3PH transformer blows breaker on power up

[ptonsparky]

211209-1427 EST

FOG1:

You need to study ferro magnetic theory to be able to understand why you see what you describe.

Ferrous magnetic cores in combination with a coil have an electrical characteristic that is known as hysteresis. This shows up as the magnetic flux in the core being different at a certain current input to the coil compared to some other time for the same current.

Where you are on that curve at the time you disconnect power will determine where you stop on the hysteresis curve. Where you are is defined by prior history and when current becomes zero. When you open that switch could be any time. If current is present at the time of switch opening, then by some path ( possibly arcing at the switch contact ) current will continue to flow until stored energy in the magnetic field is dissipated. The energy in the magnetic field has to be expended.

If the switch is next closed at a time that tends to increase core flux in the same direction, then the core will be driven way into saturation compared to an ordinary cycle. This results in a large current pulse. The peak of this pulse occurs at the next voltage zero crossing.

The characteristics of this pulse are in part a result of the shape of the magnetic core material. Going to a core material that increases the efficiency of the transformer probably results in higher magnetizing current at design operating levels.

To experimentally measure magnetizing current probably needs a scope, and experimental efficiency is increased, if you detect the residual flux level and control turn on time of the next test.

Experiments on a small power transformer, 175 kVA at 120 V, produced a peak magnetizing current of about 60 A, but not very often. Full load peak current would be about ( 175/120 ) * ( 1/0.707 ) = 1.46 * 1.414 = 2.06 A. So 60/2.06 = 29. In other words on the first half cycle the peak input current could be 29 times larger than its steady state peak current. this would likely trip a Sq-D breaker.

.

Thank you.

 

[gar]

211209-2415 EST

At least one error in my last post. kVA should be VA. The transformer is small, less than 10#.

.

.

 

[FOG1]

211209-1427 EST

FOG1:

You need to study ferro magnetic theory to be able to understand why you see what you describe.

Ferrous magnetic cores in combination with a coil have an electrical characteristic that is known as hysteresis. This shows up as the magnetic flux in the core being different at a certain current input to the coil compared to some other time for the same current.

Where you are on that curve at the time you disconnect power will determine where you stop on the hysteresis curve. Where you are is defined by prior history and when current becomes zero. When you open that switch could be any time. If current is present at the time of switch opening, then by some path ( possibly arcing at the switch contact ) current will continue to flow until stored energy in the magnetic field is dissipated. The energy in the magnetic field has to be expended.

If the switch is next closed at a time that tends to increase core flux in the same direction, then the core will be driven way into saturation compared to an ordinary cycle. This results in a large current pulse. The peak of this pulse occurs at the next voltage zero crossing.

The characteristics of this pulse are in part a result of the shape of the magnetic core material. Going to a core material that increases the efficiency of the transformer probably results in higher magnetizing current at design operating levels.

To experimentally measure magnetizing current probably needs a scope, and experimental efficiency is increased, if you detect the residual flux level and control turn on time of the next test.

Experiments on a small power transformer, 175 kVA at 120 V, produced a peak magnetizing current of about 60 A, but not very often. Full load peak current would be about ( 175/120 ) * ( 1/0.707 ) = 1.46 * 1.414 = 2.06 A. So 60/2.06 = 29. In other words on the first half cycle the peak input current could be 29 times larger than its steady state peak current. this would likely trip a Sq-D breaker.

.

Thank you for your reply!

 

[winnie]

I am going to discuss another aspect of inrush which adds to Gar's point.

The AC voltage applied to the primary of the transformer creates an alternating magnetic flux in the core. A key aspect of this is that the instantaneous voltage is not directly connected to the flux. Rather the applied voltage is proportional to the rate change of flux.

If you were to apply even a very low DC voltage to the coil the flux would climb higher and higher until the core saturated.
The alternating core flux is essentially the integral of the applied voltage. This works just fine because we apply an AC voltage. During the positive half of the AC cycle the flux climbs from the negative maximum to the positive maximum. During the negative half of the AC cycle the flux goes in the opposite direction from positive max to negative max.

And this is where we can see inrush saturation without any residual magnesium in the core. Imagine that the core is at 0 flux, but start the AC at a zero cross. Over the entire positive half of the cycle the flux is supposed to go from -max to + max. But instead it goes from 0 to beyond + max. Bang, saturation and inrush.

Take the same core but start the AC at a voltage peak and you get no inrush.

Thus the timing of the AC cycle plays a critical role in inrush magnitude, and you get minimum or no inrush if the applied AC matches the residual flux correctly.

Jon

 

[FOG1]

I am going to discuss another aspect of inrush which adds to Gar's point.

The AC voltage applied to the primary of the transformer creates an alternating magnetic flux in the core. A key aspect of this is that the instantaneous voltage is not directly connected to the flux. Rather the applied voltage is proportional to the rate change of flux.

If you were to apply even a very low DC voltage to the coil the flux would climb higher and higher until the core saturated.
The alternating core flux is essentially the integral of the applied voltage. This works just fine because we apply an AC voltage. During the positive half of the AC cycle the flux climbs from the negative maximum to the positive maximum. During the negative half of the AC cycle the flux goes in the opposite direction from positive max to negative max.

And this is where we can see inrush saturation without any residual magnesium in the core. Imagine that the core is at 0 flux, but start the AC at a zero cross. Over the entire positive half of the cycle the flux is supposed to go from -max to + max. But instead it goes from 0 to beyond + max. Bang, saturation and inrush.

Take the same core but start the AC at a voltage peak and you get no inrush.

Thus the timing of the AC cycle plays a critical role in inrush magnitude, and you get minimum or no inrush if the applied AC matches the residual flux correctly.

Jon

So is the design of the core on the newer transformer vs the older 75kva transformer the reason for the increase inrush current?

 

[winnie]

While it is _possible_ that the core design is causing the higher inrush current, IMHO there is still something wrong here. The much higher no load current for the smaller transformer suggests a different issue is involved.

The install is marginal, in the sense that the primary OCPD leaves very little room for inrush current. NEC permits primary OCPD well in excess of the transformer primary current rating, but your installation can't take advantage of this (you mention a 200A main). So you start with an install that has a high chance of tripping on inrush, and then add something else (as yet undefined) wrong and you end up with major startup issues.

Have you contacted the manufacturer to find out what the _expected_ no load current is?

-Jon

 

[FOG1]

Federal Pacific is like talking to a two year old..........they act absolutely clueless. They have been of no help at all. DMG which is the machine manufacturer who also has the transformers built and provided it, they have done some testing as of Wednesday this week. I do not know their findings. DMG's Chief electrical engineer believes there is something amiss with the transformer.

 

[winnie]

Given the information so far I'd also suspect the transformer. Something like a shorted turn or wrong turn count on one coil.

Confirming this one way or the other would require more detailed testing.

Jon

 

[gar]

211210-1045 EST

winnie has provided some good additional information. I also think there may be something fishy about the higher no load current.

Going back to an old textbook "Electric and Magnetic Fields, Attwood, Wiley, 1949 and to the Chapter 13, Ferromagnetism, one finds Fig. 13.8, where U.S.S.Transformer is a fairly soft type of curve, whereas 78.5% Permalloy is a fairly square shape. Thus, if core material U.S.S. Transformer is used in an older transformer over voltage to a coil is not going to rapidly increase magnetizing current, quite different for Permalloy.

Could be much more square loop materials are being used in newer transformers, and this particular transformer may be operating at slightly too high an input voltage.

.

 

[synchro]

I wonder if these transformers are wound with the primary on the inner windings and the secondary on the outer windings (instead of vice-versa as it's commonly done). I think that would make the construction more conventional, because in this particular transformer the taps are on the secondary, with jumpers on the primary. I think an inspection of the transformer should make it evident whether it's built this way or not.

If the primary is on the inner windings then that might accentuate the inrush current as a result of a lower leakage inductance, etc., just as when a conventional transformer is reversed (as MTW brought up earlier in this thread).

I think gar has a good point about the possibility that the transformer operating voltage may be toward the high end, so this should be checked and considered with the other available into. Because of the extra cost of higher performance core materials used to improve efficiency, perhaps the core is somewhat marginal in its cross section due to economic reasons, making it operate closer to saturation. The lack of primary taps to accommodate different input voltages, and the fact it's being operated at 240V vs. 208V also tends to make it more vulnerable to saturation.
Then again, their could be a defect like a shorted turn as winnie mentioned.

 

[gar]

211210-2041 EST

From post number 2 it appears the transformer is only 240 V. Since the transformer is being used in reverse mode it means the transformer should be more tolerant of a given input voltage. This is because a step down transformer is wound with more secondary turns than the theoretical value to account for full load voltage drop. Thus, when in reverse mode the core flux density will be lower, and less peak inrush current. Also the output voltage will be less than expected from the nominal voltage rating when run in the forward mode.

.

.

 

[jim dungar]

211210-2041 EST

From post number 2 it appears the transformer is only 240 V. Since the transformer is being used in reverse mode it means the transformer should be more tolerant of a given input voltage. This is because a step down transformer is wound with more secondary turns than the theoretical value to account for full load voltage drop. Thus, when in reverse mode the core flux density will be lower, and less peak inrush current. Also the output voltage will be less than expected from the nominal voltage rating when run in the forward mode.

Actually, a transformer run in reverse usually has a higher inrush, because of the placement of the energizing windings in relation to the magnetic core.

 

[drcampbell]

... Something is wrong when the lower kVA higher efficiency unit draws greater idle current with the secondary disconnected. ...

I don't think you can say that without knowing either both the real and reactive idle currents, or the total current at full load..

If the higher-efficiency transformer's primary windings have lower resistance and lower inductance, the reactive current at idle would be higher without indicating a problem.

 

[gar]

211211-0505 EST

Why would a higher efficiency transformer have a lower shunt inductance?

A higher efficiency transformer will have lower series resistance, possibly higher shunt inductance, but a transformer is not a linear device. In the design of a transformer how the core saturates has to be considered. The number of primary turns will be selected to avoid too much saturation of the core at maximum expected primary voltage.

A scope looking at excitation current would probably indicate the cause of the problem.

See my scope plot of magnetizing current of an unloaded transformer at P8 of:

BETATRONICS&reg, Electrical photos, communication, Long Cable, High Baud Rate, RS-232, fast, I232, vendor, supplier, manufacturer, industrial

Electrical photos.

beta-a2.com

The pulses in the waveform are the core going into saturation, and occur at excitation voltage zero crossings.

.
.


.

 

[gar]

211211-1058 EST

Now if I continue the experiment and add a shorted turn to the transformer I still have the saturation pulse plus there is a small sine added that is much more in phase with the applied voltage. Thus, looking at the unloaded transformer, and the current waveform in relation to the applied primary voltage you may be able to detect a shorted turn or turns.

.

 

[synchro]

gar makes very good points about the insights that can be gained by viewing the current waveforms on a 'scope.

Something that would take a little effort, but could aid in diagnostics, would be to temporarily configure the jumpers on the primary for 480V, and then measure the no-load current draw with your 240V supply. If the current drops significantly more than 50%, then that would be highly suggestive that saturation is occurring with the volts-per-turn applied when the jumpers are configured for 240V.

 

[gar]

211211-2111 EST

Looking back thru the various comments it appears the transformer is designed as a step-up.

I think the no load magnetizing current is too high, why? Possibilities, wound incorrectly or labeled wrong.

We need to see a scope plot of the magnetizing current for each primary winding separately with no secondary load.

For several years I have been unable to supply pictures to this site. Thus, it would be useful if others could provide typical plots for transformer inrush.

,

 

[MTW]

Here's the picture attachment

 

[gar]

211213-1526 EST

Some additional data on this A41-175-16 transformer magnetizing current.

As supply voltage changes the magnitude of the current spike changes, but the general waveform is about the same. Following are peak current values vs supply voltage:

Voltage .......... Peak Spike Current A

090 V RMS ...... near 0
110 V RMS ...... 0.48 A
120 V RMS ...... 0.60 A
130 V RMS ...... 0.90 A
140 V RMS ...... 1.20 A

Full load current for this transformer without magnetizing current is: 175/120 = 1.46 A RMS, or peak of 2.06 A.

So it is clear the peak magnetizing current at 120 V input is a large fraction of full load current. This is relatively illustrative of how power transformers are designed.

.

 

[Ragin Cajun]

Can you get a TTR test run on the transformer? TTR - Transformer Turns Ratio.