CFL as I would guess that the current leakage will cause the dreaded flashing! (Noting that the commercial unit claims to be suitable for CFLs)

Just a thought. The exact same idea must be in use in many commercially available single pole illuminated switches...

CplSyx:
... in the UK. The problem with this idea is that RCD fuse panels are standard here ...

Not so common in the US. We like to say, "GFCI," for "ground-fault circuit interrupter," where you say, "RCD." They're required on receptacles outdoors, in garages, and close to water faucets, but generally not elsewhere, and they're typically not required for lighting. They're normally implemented in the receptacle itself, rather than in the panel. We get away with it, I think, because the mains are 120 volts, roughly half what yours run. We can electrocute our citizenry more often, and with fewer consequences.

Here's a link for a no-neutral-required wall-mounted combination occupancy sensor and wall switch: Danlers PIR. No schematic, but the usage instructions are telling:

  • Can't use it where lights may be on for more than 12 hours per day, so that the battery can have time to recharge
  • The load has to be able to pass a "parasitic current."

That suggests that something closely akin to the solution proposed by mauried, above, is in use in that device.

Danlers also makes a similar device, with the notation, "neutral required."

They also provide some kind of thing to raise the apparent load of a connected device, where the load is below some minimum level. They call it a "CAPLOAD." I suspect that it's a capacitor, rated for line-to-neutral AC voltage, that draws about 2 VA in that application. They reference "CAPLOADs" on pages related to devices called, "grid time lag switches - grid module," and "high frequency controllers - grid module." I don't recognize those terms, so most of what I write in this paragraph is pure conjecture.

[Edit: capitalization]

Thank you - the more digging I do the more devices I'm turning up that are commercially available doing very similar things; just not able to find much in terms of instructions or schematics! Clearly commercial secrets :wink:

There are two options that manufacturers seem to take; either by using a rechargeable battery so that when the "switch" is enabled the battery can charge, or by having a current draw to the logic circuit at all times.
For ease I'd prefer the latter but I will definitely look into the battery option.

In terms of the "capload", it appears that this is mainly used in dimmer circuits where the load isn't high enough - for example when using LED bulbs. I guess the same applied in the context of the PIR switch.