Emergency Light Monitoring System

Been installing new emergency lights at work. The ones that are connected to mains which charges a big lead acid battery and when the electricity goes off, the battery will power 2 10W LED lights to illuminate parts of the warehouse.

I was thinking it would be cool if you could create a system that could

  • checks that the unit has power
  • monitor the health of the battery
  • perform test at scheduled times to see how long the lights actually work ( turn off electricity to the unit ie via a relay )
  • during the test checks that the LED lights actually work.

This system can then track any events and test and if there is ever a question about the lights, then there is plenty of data and tests that would help answering those questions. Also it would reduce the amount of work to test the system as this would be automated.

I see it that there would be a small arduino in each emergency light. Mains to the unit would go through a relay controlled by the arduino. The arduinos somehow need to send data to a server and also request data from the server.

So my questions would be:

a) Does this sound like a feasible project?
b) What arduinos would be best to be used? Would ESP32 a possibility or would be a uno with ethernet shield better? ( trying to keep cost down of course )
c) How could I monitor the battery health?
d) how can I check if LED lights actually work? Just check the current it draws during the test?

Is there anything that could be added?

As those units have sealed lead-acid batteries, how do you propose to check the health of the batteries? Just turning on the LED light sure won't do it. Perhaps time to recharge after extend use would do it. Can't do it with just the lights as that could take days, leaving the system inoperable for a real emergency.

Paul

Depends which country you are in but these have already been around for many years.
Monitoring battery health is done normally by discharge and any not meeting the 90 minute minimum are disposed of.
Usually a building and government requirement.
Test activation is done via a dedicated timer at the switchboard which isolates power from the circuits in their respective areas.
We use the ones you mention in hospital eating areas and the like where there is usually a large area to cover.
Egress areas also can come under their control.
I have seen addressable units used, also in hospitals, where discharge tests monitoring are done via a maintenance crew PC.
Each unit is linked via cabling not unlike network Cat5 except flexible and I think, shielded.

Agree , by adding a system you are moving away from manufacturers instructions for test - which effectively makes your emergency lighting invalid .
You will need to break into the light and modify it , removing any approval it has .

hammy:
Agree , by adding a system you are moving away from manufacturers instructions for test - which effectively makes your emergency lighting invalid .
You will need to break into the light and modify it , removing any approval it has .

If the tests are more simple then no need to break into the units.
Have the power relay on the cable going into the box and a light sensor attached where the emergency light shines on it and ambient light is mostly shielded.
Cut the power and check the emergency light remains lit. Monitor the light and after the required testing time (or the light going out) power on the relays. If the light remained lit after the time then all is good but if it died during test then it's a fail.

Riva:
If the tests are more simple then no need to break into the units.
Have the power relay on the cable going into the box and a light sensor attached where the emergency light shines on it and ambient light is mostly shielded.
Cut the power and check the emergency light remains lit. Monitor the light and after the required testing time (or the light going out) power on the relays. If the light remained lit after the time then all is good but if it died during test then it’s a fail.

That might be fine and dandy for your back shed but when it comes to anything in the real world, they have to follow a certain protocol, especially where life saving systems are in place.

There should in any case be a visual inspection for any damage or whether it needs cleaning - imo it’s a non starter.
Imaging yourself in Court after an accident “well I thought of it myself and thought it a reasonable thing to do ...”

I would be amazed if a company would allow one of its employees to 'modify' emergency lighting systems.

Think of the insurance implications ...............................

Couple of points.

Here in the USA monthly testing is required.

Also the curve of battery voltage should be predictable.

No reason you cannot add a circuit to test the the units for voltage.

As a note normal emergency use (when power fails) drains the batteries fully dead and often batteries are a one use device in this application.

SIMPLE testing monthly is to shut off the circuit breaker and visually inspect what lights don’t work. Takes 10 15 minutes. Easy on the batteries.

ANNUAL testing is usually 90 minutes and see what survived.

It seems the concern is about automatically shutting off power and guaranteeing with someones life that it comes back on.

Thanks for all the feedback. Moving on to think about a new project.

dave-in-nj:
No reason you cannot add a circuit to test the the units for voltage.

Well I’d have to disagree there.
I can imagine the look of shock/horror on the face of an insurance accessor after some defect to find stray wiring leading to a mod board from any emergency fitting.
I’d say the “strike out” pen would come out faster than midnight.

stephan2307:
Thanks for all the feedback. Moving on to think about a new project.

I am thinking about doing the project as you outlined.
test for voltage.
the 'emergency circuit' should be clearly labeled in the circuit breaker panel.
once a month, just shut off the power, the lights come on. give them 10-15 minutes and you should see the curve of the battery drain.
remember, the batter is designed to provide light for 90.01 minutes and not to 90.02 minutes. so, if 10-15 minutes, you should see enough drain to know if the battery is still good.
Also, your first instinct is right. we find out about bad EM lights during the emergency, when you need it most.
I replace a couple dozen EM lights, Exit signs, combo EM/exit lights, battery backup ballasts, etc every week. and 99.9% are from total failure, will not come on at loss of power.
so, putting a sensor on the thing is smart and does not violate any laws, and it would help keep them all workings.
the automatic power loss testing.... maybe not worth the bother.

bluejets:
Well I'd have to disagree there.
I can imagine the look of shock/horror on the face of an insurance accessor after some defect to find stray wiring leading to a mod board from any emergency fitting.
I'd say the "strike out" pen would come out faster than midnight.

insurance for what ?

There is no way to say that the fire or roof collapse was anyway affected.
your concern is about the same as saying that a parked car cannot collect from being hit, because it had non-manufacture tires.

Testing would not alter the primary design, or fundamental use.

if your arduino were to start the fire that burnt down your house... THAT is a much more realistic and MUCH more valid concern. and EVERY Arduino project that has power, has that risk.

Can you imagine the insurance company said 'sorry, your house got hit by a fallen tree, but we found you had tested the voltage of your toaster, so...... '

You will find these emergency fittings are built to a code and are certified as such.

If you think any modification is an ok thing to do then you are in for a rude awakening.

Not only that, but telling others it is an ok thing to do is highly irresponsible.

Your idea may well be, a good idea, BUT until such changes are approved and certified in the fittings used, ANY insurance accessor will reject it, and that's a fact.

the building code, nay, fire code, requires that an emergency light [in most jurisdictions in the USA] be lit for 90 minutes after power failure
the fire code also specifies the light level available over an area and typically about a foot above the floor.
the fire code requires that fixtures be inspected monthly
fire marshals perform site visits about once a year for inspection.
each jurisdiction, can and does alter that to some degree. if you want, I will link or post a fire codes in one of the states that I routinely work in, with fire marshals. Part of my job is to bring buildings up to fire code.

That is all there is to it. it is not irresponsible to say that measuring a device for operation is in any way a bad thing to do.
in point of fact, the OP's idea is a very sound and reasonable approach to MEET the fire code.

I did go on record saying that doing anything that might disable or otherwise render a device nonfunctional is a really bad thing. the automatic disconnecting of power to allow the device to drain, then measure battery level, either as a function of a curve of depletion of battery energy or a timed function of battery after X minutes or seconds, although a sound engineering CONCEPT, might, MIGHT, potentially cause a unit to drain completely if the control were to fail, and thereby rendering the device useless in an emergency.

so, measure YES there is no possibility of a violation of law as you are not rendering the device inoperable.

interrupt power automatically, NO, there is a potential that the circuit may not restore power.

the simple thing is to measure, and manually, perform the actions required by code to test the devices.

bottom line an automobile turn signal or brake light is a safety device. the EM light fixture has about the same value.
if one goes out, for any reason, there is no insurance risk or civil penalties, if the Fire Marshal finds you have one or more out, you get a 30 day notice. if you do not effect repairs in 30 days, you get a second 30 day notice.... if you do not effect repairs in 60 days..... can I tell you I have brought buildings up to code after years of failure to comply ?
the fire marshal has the duty and right to close any building for safety, so it is the best interest for all concerned to use their expertise to make the building as safe as possible at the earliest opportunity.

The proposed concept does just that.

As promised, links to the NRPA code for the USA.

interesting to read is that 'self test' is part of the code. so even that part is allowed. so the OP was on the right track for both automated testing and level testing.

overview

here is something to support the concept as not only being totally legal, but also being in the code

The option you choose for testing your emergency lighting system will depend on the type of system you have – whether it is or is not a battery-operated, self-testing and self-diagnostic system and if it is, whether or not is computer controlled.

Regardless of the type of system you have, it must be tested for a duration of 30 seconds, generally, once every 30 days and annually for a duration of 1.5 hours. Additionally, all emergency lighting must be fully operational throughout the duration of all tests conducted.
Testing for battery-operated, self-testing and/or self-diagnostic systems should include a diagnostic routine. For systems that are not controlled by computer, you must perform visual inspections once every 30 days, and your system must be equipped with a status indicator to alert you to any failure.

Here is part of the code NFPA 7.9.1 general

NFPA 7.9.1 GENERAL

and if you want to read the full and most current code, the official USA NFPA link
NFPA LINK

in reading all of this, I found that the single requirement in every link was that it passes the testing in the building where it is installed.
light levels seem more important and are described in detail.
The actual devices are not listed or referred to or described how they are designed, just that they pass the test.

The bottom line is that alteration that renders a device to be non functional would make the unit not pass testing and that would be a failure.

As for the legal disclaimer : before you attempt to modify anything, read all regulations, codes and laws to make sure you do not violate any codes, regulations or laws in your country or jurisdiction.

My OPINION and not a recommendation, is that monitoring a device in order to verify it is working goes to the spirit of all the things I have read. also, that the addition of an Arduino or other micro-controller can increase safety by catching failures early as well as the required documentation.

stephan2307:
I see it that there would be a small Arduino in each emergency light. Mains to the unit would go through a relay controlled by the Arduino.

This latter sounds quite reasonable. Note that the lights will be connected to the normally closed relay contacts, so they will be operational when the relay is not actuated, it will "fail safe".

stephan2307:
How can I check if LED lights actually work? Just check the current it draws during the test?

No! You need to check that the LED lights actually work. This means that you check that they produce light as has been explained already! :roll_eyes:

In other words, you use light sensors.

Note now that this means there is absolutely no modification to the emergency light units themselves, they are not touched in any way whatsoever and the monitor is not even nearby.

In my building, the emergency lights are on the common lighting circuits, one for each floor. The test is to switch off that light circuit; it becomes dark except for the emergency lights. So if you use light sensors and particularly if you arrange to perform the test at night ("out of hours") there is no confusion as to whether the light comes from the emergency lights or the common lights.