LED light in bathroom flashes when extractor fan TP pole switch is closed.

[Lucien Nunes]

But the snubber load will probably work OK though. You could fit it in the isolator, lamp or humidistat, since all have access to SL and N.

One reason we use a snubber (RC network) is that for a given ability to dispose of leakage current, it dissipates less heat when the SL is energised. The capacitor stands most of the 230V, passing a current 230/Xc where Xc is its capacitive reactance. The resistor limits peak currents at sudden changes in voltage e.g. to stop it welding switch contacts. As per the OP, it is possible to use pure resistance instead, but to achieve the same leakage-sinking ability there will be more heat in your back box when live.

 

[pc1966]

There's another aspect to using a changeover (2-way) switch to short the lamp which is that L and N are present on N/O and N/C contacts (L1 and L2). In a functional switch that is not designed as an isolator, and made for AC only, the contact separation when open is very small, only a millimetre or two; too small to guarantee that the separation alone will extinguish any arc that forms when a load is switched off. This is deliberate, because the arc will extinguish itself at the next current-zero in the AC cycle and the smaller it is in the meantime, the less heat it will dissipate.

But said switches are rates at 10A, so there could well be a load on the other pole that would sustain an arc of at least 10A when cold, possibly more. So while it is a risk with a historic precident I would be most surprised if any switch rated at 230V/10A was so designed that it could not safely switch from pole to pole without an arc jumping them.

 

[Lucien Nunes]

In the scenario you describe, if the switch only handles the line, the arc will always be in series with one or other of the loads, limiting its current to within the switch rating. OTOH if L & N are both present on one pole, any arc that forms can transfer to directly between them, bypassing the load resistance, therefore could rise to the short-circuit current.

I agree that very little arc is likely to form in this low-load scenario but if it does, there is nothing to stop it escalating.

 

[Lucien Nunes]

It's not altogether unlike crossfire in a controlled bridge. Doesn't often happen but can cause immediate destruction if it does. Crossfire can happen in Mercury Arcs too under both controlled and passive rectification. They usually had surge protection across the anodes to minimise the risk of a transient setting it off.

 

[HappyHippyDad]

Well the 100nF 100r snubber didn't work so I changed the design as per @Pretty Mouth 's diagram. Thanks again PM

 

[Lucien Nunes]

Excellent. It's a pity there are so many products that pass the buck as far as compatibility is concerned. You would hope that the makers of the humidistat would acknowledge that low power LED lighting might be connected to the SL and take the necessary (simple) steps to prevent the unit leaking current back out of its control input, and that lighting manufacturers would acknowledge that sometimes the SL gets a bit of inward leakage that should not be allowed to start the lamp. But they are under pressure to squeeze every possible component out of the design to save a few pence, so it becomes a 'system-level' problem to solve.

 

[Pretty Mouth]

Well the 100nF 100r snubber didn't work so I changed the design as per @Pretty Mouth 's diagram. Thanks again PM

Glad it worked for you mate

 

[HappyHippyDad]

I thought I would ask 'why' the snubber didn't work? I have just read another thread by @gazdkw82 where he fitted what looks like the same snubber, but in his case it worked and stopped the LED flashing. Why did it not work in my case?

This is the contact suppressor I fitted, which looks the same as gazdkw's and has the same ratings.

https://www.ebay.co.uk/itm/AMPOHM-W...K:MEBIDX:IT&_trksid=p2057872.m2749.l2649#rwid

 

[pc1966]

Do you have a multimeter that can measure capacitance? It might be worth checking that the snubber you got is actually working!

 

[HappyHippyDad]

Do you have a multimeter that can measure capacitance? It might be worth checking that the snubber you got is actually working!

Just measured it. It was 95NF so 5% less than it's rated farad which I'm guessing is fine?

 

[pc1966]

Just measured it. It was 95NF so 5% less than it's rated farad which I'm guessing is fine?

Yes, mostly they are 10% or even 20% tolerance.

 

[HappyHippyDad]

Yes, mostly they are 10% or even 20% tolerance.

So any thoughts on why the snubber did not work in this case. Would you then need more than 1 snubber or a different rated one?

 

[pc1966]

Did the snubber change anything?

Some DC leakage by a controller would not be shunted off by a snubber, so this might just be a reason for it.

 

[HappyHippyDad]

Did the snubber change anything?

Some DC leakage by a controller would not be shunted off by a snubber, so this might just be a reason for it.

No, it did absolutely nothing. Thanks for all the input pc1966