Strange fault

[gazdkw82]

Installed some new floodlights/pir and switches today. All tested fine, however on power up it was evident I had got something wrong. Upon investigation I had the neutral in the switched live of the PIR. Kicked myself but mistakes happen.

Corrected the mistake but still nothing was working as it should. Spent over an hour chasing the fault. Got to the metal clad RCD spur I had installed, live - cpc =245v, live - neutral = 110v, neutral - earth = 38v.

I got to the point of just having a piece of flex on the load side of the RCD and still getting the same results. Measurements are fine if you disconnect the cable. Fair to say I was stretching my head. Disconnected the RCD F/spur and tested the circuit again. All comes back ok. Wired all the lights up (bypassing all sensors/switches) and all works ok.

Has anyone come across a RCD behaving like this? Could my initial neutral/live fault have broken it? If so, how come a simple piece of flex causes these readings but no flex doesnt?

 

[gazdkw82]

Thanks, on my previous post I uploaded a photo of the setup.

I have 2 of these set ups running to 2 independent 3 position switches. 1 for the side and 1 for the back.

I cannot see any issues with the design of the set up?

 

[Lucien Nunes]

So you have a change-over switch with centre-off position? Changing over between permanent L and PIR SL out to feed the lights? So far so good. You say L1 was not apparently making contact to common as the PIR output was not reaching the lights. Did you check its resistance?

The RCD tripping seems to be a completely separate fault. Your tests seem to show an intact circuit so either you've misread something, or there's an intermittent, or you clear the fault when manoeuvring the switch etc to get the test leads on.

 

[gazdkw82]

So you have a change-over switch with centre-off position? Changing over between permanent L and PIR SL out to feed the lights? So far so good. You say L1 was not apparently making contact to common as the PIR output was not reaching the lights. Did you check its resistance?

The RCD tripping seems to be a completely separate fault. Your tests seem to show an intact circuit so either you've misread something, or there's an intermittent, or you clear the fault when manoeuvring the switch etc to get the test leads on.

Thanks Lucien, I didn't check resistance on the switch, just continuity which was open circuit no matter what the switch. Switch must be a duff. Could have been caused by the original fault.

I am back there on the 12th and plan to find this fault.

This is my method of attack.

Break circuit down.

Test and confirm the supply to switch in off position. Confirm rcd is holding in. Gradually introduce circuit a little at a time. Testing and then energizing. Side lights through switch in perm position, front light through (new) switch in perm position. Then both sets of PIRs.

One thing I did notice, the robus PIRs have a max LED loading of 150w. On the fronts I have 4x50w LED floods. Obviously this is over the max rating per PIR, would having 2 PIRs in parallel cause an issue like this?

 

[Lucien Nunes]

this is over the max rating per PIR

Not good. If the inrush is too high, the relay in the PIR might weld closed, rendering it useless. The 150W rating is a bit arbitrary because unlike tungsten lamps (which all behave much alike, allowing the makers to calculate what tungsten load their product will handle) there's no direct relationship between LED wattage and inrush. Manufacturers can only estimate and prove by testing that 150W of typical LED drivers will create a level of inrush that the relay can be survive for an adequate number of switching cycles over the expected lifetime of the PIR. With your 200W load, it might last its full rated lifetime, or it might weld after 6 months. It probably won't fail tomorrow but if and when it does, you can't send the PIR back as faulty with a clear conscience, although I suppose a lot of people try.

There's no significance to how many PIRs are in parallel. The first unit to trigger when the lights are off has to bear the full inrush; the other unit might as well not be there.

 

[gazdkw82]

Not good. If the inrush is too high, the relay in the PIR might weld closed, rendering it useless. The 150W rating is a bit arbitrary because unlike tungsten lamps (which all behave much alike, allowing the makers to calculate what tungsten load their product will handle) there's no direct relationship between LED wattage and inrush. Manufacturers can only estimate and prove by testing that 150W of typical LED drivers will create a level of inrush that the relay can be survive for an adequate number of switching cycles over the expected lifetime of the PIR. With your 200W load, it might last its full rated lifetime, or it might weld after 6 months. It probably won't fail tomorrow but if and when it does, you can't send the PIR back as faulty with a clear conscience, although I suppose a lot of people try.

There's no significance to how many PIRs are in parallel. The first unit to trigger when the lights are off has to bear the full inrush; the other unit might as well not be there.

As usual, golden pieces of information

 

[telectrix]

might need to fit a contactor/relay to switch the lights while putting a much lower load on the PIR/s

 

[Dave OCD]

I think Tel has a very good idea there, you can buy a simple '1 module' 20 amp contactor for not much money then feed the coil from the linked PIR switched live, you can mount it in a simple plastic 2 module enclosure with a blank over the spare 'way' which will give you plenty of wiring space.

 

[gazdkw82]

I think Tel has a very good idea there, you can buy a simple '1 module' 20 amp contactor for not much money then feed the coil from the linked PIR switched live, you can mount it in a simple plastic 2 module enclosure with a blank over the spare 'way' which will give you plenty of wiring space.

I agree. It's a good idea