Switch on surge.

[Dave Plowman (News)]

A pal has a light fitting comprising 15 20 watt 12 volt lamps. Fed from a pretty massive toroidal transformer in the base. After having had a new consumer unit fitted, it tripped the MCB. Every time. It's on a 6 amp lighting circuit.

The sparks who installed the new CU doesn't want to know - he was sort of supplied by the people who fitted a new bathroom with some form of whirlpool bath - and they insisted it had to be upgraded. I've a feeling they also installed/split the lighting load into two - it's a small house and originally only had one circuit. My pal is a bit vague about such things.;-) The fitting has been replaced by a single bulb pendant which works ok. Other clue was the wall switch sparked when this fitting was switched on.

My guess is the inrush current is simply too great for a standard MCB. I have the fitting to check out - but no bulbs with it as yet, as he couldn't find them.

Anyone have an idea what the peak current demand would be this sort of arrangement? I could obviously just fit a type C MCB - but am curious about the sparking light switch too.

There's plenty of room in the base for two or more electronic transformers - but altering the wiring inside the fitting to split the load might not be so easy.

[The Natural Philosopher]

Massive. Toroids are very efficient magnetically. So a few turns of thick wire is all they are using..

Leakage inductance if they are under load is not huge So there is piss all to stop massive inrush currents.

Things that may help are:-

Uprated MCB to a 'slo blo' type. I have no personal experience. I hope others here do.

A thermistor in series, that starts high resistance, gets hot and goes low. We used to use these occasionally on big power amps. For exactly this reason.

Oddly, long thin leads to the LV bulbs. These have resistance which limits the cold resistance to a minimum value. I found this by accident here - I have three lights fed from a remote transformer: They take a second to come up to brightness.

The last two solutions will limit the peak cold load on the thing, and reduce inrush current massively. Its not unusal to see partial core saturation when driving massive loads..for a brief instance. That screws even the leakage inductance.

There may be a slow start switch using either a zero crossing switch or a dimmer type triac..again, I know its theoretically possible, but don't know more than that.

Essentially you have to find some way of limiting peak current, or coping with it. I prefer the former.

I'm not sure those will be better. Toroids are very tough good things. Once the fact that they ARE sodding efficient has beentaken out of the equation..

[Rob G]

A useful and interesting reply.

I was under the impression that one of the significant advantages of the electronic transformers was that they were slow start. I've got a

4 light track with an electronic transformer and subjectively that comes on noticeably slower than the mains driven GU10 bulbed one I had there previously and blew bulbs regularly. The fact that the 12v system hasn't blown any bulbs I put down to the transformer being slow- start.

Rob

[Dave Osborne]

Humongous. Particularly if the source impedance of the supply is low and the tranny is connected on a short cable from the consumer unit.

I could obviously just fit a type C MCB

Yes. this would be my first port of call. You may find that a C type still trips out, in which case you should consider ditching the toroidal for electronic transformer(s).

[Andrew Gabriel]

Inrush is inherent with toriodal transformers for two reasons; they have no air gap, and they are made with different type of steel core. Transformers made with EI stampings always have an airgap where the E and I stampings meet, however small. The core material for mains frequency toroids has to be something which can be coiled up, and the type of steel used retains magnetism from the moment the transformer was last switched off, which interferes with the next switch-on.

No air gap means the transformer has virtually no tolerance to DC current. When you switch on, for the first 100th of a second, before the AC mains gets to reverse polarity, the transformer is effectively presented with DC, and current is limited mainly by the resistance of the primary windings. It actually takes a few mains polarity reversals before the transformer is behaving as though it has an AC supply, so a typical toroidal transformer might draw something in the 20-50A region for the first half cycle, which decays over the following few half-cycles.

Secondly, the stored magnetism in the core left from when the transformer was last switched off means that the core is quite likely to saturate at switch-on until the AC field counteracts it (and I suspect this is what makes the no air-gap problem take a few AC cycles to clear). This varies each switch-on, depending on the polarity and strength of the field left at switch-off verses the polarity of the supply at the instant of switch-on.

This element of the inrush is nothing to do with any load on the secondary -- the toriod doesn't even generate any significant secondary output until the primary inrush is decaying down, and the primary inrush will happen even if there's no load. However, a load such as LV filament lamps, when they start getting power from the secondary, will generate a load surge too, which may be the straw which breaks the camel's back after the toriodal primary inrush surge. The toroidal inrush surge and its decay as the transformer starts working properly actually results in the secondary voltage ramping up quite slowly (compared with an EI transformer), and that probably goes some way to limit the filament inrush.

[John Rumm]

Many of them are...

Even without the slow start xformer they will still seem a little slower than the mains equivalent since the filament is thicker and has more thermal mass.

It certainly helps. however 12V lamps are far less fragile than the mains halogens, and tend to last much better regardless of how you power them. They give a better light quality as well.

[John Rumm]

I had to help diagnose a trip problem with moderate sized RF power amp once (10kW)

It would occasionally trip its at least one of its 40A MCBs (one on each phase). This had your more conventional EI style transformer (although it was used as a step-up transformer and was fairly large - probably about 3' tall) to generate the 10kV required to run the main valve. In order to power it up, it had to do a "stepped start". using a pair of contactors under software control. The first would apply mains, but with an inline dropper resistor to limit inrush. This however would get hot rather quickly, so a second contactor would short it out 500ms later. Even so we still got the occasional trip. We did various measurements with a clamp meter hooked up to a storage scope, and found some quite substantial inrush currents could occur even with the stepped starting. Much depending on where in the mains cycle the waveform was when the on switch was thrown. IIRC they had to go up a circuit breaker rating, and a letter to guarantee you would never get s switch on trip.

[meow2222]

Toroidals are famous for bad surges at start. Simplest solution, if posible, is to replace the MCB with a fuse. Other possibles are:

- move the fitting onto a 32A circuit (and provide thermal protection if lacking)

- fit a 3 position switch to power it. The halfway position supplies the toroidal via a filament lamp to limit initial surge

- run it from an isolating transformer, the resistances of which will limit i.

- run it via a 10v step up transformer and dropper or choke. etc etc

NT

[Dave Plowman (News)]

[snip]

Nice info Andrew - but doesn't help with the problem. ;-)

[Dave Plowman (News)]

Would that conform to 17th edition regs - the CU does.

It's a hall light.

Two way switched too.

Thanks for the most impractical suggestions this far. ;-)

[Andrew Gabriel]

Change to an electronic transformer. Better regulation, dimmable, etc.

I have a 100W toroid to which I fitted a NTC thermistor, but this isn't the sort of modification I would suggest someone does unless they are fully familiar with designing safe mains appliances. e.g. a thermister in series with a toroidal transformer could in theory burn out and sustain an arc for long enough to ignite nearby materials with the transformer acting as a ballast and preventing any fuse blowing, and you need to design a solution which removes the fire risk from that failure mode.

[Dave Plowman (News)]

Just looking at TLC, the largest they do is 150 watts. And with 15 20 watt lights that doesn't compute for two of them. And as I said it might be difficult to rewire it to split the load.

[meow2222]

yes, but your 6A circuit will become 5A.

and...

I assume youve heard of relays

Hardly difficult. 15x20w = 300w = 1.25A at mains, so it would require a 12v 15w transformer. If thats too challenging you may as well give up.

NT

[The Natural Philosopher]

And the connection between cause and effect is plain wrong.

[The Natural Philosopher]

I cant quite get my head riound it, but I am sure there is something comprising a big triac, and a diode capacitor and a few components that acts as a 'startup dimmer'

Also a zero crossing switch is possible, but again memory fails me...

[Charles Fearnley]

I've added this link to a soft start circuit for toroidal PSUs on power amps - so should do what is needed technically. However, it may be over complicated for what you want, and especially in terms of getting new circuitry into the enclosure of the lamp.

F

[Dave Plowman (News)]

Right. I thought MCBs were mandatory these days. A quick glance at TLC doesn't show them available for modern CUs.

Are you suggesting re-wiring the entire lighting circuit in 4mm and using a 32 amp MCB? Can't you see this is slightly impractible? For a start the fitting wouldn't take cable of that size - and in any case it wouldn't conform to regs as the only protection on a radial is the MCB, etc.

Well as I said it's not my house. And I've no idea if there's room for a bodge like that.

You've lost me there. Where does 12v and 15 watts come in? The fitting takes 300 watts.

>

[Dave Plowman (News)]

relay connected directly across the supply would provide a long enough delay to prevent the MCB tripping? After all, they don't operate immediately. I do have some 10 amp 240v coil relays lying around and there would be room for one plus resistors inside the 'base'. Making up the timer section would be too costly, time wise.

[meow2222]

Fuses are still fully compliant, but they are out of fashion. MCBs are more about convenience than safety.

no

Lots of plug-in transformers have no fuse. The mains plug fuse protects as far as the transformer itself, and the thermal cutout protects from there on. The same principle can be applied here. Whether a higher current feed is available nearby in this case only you can tell us.

Its effective, safe, reliable, cost effective, and the finished item looks fine. Where's the problem?

OK, first we increase the mains voltage by 12v. This is done with a

12v transformer. Then we add a dropper to bring it back to 240, ie a capacitor, choke or less likely a resistance. The point of this is a lot of series limiting reactance has been introduced, and the switch on surge is thus much reduced. 300w is 1.25A roughly, so you'd need a 12v 1.25A transformer.

I suspect another option would be simpler though: ignore the built in toroidal and run it off an external EI transformer rated 12v 300w.

NT

[dennis]

Have you thought of wiring a soft start dimmer in the supply? Just leave it on max and hide it somewhere. Make sure it can handle inductive loads though.