# Transformer switch off spike

From Wikipedia:

When a transformer, electromagnet, or other inductive load is switched off, the inductor increases the voltage across the switch or breaker and cause extended arcing. When a transformer is switched off on its primary side, inductive kick produces a voltage spike on the secondary that can damage insulation and connected loads.

How to avoid transformer switch off spike and protect the circuit?

Is it something that we should be worried about for low power designs (e.g. step down 2A transformer)?

• Not sure if this claim is correct. It could produce a voltage spike on primary side, but not so sure that the spike will be seen on secondary. Commented Jan 20, 2020 at 9:09
• @MarkoBuršič The inductive kick happens in the field so it affects every winding. That's how flyback power supplies work, for example. If there's a rectifier and filterig cap in the secondary the kick will be absorbed and nothing special can be seen. If the secondary hasn't a capacitor a damaging peak can occur.
– user136077
Commented Jan 20, 2020 at 12:30

How to avoid transformer switch off spike and protect the circuit?

It's a bit of a red herring for most loads because the induced voltage on the secondary is usually adequately "snubbed" by the load itself. This is how a flyback transformer works; energy is stored in the magnetic field of the core during the 1st half of the cycle and is released into the secondary on the 2nd half of the cycle.

So, a load would naturally "contain" the small amount of energy in the transformer's magnetic field should the primary become suddenly disconnected.

Having said all of the above, the wiki page linked in the question is entitled: -

Inrush current


So, given that the inrush current in a transformer can be significantly higher than its normal running current, some consideration should be made.

If the core is not saturating, the inrush current of the primary can be up to twice$$\^{\text{NOTE 1}}\$$ the normal-running peak magnetization current so, we should be wary of this doubling of current and note that if the supply were disconnected after half a cycle of the voltage being applied (that voltage being applied at a zero-crossing point for worst effect) we would witness a peak magnetization current twice as high as the normal peak of magnetization current.

Making a comparison is probably a good idea to put things into perspective. Imagine we have a 1 kVA 1:1 transformer supplying 250 V AC to a 1 kW load. The load current will be 4 amps. For a transformer of this size the magnetization inductance might be 10 henries. This implies that the "normal" magnetization current will be about 80 mA RMS. This has a peak value of 113 mA and, with the 10 henries inductance represents a peak stored energy of 63 mJ.

Given that the load (1 kW) is "handling" energy at a rate of 1 joule per millisecond I would say that most normal loads will survive the extra surge energy of 63 mJ (or 4 times that if the primary is disconnected at the worst possible moment during the worst case inrush scenario.

Is it something that we should be worried about for low power designs

I'd be concerned but not worried AND I would simulate the worst case scenarios to see what problems might occur. This is the power of simulation.

$$\\text{NOTE 1}\$$ - if the core saturates, then the inrush current can be very much higher than twice but, because the peak flux is clamped by the effect of saturation, it is sufficient to analyse the non-saturating situation (in my opinion).

• Is it good practice to always consider having a varistor in the design? Commented Jan 20, 2020 at 9:27
• A lot of series-connected NTC varistors are used in switch mode supplies and those circuits don't use a conventional transformer. Having said that you can use NTC varistors in conventional supplies if there is a possibility of the overall inrush surge current might trip a breaker. Considering the possibilities is good practice in all designs. Commented Jan 20, 2020 at 9:41
• I know note1 corrects the impression, but the basic effect of the words The inrush current of the primary can be up to twice is to suggest just what it says, twice. The inrush of a primary can be orders of magnitude greater than the normal magnetisation current, or even the normal running current, so I'd be inclined to remove the spurious ratio of 'twice' for clarity. The flux tries to go to twice during a switchon that creates inrush, the current can be huge. Commented Jan 20, 2020 at 12:30
• @Neil_UK I'll make it clearer. Commented Jan 20, 2020 at 12:43