"New" starting method of 3-phase motors 1

[zlatkodo]

Recently, many motors with 9 terminals have been appearing in workshops.
These motors are designed for one voltage and one speed.
For more information, please visit the following link: [URL unfurl="true"]https://winding.wixsite.com/design/post/unusual-example-from-practice[/url].
This new motor design appears to use a different starting method (hopefully will be useful for winders).
During the start, the motor is connected in a serial YD configuration, as shown in the attached diagram.
During the running period, the motor is connected in a simple 1Delta configuration.
Can someone explain this theoretical case, and how much the starting current decreases compared to other starting methods such as DOL, Y-D,or PWS?

 

[waross]

A similar connection has been discussed on Eng-Tips.
There were unanswered questions.
Can you give us a vector sketch of the connections?
Most of us lack your familiarity with the type of drawing that you have provided.
Thanks.

--------------------
Ohm's law
Not just a good idea;
It's the LAW!

 

[zlatkodo]

I hope this helps simplify the connection.

More on Winding Design and Motor Repair

 

[waross]

I see an electrical phase shift without a matching mechanical phase shift.
U1-U0 is no longer in phase with U0-U2.
The electrical phase angle of U1-U0 is the directed sum of U0-U2 and W2-W0.
This will introduce a 60 degree phase shift relative to U0-U2, but there is no corresponding mechanical phase shift.
Furthermore, the back EMF phase angle of U1-U0 will be influenced by the physical location of U1-U0.
I suspect that the impedance of U1-U0 will be quite low, but nevertheless enough to reduce the current through U0-U2 and W2-W0.
The torque developed by the current through U1-U0 will be displaced 60 degrees in the time domain from normal.
What effect may this have on gross motor torque?
A further complication, the current in U1-U0 will be 1.73 times higher than the current in U0-U2, suggesting 1.73 PU torque.
However the torque developed by the current in U1-U0 will be over only 1/2 of the number of poles, for a possible net torque of .866 PU.
Possibly the 60 degree shift will cause other interactions that I have not anticipated.
Conclusions;
I have a lot of experience doing it right and not much experience doing it wrong, so I don't know what will happen.
I think that we may have to wait for someone to do practical testing of this connection.

--------------------
Ohm's law
Not just a good idea;
It's the LAW!

 

[waross]

As I recall, the previous discussion pertained to a dual voltage motor connected with a star outside the delta.
All windings were equal, not in a 2:1 ratio.
I have reviewed over 400 previous posts looking for the other thread.
Possibly someone else will have stronger Google skills.

--------------------
Ohm's law
Not just a good idea;
It's the LAW!

 

[petronila]

Hi Zlatkodo,

What you are describing is not a starting method intended for reducing the inrush current this starting method is used for increasing the torque during the start, considering the rated current is Delta the motor will be starting as a "2Y connection" and the inrush current will be much higher and the motor will be producing about a 15% more of torque. This method maybe used for applications with high starting loads, for instance when you have an application with a 75 kW motor but once start will be using only 50 kW, then the 75 kW motor maybe replaced with a 55 kW saving energy.

My opinion is that starting motor should be avoided for applications with frequent starts because the windings will be fried very quickly.

Regards

Petronila

 

[zlatkodo]

the rated current is Delta the motor will be starting as a "2Y connection"

There is no 2Y connection here at all.
It is about what is the starting current ratio for serial YD and 1D starting of same winding with 9 leads, as mentioned earlier.

 

[electricpete]

I'm not familiar with single layer windings and I'm not sure I understood all the previous comments.

I'll say in my own words how I think this is supposed to work. (Please correct me if it sounds wrong or needs clarification).

Starting is accomplished with power applied to V1, U1, W1, AND V2 connected to W1, W2 connected to U1 and U2 connected to V1. This would be a delta winding including all coils. I'll also mention that all coils in a leg are in the proper polarity.... Specifically for the blue phase, we notice 13 and 31 have the opposite electrical polarity from the others, and that is required since they are approx 9 (out of 36) slots away, which means they are 90 degrees away in a 4-pole motor, which means the flux should be flowing at the opposite direction at those two locations. So during start, the blue phase in this 4 pole motor has 4 pole phase groups, differing in number of coils per group.

Running is configured with power applied to V0, W0, W2... AND U2 connected to V0, W2 connected to U0, V2 connected to W0. On the circuit diagram this is just a smaller delta including less coils. On the blue phase, we now only have 2 out of 4 poles represented (18 slots apart from each other = 180 mechanical degrees apart). You might ask how that works for a 4 pole motor. The alternate magnetic poles (90 degees away from the first two poles of the blue phase) would be formed by combination of the red and black phases. So the run configuration would effectively be a consequent pole winding where 2 electrical poles per phase are arranged in a way to give us 4 magnetic poles.

So the design objective is the same as delta wye start. The positioning of coils is a little tricker.

Again, that's just my thoughts, open to correction.

 

[electricpete]

During the start, the motor is connected in a serial YD configuration, as shown in the attached diagram.

Call me skeptical. Are you're saying there is a permanent connection from U2 to V0, from V2 to W0, and from W2 to U0?

 

[waross]

Pete:
U2 to V0 for starting.
U2 to V1 for running.

To make a point, what may happen if we ran a two phase motor (phase windings displaced 90 degrees) on two phases of a three phase supply (electrical phases displaced 120 degrees)?

What your diagrams do not make clear is the electrical phase shift of the current through U1 to U0.
The diagram illustrates the physical phase relationship of the windings, but not the electrical relationship.

With a star delta start, there is a 30 degree shift in the applied current vs the back EMF comparing star to delta.
We know the importance of arranging the connections and tweaking the timing so that the motor drops back 30 degrees rather than being forced to accelerate ahead 30 degrees.

Consider this arrangement to illustrate my concern.
We take a standard dual voltage 230/460V motor.
It may be either delta connected or star connected internally.
Now we will feed one 230 Volt winding with 230 Volts from a delta/star transformer.
We will feed the second identical winding from a delta/delta transformer.

This motor will probably run but can we expect it to draw excess current?

I know that this series star arrangement is more complex for at least two reasons.
1) The current in the star portion will be 1.73 times the current in the delta portions.
2) The delta portions will have twice the turns and twice the impedance of the star portions.

Is anyone in a position to try this in a shop and see what happens?

--------------------
Ohm's law
Not just a good idea;
It's the LAW!

 

[electricpete]

You might be right, but I'm not sure. It seems it would be tough to predict how that winding would act during start in that described starting configuration ("series wye delta"). I think you're analysing it a little bit like a transformer winding, but the phase of the voltage induced in a coil group depends not just upon its position within the electrical circuit but also upon its mechanical position around the circumference.

For the blue phase, the proper 4-pole-winding phase relationship for the 13 / 31 coil groups with respect to the 21/22/3/4 coil groups would be obtained IF power was applied between V1 and V2 (a big delta including all coils). Connecting U2 to V0 etc would disrupt that phase relationship. so I'm doubting the phase relationship would still be correct in that "series wye delta" connection

Bill / Zlatkodo - Do we have any evidence of the intended start sequence connections apart from the winding itself?

 

[waross]

In support of the arrangement:
Locked rotor current will be less.
Once the rotor is turning fast enough to develop back EMF, all bets are off.

--------------------
Ohm's law
Not just a good idea;
It's the LAW!

 

[electricpete]

Maybe so. If the phase relationships among groups get screwed up the way I'm thinking, I'd imagine the starting torque per current would be a lot less here than with normal wye delta start.

Either way I can't begin to analyse / imagine what the torque or the current would look like.

Let me try an overly simplistic circuit analysis of the start configuration which just treats the coils as impedances (the kind of thing I warned against before, since it ignores their position in the core and proximity to other phases) in a half-@ssed attempt to come up with some number for the starting current. (no guarantees that it's right, even if I don't make a math mistake).

Let's say Z0 is the impedance of each group of 2 coils (each box in the figure has an impedance Z0).

So we have in the center of the configuration a delta with impedance 4*Z0 in each leg. Outside of each corner of the delta we connect to the line through 2*Z0 in a wye-like manner.

Let's do a delta-wye transformation of that inner delta. The equivalent impedance in the wye leg is a factor of 3 less than the original impedance in the delta leg. The delta with 4*Z0 in each leg transforms to a wye with (4/3)*Z0 in each leg.

Now assemble it with the outer wye and we have a wye with 2*Z0 + 4*Z0/3 = (10/3) Z0 in each leg of that bigger wye.

Compare that to a normal wye/delta start where during start the wye would have 6*Z0 in each leg during the start. The series-wye-delta is going to have higher current than a wye connection by a factor of 6/(10/3) = 18/10 = 1.8. In other words if the wye/delta reduces the starting current to 1/3 of what it would have been if started directly in delta, the series-wye-delta/delta reduces it to (1/3)*(18/10)=3/5 of what it would have been if started directly in delta. On that count it wouldn't be as good as the wye-delta if this answer is to believed. BUT, I'm not claiming that answer for current is the correct answer, just a math exercize whose foundations are pretty dubious (even if I didn't make any math mistakes along the way).

 

[zlatkodo]

Hi Pete,

Your thoughts are very similar to mine regarding the starting torque (60%). To confirm our findings, it would be helpful if you could provide your estimation for the same winding but connected in a way where each internal D-leg has 2 coils and the external Y-leg has 4 coils.

If your result matches mine (approximately 43% of nominal torque), then we are on the right track. However, we should keep in mind that this doesn't necessarily mean that our calculations are completely accurate. It would be beneficial to hear from other experts on this forum.

Please let me know your thoughts.

Best regards
ACW

 

[electricpete]

Yes, 0.43 is the same answer I come up with for that new configuration. Below I repeated the previous analysis using the new configuration

For that configuration the inside delta has 2*Z0 in each leg. Outside of each corner of the delta we connect to the line through 4*Z0 in a wye-like manner.

Let's do a delta-wye transformation of that inner delta. The equivalent impedance in the wye leg is a factor of 3 less than the original impedance in the delta leg. The delta with 2*Z0 in each leg transforms to a wye with (2/3)*Z0 in each leg.

Now assemble it with the outer wye and we have a wye with 4*Z0 + 2*Z0/3 = (14/3) Z0 in each leg of that bigger wye.

Compare that to a normal wye/delta start where during start the wye would have 6*Z0 in each leg during the start. The series-wye-delta is going to have higher current than a wye connection by a factor of 6/(14/3) = 18/14 = 1.286. In other words if the wye/delta reduces the starting current to 1/3 of what it would have been if started directly in delta, the series-wye-delta/delta reduces it to (1/3)*(18/14)=6/14 = 43% of what it would have been if started directly in delta

It's a number calculated from circuit analysis assuming the coil groups act like simple impedances. As we both noted, it's not clear whether or not that is a completely valid approach here. I tend to think the accuracy of the answer rests on how faithfully the winding creates a roughly sinusoidal 4-pole variation in space (neglecting high order spatial harmonics from slotting). If it doesn't create that type of pattern, then the voltage induced by the airgap flux into the coil groups is going to vary among groups in the same phase. Since the full delta creates a sinusoidal pattern with this physical coil layout and the series-wye-delta uses the same physical coil layout with weird (hard to analyse) electrical phase shifts, I think the series-wye-delta gives a lot less sinusoidal pattern than the full delta.

I'd also be interested to hear what others think about it.

 

[electricpete]

A note about the direction of the error....

I think if the flux pattern is non-sinusoidal, then the above estimate is most likely low.

i.e. The actual starting current would probably be higher than 60% (original starting configuration) or 43% (revised starting configuration) of the starting current that would occur in delta.

 

[zlatkodo]

It is important to note that each leg, either in the internal Delta or external Wye, must have a real 4-pole winding ( in this particular example). Therefore, it is not possible to achieve this winding pattern using only 3 (the same number) coils in each leg. If we wanted to do so, we would need to use another type of winding, called a mixed or double-layer winding. At first glance, such a connection will provide approximately 50% of the starting torque.

As it is evident, this enables us to better customize the motor to the specific load during production or repair: 33%, (43%, 50%, 60%), or 100%. It should not be overlooked that any increase in the starting torque results in an increase in the starting current.

In fact, it is evident that these mentioned connections are very similar to the standard Y-start, D-run starting of the motor. By reducing the number of coils in the internal Delta, we move towards the classical 1Y connection (when there are zero coils in the internal triangle) and vice versa. The reduction of coils in the external Wye leads to the classical 1D connection.

Hopefully, this information will be useful.
AC Windings

 

[petronila]

Dear All,

These conection is also used for running motors delta rated @ 460 V at 575 V with a 15 % torque reduction. See the technical article: A closer look at winding conversions by reconnection.

Regards

Petronila

 

[electricpete]

yup, that looks very similar. I was able to download that article by clicking the pdf link here

https://easa.com/training/private-webinars/a-closer-look-at-winding-conversions-by-reconnection

which opened the following document:

https://easa.com/DesktopModules/Eas...0&moduleid=3015&articleid=1721&documentid=258

 

[zlatkodo]

petronila nad electricpete,
Yes, that is very similar to the original post, except that the original post refers to starting currents and torques of single-voltage windings.
By the way, the diagram you provided (the diagram on the left) doesn't seem to be the most suitable, as it can be misinterpreted as implying a phase shift between parts of the same phase, such as 1-4 and 7-10, which is incorrect.
The diagram on the right shows the way we represent such connections.

ACW