Transformer Inrush Currents 4

[Vladpl]

Hello all and happy new year.

I have a problem which is inrush current.

The main problem I have that the transformer is being used as a power supply with mains outlets (you can plug in any mains powered equipment) the unit is 240V and rated at 10A. As you all know when you turn on a power supply / transformer there are very high inrush currents for a few cycles that can go up to 1000A in this case. This would all be ok if the hospital where this is going didn't install C-curve circuit breaker which trips when there are high inrus currents. Have they put curve-D breaker there would be no problem since it has motor start characteristics. And as you have guessed if you power up the power supply it will tripp the circuit breaker due to the high inrush currents.

One way this can be avoided is to put a resistor, capacitor and a relay and create a short timing circuit to stop high inrush currents.

My question is this.... Have any of had a similar problem and how did you go about solving it.

Thank you in advance for your help.

 

[Skogsgurra]

Sorry cbarn,

What you describe is the decaying transient DC component known from connecting transformers and generators. It does, as you say, amount to twice the steady state current and it does not blow fuses or trip circuit breakers.

The inrush current is, as I say, something else. It has its origin in the DC component, yes, but it needs a saturated core to develop.

As the OP said " The main problem I have that the transformer is being used as a power supply with mains outlets (you can plug in any mains powered equipment) the unit is 240V and rated at 10A. As you all know when you turn on a power supply / transformer there are very high inrush currents for a few cycles that can go up to 1000A in this case."

As you can see, a current being up to 1000 A in a circuit that is expecting (fused for) 10 A is something quite different from the transient DC component.

All this is classical knowledge and you can find good support for the saturation effect in many text-books on basic electricity. I googled "inrush current" and got this as link #1:

http://www.allaboutcircuits.com/vol_2/chpt_9/12.html

It clearly shows the saturation and the consequential high current.

I have nothing against a good discussion, but I do not think that it should be carried on in absurdum.

Gunnar Englund

http://www.gke.org

 

[davidbeach]

skogsgurra is correct. DC offset is DC offset and inrush is inrush, and while both may happen simultaneously, they are not the same thing and they happen for different reasons. DC offset is a mere doubling of the peak of the wave and leaves the waveform relatively sinusoidal. Inrush peaks 10x or more of the normal full load current and is not particularly sinusoidal. The inrush current will have an extremely high 2nd harmonic content and the peaks are highly distorted.

 

[cbarn24050]

Well skoggs I just read your link and it says exactly what ive been saying.

 

[Warpspeed]

One thing is for sure, a suitably designed air core transformer would not have any inrush problem, it is purely due to core saturation and the drastic reduction of inductance that this saturation causes.

An air cored choke will have a current rise proportional to instantaneous applied voltage and inversely proportional to inductance no matter what. It is the iron core of a mains transformer (and remnant flux) that is the problem.

A mains transformer with a design peak flux swing of less than half the peak saturation value will not have any inrush problem. It would handle the flux doubling phenomena with ease and quickly settle down to normal symmetrical operation over a few cycles. It would be a wastefully large transformer though.

 

[davidbeach]

cbarn24050, the link from skogsurra's post DOES NOT say what you've been saying. The material at the link is talking about the saturation driven inrush of a iron cored inductor; what you were talking about is the DC offset of any inductor, air core or iron core. It takes the saturation of the iron core to get an inrush current 10x full load current. The DC offset you are talking about can only get to 2x.

 

[Vladpl]

Hi All

Thanks for helping me out on this topic.

Yes the normal transformer that you can buy (not an ideal transformer which would cost an arm and a leg) has a nominal inrush current at 10x which occurs mostly when a transformer that has a load is turned on. The highest inrush current will occur when it is turned on anywhere on the positive part on the cycle.

The easiest way to overcome this would be to change the circuit breaker that can withstand the inrus ie motor start circuit breakers with curve D. Only thing is that this is for a Hospital and you would have to change thousands of circuit breakers and that is not an option.

At the moment I am looking into making a timing circuit or using a current limiting diode to lose few cycles and prevent circuit breaker from tripping unless any of you have any better solution.

Thank you all again for your inputs

Regards Vlad

 

[Skogsgurra]

Easy now!

How would you apply the current limiting diode?

If you are going to do what I think you are going to do, then don't do it! It will DC saturate your core very effectively.

Gunnar Englund

http://www.gke.org

 

[Warpspeed]

Vlad, is this just one transformer, or a whole series of transformers feeding for example, a class A electrical area ?

If you absolutely MUST use that transformer, try fitting a negative temperature coefficient thermistor. These large thermistors are commonly used to limit the high inrush into the cold filaments of large projection lamps. See if you can get one to try.

Cold they have a fair bit of series resistance, but they quickly self heat, and the resistance then falls quite low.

A diode is only going to make things much worse, don't do it !

 

[logbook]

NTC thermistors are great BUT if the transformer has been on for a while and then gets switched off and back on almost immediately the thermistor will still be hot and therefore will give no surge protection.

 

[davidbeach]

I have no idea what a C or D curve is, sound like an IEC thing, but come on; transformer inrush is not a surprise. Ideal transformer would cost more than an arm and a leg, they just don't exist. You could throw enough iron at the problem and make a transformer that didn't exhibit inrush characteristics, but that would cost more than an arm and a leg for most application and would not be a better transformer in the steady state condition. Somebody goofed and the situation needs to be corrected, and the only practical way of correcting it is to use a breaker that coordinates with the transformer inrush; you won't find a practical means of eliminating or limiting the inrush, although the more load on the other side of the transformer, the faster the inrush will damp out.

 

[Warpspeed]

Yes that is very true, but it is simple to do, and would improve the situation significantly.

O/k then. How about combining an NTC thermistor and a suitably rated "polyswitch" in series ?

It may take a bit of experimentation to get it all working satisfactorily, but a polyswitch would certainly open the circuit fairly smartly if the NTC was ever badly caught out with a hot restart.

I am sure the thermal time constants of NTC and polyswitch could be arranged so it did not repeatedly cycle, but automatically recovered after a suitable cool down delay for the NTC.

 

[JimCook]

"On power transformers, the industry standard is to assume the RMS inrush current is 25 times RMS full load current for .01 seconds and 12 times RMS full load current for .1 seconds. Control transformers can have peak inrush currents that are 100 times the full load RMS current for 1/2 cycle."

Quote by Vince Saporita, Bussman Fuses

 

[Skogsgurra]

I had an interesting mail only a few days ago. It is from a German guy named Michael (probably a forum member, but he didn't say). Anyhow, he pointed out that a large part of the inrush current in a toroidal core is dependent on remanence. I obviously had said somewhere that remanence was a minor problem - although I cannot find where I said that.

I have now done some experimentation. I used a toroid core, a fast recorder and current and voltage transducers. I also used an SSR to switch on/off in a controlled way.

The result is that remanence seems to be contributing to inrush current in about the same degree as the phase angle where you switch on. I had thought that remanence had less influence than that.

Good to know. Thanks Michael!

Gunnar Englund

http://www.gke.org

 

[waross]

Hi skogsgurra;
The e-mail may have been refering to your 14 Jan 06 5:27 post. I was left with the understanding that remnance was unimportant. Thank you for the correction and clarification.
Respectfully

 

[Skogsgurra]

Thanks Bill,

I was searching for "toroid" - and couldn't find it. So, I didn't say anything about toroids, actually. Now, do I have to measure EI cores as well? Sigh...

I will probably be back when I done that. But there is a holiday going on. And Sweden is playing Trinidad-Tobago right now. Have to help our guys there. ;-)

Gunnar Englund

http://www.gke.org

 

[Warpspeed]

Strip would grain oriented silicon iron toroids are absolutely the very worst thing conceivable for creating an inrush problem. They combine very sudden saturation with almost no air gap. Wound and cut U cores are nearly as bad. E and I laminations of soft iron are always much more friendly at turn on.

Remnance is the real demon here. With negligible air gap and very high permeability, the iron maintains a high remnant flux after the primary is de-energised. If you try to drive the flux higher in the same direction at turn on, and then hit hard saturation, the resulting current spike will be very spectacular indeed.

 

[Skogsgurra]

OK, Sweden didn't win that match. I think it is because I was doing some inrush current experiments instead of supporting them.

This is what an ordinary transformer looks like. I was running a slight overvoltage (250 V instead of 220 V). The periodic switching was done with an SSR and it was controlled in such a way that an integral, but odd numbers of half-periods were applied. The upper trace in picture below shows transformer primary voltage. Primary current is shown in bottom trace. Note that the last half-wave is negative. The first half-wave the transformer sees in the next burst is also negative. Remanence makes the inrush current very large as evidenced by the negative current spike. The inrush current decays to normal values in about two periods.

To verify that it is really the remanent flux that causes the current spike, I arranged a simple "automatic degaussing" by connecting a 2 microfarad/250 V capacitor across the primary. The SSR was still controlled by the same bursts with odd number of half-periods.

The influence of the capacitor can be seen at the end of the burst. In the first picture, there was a typical snap-off high voltage transient when the SSR stopped conducting. The voltage spike actually goes off scale. With the capacitor, things are a lot different. Instead of getting an overvoltage, we get a decaying oscillation which goes to zero in about 30 milliseconds. The decaying oscillation effectively degausses the core, so when we next time apply the burst (same polarity) we do not get any inrush current at all. Beautiful, I would say.

Actually, I think that this could be the remedy for the OP. Adding a capacitor of suitable size does bring remanent flux to zero and thus eliminates the inrush current.

I must admit that I was wrong in my first assumption. I have really learned from this - and I hope you did, too.

Gunnar Englund

http://www.gke.org

 

[Warpspeed]

That is lovely picture Gunnar.

What happens if there is an appreciable secondary load ? Is there enough Q left to get that same beautiful damped oscillatory rundown ?

 

[itsmoked]

Well you certainly are having fun with your new logger..

Wide pics I actually had to spread it across two monitors!

Nice demo.

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[Skogsgurra]

You guessed! Hmm.. How do I reduce the picture size?

Gunnar Englund

http://www.gke.org