The beginnings and ends of phase windings 2

[zlatkodo]

Distance between the beginnings of three-phase windings is , generally, two thirds of the full pitch.
But this may not always be so, it depends on which shortened pitch is used.
For example: where are the beginnings for three phase winding, 33 slots, 8 poles, double-layer:

- the beginnings of the first, third and fifth pole-phase group or
- beginnings as shown in the attachment?
Which option is correct? Which is better?
Is somewhere I can find a detailed analysis of this topic?
Zlatkodo

 

[electricpete]

I’m not really understanding the attached diagram. I was expecting to see diamond coils in lap configuration, but it looks like something more complicated. Is this a wave winding diagram?

=====================================
(2B)+(2B)' ?

 

[zlatkodo]

It is not a wave winding. It is still lap winding. Diagram is from a book and it is also not my way of drawing a diagram. A new attachment has a little more details.
Zlatkodo

 

[koizumi]

"Distance between the beginnings of three-phase windings is , generally, two thirds of the full pitch.
But this may not always be so, it depends on which shortened pitch is used."
This is only for general estimation.
In fact, this distance depends on some factors:
1) Angle in electrical-degree in which you decide to pull out the leads (normally it is 120 degrees, but in some cases it may be 240, 360 degrees and so on).
2) The number of slot per one phase on one pole (referred to as q) and phase arrangement of windings.
3) In wound rotor of induction machines, the leads normally are arranged at 120 degrees in space (evenly distributed) for easy balancing.
4) Some other specific winding constructions.

In this example diagram, the parameter q is not a natural number.
Formula q=Z/(2.m.p) here you have q=33/(2.3.4)=11/8=1+3/8; this means each phase you have 8 poles with 3 having 2 coils and the remaining having 1 coil each.
And the electrical degree between the U-V and V-W leads here is 480 degrees.

 

[zlatkodo]

Koizumi,
Thanks for the reply.
Does your answer means that both versions are right?
Is there still some opinions from other experienced rewinders?
The main question is:
How to determine the beginnings of the three-phase windings with fractional q?
Please see the attached internal connection diagrams for both the aforementioned versions.
Zlatkodo

 

[electricpete]

I'm no experienced rewinder, but my two cents fwiw.

I think both diagrams in latest post give identical performance. If you trace the path from V1 to V2 and call current flowing "to the right" as V and to the left as V', then both connections give the same sequence:
U W' V U' W V' U W' V U' W V'....
And position of groups and number of coils per group is identical.

As far as whether some other arrangement of coil groupings (which have 1 and which have 2) might be better, the only way I know to check would be compute the distribution factor for all 3 phases using spreadsheet approach previously posted, and check which one gives highest and more balanaced distribution factor.

Those are good diagrams by the way.

=====================================
(2B)+(2B)' ?

 

[electricpete]

I will strengthen my last comment to say I am positive that the two connections shown give identical performance. The only thing the coil cares about is which phase (U V W) is flowing through it and which direction is the current flowing (unprimed or primed to right or left). Pick any coil and compare those two diagrams, the coil carries the same phase current in the same direction.

=====================================
(2B)+(2B)' ?

 

[electricpete]

I guess I should say they are identical with respect to performance (distribution factor, balance). One may involve easier connection or less copper than the other... not sure about that.

=====================================
(2B)+(2B)' ?

 

[electricpete]

Looks like all of the series jumped and pole jumpers would be the same... the only difference is the location where the T-leads are connected... I doubt that makes much difference in ease of construction or amount of material.

=====================================
(2B)+(2B)' ?

 

[electricpete]

Attached is a drawing marked up as described 10 Sep 10 15:26.

If you look at the designation for each group (U, U', W, W',V, V') and the number of coils in each group, I think there should be no question they are identical magnetically. (I am assuming that the external phases are defined by the letters, not by the colors.... which swapped... was that to trick us?)

I don't want to interfere with your thread. But after your question is answered and the discussion dies down, I would be interested to know for my own info how you came up with that particular sequence of 1-coil and 2-coil groups.


=====================================
(2B)+(2B)' ?

 

[koizumi]

To zlatkodo,

In your given winding diagrams (dated Sep 10), the upper diagram might not applicable. There are two reasons:
1) The electrical angle between the U1-V1 and V1-W1 have to be n*120 degrees (n is a natural number).
The slot-to-slot electrical angle alfa=p*360/Z=4*360/33=43.64 deg.
In the upper diagram, from U1 to V1 you have 3 slots, so the electrical angle here is 43.64*3=130.92 deg., very much difference to 120 deg. The same with V1 to W1.
2) To ensure the best balance between the 3 phases, the sequence of coil number inside each phase have to be the same.
-In the upper diagram, you have the sequence (count from the lead-in):
U phase: 2 1 1 1 1 1 2 2
V phase: 2 2 1 1 1 1 1 2
W phase: 1 1 2 2 2 1 1 1
-In the lower diagram, the sequence of the 3 phases is the same 2 1 1 1 1 1 2 2

*Conclude: the lower diagram is correct.

To electripete: I am so sorry!

 

[electricpete]

There is no reason to apologize for a disagreement.

I would certainly disagree emphatically with your statement that the bottom diagram is correct while the top is somehow not applicable, because I have already shown that these two diagrams perform identically.

These are single circuit windings. We have 11 series coils connected between V1 and V2 (and 11 between U1 and U2... and 11 between W1 and W2). The only thing that we changed between the top and bottom diagram is the order of connection of the coils (which one comes 1st after V1, which one 2nd, which one 3rd etc), without changing the polarity (or phase) of current in any coil. Now think about the implications of that:
1 – there is no difference magnetically. Each coil carries the same current with same polarity and this results in the same mmf distriction for top and bottom connection.
2 – there is no difference electrically. We have the same coils connected between U1 and U2 just in a different order, so when we add up all those series voltages the sum is the same.

Do you disagree with 1 or 2? Or have some other reason for thinking these two arrangements will act differently?

I think the bottom diagram conforms more to traditional expectations of what a winding "should" look like, but that doesn't mean it performs any differently than the top one.....they perform the same imo.

By the way, it looks like this is your first thread in the eng-tips motor forum. Welcome! Hope you'll stick around.

=====================================
(2B)+(2B)' ?

 

[zlatkodo]

I agree with Koizumi that the second version (lower diagram) is good.
But I'm not quite sure how bad is the first version (upper diagram )?
I think this topic is very important for rewinders, because (in such cases) most of them uses the rule: the beginnings of phases are the beginnings of the first , the third and fifth pole phase group.
Zlatkodo

 

[electricpete]

I may be outvoted, but that doesn't mean I'm wrong.

Let's draw the V phase electrical circuit. The number will represent the group number. We'll use unprime/prime notation to indicate the polarity of a given group.

[TT]Top Diagram:
V1 - 3 - 6' - 9 - 12' - 15 - 18' - 21 - 24' ----------- V2

Bottom Diagram:
V1 - ---------9 - 12' - 15 - 18' - 21 - 24' - 3 - 6' - V2
[/TT]
All we did was move two coils to a different part of the series circuit. We didn't change the current or polarity in any coil. Analysis of the W phase would show the same thing.

How can changing the electrical position of coil within the circuit without changing the polarity possibly affect the performance?

(imo it can't)

=====================================
(2B)+(2B)' ?

 

[electricpete]

Let's do a mental experiment.

You have 2 coils: Coil1 and Coil2 . Each has two terminals labeled as + and -.

Wire them up inside a box as follows:
A-----[+Coil1-]------[+Coil2-]---B
where A and B are the two labeled terminals coming out of the box.

Close the lid on the box and perform a set of experiments:
Apply dc voltage between A and B and measure resistance.
Apply ac voltage and measure inductance.
Apply dc voltage and measure flux pattern external to the box
Apply ac voltage and measure flux pattern external to the box.
Apply external time varying flux at certain location on the box and measure voltage at terminals A and B.

Now open the box and without changing the physical position of the coils, reconnect the coils as follows:
A-----[+Coil2-]------[+Coil1-]---B

Now reperform you experiments. You will not see any difference in results. You cannot tell from outside the box that the internal wiring has been changed!

Why is it so?
The coils still occupy the same position in space.
The terminal voltage is still the sum of the two coil voltages (V1+V2 = V2+V1)
The impedance seen from A-B has not changed (Z1 + Z2 = Z1 + Z2)
The coils still carry the same current as each other (I1 = I2 = IA).

It is a characteristic of a series circuit..... changing the order of the elements in the circuit does not change anything that we are interested in.

If you disagree, please tell me how you can possibly tell from the outside of the box (based on current, voltage, flux, or voltage induced as a result of applying external flux) whether the internals are in the first or second configuration? I say it is impossible to tell the difference.

It is the same change we made between top and bottom drawing.

=====================================
(2B)+(2B)' ?

 

[electricpete]

I say it is impossible to tell the difference.

I am excluding high frequency wave behavior (such as inject a pulse and look for a reflection)..... just talking about simple power frequency characteristics applicable to an induction motor. (magneto-quasi-static analysis).

=====================================
(2B)+(2B)' ?

 

[zlatkodo]

Koizumi wrote:
1) The electrical angle between the U1-V1 and V1-W1 have to be n*120 degrees (n is a natural number).
The slot-to-slot electrical angle alfa=p*360/Z=4*360/33=43.64 deg.
In the upper diagram, from U1 to V1 you have 3 slots, so the electrical angle here is 43.64*3=130.92 deg., very much difference to 120 deg. The same with V1 to W1.

I think this is the only difference between the two versions.
The issue is how it affects motor performance?
Zlatkodo

 

[electricpete]

As Ricky Riccardo would say: "Ay caramba!"

=====================================
(2B)+(2B)' ?

 

[electricpete]

That was an expression of frustration because you seem to have ignored everything that I said.

=====================================
(2B)+(2B)' ?

 

[koizumi]

Many thanks electricpete,

I agree with your experiment analysis. Perfectly correct with a single phase circuit.
But in three-phase circuit, you need to have a balance between three phases (electrically symmetrical).
I am a rewinder myself. I remember once I (for experiment purpose) tried to change the arrangement of a winding with fractional q (just like the example given by zlatkodo, but different parameters). Result was a big imbalance between phase currents to an unacceptable level.