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

 

[zlatkodo]

Electricpete,
No, I dont want to ignore your words. By the way, I highly appreciate your knowledge and contribution to this forum. But I tried to say ( in my bad english), that the best practice is to make a beginnings at a distance of 120 or 240 or ...degrees.

I know that even many manufacturers take the beginnings from the first, third and fifth P.PH.G.
Why? I do not know the answer.They would have no extra costs if they would produce motors at another, better way.

How many times have you heard the question: why my brand new motor, connected to an absolutely balanced voltage, has a significantly different currents in phases?

Many rewinders often ask: why my motor after rewinding have such differences among the currents, even in "no-load" conditions, although the voltage is balanced?
I wonder if it is (sometimes) the cause of all that, this issue about which we speak.

For Koizumi: I am glad you are now , included in this forum (as an experienced rewinder),and I think it would be good to hear from you often.
Zlatkodo

 

[zlatkodo]

There is another, more frequent cases :
- 30 slots, 8 poles,
- 36 slots 10 pole.
It would be good to hear the answer to the question: which slots should be taken for beginnings ?
I have my own solution, which I would like to compare with second opinion.
(We're talking about double-layer windings).
Thanks in advance.
Zlatkodo

 

[electricpete]

Thanks guys. Forgive my momentary frustration from last night. I am certainly not an expert in winding and don't have the experience that you do in general. But on this specific question, I am 100% sure that I am correct .... i.e. those two diagrams DO give identical performance.

koizumi – The mental experiment that I described corresponds to one phase of a three phase machine. We swapped electrical order of coils in the box (without changing physical position or polarity) because we swapped electrical order of coils within a single phase (not among phases). Yes, there are adjacent phases which interact. We can simulate the interaction with adjacent phases by measuring flux produced at a given location outside the box...... or by imposing a flux outside the box and measuring the induced voltage at the terminals. There will be no difference in the 2 experiments. So, when we swap coil electrical locations in our single phase box, nothing changes, including the way it interacts with adjacent phases... it is an equivalent box. We can do the same for all three phases and nothing changes.

How many times have you heard the question: why my brand new motor, connected to an absolutely balanced voltage, has a significantly different currents in phases?

One factor is that under no-load, the current unbalance is approximately 12-15 times as high as the voltage unbalance (as shown in figure 5.6 of the reference below). So it only takes a tiny voltage unbalance to create a large current unbalance.

http://books.google.com/books?id=tX...n polyphase induction motor currents"&f=false

My answer to the question: where should we attach the phase leads is as follows: If it is a single-circuit winding, then it doesn't matter where the groups are located electrically between V1 and V2, as long as every group is included between V1 and V2 and all are connected with the proper polarity. So it doesn't matter where you attach the phase leads as long as you meet those same conditions (include all groups in the proper polarity.

Now it is important to note that when you decided which groups had 1 and which had 2, you took a phase pattern (such as 2 2 2 2 1 1 1 1 1) and shifted it by 1/3 of the slots in the motor to find the location where the pattern would begin between each phase. That is an essential step to help ensure balance (having 3 identical phases shifted exactly 120 degrees). So there is special significance to the location where where the leads begin in the bottom diagram that must be considered when laying out the coil groups..... but once they are laid out you can attach your phase leads anywhere you want (as long as you include all the coils of the phase and don't change their polarity).

The fact that the 3rd group is a number of slots from the 1st which does not correspond to exactly 120 degrees is an inevitable characteristic of fractional slot windings. In an integral slot winding, we can say that the number of slots corresponding to 60 electrical degrees is q = Q/(m*p) and we observe that every adjacnet group starts q slots or 60 degrees apart... the winding is perfectly positioned with respect to the field. In a fractional slot winding you don't have that luxury, because if you did then all the groups would have to have the same number of coils (q), and we know they don't in a fractional slot winding. So the groups do not line up perfectly with the ideal sinusoidal field and the distribution factor suffers. It is a fact that exists for fractional slot windings regardless of where you connect up the leads.

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

 

[electricpete]

EASA has a "recipe" for selecting coil grouping for fractional slot winding.
To see a copy, google the following words: joliet coil grouping chart
It should be the first item on the list. For some reason it is tough to grab the link since it opens directly into word.

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

 

[electricpete]

Well, hold off on that Joliet link. At first glance it looks a little bit screwed up to me. There is a similar chart in the EASA tech manual, but not identical.... in fact doesn't match the 30-slot 8-pole case.

Here is my attempt to analyse the case of Q = 30, Poles = 8.
q = 30/(8*3) = 30/24 = 5 / 4 = 1 and 1/4

So in each phase we have 10 coils: including two 2-coil groups and 6 1-coil groups. It can be split into repeating patterns 2 1 1 1.

In phase A we have groupings starting in slot 1:
2 1 1 1 2 1 1 1

In phase B we have groupingss starting in slot 11
2 1 1 1 2 1 1 1

In phase C we have groupingss starting in slot 21
2 1 1 1 2 1 1 1

If we overlay all 3 phases, it would look as follows starting in slot 1:
2 1 1 / 1 2 1 / 1 1 2 / 1 1 1 / 2 1 1/ 1 2 1 / 1 1 2 / 1 1 1

You could put the T-leads at the beginning and end of the phases, or any other location that includes all the coils with proper polarity for this single-circuit winding.

Note since there are two repeating groups per phase, we could also put them in parallels of 2 1 1 1 per circuit (2 circuits).

May analysis above matches what shows up in the EASA Tech Manual. But it doesn't match what shows up in the Joliet link which shows a patter: 2 2 1 / 1 2 1 / 1 1 2 / 1 1 1.... it looks to me like the ratio of 2's to 1's is too high.

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

 

[electricpete]

Correction: the B phase pattern starts in slot 21 and the C phase starts in 11

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

 

[electricpete]

Q = 36 slots, Poles = 10.
q = 36/(3*10) = 36 / 30 = 6 / 5 = 1 and 1/5

Each phase has 12 slots
Pattern = 2 1 1 1 1
Can be repeated.

The number of coils/group in phase A groups starting in slot 1:
2 1 1 1 1 2 1 1 1 1

The number of coils/group in phase B groups starting in slot 13:
2 1 1 1 1 2 1 1 1 1

The number of coils/group in phase C groups starting in slot 25:
2 1 1 1 1 2 1 1 1 1

Combine them all:
2 1 1 / 1 1 2 / 1 1 1 / 1 2 1 / 1 1 1 / 2 1 1 / 1 1 2 / 1 1 1 / 1 2 1 / 1 1 1

Again in this case it can be either single circuit or 2-parallel circuit. For single circuit winding electrical position of coils within the string (which one happens to be next to V1, which one happens to be next to V2 etc) is irrelevant as long as polarity is not changed. For 2 circuit winding, electrical position is mostly irrelevant except that in addition to observing proper polarity, we also have to partition the coils among parallels correctly, for example first 2 1 1 1 1 sequence from phase A goes into one parallel and second 2 1 1 1 1 sequence from phase A goes in the other parallel.

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

 

[zlatkodo]

Hi, Electricpete,
You're given a lot of material for analysis.
1/. Q = 30, Poles = 8.
„IF WE OVERLAY ALL 3 PHASES, IT WOULD LOOK AS FOLLOWS STARTING IN SLOT 1:
2 1 1 / 1 2 1 / 1 1 2 / 1 1 1 / 2 1 1/ 1 2 1 / 1 1 2 / 1 1 1
THE B PHASE PATTERN STARTS IN SLOT 21 AND THE C PHASE STARTS IN 11.“
I agree. My choice is : U-1, V-11 and W-6
both variants satisfy the criteria n * 120 degrees
„May analysis above matches what shows up in the EASA Tech Manual.“
Unfortunately I have not seen EASA Tech Manual. What is the distance between the beginings for this case (Q = 30, Poles = 8) in this publication?
„But it doesn't match what shows up in the Joliet link which shows a patter: 2 2 1 / 1 2 1 / 1 1 2 / 1 1 1.... it looks to me like the ratio of 2's to 1's is too high.“
Apparently, there is an error. I have seen more in the Joliet link .

2/. Q = 36 slots, Poles = 10.
2 1 1 / 1 1 2 / 1 1 1 / 1 2 1 / 1 1 1 / 2 1 1 / 1 1 2 / 1 1 1 / 1 2 1 / 1 1 1
My choice is: U-1, V-25, W-13
Zlatkodo

 

[zlatkodo]

Correction:

But it doesn't match what shows up in the Joliet link which shows a patter: 2 2 1 / 1 2 1 / 1 1 2 / 1 1 1.... it looks to me like the ratio of 2's to 1's is too high.
Apparently, there is an error. I have seen more in the Joliet link . There must be:
2 1 1 / 1 2 1 / 1 1 2 / 1 1 1....
Zlatkodo

 

[zlatkodo]

By the way it should be noted that this table in the Joliet link is full of errors.
It is hard to believe that such mistakes are accidentally created.
I suggest do not use the information from the table and also do not go to this link, or there is a possibility you can get something unwanted in your PC.
Zlatkodo

 

[electricpete]

We both agree the Joliet link is screwed up. They must have made a whole bunch of typographical errors when transcribing from the EASA document to the Joliet word file.

Glad you came up with the same 30-slot 8-pole configuration.

You still mention about importance of choosing the correct terminal location. It is still not the case. For example the top and bottom windings you posted perform identically. Again, I am 100% positive of this conclusion.

What does a coil do magnetically? 2 things:
1 - it has voltage induced as a result of changing flux passing through it (v = N*dPhi/dt)
2 - it creates mmf to drive flux (obeying H = Integral I dot dL).

But neither one of these functions "cares" about the electrical position of the coil within a series circuit (as long as polarity is maintained):
1 - The induced voltage affects only the sum of voltage around the entire loop through all coils and voltage source. That sum does not depend on the relative electrical location of any coil within the series circuit. (V1 + V2 = V2 + V1)
2 - The mmf created by a coil depends only on the current. Current does not care what electrical position position of the coil within the circuit is... it cares only about the current... which is the same for all series coils within the circuit. (I1 = I2 = I).

For a set coils in specified physical positions, which have already been assigned phases, groups, and polarities that will not change, the electrical position of the coil within the series circuit does not matter.

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

 

[koizumi]

electricpete,
I think both diagrams are OK now.

But I still have one correction:
-In 30slot-8p diagram, indeed B phase starts in slot 11, and C phase in slot 21. Otherwise you could have other options, like A-1, C-6 and B-11 (zlatkodo's) or B-26, A-1 and C-6, etc.
-In 36slot-10p diagram, C phase starts in slot 13, and B-25.

Please remember that, in order to have full 6 poles per 360 electrical degrees, you have to count 6 poles in sequence A-Z-B-X-C-Y (zlatkodo please convert to UVW for me).

 

[koizumi]

We've just got an incident with a new winding.
Z=90; 14 poles;

Which slot should you pull out the T-leads?
I just put forward for you to think before telling you what happened.

 

[electricpete]

Hi koizumi

I think both diagrams are OK now.

I'm glad you agree. I was worried this thread would end without anyone else agreeing on that point, that seems relatively straightforward.

But I still have one correction:
In 30slot-8p diagram, indeed B phase starts in slot 11, and C phase in slot 21
-In 36slot-10p diagram, C phase starts in slot 13, and B-25.

I am not sure about your corrections:
For the 30-slot, 8-pole motor, I have laid out the group sequence A B' C / A' B C'... starting in slot 1 as follows (11 Sep 10 13:30):
2 1 1 / 1 2 1 / 1 1 2 / 1 1 1 / 2 1 1/ 1 [blue]2[/blue] 1 / 1 1 2 / 1 1 1
The 11th slot occurs at the bolded number and it belongs to phase C as I corrected 11 Sep 10 13:56
The 21st slot occurs at the blue number and it belongs to phase B as I corrected 11 Sep 10 13:56

For the 36-slot, 10-pole motor, I have laid out the group sequence A B' C / A' B C'... starting in slot 1 as follows (11 Sep 10 14:17):
2 1 1 / 1 1 2 / 1 1 1 / 1 2 1 / 1 1 1 / 2 1 1 / 1 1 [blue]2[/blue] / 1 1 1 / 1 2 1 / 1 1 1
The 13th slot occurs at the bolded number and belongs to the B phase as I stated.
The 25th slot occurs at the blue number and belongs to phase C as I stated.

Please remember that, in order to have full 6 poles per 360 electrical degrees, you have to count 6 poles in sequence A-Z-B-X-C-Y (zlatkodo please convert to UVW for me).

Well I am certainly aware there are 6 groups per 360 electrical degrees although I gave them a different name (A B' C A' B C'). Is there something I wrote that you think is incorrect?

We've just got an incident with a new winding.
Z=90; 14 poles;

Which slot should you pull out the T-leads?

q = 90 / (3*14) = 90 / 42 = 2 and 1/7.
It suggests that the 42 coils in a given phase should be arranged:
3 2 2 2 2 2 2 3 2 2 2 2 2 2

If we start A in slot 1, C in slot 31, and B in slot 61, and look at the sequence including all the phases, then we have the following AB'C / A' B C'/....:
3 2 2 / 2 2 2 / 2 3 2 / 2 2 2 / 2 2 3 / 2 2 2 / 2 2 2 /
3 2 2 / 2 2 2 / 2 3 2 / 2 2 2 / 2 2 3 / 2 2 2 / 2 2 2 /
Which can be done as 1 or 2 circuit winding.

The location of T-leads can be anywhere you choose as long as all coils are included in the proper phase with proper polarity (for one-circuit winding), and for 2-circuit winding allocate properly between the two circuits (first 3 2 2 2 2 2 2 from a given phase in one circuit and last 3 2 2 2 2 2 2 from that phase in the other circuit). Now you've got me wondering why you are asking about the T-leads...


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

 

[zlatkodo]

Hi, Koizumi,
We've just got an incident with a new winding.
Z=90; 14 poles;
Which slot should you pull out the T-leads?

I would propose the same winding arrangement as Electricpete, but with the leads as follows:
U-1, V-31 and W-61.
By the way, I'm just curious, which was a pitch of this winding (1-6 or another)?
Zlatkodo

 

[koizumi]

Hi electricpete,

I see your sequence in phase counting is different with mine.
Yours is A-Y-C-X-B-Z so your lead assignment is different.
But the conventional way is A-Z-B-X-C-Y.
Anyway the diagram is still OK, no problem.

Back to the Z=90; 14 poles diagram.
One of our engineer, had put the winding arrangement as below

3 2 3 / 2 3 2 / 2 2 2 / 2 2 2 / 2 2 2 / 2 2 2 / 2 2 2 /
3 2 3 / 2 3 2 / 2 2 2 / 2 2 2 / 2 2 2 / 2 2 2 / 2 2 2 /

Lead position: A - 1; B - 6; C - 11;
And the motor is now running with a 15% imbalance in phase current.

After reviewing, I see that the A - B and B - C electrical angle is 140 degrees. That's not suitable.

I see the arrangement proposed by electricpete is OK and I'll try on our motor.
Thanks electricpete and zlatkodo!

 

[electricpete]

Yes, that sequence is unbalanced. I can tell two ways.

1 - I have a spreadsheet. It shows distribution factor 0.853593736 0.875067314 0.853593736 for that configuration. For my configuration it shows distribution factor 0.896806749 0.896806749 0.896806749

2 - The three phases are not carbon copies of each other shifted by 120 degrees. Each phase has only one 3-coil group so that group needs to be exactly 120 degrees apart or some multiple. For your configuration the spread is not. alpha = 360*7/90 = 28. The your configuration 3-coil groups start at a distance 5 slots apart where 5*28 = 140 degrees. In contrast for mine the coils are 30 slots apart. 30*28 = 840 = 7*120 degrees. Note this consideration exactly 120 vs approximately 120 degrees is very important for laying out coils..... but not at all important for selecting T-lead location.

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

 

[electricpete]

After reviewing, I see that the A - B and B - C electrical angle is 140 degrees. That's not suitable.

Sorry, I didn't read carefully you had already computed this.

I'm not trying to be a know-it-all... just in a hurry.

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

 

[electricpete]

I had an error in the distrubtion factors mentioned above. Should have been 0.919, 0.919, 0.912 for the first configuration from your engineer.... 0.955, 0.955 ,0.955 for the newer configuration that we all agreed on.


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

 

[electricpete]

zlatkodo – Take a look at the attached is an excerpt from "Electrical Machinery – Volume 2 – AC Machines", by Liwschitz-Garak and Clyde Whipple, published by Van Nostrand 1946.

starting at bottom of 2nd attached page:

"In the integral-slot winding, the beginnings of the phases are displaced by 120 and 240 electrical degrees. Also in the fractional-slot windings the distances between the beginnings of the phases can be made 120 and 240. This will be the case, for example, when the beginnings are placed in slots 1, 1+N, and 1+2N, when Beta is an even number, and in slots 1, 1+2N and 1+(1+m)N, when beta is an odd number. This arrangement of the beginnings of the phases will place them far apart from each other mechanically, while it is usually desirable to have them near to each other. In order to place the beginnings of the phases near to each other, the beginnings can be placed approximately 120 and 240 degrees apart and the winding still will be balanced, i.e. the 3-phase voltages still will have the same magnitude and will be displaced exactly by 120 and 240 electrical degrees. This is due to the fact that in the fractional-slot winding the emfs of the consecutive coil groups are not in-phase, and the sequence of the geometric addition of the single emfs is of no influence on the resultant phase emf. So in the example considered, the beginnings of the phases can be placed in slot 1 which belongs to phase I, in slot 5 which belongs to phase III, and in slot 8 which belongs to phase II. The angles between the beginnings are then 150 and 262.5 electrical degrees. "

[note – bold emphasis was added by me, italic emphasis was added by Liwschitz-Garark]

When he says "the sequence of the geometric addition of the single emfs is of no influence on the resultant phase emf.", that is the same thing I meant by Va + Vb = Vb + Va.

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