Relay contacts welded

[Saull]

Hello,

with 20A relay(sanyou sfk-112dm) I am driving 3kW heater. After about a 1 year of usage(water heater) contact got welded. If I calculate current consumption for 3kW I get 13A @230V, so I tough that 20A relay will be enough.

I would like to replace that relay but I dont know what current rating relay should I chose ?

If the relay has two switching contact can I parallel them for prolonged usage ?

Thanks !

[madires]

The relay seems to be fine for that purpose, but the heater could have some inductance causing back-EMF when switching off. Back-EMF is a typical cause for welded contacts. Try to add a snubber.

[Zero999]

I doubt the inductance of the heater is significant enough to cause arcing.

That Sanyou relay is a Chinese rip-off of a good brand. Try replaying it with a real Sanyo relay and it stands more of a chance of lasting longer than a year.

[cat87]

Becausethose are mechanical relays, you can't really parallel them because the movement isn't sync'ed between the relays. One might open/close before the other  so that basically leaves you in the same position you're in now....a single relay will do all the job.

Indeed, get some reputable name-brand relays. Those tend to stand up much better to the specs on the label. Also, take care as no matter how good you relay will be, because of that much current on the  contact of the relay, it will still pit over time because of some arcing. So basically, the longevity is all down to how good the contact material is.

As an alternative, perhaps you could get some old mercury wetted relays, if switching speed isn;t something important for you. Those take a lot more punishment. Maybe you could get some off of eBay....

[james_s]

A good reputable brand power relay should last a long time. If that still isn't doing it, you can get industrial contactors for switching large loads.

[senso]

Industrial anything will be rated at at least 1 million cycles, so pay a couple bucks more and get a decent/name brand relay instead of relying on a ebay relay to shutoff your water heater, steam explosions are no joke...

[Zero999]

Industrial anything will be rated at at least 1 million cycles, so pay a couple bucks more and get a decent/name brand relay instead of relying on a ebay relay to shutoff your water heater, steam explosions are no joke...

If this is safety critical, there should be a thermal fuse, in series with the heater, to protect against explosions and fire. If it's not possible to get a thermal fuse with a high enough current rating, then an approved contactor should be used, with the fuse in series with the coil.

[rsjsouza]

with 20A relay(sanyou sfk-112dm) I am driving 3kW heater. After about a 1 year of usage(water heater) contact got welded. If I calculate current consumption for 3kW I get 13A @230V, so I tough that 20A relay will be enough.

It is a bit hard to say how much the relay would endure without knowing how many times it actually switched. The datasheet specifies at least 30k iterations, which is somewhat high for one year of usage (switching around 3~4 times per hour if the heater is turned on all year around). Also, as others have said, any load would arc either via inductance or bouncing - that is where the quality of contacts is a key factor for its durability.

If the relay has two switching contact can I parallel them for prolonged usage ?

If the relay is a dual or triple pole, it would either share the current among its poles and maybe reduce the arcing due to bouncing, as the contacts will close the circuit at different times (milliseconds apart). If anything, they would at least last 2x (if dual pole) or 3x (if triple pole) as each contact that closes earlier starts to falter...

All in all, I would definitely replace it with a quality relay.

[SparkyFX]

I doubt the inductance of the heater is significant enough to cause arcing.

When a relay opens, the tiny amount of air in between will get ionized in between contacts, so arcing even occurs without inductance, without high voltage. This heats up small spots and you get microwelding/micropitting on contacts, which will be teared apart by the larger mechanical forces. This ruins contact surfaces in the long run, bringing them out of flatness, which results in even more microwelding/-pitting.

One good method around the problem is to use relays with high quality contact surfaces, with e.g. nickel alloys or switch to SSR.

[floobydust]

The relay does not last due to the heater's inductance, which makes the contacts arc badly. It is significant because of the power levels, where even a few uH cause trouble.

Paralleling relays does not help because contacts are not identical and one contact  makes-first and another breaks-first, so they wear just like a single relay.

Fix is to add a RC snubber network, or MOV across the contacts.

[Saull]

Hello,

Thanks for all the feedback. Right now I ordered replacement relay from TE Conectivity 25A, I will add one external 40A from same brand(for easier serviceability). I would like to avoid adding stuff like RC snubbler.  As far the safety goes the system has pressure safety valve + thermal fuse.

Best regards Saull.

[SparkyFX]

The relay does not last due to the heater's inductance, which makes the contacts arc badly. It is significant because of the power levels, where even a few uH cause trouble.

Inductance (or better said the fly back when switching off) might lead to more arcs forming and staying on at higher distances, but striking an arc itself is possible at DC with low voltages like 24V, like in technical processes as in welding.

That small arc will stop to exist at the zero crossing or above a certain distance, but as a mechanical relay can not be timed to switch at zero crossing its contact surfaces will only have a certain lifetime in them.

At opening time, the voltage drop across the switch is small, getting bigger the more it opens. At some point its resistance will be relevant to the application and only limited by the insulating properties of air, arcing caused by flyback voltage could pinch through that air gap. The density of the metal and therefore the overall flatness will change slightly at the point of an arc start if it melts there, which is why contact alloys with higher melting point might last longer. As relays bounce when making contact this happens several times when opening or closing a contact. Additionally the contacts might cold weld under certain circumstances.

Kettles however are resistive loads, without much inductance (except... it is an induction heater...).

Fix is to add a RC snubber network, or MOV across the contacts.

Which would be useful if the arcing is caused by high voltage spikes, not if the voltages involved are lower.

[drussell]

If a "20A" relay will not switch a 13A resistive heating load off and on for many, many years without fault, then the relay is bogus garbage. 

Buy a decent relay, or even better, an industrial  contactor like is used for motor loads (even an old, crunchy, crusty, used one will do better than a Wan-Hung-Lo brand garbage relay) and you'll be fine.

A heating element is about as nice as you can possibly be to a relay.  You should not need snubbers, arc supressors or any other kind of hooey on a heating element.  The relay that you are currently using is obviously overrated and is obviously unsuitable for the task at hand.

Well, unsuitable for any purpose, apparently...    Get a quality relay.

[Zero999]

The relay does not last due to the heater's inductance, which makes the contacts arc badly. It is significant because of the power levels, where even a few uH cause trouble.

I doubt it. Just look at the figures. What inductance and parasitic capacitance do you think is reasonable for a heating element? More importantly what's the resistance? Unless the element is wound on an iron former, the RLC circuit will be over-damped, so there will be no inductive voltage spike.

I doubt the inductance of the heater is significant enough to cause arcing.

When a relay opens, the tiny amount of air in between will get ionized in between contacts, so arcing even occurs without inductance, without high voltage. This heats up small spots and you get microwelding/micropitting on contacts, which will be teared apart by the larger mechanical forces. This ruins contact surfaces in the long run, bringing them out of flatness, which results in even more microwelding/-pitting.

One good method around the problem is to use relays with high quality contact surfaces, with e.g. nickel alloys or switch to SSR.

Yes that's true, there will be arcing, even without any inductance and at the rated voltage, but that wasn't the point. If the relay is designed to switch a 25A restive load, at 250V, 6.25kVA, then it should be able to switch a 13A, 230VAC resistive load (under half the rated power) without the contacts welding. The relay is shit. It's most likely a much lower rated unit, with a false rating printed on it and needs to be replaced with a decent brand.

[floobydust]

I hear your doubt and invite you to measure the inductance on some heating elements, it's not zero.
The heating element has inrush current and inductance, especially if coiled, or near magnetic metals (ferritic stainless steel).
Also include the appliance wiring inductance.My point is people think it's a "resistor" and the relay arcing says otherwise.

Interrupting a 13A 230V load with 10's uH inductance will easily make a few kV and flash the relay contacts coating.
This is what I find happens.

[james_s]

You don't need inductance to arc when dealing with sufficiently high voltages. When I was a kid I made a crude arc furnace using a pair of carbon rods from D batteries with an electric iron in series from 120V. A heating element makes quite a good ballast for sustaining an arc without tripping the breaker.

[NiHaoMike]

240V double pole contactors are about $15 at a local HVAC parts store. Or get two single pole contactors plus a high current rectifier diode to bypass one in order to allow selection of full or half power.

[Zero999]

Again, the point is being missed and a lot of irrelevant information is being posted, by people with the best of intentions. 

Please read the relay data sheet, before posting anything about arcing, inductance etc. The relay is specified to switch slightly inductive load, cos?=0.4, at a reduced current and voltage. Even if the heating element is slightly inductive and the circuit is under-damped, it will not be that inductive. A quality relay, specified to switch a 25A resistive load at 250VAC, will be able to withstand the arcing created by a very slightly inductive 13A heating element. Another thing to note is that it's a water heater, so the inrush current will be negligible, because the resistance of heating element wire doesn't change that much, between 5^oC and 100^oC.

http://www.sanyourelay.ca/public/products/pdf/sfk.pdf

I repeat, arcing, inrush current and inductance are all irrelevant here and discussing them will not help. The solution is to buy a decent relay.

240V double pole contactors are about $15 at a local HVAC parts store. Or get two single pole contactors plus a high current rectifier diode to bypass one in order to allow selection of full or half power.

Yes, that's a much more sensible solution, even if it's a little overkill, you can be sure it'll be reliable. Be careful about half wave rectification of large loads though, as an upstream transformer won't be happy, although I suppose 3kW isn't big for a distribution transformer.

[richard.cs]

How often are you switching the water heater? I don't see that mentioned anywhere. Relays do have a cycle count limit which can be exceeded fairly easily if you are using the relays to control power (as a very-low frequency PWM) rather than just turning the heater on and leaving it. One switching event per minute is over half a million per year.

[floobydust]

Arcing aside, we could argue the reasons a relay can have a short life.

Safety standards for heaters, boilers, furnaces etc. have an endurance test in IEC 60730 part two's. It's typically 100,000 cycles the product is operated.

So OP's hot water heater should've already had a decent relay- not this crap <25,000 cycle part.

[drussell]

Arcing aside, we could argue the reasons a relay can have a short life.

Safety standards for heaters, boilers, furnaces etc. have an endurance test in IEC 60730 part two's. It's typically 100,000 cycles the product is operated.

Yeah, and that relay is specified for something like 10m cycles mechanical (heh.. heh..) and 100k electrical, according to the datasheet.  (And we all know that the datasheets for cheap Chinese electromechanical devices are always totally truthful and never embellished at all, of course...  : )

So OP's hot water heater should've already had a decent relay- not this crap <25,000 cycle part.

I am assuming, by the OP's phrasing, that HE has installed this relay, not that it was a part of the original apparatus.  This is why he seems to need a more robust relay.  (Not that manufacturers don't tend to often use poor quality components from the factory, but that is another issue, I suppose....)

[floobydust]

Another point about power relays is their contact resistance increases with age.

Ref. Fujitsu FTR-K3 datasheet which is compatible with OP's part but rated 100K cycles. Contact resistance is in milliohms; graph labelled wrong.
I've used this model relay, they are very compact but can't get rid of their own heat. Then I saw the lifetime curves and foresaw the product's eventual meltdown 
An old relay >50K cycles has triple the contact resistance and triple the (contact) heat. The PCB ends up being a heatsink, looking at thermal IR pics.
I would see 60-100°C operating temperature as the norm; so I design with larger parts now.

Power relays in the 16-25A class, a coil is 0.4W-1W and adds 20-50°C temperature rise plus heat from the contact-current adds another 10-15°C.