Thoughts on 'snubbing' an AC Shower relay to stop it sticking?

[jkn]

Hi all our electric shower 'periodically' makes the water too hot. It has just started to do it again.

From previous inspection & repair, I think one of the relays (there are more than one, depending on what you set the 'hot/efficiency/cold' setting to) sometimes gets stuck in the 'on' position. Presumably the contacts weld themselves together, perhaps only for a while, so that it remains heating the water even when the coil has no voltage across it.

I've replaced the relay more than once but since the problem keeps occurring I am wondering about an improved approach. The relays are open frame 'T&E' relays like this:

I realise that I can't 'snub' these AC contacts but am wondering of any other approaches that might be worth trying.

Ideally of course, a bigger spec'ed relay might be the way to go. There is probably room inside the shower, if I can find an alternative device...

Thanks for any thoughts J^n

[The Natural Philosopher]

Why not? DC is only *very low frequency* AC. :-)

Normal R and C in series cross the contacts should work OK

[Paul]

On Central Air Conditioning, the compressor is switched on with a "contactor".

Heavy loads need a device with a greater contact rating, to prevent welding.

A contactor might be used with an inductive load like a motor, where the motor starting current is 10x the running current flow level.

You would think a resistive load, would be easier to control without welded contacts.

Maybe a clamp-on AC ammeter, would help verify what the starting currents are like on the shower. I have one here, that measures up to 400 amperes, without making an electrical connection. You just put the jaws of the meter, around a single conductor, and the current is implied from the level of mag field. That's why no wiring need be done to use the meter.

That won't tell you everything you need to know, but it's a way to figure out whether something is not right with the installation. Even if the thing said it was 12kW, it would be nice to have some verification that is what the device is.

Paul

[Brian Gaff]

Yes well the issue may be that the spring that opens it afterwards is too weak. I know those in some oof these remote controlled outlets can have the same stuck on issue. They stick on and a thump to get them to open. Brian

[Brian Gaff]

You used to see those in washing machines a lot as I recall. One used some type of contact that slid as well as moved in and out, due to the spring in the contact holder and that never jammed but was not the cleanest from the radio clicks it could make. Brian

[jkn]

Actually, I confess to a bit of a brain fart here. I was writing 'I can't use a 'snubber' ... when in fact I was thinking of a reverse flyback diode, a totally different thing!

So having checked a bit more, it looks like something like a 100Ohm resistor + 150nF X2 cap might be some good values to try.

It also looks like you can get 40A contact relays in the same range for not too much more money ... I might fit one of those at the same time.

Ho hum...

[Fredxx]

A shower heater is resistive, so no snubber is realistically going to save the contacts.

I would suggest a heavier relay if it's available.

[jkn]

Are you sure? I haven't yet measure it but I would expect that the turns of the heating element wire would be likely to give some inductance. I stand to be corrected though.

As mentioned in my second post, I am thinking of adopting both approaches.

[Fredxx]

Fairly sure. The energy associated with inductance would be very low. The only way to be certain would be to measure it. Heating associated with contact resistance would be significant.

I'd wager if you took your your relay apart you'd see signs of heat and springs that lost their temper.

I suppose it wouldn't do any harm.

[Animal]

Resistive heating has only trivial inductance. But you still get sparking as the contacts open, a snubber does help with that.

0.1uF + 100R is a common snubber for small loads, for many kW I'd use a fair bit more C and less R, otherwise it won't be effective.

Bigger contacts is definitely the first place to start though.

Low relay coil voltage can also cause this, due to reduced contact pressure, might be worth checking.

A snubber compromising the isolation is not permitted, so best check your dying relay is not doing the isolating job as well as just switching.

[The Natural Philosopher]

All conductors - even a straight piece of wire - have a certain amount of inductance.

In the case of a spiral heating element its simply not very much. So its pretty safe to ignore it unless you are building a MHz transmitter with one.

[Fredxx]

Spot on, and it's really down to energy being dissipated when the contacts open. A snubber won't do any harm but at 240V with rated amps passing, the contacts will arc in their own right. Quality of contact is everything.

100R in series with 0.1uF seems pretty pointless. Any inductance will maintain the current, say 30Amps so the resistance would have to be sub 8R for the capacitor to have any meaningful quenching effect.

[Paul]

The snubber conducts (damps) the high frequency component. The longer-term low frequency component, is still an issue, but may not afford the voltage stress of that ringing waveform. That's why the tau of the snubber, always looks like it's some kind of "transmission line component". Nobody would reasonably expect a 100 ohm quarter watt resistor to eat a "plasma storm", without it turning into a plasma storm itself :-)

And there's lab techniques for dialing in a value empirically. These are the things they don't teach you in school.

But you really need a scope and proper probing technique (to keep you safe), to pick a value and be satisfied with the result. For people who don't know what they're doing, a battery-powered scope is a start. Another thing a well-equipped lab will have, is 10:1 and 100:1 probe, so if the voltage rating of the scope is not sufficient, you can use a 100:1 probe to extend the voltage measurement range. To use such probes, you need to know the equivalent circuit, and configure the scope input termination properly to make it work. The probe expects the scope to have a certain input impedance, as the scope is "part of the circuit".

If you needed help, if there are any old farts left in sci.electronics, you could likely find someone to help. Someone who has actually tuned one. There was one guy in sci.electronics, who wrote a text book, and if that guy is still alive, he can answer a lot of questions satisfactorily.

Paul

[Fredxx]

The old farts on sci.electronics show more interests in gun control and extreme right wing politics!

I think we need to assess the Frequency of interest and whether a T+E cable is going to exhibit the properties of a transmission line.

At the frequencies of interest, say a few MHz the cable length would have to be over wavelength / 4. At say 10MHz, that would mean 1/4 wave would be 7.5m. If short, the line can be discounted and the heater treated as a low value resistor and series inductance.

A snubber is ideal where there is a great deal of stored inductive energy. In this case it's going to be negligible.

[Animal]

The problem is not inductance, it's the time when the switch is closed but not firmly & the contacts are still moving a bit. That's what causes poor connection which generates the hash, and damages the contacts. /If/ your snubber can reduce that, great. But there's an inherent problem with RC snubbers. While they divert current away from the contacts during opening events during the hash, they inevitably then dump the current back into the contacts when they close again. It's not a free lunch. If you have a 30A shower, you need the snubber to eat 30A (rms), and it will then dump a lot of current into the contacts when they remake.

Basically the OP's relay is not capable of the job, and one is needed with larger contacts, faster switching (the 2 directly contradict each other of course), and maybe better contact construction/materials. When that's not practical one resorts to icky workarounds like snubbers for noninductiveloads, parallelled contacts to reduce total contact burn, momentarily connected resistors etc. They can help but they don't properly solve the problem. Get a better relay.

[Adrian Caspersz]

...

Solid State Relay?

[jkn]

General update FWIW - I have recently taken delivery of some upgraded relays - same footprint but 40A instead of 30A.

'Interestingly' - these are potted, instead of the open frame style of the original. I am slightly wondering if the enhanced rating is entirely due to this?!

Perhaps they are potted in an inert environment ... or after a few sparks, is becomes somewhat inert?? Or, the enhanced rating is a bit dubious - that is also possible, we shall see.

BTW, from another old but similar thread on uk.d-i-y I got this nice link on snubber design:

Cheers, J^n

[Animal]

interesting, but it looks at solving different problems to this one.

Let's hope your slight upgrade works.