Low voltage power from a UK live only light switch

I am trying to find a way to power an Arduino and RF transceiver behind a UK light switch which just has the live and switched live present. This will run on 5v and 500mA and it would be in series with the light.

I know this is possible as LightwaveRF light switches draw some power from the lighting circuit to power the electronics inside the switch. They seem to keep a low current running through the light, but not enough to light it up. The unit also dims the light but will only go up to about 90% of the normal brightness as the switch is using some of the power.

As far as I can tell it seems to involve a digital AC dimmer and an AC-DC convertor. I have been searching for a way to do this for a while when I came across this site. Any help would be greatly appreciated.

 

So in the back box, as well as the switch innards, you're hoping to get an Arduino (in a case, of course), an RF transceiver and a 5v power supply.

I don't think so.

This forum is awash with posts from people who will assure you that that design does cause some lamps to glow, or periodically flash. It's a daft design.

Buy an off-the-shelf system to do the remote control and/or automation which you want?

But if you want the fun of DIYing, do not try and cram that stuff into a back box and do not try and power it by putting it in series with the lamp and passing a current even when "off".

 

The 5 ~ 20 mA system has been used in industry for a long time and what you are looking at is something similar. However 5 mA drain may well not light any lamp but 500 mA is going a bit OTT.

Many RF light switches use small batteries the humble door bell has a button cell and lasts for years this points out how little power is used. I would expect bulb manufacturers do allow a small amount of current to leak through the bulb to dissipate any voltage build up due to capacitive or inductive linking with parallel cables.

Put a standard ohm meter on a CFL and often you get infinity reading only as the voltage builds up with current flow. In the old days of tungsten you may have got away with it but today with LED and CFL it's not really going to work.

To draw even 5 mA that's nearly 12W at 230 volt it would require a double system one with lights off with a voltage transformer and one with lights one using a current transformer. Your looking at 2.5W to get 500 mA at 5 volt so with a 3W LED bulb it would be clearly impossible.

Do the maths you state 90% so volt drop of 23 volts so with a 60W bulb you can get 6W = 1200 mW as the bulb wattage reduces so does the power available so allowing for a 3W bulb means 60 mW maximum.

If you are using RF why do you need any switch wire to the lamp? Building the unit in the ceiling rose will mean you have a spare wire you can use as a neutral.

The Arduino is cheap compared with full PLC but compared with a simple PIC it still expensive so firms can make complete units for less than it costs you using an Arduino. I found the same with my project to keep my brewing temperature constant I could buy whole thing for £9 so was just not worth making.

 

Thank you for taking time to reply. Just to clarify I am not sure what the current draw would be and I would be keeping it as simple and as low as possible. I know the units can be battery powered easily but I am looking for a solution that can be installed and not maintained.

I am aware that squeezing an arduino and rf and a switch would be mad. I am looking at an arduino micro or nano and rf unit soldered to a custom built pcb control board and replacing the traditional switch with a blanking plate with a capacitive touch sensor. With one or more touch sensitive areas.

These boards would only use about 2cm x 3cm x 0.5cm leaving enough space for the high voltage side. I am looking at getting this professionally designed if the basic idea works and it will be fully cased in, double insulated and low voltage separate from high voltage.

I really just need to know at this stage what extra components would give the desired result and how they would be wired so that I can look at creating a sample and then make it into a pcb.

This may be a little off topic as it is a diy site and this is slightly out of the realms of most diy projects but I would really appreciate any help.

 

Just to add 3.3 volts would be enough for the low power side and I will look into the current requirements.

 

After a quick internet search it seems I was way out with the current. It seems to be up to 50mA and I am not against redesigning the controller to a lower power version.

 

I presume you mean 1.2W, not 12W?

Kind Regards, John

 

Any device that is dependent on obtaining its power from a low current through the lamp is going to be problematic. Both in design and in operation.

The current through the lamp could be very low and with incandescent lamps the filament will not glow. But modern lamps such as LEDs and CFLs will, as Eric has said, either continuously glow dimly ( LEDs ) or store the energy until there is enough to light the lamp for a fraction of a second, this repeats causing the lamp to flash ( CFLs ).

Obtaining power this way was an easy way to obtain power for dimmers and "intelligent" switches without having to re-wire to bring a Neutral to the switch. This method worked with incandescent lamps and this lead the general public ( often mislead by commission based sales people ) that it will work with any type of lamp. The simple truth is that it will not work with all types of lamp..

 

As eric has implied, it would be impractical to get even 50mA from an in-line supply relying on current through the lighting load. Even with incandescent lamps, that 11.5W would probably result in appreciable light output (when they were 'off'), but with CFLs or LEDs, that current could result in their being almost at 'normal brightness' (when they were 'off'). 'A few mA', at most, is the maximum you could hope to get by this method, and even then you could end up with glowing or flashing 'off' CFLs/LEDs.

If you wish to DIY this project using the sort of technology you propose, I cannot see how you are going to be able to avoid getting a neutral to the switch location (or, alternatively, siting the electronics where there was a neutral, somewhere remote from the actual 'switch').

Kind Regards, John

 

50 mA at 5 volts is 250 mW which can be obtained from just over 1 mA at 230 volts if the convertor was 100% efficient. 80% is a more realistic efficiency figure. That is while the lamp is OFF. To power the unit while the lamp was ON would require a means to store power. This could be a rechargable battery or super capacitor charged while the lamp was OFF .
The ratio of ON - OFF timing would need to be determined in order to size the storage necessary to have adequate power while the lamp is ON.

 

Yes, that could be engineered, but (for the very reason you mention) it's not how the power for the electronics of 'two wire' dimmers etc. is usually generated, is it? The electronics of such devices only require significant power when the lamp is ON, and I've never heard of a dimmer stopping working because they had been left ON for too long, as a consequence of which a battery or capacitor had completely discharged, have you? For that reason, I'd always assumed (perhaps wrongly), that they utilised (some of the) in-line current when the load was ON.

Kind Regards, John

 

Thank you all so much for your feedback. I am aware that there are limitations with this method and the type of light bulbs that can be used, the LightwaveRF version only uses dimmable bulbs.

I would like to know the best components to use for this and the best way to connect them and take it from there. A capacitor would be useful so that power is not lost as the light changes from off to on (or 1% to 90% light).

I am looking to make this in addition to a version for light switches that do have neutral connected (which is a much easier circuit).

 

I would think it would need two separate power supplies feeding the same capacitor or battery.

With light off a voltage transformer across the switch contacts and with light on a current transformer in series with the light.

So if we assume a 5W LED lamp and work on that. Many will run on 150 volt so we will consider it having 46 mA so at 70 volt that's 3.2 watts maximum power with light on so at 5 volt 650 mA is absolute tops with no loses.

Now with same light off. For the lamp not to light we need the voltage down to at least 50 volts this is also the voltage which would be considered safe to earth so at 150 volt it takes full 5W so at 50 volt around 1.8W that equals around 36 mA at 50 volt so around 350 mA at 5 volt.

So I would say by time you include losses the maximum draw is around 300 mA at 5 volt. This is selecting exactly the correct transformer so in real terms if you need over 100 mA at 5 volt then the whole idea is dead.

But the whole idea seems daft anyway as using RF so can build into ceiling rose instead.

 

The two wire dimmer does not need continuous power to operate a wireless receiver so it can "steal" a volt or two for a power supply source.

Some wireless ON-OFF remote control switches use a remnance relay which only requires a short pulse of power to latch it ON or OFF and thus this can operate for many months on a button cell battery. The wireless receiver is designed to use the minimum possible current. Many turn on the receiver for a few milliseconds about ten times a second. If there is no wireless signal the receiver turns off. Thus the receiver is only on for about 1/10 of the time which extends battery life by almost 10 times.

 

I suppose that may be the ideal (albeit rather complicated and maybe bulky), but would have the advantage of requiring a minimal sized capacitor/battery (since light is always either on or off!).

Kind Regards, John