What happens when you overvoltage transformer primaries

[disorder]

I was recently working on an amp that has a switching module that selects between 120/240 voltages and I tried hot switching the line voltage from 120 to 240 without powering down the amp in between and I blew up some protection varistors. My question lies in the fact that these varistors see 240VAC all the time and in this scenario they BLEW THE FRIG UP, loud pop, spark, and more fireworks after that. An educated guess points to the "auto line voltage selector" not acting quick enough and the transformer primaries (that should be wired in series for 240V operation) were still wired in parallel leading to core saturation. When you overvoltage transformer primaries is there some sort of physics at play that puts a high voltage/current back onto the primary which would toast these varistors (MOV type) ??

[eschertron]

My completely uneducated guess is that the switch broke contact while switching, and the line protection devices saw a couple of big spikes. I wouldn't want to bet my star fleet academy scholarship on that, though.

[R.G.]

When you apply a vastly too-high voltage to a transformer primary, the core saturates on the peaks - well, all of the AC waveform significantly above the usual peaks. Say, everything above 150Vac for a 120Vac rated trannie. That is guess about the voltage level, and each trannie will be a little different, but it's representative. What "saturation" means is that the iron atoms are all aligned and cannot any longer enhance the inductance of the coil charging them, and so the coil looks to the outside world like the primary has suddenly changed from many henries of inductance to millihenries. That in turn means the AC core magnetizing part of the incoming current goes from being limited by a many-henry inductor to being limited only by the resistance of the wire itself, and current skyrockets.

Note that it's not putting a high voltage/current BACK into the primary, it's just letting the AC wall current pour through un-limited by the normal primary inductance. Same effect though. Current skyrockets. Probably overvolts the varistors, which are also probably rated lower, and they overheat and pop. They can die another way, too. When the peak of the incoming AC overvoltage passes, when the voltage gets back UNDER saturation, the inductance comes back up and tries to limit the current again. But the leakage inductance in the coil has already been charged up, so it puts out a big flyback pulse of voltage to try to keep the overcurrent flowing. That could pop varistors too - and maybe puncture insulation in some cases.

Yeah, it's believable that the varistors suddenly became smoke- and light-emitting.

It's also possible that the PT did not overheat enough to be damaged - but you wouldn't want to do this kind of thing a lot.

[g1]

I'm a little confused about the equipment here.
I'm assuming it's a switchmode supply that automatically senses the incoming line voltage?
And somehow you are able to switch your line voltage from 120V to 240V (can't do that where I am, care to comment ).
Have I got it right?
If so, I would guess the switchmode supply is sensing the voltage at power up, and adjusting accordingly, but is unable to make the adjustment after power up.

Edit: Upon re-reading, I see you are talking about transformer primaries so I guess this is a linear supply, not switchmode. I was not aware they made linear supplies with auto voltage selections. What is the unit? I guess those varistors saved the transformer?

[R.G.]

I took "switching module" to mean "a manual switch that picks 120 or 240, and lets you choose" like the ones on PC power supplies before they got robust enough to run on either 120 or 240.

I didn't think it was auto switching.

[Alan0354]

I don't think he meant switch mode power supply, just a switch to select between 110V and 220V.

Where is the protection diode. A schematic would help.

[g1]

has a switching module that selects between 120/240 voltages...
"auto line voltage selector" not acting quick enough

Sounds like some kind of automatic switching to me.
Maybe he'll return one day with more info.

[Enzo]

If it has an "auto line voltage detector" why would it also have a line voltage switch? If it is an SMPS, there are two kinds. One uses jumpers to convert between 120 and 240v input, and sometimes they do just mount a switch. But in that case, it would not be autodetecting in any way. Alternatively there are plenty of SMPS that are 'universal" in that they will work off of anything from 100v to 250vAC. As far as my experience goes, those do not really detect the incoming voltage, they just have the range in their controlling circuits to do the job.,

So on the the OP. I wonder if perhaps he is confusing types. Certainly we might find a switchable SMPS, but then it would not have a power transformer and primaries. Not in the sense we have been discussing.

[gui_tarzan]

The first thing that popped into my mind was why would you switch to begin with unless you're using a 240v source?

[disorder]

Some clarification: the DUT has a circuit that detects incoming line voltage and properly sets relays such that the transformer primaries are either in parallel or series. The quick/instant jump from 120 to 220 was done through a (probably illegal) switch box that has my 120 and 220 line coming in and selects one of the two to spit out to my power cord. Upon switching from 120 to 220 there were no relay clicks heard (these relays are loud I would have heard them, I SHOULD have heard them switching to put the transformer primaries in series but I heard nothing) and also there was about four seconds before the varistors blew up.

RG, going on what you said regarding the trans core hitting saturation... The varistors are in parallel with the transformer primaries so wouldn't it be the windings that see the current, not these guys? My understanding is the varistors are for large voltage transients from line voltage spikes and decrease their resistance as the voltage across them incresses. They aren't current affected in the way that thermistors or fuses are in this application.

[g1]

Thanks for the clarification. My guess would be that once the relays are latched, the power would need to be disconnected before the sensor circuit would re-sense the voltage.
The sense circuit would be designed for power-up, I don't think a voltage event like you created could occur in real life, so they wouldn't need to design around that. A little more delay in your line voltage switching mechanism may have done the trick.
Hopefully R.G. will comment on the varistors, but without a schematic it may be difficult.

Edit: If you didn't hear the relays, then you did have 220 going to the 120V primary for those 4 seconds, and the varistors probably did save the PT.

[R.G.]

Ah. I understand better what you did now. G-one is probably right - the varistors likely *did* save the transformers.

...Upon switching from 120 to 220 there were no relay clicks heard (these relays are loud I would have heard them, I SHOULD have heard them switching to put the transformer primaries in series but I heard nothing) and also there was about four seconds before the varistors blew up.

It's possible the high current stuck the relay contacts together. That happens sometimes. But it's just about as likely that the input switching setup does only work at power up, as noted.

RG, going on what you said regarding the trans core hitting saturation... The varistors are in parallel with the transformer primaries so wouldn't it be the windings that see the current, not these guys? My understanding is the varistors are for large voltage transients from line voltage spikes and decrease their resistance as the voltage across them incresses. They aren't current affected in the way that thermistors or fuses are in this application.

No. Varistors come in all sizes and voltages. Varistors *may* be for big spikes, but the most common varistors are rated for just a bit over the peak voltage of the AC power line. A varistor for use on a 120Vac line might be rated for the equivalent of 150Vac, for instance.

They do not decrease resistance in any smooth way, although long ago, early varistors did. Today, they snap over like zeners and start conducting heavily when their voltage is exceeded. They're like bi-directional, high power zeners. So if the switching you did wound up suddenly putting 220 on the paralleled 120Vac primaries, the varistors would in fact start conducting heavily. This current would be in addition to the currents pulled by saturation of the primary windings. I imagine they did blow up.

In any case, I think you've discovered another thing to write in the "don't do this" section of the instruction manual.

[disorder]

This makes total sense now. As always, thank you for the help guys! Looking at the circuit now I see that there is ONE varistor across each winding so the relays were still set to have the windings in parallel and because of this EACH vari saw ~220-240VACrms and became toast. RG you're explanation of varistor function explains why they were so badly damaged.

I knew something bad might happen when I did this so I was a safe distance from the DUT and my incoming voltage box on my bench is fused on both lines along with being on a smaller 10A breaker. Still dangerous yes but this is how I learn sometimes

Ah. I understand better what you did now. G-one is probably right - the varistors likely *did* save the transformers.


It's possible the high current stuck the relay contacts together. That happens sometimes. But it's just about as likely that the input switching setup does only work at power up, as noted.

No. Varistors come in all sizes and voltages. Varistors *may* be for big spikes, but the most common varistors are rated for just a bit over the peak voltage of the AC power line. A varistor for use on a 120Vac line might be rated for the equivalent of 150Vac, for instance.

They do not decrease resistance in any smooth way, although long ago, early varistors did. Today, they snap over like zeners and start conducting heavily when their voltage is exceeded. They're like bi-directional, high power zeners. So if the switching you did wound up suddenly putting 220 on the paralleled 120Vac primaries, the varistors would in fact start conducting heavily. This current would be in addition to the currents pulled by saturation of the primary windings. I imagine they did blow up.

In any case, I think you've discovered another thing to write in the "don't do this" section of the instruction manual.