Autotransformer output clipping

[GoldDigger]

Seems a little tenuous to me.

This, for example, is my voltage at home:
...
I've seen this and worse on numerous sites unrelated to ferro-resonance.

But have you seen a "good" waveform at the input to two identical transformers and a distorted waveform and lower than expected voltage on the unloaded secondaries of those same transformers?
That is the specific condition we and the OP are trying to understand.
AFAIK the OP has not been able to check the loaded output of the transformer(s), although that sounds like a promising thing to try.

Oh, and is there anything which may be related to the fact that these are autotransformers (boost)?

 

[Besoeker]

But have you seen a "good" waveform at the input to two identical transformers and a distorted waveform and lower than expected voltage on the unloaded secondaries of those same transformers?
That is the specific condition we and the OP are trying to understand.
AFAIK the OP has not been able to check the loaded output of the transformer(s), although that sounds like a promising thing to try.

Oh, and is there anything which may be related to the fact that these are autotransformers (boost)?

As I said in post #11, I don't have an answer to the problem.
Neither saturation nor ferro-resonant effects really convince me.
For saturation to occur would need the voltage to be higher than design rating.
Ferro resonance would normally improve waveforms as in the CVT.

 

[GoldDigger]

As I said in post #11, I don't have an answer to the problem.
Neither saturation nor ferro-resonant effects really convince me.
For saturation to occur would need the voltage to be higher than design rating.
Ferro resonance would normally improve waveforms as in the CVT.

Unless specific (expensive) design steps are taken to deal with it, a CVT will improve voltage stability at the expense of high harmonics

 

[Besoeker]

Unless specific (expensive) design steps are taken to deal with it, a CVT will improve voltage stability at the expense of high harmonics

They used to have a reputation for that.

 

[LMAO]

It's more likely BECAUSE there is no load on the transformer. With no load, the transformer magnetizing current is the only thing that will show up and it will distort the reading of the voltage because once the core is energized, it can go into a state of ferroresonance. If you put a load on the autotransformer, the magnetizing current become such a small percentage of the total that any ferroresonant effect it has on distorting the voltage waveform gets buried in the overall effect of the load. If you need help to fall asleep at night, download some of the copious papers on "ferroresonance in unloaded transformers" that you will get if you Google that specific phrase.

But bottom line, don't worry about it unless you are for some reason going to leave that transformer unloaded and are concerned for power quality on the output side....

I got the same result with load; load was anywhere from 300 to 1000A (rectifier).

 

[GoldDigger]

I got the same result with load; load was anywhere from 300 to 1000A (rectifier).

A rectifier load is a very non-linear load and can create high harmonics in the current waveform (normal) and moderate harmonics in the voltage (lower percentage and caused by the impedance of the transformer and its corresponding voltage drop reacting to the current waveform.)

 

[LMAO]

A rectifier load is a very non-linear load and can create high harmonics in the current waveform (normal) and moderate harmonics in the voltage (lower percentage and caused by the impedance of the transformer and its corresponding voltage drop reacting to the current waveform.)

yep, I know that. A lot of harmonics are introduced by connecting a 6 pulse rectifier but the clipped peaks did not change; it just got worse (because of harmonics caused by loaded rectifier). Still does not explain the clipped peaks. I have sent the waveforms to manufacturer and asked for explanation. My guess is saturation.

 

[Besoeker]

yep, I know that. A lot of harmonics are introduced by connecting a 6 pulse rectifier but the clipped peaks did not change; it just got worse (because of harmonics caused by loaded rectifier). Still does not explain the clipped peaks. I have sent the waveforms to manufacturer and asked for explanation.

Good move.
Will you let us know the results?

 

[ggunn]

Perfect, thank you.
It is nice to see the explanation with accompanying pictures. ("And circles and arrows and a paragraph on the back...")

Aha! I thought you might be an old fart like me!

 

[gar]

131002-2043 EDT

To repeat. A simple normal single phase transformer core DOES NOT saturate at the peak of the applied primary voltage. dPhi = e dt. Integrate this to get Phi and you see that maximum flux occurs at the voltage zero crossing. Only when e changes polarity does Phi start to drop from its peak.

This is easy to see with a scope setup to display primary voltage and current as two traces synchronized to the voltage zero crossing. The peak current, indicating saturation, occurs at the voltage zero crossing. If you look closely you can see this at the second current pulse in photos P6 and P7 at my site http://beta-a2.com/EE-photos.html .

This saturation occurrence has nothing to do with whether or not the transformer is an isolation or auto type.

There is something missing in the description in the original post. At this point I have no idea what is missing in that description that causes the distorted output voltage waveform.

.

 

[Jraef]

I got the same result with load; load was anywhere from 300 to 1000A (rectifier).

OK, well, theory blown then.

 

[gar]

131003-0830 EDT

LMAO:

Core saturation is not your problem in a simple transformer. Go back to your books. With sine wave excitation to the primary the core flux with low excitation levels is a negative cosine. The peak flux occurs at the excitation voltage zero crossing. Increase the flux level and you can see that maximum flux, the cause of saturation, occurs at the input voltage zero crossing.

The distorted output voltage is somewhat similar to what occurs with a secondary load consisting of a diode, capacitor, and resistor to drain charge off of the capacitor between cycles. But this load starts to draw current before the voltage peak, and stops just after the voltage peak.

To test whether a load would eliminate your distortion you would not use a diode rectifier load. You would use a resistive load. A diode capacitor resistor load creates a problem directly in the region in which you already have a problem. In your yesterday afternoon post you mentioned a 6 diode rectifier as your load. Is this transformer you are talking about a three phase transformer?

.

 

[steve66]

131003-0830 EDT

LMAO:

Core saturation is not your problem in a simple transformer. Go back to your books. With sine wave excitation to the primary the core flux with low excitation levels is a negative cosine. The peak flux occurs at the excitation voltage zero crossing. Increase the flux level and you can see that maximum flux, the cause of saturation, occurs at the input voltage zero crossing.

The input voltage crossing zero would occur at the same time the output voltage peaks. The input and output are not in phase - at the peak input, di/dt is zero, and the output voltage is zero. Conversley, as the input crosses zero, di/dt is max, and the output voltage peaks.

I think you are simply seeing that the non-linearity of the BH transfomer curve. I don't know if I would call it saturation - that occurs at very high values of flux. But even at lower values, there is a smaller amount of non-linearity.

I don't think it is anything that you have to worry about.

 

[Besoeker]

The input voltage crossing zero would occur at the same time the output voltage peaks.

That infers a 90deg phase shift.
I don't think an autotransformer would do that.
Nor a double wound Dd either.

 

[GoldDigger]

That infers a 90deg phase shift.
I don't think an autotransformer would do that.
Nor a double wound Dd either.

Magnetizing current in the coil is 180 degrees out of phase with applied dV/dT. Induced voltage is in phase with dI/dT. So the back EMF is in phase with and opposite to the applied voltage, and the induced voltage in the "secondary" part of the autotransformer is in phase with the voltage applied to the "primary" part. QED.

Another simple way to look at it is that the transformer will try to minimize the magnetic field resulting from the primary and secondary current together. That requires that the two currents be in phase, but with opposite resulting flux density.

The problem people seem to be having is that the magnetizing current is not in phase with the applied voltage, but the load-related current IS.
That is why an unloaded transformer presents a nearly 100% reactive impedance, but gets closer to a PF of 1 the more heavily it is loaded.

 

[Besoeker]

The problem people seem to be having is that the magnetizing current is not in phase with the applied voltage, but the load-related current IS.
That is why an unloaded transformer presents a nearly 100% reactive impedance, but gets closer to a PF of 1 the more heavily it is loaded.

I mostly agree with that. The load current is reflected back to the primary and normally not in phase with the magnetising current so, as a rule, power factor on load is better than at no load.
If the load on the transformer is a bunch of cage motors with no PFC it might never be better than 0.8 pf particularly if the motors are running at below rated capacity.

 

[gar]

131003-1952 EDT

GoldDigger:

Magnetizing current lags applied voltage by close to 90 deg. Going somewhat into saturation causes a large current peak at the voltage zero crossing, and at the peak point of the magnetizing current.

.

 

[GoldDigger]

131003-1952 EDT

GoldDigger:

Magnetizing current lags applied voltage by close to 90 deg. Going somewhat into saturation causes a large current peak at the voltage zero crossing, and at the peak point of the magnetizing current.

.

Which is the zero crossing of both the input and the output voltage. Right.
Now, what effects if any will hysteresis have? Those will happen at or near the voltage peak, where the flux is near its own zero crossing, yes?

 

[gar]

131003-2122 EDT

GoldDigger:

Which is the zero crossing of both the input and the output voltage. Right.

Yes. I was only correcting your comment of 180 deg.

Now, what effects if any will hysteresis have? Those will happen at or near the voltage peak, where the flux is near its own zero crossing, yes?

Hysteresis is basically a power loss and therefore can be viewed as a resistive load, but not linear. Its current component will be in phase with the applied voltage.


steve66:

You said the output voltage was 90 deg shifted from the input voltage. This is just not true for a closely coupled transformer. The magnetizing flux is shifted approximately 90 deg from the applied voltage. This flux in turn induces a voltage in the primary coil that opposes the applied voltage, and in turn limits input current. Called a counter EMF. If a second coil, the secondary, is linked to the same flux and wound in the same direction, then the induced voltage is in phase with the primary voltage.


LMAO:

What is special in your autotransformer that you have not described? Is it really just an autotransformer? That is a single magnetic core with a single winding and one tap somewhere along the winding, where the flux density is relatively uniform throughout the core.

.

 

[GoldDigger]

131003-1952 EDT

GoldDigger:

Magnetizing current lags applied voltage by close to 90 deg. Going somewhat into saturation causes a large current peak at the voltage zero crossing, and at the peak point of the magnetizing current.

.

Read what I said again. Current lags voltage by 90 degrees. And voltage lags dV/dT by 90 degrees. So current lags (or leads, flip a coin) dV/dT by 180 degrees.
For an ideal capacitive load, on the other hand, current leads voltage, but is exactly in phase with dV/dT.

My statement may not have been well chosen for clarity for non-mathematicians, but I believe it was correct.

In addition to causing a dissipative load, hysteresis also causes the current versus flux curve to be non-linear (and two valued!) in the hysteresis region. That has to have some effect on the behavior of an unloaded transformer.