Monitor a breaker to detect trips

[rthorntn]

Hi,

I have 10 large outdoor batteries that occasionally trip, they have a switch on them like the one attached, I want to detect when one trips and I cannot modify the battery in any way, the switch has a rubber hood over it.

I'm stuck with the occasional trips as I have crappy inverter chargers that I cant change..

How would you do it, some sort of sensitive weatherproof vibration sensor that I can stick on each battery near the switch that I can read with a microcontroller?

Thanks.
Richard

[Andy Chee]

Using the same principle as indicator fuses, what about connecting an optocoupler LED (and resistor) across the breaker?  The optocoupler transistor can then signal whatever alert circuitry you want.

[rthorntn]

Thanks, I can't modify the switch though, the batteries are very expensive and warrantied for 20 years.

Are you talking about creating a light beam and detector with the beam getting broken when the switch moves?

The moulded rubber shrouds over every switch are, for want of a better word "frosted", I'd probably have to cut the shrouds off and weather seal the switch some other way to guarantee a beam could pass through.

[Andy Chee]

Not a light beam.

Study the Bigclive video to see how the "blown fuse" indicator LED works.  Basically I propose the same circuit topology, but using an optocoupler LED.

Retrofitting this technique requires access to the terminals behind your circuit breakers.

[rthorntn]

Thanks, I wont tamper with a $3K battery that still has 17 years of warranty left, I guess I could ask the manufacturer for permission but its unlikely that they would give it (the batteries do 40A at 48V).

[Andy Chee]

I don't understand.  I thought the breaker was separate to the battery?

Can you post a photo of the breaker/battery combo?  (instead of the breaker by itself in your previous picture)

[rthorntn]

Apologies, I should have been more specific, the breaker is part of the battery.

[rthorntn]

This is a drawing of a slightly newer revision, my batteries do not have the LED.

[Phil1977]

Can you glue a small magnet on top of the breaker switch and put a reed switch / hall sensor next to it?

[rthorntn]

Thanks, I think this could be the ticket!

[Doctorandus_P]

I have 10 large outdoor batteries that occasionally trip,

Are those long trips, or just short hikes?
Maybe too much LSD?

Apparently this is some kind of waterproofed outdoor stuff. What does the fuse box look like? Are those 10 batteries close together, or are they completely separated installations? Can you make any sort of electrical connection to those systems?

A common way is to put a transistor with series resistance over the fuse. As soon as the fuse breaks, the secondary side will lower it's voltage (assume the fuse is in the positive lead) and this will pull a base current through the BJT (or put gate voltage on a FET), so the transistor opens, This can turn on a LED, or give a signal to a uC device that sends an RF message or switches a signal horn (Powered via the battery and another fuse ) Simpler is to just wire an LED and series resistor over the fuse. But this assumes you can make electrical connections to the whole thing.

If you can't make electrical connections, you can put HALL sensors on the cable, These can not measure voltage, or a direct break of a fuse, but they can detect whether current is flowing though the cable.

[inse]

Simply monitoring the presence of battery voltage is not sufficient?

[rthorntn]

Thanks!

LSD 

Ten 48V batteries stacked, ten cables, connected in parallel.

Are you saying I can put a hall sensor on each cable and sense no current flowing, can it detect tiny amounts of current, mA, because if the sun is shining and the batteries are charged very little current flows to/from the batteries?

[Andy Chee]

Hall sensor won't be able to tell the difference between; load switched off normally, and tripped breaker.  In both cases there will be zero current flowing.

[rthorntn]

I don't think I've ever seen zero current flowing, watching my shunt data I can see at least 10W flowing constantly, so around 1W per battery?

[Doctorandus_P]

Uni-T UT210D is a current clamp meter that can measure DC current with a 1mA resolution, but it is not very reliable in those low current range, both drift and the the earth magnetic field cause noticeable errors. For example rotating the meter 90 degrees throws off the least significant digits by a few counts. It's also mentioned in the manual, you have to zero the meter before each use, and make the measurement in the same orientation as when you zeroed it. Because of these issues, and because it can't see the difference between a normal disconnected load or a tripped fuse, it would be my last ditch option, because it "mostly works" without any physical connection to the system.

If your automatic fuse has something that moves on the outside, then interfacing with that may be a better option.

[Jeroen3]

This feels like an X-Y problem.
The circuit breaker in the battery should never trip.
Just look at this chart, how long you need to overload it for the breaker to open! (I assume it's aeon battery based on the picture)



Can you add secondary, lower, circuit breaker or overload detection on the whole unit?

[rthorntn]

Thanks, the breaker switch physically pops out when tripped, so a tiny magnet glued to the switch and have reed switch / hall sensor mounted above where the tripped switch "ends up" should work.

It's not high current that's tripping it, it's over voltage, my crappy buggy solar inverter charger occasionally overcharges, usually when the battery is charged and full sun quickly appears from behind heavy clouds.

The ten batteries can support 400A combined and I can't even push 60% of that.

[TERRA Operative]

Overvoltage won't cause a breaker to trip, only overcurrent.

[rthorntn]

Thanks, the isolator switch is tripped by an internal BMS that monitors:

Current
Over/Under Temperature
Over/Under Voltage

The BMS has no external interface.

[Siwastaja]

Every time a fuse or breaker blows, you should be saying to yourself:
"Phew, that was a close call, this safety device prevented a fire!"

Then you go and investigate what nearly caused the fire and fix that. After THAT you replace the fuse or turn on breaker again.

So two key questions:
1) why are the breakers tripping?
2) what's wrong with monitoring power availability to load instead, e.g. by monitoring output voltage after the breaker?

[rthorntn]

Thanks.

It's complex, it trips during a sunny day so no real battery draw plus there are ten batteries and if only one trips its very difficult to detect losing one battery in parallel as it doesn't drop the voltage, the house current will deplete the pack 10% faster but that's pretty hard to detect at the 0.4KW house base load (its a 19.2KW battery).

[Siwastaja]

Oh, outputs are in parallel, now I understand why you want to monitor the fuses. Really, you would need to try to work out why they trip but I can see that monitoring the status (knowing when it happens) would be useful for time being, for both system uptime and troubleshooting.

Does manufacturer allow paralleling of outputs of multiple batteries, so is what you are doing normal? It would be helpful to know the battery model to see the datasheet/manual.

It is hard to non-intrusively sense very small currents, but using hall effect DC current sensor (like those by LEM) you could at least detect the condition where you are charging/discharging at significant power and one battery is tripped, before it causes large differences in state-of-charge between the batteries.

[Jeroen3]

Based on the photos is should be possible to hot-glue 3D print a bracket for a microswitch to detect the fuse is not tripped.

(you want this fail-safe, so that if the microswitch falls off it doesn't do so unnoticed)

[BeBuLamar]

How much does the breaker move when it trips? Do you have to turn it to off and then on to reset it? A lot of breakers when trip only move very slightly (and thus difficult to detect) not all the way to off position.