# component to limit inrush current in a transformer

#### strantor

Joined Oct 3, 2010
5,369
I have a 480V>120V control transformer that keeps tripping a 6A breaker (installed on the primary) when I turn it on. I installed a temporary fast acting 6A fuse to make sure the breaker is not faulty and the fuse blows, so I know inrush current is >6A and breaker is good. If I hold the breaker in for a couple of seconds it will stay on and I read 1.5A with clam meter, so steady state current with maximum load is well below the breaker trip point. I need something to limit the Inrush current. I was thing a NTC resistor, sound good? I found this one but I'm not sure it will work. does not specify a max voltage, but mentions capacitance at 440V. datasheet says "Thermal Time Constant 194 seconds"; does that mean it will take >3min before it is effectively a short?

I've never used one of these and if someone could recommend me one that would be great. thanks

#### praondevou

Joined Jul 9, 2011
2,942
"The thermal time constant for a thermistor is defined as the time required for a thermistor to change 63.2% of the difference between the initial temperature of the thermistor and that of its surroundings when no power is being dissipated by the thermistor. The value of t defines a response time for the thermistor when it has been subjected to a step change in temperature. For example, a thermistor that has been in an ambient temperature of 25°C for a period of time long enough for it to reach equilibrium, is then moved to an environment where the temperature is 75°C. The thermistor will not immediately indicate a resistance corresponding to the new temperature but rather will exponentially approach the new resistance value."

2 Ohms sounds as if it is not enough. Is there a way you could measure the inrush current?
Also, will you be turning the breaker on and off several times in short intervals? In this case the NTC doesn't fulfill its function, because its body temperature has already increased / resistance decreased when you turn it on the second time.

#### praondevou

Joined Jul 9, 2011
2,942
Also, if the load is already connected and its a SMPS with a simple fullbridge rectifier on the AC input this could contribute to an increased inrush current when its primary side cap is discharged.

#### strantor

Joined Oct 3, 2010
5,369
" The thermal time constant for a thermistor is defined as the time required for a thermistor to change 63.2% of the difference between the initial temperature of the thermistor and that of its surroundings when no power is being dissipated by the thermistor. The value of t defines a response time for the thermistor when it has been subjected to a step change in temperature. For example, a thermistor that has been in an ambient temperature of 25°C for a period of time long enough for it to reach equilibrium, is then moved to an environment where the temperature is 75°C. The thermistor will not immediately indicate a resistance corresponding to the new temperature but rather will exponentially approach the new resistance value."
got it, thanks!
2 Ohms sounds as if it is not enough. Is there a way you could measure the inrush current?
I don't have any fast acting fuses between 6A and 25A. I have a logging meter but It's fused @10A and I don't want to risk it. my clamp meter is too slow as mentioned; so I really don't know how I would measure that spike. any suggestions?
DC resistance of the transformer primary is 1.5Ω; so 2Ω would more than double it. What resistance would you think proper for this application?
Also, will you be turning the breaker on and off several times in short intervals? In this case the NTC doesn't fulfill its function, because its body temperature has already increased / resistance decreased when you turn it on the second time.
No, once it's on it will stay on for hours or days. It's just that every time it gets turned off one day, and then on the next day, I have to go hold the breaker in for a sec. only a minor inconvenience, but it's not proper and I would prefer it be proper.

#### strantor

Joined Oct 3, 2010
5,369
Also, if the load is already connected and its a SMPS with a simple fullbridge rectifier on the AC input this could contribute to an increased inrush current when its primary side cap is discharged.
There's all kinds of stuff on the secondary; several PID heater controllers (probably each have their own internal SMPS), pilot lights, et. al.

#### praondevou

Joined Jul 9, 2011
2,942
I have to go hold the breaker in for a sec. only a minor inconvenience, but it's not proper and I would prefer it be proper.
After doing this a few times it can also damage the breaker, so it's better to find a solution.
Here you find some info on how to scale the NTC. But I really think you need to know how much you want to damp it and therefore you need to know how big and how long is the current spike.
http://www.epcos.com/web/generator/...operty=Data__en.pdf;/PDF_Applicationnotes.pdf

Well , I would go with a Hall sensor and an oscilloscope to measure it. If no Hall sensor, do you have an oscilloscope? Then you could at least put a low-ohm power resistor in series (for a short moment) and measure the voltage drop on it. But beware of the ground of your oscilloscope in case it's not isolated.

#### gootee

Joined Apr 24, 2007
447
How about a big MOSFET in parallel with a current-limiting resistor, probably in series with the supply or ground rail on the secondary side, with a simple control circuit that doesn't turn the MOSFET fully on until the inrush subsides?

#### praondevou

Joined Jul 9, 2011
2,942
How about a big MOSFET in parallel with a current-limiting resistor, probably in series with the supply or ground rail on the secondary side, with a simple control circuit that doesn't turn the MOSFET fully on until the inrush subsides?
One Power MOSFET is not enough to switch AC. They have a diode in parallel which would conduct every second halfwave. We also don't know what has more influence on the current spike, the transformer or the load, so we don't know if it's sufficient to put it on the secondary side.

In general the idea is good, limiting initially with a resistor or NTC and then closing a contact in parallel with the resistor/NTC, which might be a relay (classic or solid state), or any other suitable device. I attached a document giving more info on AC switch design.

However, I still think to correctly scale the NTC/resistor one would need to know the maximum peak current at the input.

#### Attachments

• 67.1 KB Views: 42

#### gootee

Joined Apr 24, 2007
447
One Power MOSFET is not enough to switch AC. They have a diode in parallel which would conduct every second halfwave. We also don't know what has more influence on the current spike, the transformer or the load, so we don't know if it's sufficient to put it on the secondary side.

In general the idea is good, limiting initially with a resistor or NTC and then closing a contact in parallel with the resistor/NTC, which might be a relay (classic or solid state), or any other suitable device. I attached a document giving more info on AC switch design.

However, I still think to correctly scale the NTC/resistor one would need to know the maximum peak current at the input.
Yes, I forgot to mention that it would have to be just downstream of the rectifiers in a DC power supply. But I guess that wouldn't work in the OP's case because he has multiple other power supplies after the transformer, if I understood correctly.

Yes, the peak inrush current would need to be known, preferably with the worst-case magnetization of the transformer due to the timing of the previous power off with respect to the AC phase at that moment.

Joined Jul 7, 2009
1,583
Why not measure the inrush current and get a circuit breaker that's rated for the inrush current, but protects in the longer term as the circuit demands? Your local electrical supplier should be able to help size the thing.

#### praondevou

Joined Jul 9, 2011
2,942
Why not measure the inrush current and get a circuit breaker that's rated for the inrush current, but protects in the longer term as the circuit demands? Your local electrical supplier should be able to help size the thing.
Well, he WAS measuring it with fast acting fuses, from what I understood.
I agree, first the inrush has to be known, then you can take the appropriate measures....

#### strantor

Joined Oct 3, 2010
5,369
Why not measure the inrush current and get a circuit breaker that's rated for the inrush current, but protects in the longer term as the circuit demands? Your local electrical supplier should be able to help size the thing.
I would assume (*I did't design the machine) that the breaker is already sized to protect the circuit. I'm not sure if it would be ok for me to put a larger one in.

I have an idea; what If I put a time delay relay on the secondary side? That would give the transformer core time to saturate before trying to power the loads on the secondary. Right now I think the problem might be that the core is saturating at the same time as all the caps on the secondary are trying to charge. Think that would work? is my logic correct? I already have a time delay relay.

tomorrow I'm going to try hooking a current clamp to a fluke scopemeter and see if I can capture the peak inrush current.

#### awright

Joined Jul 5, 2006
91
The manufacturer of the NTC thermistor you provided the link to: <http://www.ametherm.com> has lots of tutorials and FAQs on surge protection NTC thermistor selection. I'd suggest you peruse them. They also offer free samples.

awright

#### jaclement

Joined Apr 15, 2009
34
When you turn on the breaker/switch the voltage across the transformer primary can be quite high on the sine curve,which can cause a transient condition that can saturate the transformers primary. If you have a zero crossing switch, this will never happen.

#### praondevou

Joined Jul 9, 2011
2,942
When you turn on the breaker/switch the voltage across the transformer primary can be quite high on the sine curve,which can cause a transient condition that can saturate the transformers primary. If you have a zero crossing switch, this will never happen.
@jaclement: That's not correct, because current is lagging

http://www.opamp-electronics.com/tutorials/inrush_current_2_09_12.htm
Especially the paragraph below the first picture and below the 3rd picture.

#### jaclement

Joined Apr 15, 2009
34
Looks like you should turn the the circuit at the peak.

#### THE_RB

Joined Feb 11, 2008
5,438
@jaclement: That's not correct, because current is lagging
...
Doesn't current lag because of the existing magnetic field in the transformer?

At turn on there is no existing field. There will be zero AC Vin and zero current into the transformer winding.

As for a simple solution, the OP could connect a resistor with a high PTC factor (like a jug element) in series with the primary.

Last edited:

#### gootee

Joined Apr 24, 2007
447
Some of you seem to be thinking partially in terms of steady-state operation's characteristics. Startup is a little different. Did no one else read the nice link that praondevou posted?

And no one has mentioned that it can also be heavily dependent on the conditions at the exact moment when the tranformer last had the AC power REMOVED. There is a worst-case core-magnetization condition that must also be taken into account.

#### praondevou

Joined Jul 9, 2011
2,942
Doesn't current lag because of the existing magnetic field in the transformer?

At turn on there is no existing field. I think you should connect the AC mains to the transformer at the zero voltage (ZC) point... There will be zero AC Vin and zero current into the transformer winding.

As for a simple solution, the OP could connect a resistor with a high PTC factor (like a jug element) in series with the primary.
did you have a look at the link I posted?

I can't remember it very well, because it's been a while, but I did a test on exactly this phenomen, and as far as I remember I was surprised seeing a bigger inrush current when the transformer was connected at zero voltage. Maybe someone should test this again...