Breaker tripping when load is switched off

[David_AVD]

An associate has a system whereby a relay switches the primary of a large (240V 1800W) toroidal transformer.  The transformer is powering a mostly resistive load.

There are actually 2 relays; one switches on first and has a few NTCs in series to limit the turn on surge.  After about a second, the other relay pulls in and bypasses the first relay (and NTCs).  After another second the first relay drops back out again.   This works fine when switching on.

The issue is when the "full power" relay turns off ("soft start" relay is already off).  Occasionally this causes the breaker in the switchboard to trip.

I believe there is a 150R resistor in series with 100n X2 series cap across the switched output (A - N).  Nothing across the relay contacts.

Can anyone think of what might be the cause?

[tautech]

Sounds like back EMF to me?
Is the "breaker" a MCB or RCD type?

Edit
Maybe an MOV needed?

[HackedFridgeMagnet]

I believe there is a 150R resistor in series with 100n X2 series cap across the switched output (A - N).  Nothing across the relay contacts.

Is that contradictory?  Maybe just restate or draw a diagram.

What nominal currents are the relays, and the breaker current too.

[NiHaoMike]

If it's one of those new AFCI breakers, it could very well be false triggering on the arc in the relay contacts.

[David_AVD]

To clarify, the R/C network is effectively across the transformer primary.  There is nothing across the relay contacts.  Maybe it was meant to be?

The arc upon opening is what I'm suspecting is the cause.  The inconsistency could be due to the random nature of the turn-off vs mains cycle.

EDIT: There is a 275V MOV across the incoming active - neutral.

[SeanB]

Place another MOV across the relay contacts, or a series snubber of 22R 0.5W and 0.22uF class X capacitor. Likely arcing of the contacts is doing this, with the very large toroid acting as an almost pure inductor.

I had the same with my variac until I did this, as it is also a toroidal transformer of around 2kVA. Turn on surge would trip the 30A mains breaker.

[David_AVD]

Thanks.  Will give that a go.

[ConKbot]

If it's one of those new AFCI breakers, it could very well be false triggering on the arc in the relay contacts.

+1 to sounding like garbage AFCI shenanigans

[David_AVD]

Apparently the 220n + 22R snubber had no effect.

Placing a 275V MOV across the relay contact did reduce the incidence of breaker tripping, but did not eliminate it.

I could ask if they can change the breaker for another type, but not sure what type would be best?

[NiHaoMike]

One that is not an AFCI, assuming building code does not require it. If code does require it, try a different brand of AFCI. The burst of noise could also trip a GFCI that doesn't have sufficient filtering.

You can also try adding a LC line filter before the relay.

[David_AVD]

Both good ideas.  I'll advise my associate.  Thanks.

[David_AVD]

Some more to the story:

The tripping of the MCB/RCD breaker has been shown to be on the over-current portion.  The RCD indicator flag on the MCB/RCD is not active after the breaker trips.

A camera was set up to record the event and it showed that the breaker tripped just (< 1 second) after the thermostat turned the relay (and transformer primary) off.  This only happens once every 10 - 100 switch off events though.

There is currently a 275V MOV across the incoming mains supply and another across the relay contact.  There is also a 220n + 22R series R/C across the load (transformer primary).

Originally there was a 100n + 150R series R/C snubber across the load.  The addition of the 220n + 22R did *seem* to help but since the trips are not deterministic it's hard to say for sure.

So, how does one calculate the optimum R/C values for a given transformer inductance?

[tautech]

So, how does one calculate the optimum R/C values for a given transformer inductance?

That leads to the question how does one calculate how large the back EMF is.....

Your snubber C value needs increasing IMO.
Double it.
What would you have to loose?

[David_AVD]

Do I need to know the value of the back EMF?  Aren't the snubber values calculated from the supply voltage and load inductance?

[tautech]

Do I need to know the value of the back EMF?  Aren't the snubber values calculated from the supply voltage and load inductance?

But the Back EMF value is related to AC sine wave switching timing and back EMF is infinitely variable with a relay doing the switching.
Ever changing, thats why your tripping is random.
Other alternative is to replace with zero crossing switching, then the back EMF should be a more constant value, but then you will still have to protect the Silicon doing the switching.

So snubber value must be large enough to suppress the worst case senario in the current setup.

[David_AVD]

I did find the formula : C = I² / 10



The steady current of the transformer is around 7.5A so that comes out to approx 5.6uF - much larger than I've been thinking.

If I'm reading the 2nd formula correctly, R comes out as about 0.4 Ohms assuming 240Vac and 7.5A load current.

[tautech]

Had a chat to my sparky mate re your problem.
He had encountered this sort of thing before.
2 issues to consider:

Degradation of the MCB after many trip cycles, a common problem he encounters.
To the point MCB's will not hold 50% of their rated load.

Replace MCB with a "Motor Rated" type. C or preferably D class.
That was his usual fix as they are much more tolerant to overload and spikes in his experience.

[David_AVD]

Thanks for that.  I did say to them that a slower D type breaker may be worthwhile.  I'm not sure if you can get them as an MCB/RCD combo though, but it's certainly worth pursuing.

I tried a 5uF 450Vac capacitor across the transformer primary and the 'scope trace (current draw channel) does show the current dropping off faster after the relay opens now.

Some extended testing will happen over the next few days so we'll see what happens.