wye wye transformer neutral impedance

[winnie]

Instead, I am taking Z (pri) = N^2*Z(sec) (thanks for the reminder; I knew that one but wasn't thinking down that path), and adding that when coils are on two separate legs only half the flux will couple from primary to secondary, essentially halving the turns ratio.

Of course this half coupling works both way. So I don't know what the turns ratio is in both directions.

Have to ponder more.

-Jon

 

[gar]

181213-1016 EST

winnie:

Do you have a small three phase transformer, possibly 1 kVA, so that it is easy to experiment?

.

 

[Sahib]

Agreed. But what I am trying to figure out is the what happens to the neutral on the secondary side. For simplifying I am assuming the primary line-line voltage is constant.

There is neutral shift on secondary side too.

 

[winnie]

There is neutral shift on secondary side too.

Yes, I know. Can you point me to a calculation of how much, in the context of a 3 leg common core transformer?

 

[winnie]

181213-1016 EST

winnie:

Do you have a small three phase transformer, possibly 1 kVA, so that it is easy to experiment?

.

Unfortunately no. The experiment will be run by someone else (grr.) using single phase excitation of a 30kVA transformer. If I had a 3 phase variac...

-Jon

 

[gar]

181213-1330 EST

winnie:

I an ideal transformer load current flux exactly counterbalances supply current flux. Thus, no net flux.

In a non-ideal transformer there is some leakage flux, and therefore some net flux.

.

 

[gar]

181213-1423 EST

winnie:

Some additional.

Do you need a Variac? That is do you need variable voltage?

My guess is that if you used a 1/4 supply voltage that this would keep you low enough so that magnetization and saturation were not major factors.

In a real transformer if there is no load on the secondary, then the impedance looking into the primary is determined by the primary shunt inductance, primary resistance, and core losses. So an unloaded transformer still presents some load.

From my previous post there is no coupling from transformer 1 to 2 or 3 from transformer 1 load current. But there is some because there is not complete flux cancelation in transformer 1. This flux difference is possibly in the few % range.

The leakage flux is that which is not contained within the magnetic core. Some of this will interact with any coil within its field. That field being outside of the core, but can interact with a coil that is around the core.

This may sound like nonsense, and possibly it is, but at this point I have not created a real intuitive model in my head.

.

 

[winnie]

Do you need a Variac? That is do you need variable voltage?

For operational use I do not need a Variac.

But for measuring the cross coupling between coils on different legs I want to measure open circuit voltage and short circuit current; and it seems that the safest way to do that is to reduce the primary voltage until the short circuit current equals the normal load current (as is done for measuring transformer impedance).

181213-1423 EST
My guess is that if you used a 1/4 supply voltage that this would keep you low enough so that magnetization and saturation were not major factors.

In a real transformer if there is no load on the secondary, then the impedance looking into the primary is determined by the primary shunt inductance, primary resistance, and core losses. So an unloaded transformer still presents some load.

Agreed.

I an ideal transformer load current flux exactly counterbalances supply current flux. Thus, no net flux.

In a non-ideal transformer there is some leakage flux, and therefore some net flux.

From my previous post there is no coupling from transformer 1 to 2 or 3 from transformer 1 load current. But there is some because there is not complete flux cancelation in transformer 1. This flux difference is possibly in the few % range.

The leakage flux is that which is not contained within the magnetic core. Some of this will interact with any coil within its field. That field being outside of the core, but can interact with a coil that is around the core.

Hmm. That particular transformer model assumes 'infinite' core permeability and thus no magnetizing current. The model that I use has the 'magnetizing branch' which produces flux in the core, and then the load current produces flux which is canceled by the corresponding 'load supplying current' in the primary. The way I imagine things, the AC applied to the primary creates an AC magnetic flux in the core. Load current on the secondary acts to reduce this flux, which demands more current in the primary to maintain the flux. So you have 'load' and 'supply' flux terms that cancel out, and a 'baseline' flux term that is there because the transformer is energized.

-Jon

 

[gar]

181213-2259 EST

winnie:

A simple experiment that can be performed is:

1. Take and EI transformer with both primary and secondary windings on the center magnetic leg.
2. Add a turn or more around the coils. Or get a plate transformer with a filament winding, or some other small transformer where you can get essentially two secondaries. All secondaries on the same magnetic leg as the primary.
3. One secondary will serve as a test pickup coil. This will have no load except a voltmeter.
4. The other secondary will have a variable load applied.

This will allow us to see how the test coil voltage varies with loading of the transformer. I expect it will drop approximately as source voltage to primary - voltage drop across primary resistance and primary inductance varies with primary current from secondary load.

.

 

[Sahib]

Yes, I know. Can you point me to a calculation of how much, in the context of a 3 leg common core transformer?

One thing I wish to state as a starting point is there would be no difference in their performance for common or separate core transformers with same KVA and voltages, if they have same percentage impedance.

 

[winnie]

One thing I wish to state as a starting point is there would be no difference in their performance for common or separate core transformers with same KVA and voltages, if they have same percentage impedance.

What is the basis for that statement? There is clearly zero coupling between the cores of the separate transformer set, and must be some coupling between the core limbs of the common core transformer, and the implications of that coupling are exactly what I am trying to understand.

I can clearly describe test situations where the two versions (3 separate cores versus a common core) will give different results; for example if I energize a single coil on the primary I will measure open circuit voltages on all 3 of the secondary coils in the common core configuration.

However with the constraint of a wye connected primary that has no neutral connection, there is no way to energize just a single core.

Thanks
Jon

 

[winnie]

181213-2259 EST

This will allow us to see how the test coil voltage varies with loading of the transformer. I expect it will drop approximately as source voltage to primary - voltage drop across primary resistance and primary inductance varies with primary current from secondary load.

That fits with what I expect. The test coil has essentially no current flow, and is measuring the open circuit voltage created by the magnetic flux created by the magnetizing branch of the primary...and the voltage across the magnetizing branch drops as more current flows through the primary resistance.

-Jon

 

[Sahib]

What is the basis for that statement? There is clearly zero coupling between the cores of the separate transformer set, and must be some coupling between the core limbs of the common core transformer, and the implications of that coupling are exactly what I am trying to understand.

I can clearly describe test situations where the two versions (3 separate cores versus a common core) will give different results; for example if I energize a single coil on the primary I will measure open circuit voltages on all 3 of the secondary coils in the common core configuration.

However with the constraint of a wye connected primary that has no neutral connection, there is no way to energize just a single core.

Thanks
Jon

If common and separate cores transformers have same KVA, voltages and percentage impedances, their neutral shift would also be same for unbalanced loads with neutral disconnected. It is just a matter of design and POCO sometimes replaces a bank of 3 single phase ttranformers with a single 3 phase transformer or vice versa and they do not seem to face any issue.

 

[winnie]

Okay, I finally found a reference on topic:

https://apps.geindustrial.com/publibrary/checkout/GET-3388B?TNR=White Papers|GET-3388B|generic

From GE, titled 'The Wyes of the Wyes'.

In particular:

The first problem discovered was that in three-phase banks of single-phase units and in three-phase shell type units, although the line voltages conformed to turn ratio, the line-to-neutral voltages were not 58% of the line voltages but about 68% at no load and diminished very rapidly when the bak was progressively loaded line to neutral.
[...] The line to neutral voltages of these banks had about 60% third harmonic component. This explained the overvoltage at no load[...]
The three-phase units with three-legged cores were vound to behave quite differently in these matters. Their third harmonic voltages were negligible (according to the standards of those days), and the line to neutral voltages were practically 58%, and their voltage regulation for loads to the neutral, though poor, was not altogether intolerable.

It goes on to describe the impedance in terms of positive and zero sequence impedance. There is an entire chapter on the three phase common core units. This is mostly written from the point of view of utility distribution, where line capacitance and its interaction with the transformer inductance is a huge concern.

-Jon

 

[Sahib]

:thumbsup:

 

[kwired]

Not the schematic where the coils are drawn with their phase angle, but instead the actual position of the coils on the steel core.

For common small transformers you have 3 coils on 3 'legs' with the legs joined top and bottom with cross pieces. One leg is in the center, and presumably couples differently to the outer legs than the outer legs with each other.

I've simply decided to test this using a single phase variac and measuring output voltage (and short circuit currents) on the various terminals.

-Jon

Thanks, makes sense now.

 

[gar]

181214-2501 EST

winnie:

An experiment with a Stancor P8668 Transformer. Secondary rated 28 V at 2 A. No load infinite, load 25 ohms. At this time of day primary voltage is moderately stable, and I did not monitor it.

Used #24 tinned copper wire to make two additional secondaries for voltage measurement only. Used a 10 megohm meter, Fluke 27. Two 1 turn coils were created, one around the center spoke, and the other around an outer spoke. This transformewr is an EI core with the primary and secondary wound over the center spoke.

Results:

Secondary no load was 32.4 and loaded 31.0 . This is a 4.3% change.
Test secondary on center spoke was no load 244.5 mV and loaded 240.1 . This is a 1.8% change. 42% of the load secondary.
Test secondary on an outer spoke was no load 121.9 mv and loaded 119.7 . This is a 1.8% change.

About 1/2 half of the transformer series internal impedance is associated with the primary as is expected and verified by the 42%.

These tests correlated with what would expect theoretically.

I think these 1 turn test coils could be useful to you on the 3 phase transformer.

I never had a course in either physics or EE that performed this type of test.

.

 

[winnie]

gar,

Thanks for your test on the 'side legs' of the transformer. They were in line with my eventual understanding of the interaction between all of the coils of a 3 leg 3 phase transformer.

In any case, I can report the results of the actual test. I only got minimal data since we were looking for a 'go/no-go' on the neutral impedance.

The unit being tested was a 30 kVA 208 delta to 480/277V wye three leg dry type transformer.

The 208V delta coils were reconnected in star, using the 190V taps. When energized we saw roughly 194V l-n and 335V l-l.

We used a 47 ohm resistor to place a single phase l-n load on the transformer. Using a simple clamp meter we measured 4A of current.

So we saw roughly a 3% voltage drop with about 8% loading. This is pretty high impedance, and would not be acceptable for running l-n loads. However we only have l-l loads, and the impedance is low enough for grounding the secondary. The transformer feeds test equipment via an RCD system that trips at 1A, and we confirmed that the neutral impedance is low enough to reliably trip the RCD through our grounding resistance.

-Jon