Sensor to detect when a light is turned on

[aurumdeus]

Hi,

I'm looking for something rather specific and I don't know if it exists or not.

I'm trying to plan a system that allows the lighting circuits in a house to be powered by a solar system and a small inverter when the battery has enough power, and will revert to a mains supply if the battery is empty. I have all this designed already.

The light bulbs will all be LED filament designed for mains AC to allow switching between the two.

When the inverter is turned on it still draws just under 5W even with no load. I'm hoping to find something that can detect when an LED light is turned on and send a signal to a relay that will turn the inverter on, that way it will only be powered up when a current is required.

Some inverters have an ECO mode where they will check for current draw every second or two, but that is no use since when the inverter is off the lights will simply draw from the mains.

If it was just normal filament bulbs then something like measuring resistance or continuity across live and neutral would work, but since the LED bulbs will be using a transformer and rectifier in each bulb I'm not sure this would work.

1. Does anyone have any ideas? Is this possible?

All I've been able to come up with so far is measuring current drawn through the circuit and having a relay turn the inverter on when anything is drawn.

2. I can't run the whole circuit (max probably 250W) through the coil in series, so would I need a micro controller connected to an ammeter to measure and activate the relay or is there a more direct way of doing this?

When a light is first turned on it would be powered by mains for a moment before the inverter kicks in and takes over. This might cause a tiny dark flash.

3. Could this be avoided, possibly by placing a capacitor across live and neutral (with a discharge resistor of course)?

Thank you in advance.

 

[Lucien Nunes]

You want to avoid inverter standby losses in two situations: When the mains is on, and when the mains is off but no lamps are switched on. The first is easy - put a relay with a mains-voltage coil in the control circuit of the inverter. The second situation requires two separate detection circuits. #1 uses a low-power source, perhaps ELV DC, to detect switch-on events, which then starts the inverter, disconnects itself and connects the inverter to the load. Then #2 monitors the load current and when it falls to zero, switches off the inverter and reconnects the lower-power detection circuit instead to detect the next switch-on.

The problem as you have spotted is that the SMPSU or capacitive dropper in an LED lamp may be difficult or impossible to detect e.g. with battery volts, or might cause the lamp to light or flash if anything more than a whiff of power is fed in by the detection circuit. You might get results with a resistor across the lamp, but it would have to be of high value to avoid excessive dissipation, and then you need sensitive detection and it might be fooled by leakage. 100KΩ would dissipate about 1/2W and might be acceptable.

If you are trying to save the odd watt here and there, such as inverter standby loss, then inverting to 230V AC and using 230W lamps is not the most efficient way to go. I don't know how practical it would be to convert all the fittings to 12V and run the entire circuit like that, permanently? Otherwise, separating the switching from the lighting power circuit is probably the most effective if you want to take control of the inverter.

BTW, if you have made the transfer switch (mains / inverter selection relay etc) then you must ensure it meets standards for electrical separation between supplies. Failure to do so could send mains voltage from the inverter back out into your supply and shock someone working to repair the interruption.

 

[Spoon]

I'm trying to plan a system that allows the lighting circuits in a house to be powered by a solar system and a small inverter when the battery has enough power, and will revert to a mains supply if the battery is empty.

Also power cuts aren't that unusual, especially in winter, so having a lighting system that won't die along with everything else is another benefit.

It sounds like you want to run the light off the battery most of the time, and only use mains when the battery is empty.
If power cuts are common place then what happens when there is a power cut and the battery is empty?

 

[marconi]

deleted

 

[marconi]

marc8 at#8 makes a promising suggestion. I think though, from the last time I studied how these devices work (not immersun in particular) they rely on the mains being live so that during a 1/50 second period or in bursts of them, energy can be drawn from it to power the immersion element and then paid back (in part or full) by PV. Immersun might operate differently - I have not checked. There is then no direct transfer of power between PV and output. Worth checking first.

 

[aurumdeus]

Regarding feeding into the grid, that starts getting far more expensive than I'm planning. I'm aiming for a parts budget of max about €600. From what I've read, simply connecting in can cost €1500 let alone the cost of parts, multiple thousands with a long long time to break even.

From the looks of it adding complexity won't significantly increase the cost of it, it's just extra labour. I'm going to add the complexity but in such a way that if it fails the system will still carry on working.

I'll replace all bulbs with LED ones and use a low power pure sine inverter hooked to a leisure battery. I'll use the load output from the charge controller to trigger a relay on the remote wire for the inverter, so it will only power on when the battery has enough charge.

It' better to connect to the remote line rather than battery positive as the currents are far smaller, and are constant with no big peaks.

I'll run the AC output from the inverter through a DPDT relay along with the mains AC from the consumer unit. I'll have mains connected to NC and the inverter to NO. I'll also run the inverter AC to the coil (AC coil) on the relay, so the relay will only trigger when the inverter is already running. I'll also run the inverter AC through an RCD and MCB before the relay, to ensure everything is safe.

I'll also run the AC going to the coil through a switch, so it's possible to force the system to use mains supply if wanted for any reason.

That system on its own is enough to work correctly and should be pretty robust.

Now for the interesting parts.
The inverter being on 24 hours a day will cost roughly 10 Ah from the battery, so if I can avoid that then I will. Assuming it's on for 8 hours a day this is reduced to 3.33 Ah, much more reasonable.

I plan on running the AC line (after the relay) through a non contact ammeter connected to an arduino. The arduino will connect to another relay on the remote wire, in series with the one from the charge controller. I'll have it connected to the NC port so that if the arduino fails then the system will still function correctly.

I'll also add 2 wattmeters, one measuring the mains AC and the other the inverter supply. They send out little pulses which the arduino can use to tally up energy use from each of the sources and display the performance of the system, and also I can calculate an estimate of the energy (and money) saved. They do use a little bit of energy, but the one measuring the battery will only be on when the inverter is in use anyway. I'll power the arduino and relay board from the battery (through a voltage stabiliser), or possibly add an extra little SLA battery fed from the main one to ensure that it doesn't lose power even if the main battery is discharged.

This system will keep the inverter turned off until it is needed, and if it fails the inverter will just be on constantly so no big problem. I'll add little indicator lights to show which one is supplying the power, and possibly a multifunctional meter reading the output of the DPDT relay like this AC 220-380v 100A 45-65Hz Din-rail digital AC voltmeter ammeter frequency meter | eBay - http://www.ebay.co.uk/itm/AC-220-380v-100A-45-65Hz-Din-rail-digital-AC-voltmeter-ammeter-frequency-meter-/182623248434?hash=item2a8531a032:g:RDgAAOSwz71ZQ6U1

All the extras will make it a lot easier to monitor and improve system performance, and while adding complexity they aren't too expensive. Under €100 for the lot I reckon. If it all fails the system will still work correctly, albeit not as efficiently.

It's for a friend who runs a B&B, so this means that his guests will still have lights in the event of a power cut. Also his highest usage of lighting will coincide with peak guests which is in the middle of the summer, when there is peak power from the sun.

If it works well then I'll do the same thing at my mother's house for her to keep costs down (also a B&B). It should be good fun!

 

[Pete999]

Where's the painkillers love? my head hurts.

 

[aurumdeus]

Huh, some of the replies only appeared after I posted that last message.

Spoon is right in that I want to use the battery as much as possible, and only use the mains as a backup. I'm hoping that I can get peak usage to use maybe 80% of the battery, so there is still a little left over. Power cuts aren't super regular, but a few times a year. If they happen during the day then no problem as lights aren't being used and the panels will be charging the battery. If they happen during the evening then the lights should already be running from the mains.

I had considered changing the lighting circuits to 12V and directly powering LED's that way. The main issue is the size of the wiring, and getting enough current down the not particularly thick wires to light up a very large house. Also if during winter the batteries run empty then I would have to convert the mains supply down to 12V and send that round the circuits. Given that the person I'm doing this for isn't very technical, I think it might be easier to go the AC route. Also if a bulb goes and someone screws in a normal halogen bulb (or there are no LED ones to hand a a guest is about to arrive etc), the system will still operate.

 

[marconi]

Will you be conforming to what the French/EU regulations for safe use of electricity say about an electricity supply derived from two sources of energisation with automatic switching between the two?

Is your proposed scheme safe for all users?

Remember you are responsible and liable for this design and installation.

 

[aurumdeus]

Lucien that's a good point about separation of the circuits. They will be separated within the relay and the supply from the inverter will go through an RCD, but I don't want to risk shocking a worker (even if only a little). I wonder if there is some kind of heavy duty contact breaker that can be controller like a relay. The mains could power that to keep itself connected to the DPDT relay and the rest of the system as long as it's energised, and if there is a power cut then it will separate itself better than just by a relay. I am friendly with a guy who is an electrician here, so I will check with him exactly what is needed in order to keep everything and everyone safe.

Regarding the short power lag while the relay switches sources, the relays I am looking at have roughly 15-20ms switching time so it should be little enough to not be noticed.

 

[aurumdeus]

This is all assuming I don't encounter anything too problematic when I get stuck in, like shared neutrals between lighting and other circuits, sockets wired into lighting circuits, and so on.

At one point while replacing two adjacent sockets I found 4 different colours of wires being used for live, red, black, blue, and green/yellow striped! This was on a house built in the 1970's, so hardly that old either.

 

[Lucien Nunes]

They will be separated within the relay

You must account for such events as transit arcs, inverted phase and contact welds, so you need to know how the regs and design practice inform the choice of relay / contactor. FWIW there can be up to 650V across the contacts.

and the supply from the inverter will go through an RCD

Have you considered the different grounding arrangements in different inverters - C/T vs. IT vs. earthed neutral, whether the DC input is electrically separate from the output, and how these will affect the functioning or not of the downstream RCD? Otherwise the RCD might appear to be offering protection when it actually is not, and will not trip in the event of a line-earth shock; even the possibility of getting a mains voltage shock off the ELV DC side.

 

[marconi]

http://www.----------------------------/mediafile/100117573/Best-Practice-Guide-2.pdf

Nota bene

 

[Pete999]

http://www.----------------------------/mediafile/100117573/Best-Practice-Guide-2.pdf

Nota bene

Are they any better than Heinz beans?

 

[aurumdeus]

I was looking at this inverter https://www.amazon.co.uk/Phoenix-In...coding=UTF8&psc=1&refRID=15DVA1318QJJZW1TDMH2

There is an earth connection on the case itself. I can connect this to the main earth line for the property, or if it would be better drive an earth stake into the ground just for this system.

 

[PEG]

I was looking at this inverter https://www.amazon.co.uk/Phoenix-In...coding=UTF8&psc=1&refRID=15DVA1318QJJZW1TDMH2

There is an earth connection on the case itself. I can connect this to the main earth line for the property, or if it would be better drive an earth stake into the ground just for this system.

Hi,a great deal of the yes/no,of this system,will not be decided by theory or function.
It will be decided by the regulations and recognised specification,in force,where the installation is.
This is especially so,as the properties are open to fee paying,members of the public.

Whilst being enthusiastic,regarding such experimenting,you may be hampered by issues such as the letting insurance,not being able to have standard testing carried out,on this installation.

As an aside,i have experienced systems,working fine...until members of the public turn up,and carry out their own "unique" form of testing,which often involves completely ignoring instructions,followed by indiscriminate and clumsy attempts at a remedy