Slightly OT- Electrical

It depends on how small small is....And how electrically "stiff" the supply wiring is. If those transformers are in the 5 to 15 watt size range I wouldn't expect you'd get a spike any worse than that from a switching a 1/2 HP motor on and off.

If you're really worried, you can always wire a "whole house" surge arrestor across the circuit involved.

Jeff

Unlikely as the contactor's contacts will arc over until a current zero occurs(every 8.33 msec. on 60Hz). At that point, the arc will go out and the circuit will be broken.

Randy

The 50 small transformers will act as if they are one large transformer providing the same output as the total of the 50 little ones for such things. Also, what are you thinking of as a line spike? Transformers are more designed to minimize the inductance of the windings as this just sucks power.

-- Bob May Losing weight is easy! If you ever want to lose weight, eat and drink less. Works every time it is tried!

The 'real short answer' to this question is yes.

Jim

================================================== please reply to: JRR(zero) at yktvmv (dot) vnet (dot) ibm (dot) com ==================================================

The short answer is yes, you most likely will have a spike on both sides of the transformers as the field(s) collapse. To quantify this would take a lot more analysis and involve a lot of information I've forgotten years ago. The nature of the spike would depend on the exact phasing of the transformers, the type of load, the transformer ratios, and at exactly what point in the AC cycle the power was interrupted. Pretty complex problem...

Bob

But if you say that the L(di/dt) rule holds, you simply measure the inductance of the transformers (in parallel) and figure that worst case it will switch at peak current flow. Then figure that di/dt will happen in less than 1/100 power line cycle or so, that should give a rough guess for the induced voltage.

Anyone who's ever powered up a large solenoid and then rapidly switched it off while touching the terminals will know that this can be hundreds of volts.

Jim

================================================== please reply to: JRR(zero) at yktvmv (dot) vnet (dot) ibm (dot) com ==================================================

You cannot collapse a magnetic field instantaneously, which is what is attempted when you interupt power to a transformer. Generally you will get a small spark when disconnecting an inductive load like a transformer or motor. It's caused by the stored magnetic field releasing its energy.

Dave

The next question is likely to be how much of a line spike and how can it be prevented. The size of the line spike will vary with the Open Circuit Inductance of the transformers and where in the cycle the current is interupted. There are various ways to limit the spike. One of the simplist is to put a MOV across the line in parallel with the transformers.

Dan

Yes, both from the transformers and the contactor. Any time you break an inductive circuit, the energy stored in the magnetic field has to go somewhere.

I'm not sure how relevant it will be, but here's a link to an application note I wrote on the subject:

One of the other things that goes on as the field collapses is that the current appears to the transformer as a much higher frequency. At much higher frequencies, the coupling through the transformer changes from being predominately magnetic to capacitive, and its apparent transformation ratio can alter, sometimes significantly.

I learnt this the hard way by nearly being killed by a poorly designed high voltage test transformer, whose low voltage side developed about 5000 volts when the high voltage flashed over an insulator on the test object. Subsequent testing of the test transformer showed a radical alteration in its transformation ratio when excited by the kind of high frequency that can happen during the field collapse.

Another problem you may get is inrush at switch on. While the arcing at the contactor will slow the rate of change when turning off, at switch on you will get what ever point on the wave you hit, and you may get some contact bounce as well just to confuse things further.

Having said all that, my experience is all with much bigger transformers, your set up may work fine. Try it and see, just wear your peril sensitive glasses when you do it. If you run into trouble, a trick is to break the transformers up onto a couple of seperate smaller contactors, and have each contactor turn on the next one, to give a slight delay so they don't all spit the dummy at the exact same instant. Another is to put a small value capacitor between the 110 close to the transformers, the 60 Hz won't go through it, but the high frequency current will find it an easy path and this will limit the voltage rise. Voltage rating and type of capacitor is important.

When these types of circuits get exciting, sometimes the only way to 'see' what is happening is with a signal generator and a cro.

hope this helps

regards,

John

Bit difficult to be specific because you don't specify the power level of the circuit or the size of the spike that may bother you. Just guessing, since you're working at the 110VAC level, the load is probably a fairly small number of kilowatts. Multiple paralleled transformers sounds like drive for a forest of low voltage quartz halogen incandescent lamps

If this is the case you are very unlikely to have any problem.

50 small paralleled transformers will be less demanding to switch than a single large transformer of the same total power rating. This is because multiple small transformers are less efficient and have higher resistive losses. This means that the worst case peak inrush current is reduced. The switch off voltage transient will be a bit smaller for the same reasons although there will not be a major difference.

If you are driving a multiple incandescent lamp load the tranformers will be so heavily damped by this resistive load that the switch off voltage transient will be small. The worst case would be a switch off at zero input voltage (i.e a magnetising current maximum) with all transformers unloaded. However there would be far less energy in this spike than when switching a motor of similar power rating.

There will be the usual switch on current transient caused by the lamp cold resistance - about six times normal full load current. Unless the there some unusually stringent spike requirements normal lighting contactor switching should be all that is needed. Although some spike generation is inevitable the energy level is unlikely to be high enough to bother properly protected user equipment. Any further protection needed is better placed at the user end.

Jim

Not just *a* spike, you'll get multiple ringing spikes as the current surges back and forth among all those windings. But the spikes will dampen very quickly due to the lighting loads on the transformer secondaries.

On the hot side of the contactor, you'll get a line spike equivalent to an inductive switched load which is the sum of all the transformer loads. In other words, if each transformer is rated at 15 watts, 50 of them will give a line spike about equal to what you'd get switching a 1 hp motor.

Any sensitive equipment upstream of the contactor should have adequate surge suppressors installed. But then that should be true even if you aren't switching LV lighting on and off of that circuit. Normal household power has a lot of nasty surges and spikes. The situation you present is not much, if any, worse than the sorts of spikes you get from normal air conditioning and appliance switching.

Gary

Keep in mind these transformers are spaced apart. The chance of everything kicking back in phase and in time sync is pretty remote.

Kevin Gallimore

As long as they're wired right in parallel, they'd have to be several miles apart physically to have them not be in phase.

Jim

================================================== please reply to: JRR(zero) at yktvmv (dot) vnet (dot) ibm (dot) com ==================================================

=====

The phase of the spikes generated generated by the collapsing field, Jim. It will be a short ringing spike, and the time skew between the individual spikes will be significant.

Kevin Gallimore

If you are switching transformers, (inductances), there will be 'spikes` unless you switch the at the `0` or 'crossover` point. If the individual loads are small enough there are solid state switches available which can do this and be electronically controlled, (ganged), to use instead of the relay. Is the total load big enough for you to worry about the spike? If its all on one 120V. circuit, its probably not big enough to be a problem. If on multiple circuits, a very rough idea can be had by comparing total transformer KVA to motor horsepower.

Pragmatist- "It won't fit? - Use a bigger hammer!"

In this era we are ruled by a George II. Doesn't history have a fine sense of irony?

The propogation delay is around a nanosecond per foot, so given that the spikes are probably around microsecond risetime they could still be separated by considerable distance and still all add up.

Jim

================================================== please reply to: JRR(zero) at yktvmv (dot) vnet (dot) ibm (dot) com ==================================================

That THHN insulation must be a real good dielectric. Better than polystyrene foam coax.

so

Can't be given. What are you using for a model?

Perhaps. Perhaps not. But the only way to determine an answer would be to measure it, so it _might_ be anything. Your experience says the spikes of voltage will probably add, my experience says they probably won't. It depends on the characteristics of an undefined system. I'm sure we are both right for the system we think it is. :)

Kevin Gallimore

Quite right. The nS per foot is not a bad approximation for open wire line which is what the the zip cord is. Sure it might be two or three but it's going to be real close.

My bet is that the trouble will add if they're within ten feet of each other.

Jim

================================================== please reply to: JRR(zero) at yktvmv (dot) vnet (dot) ibm (dot) com ==================================================