Charging Coils in a Three-Phase AC Induction Motor System

It seems you're getting bogged down on the VFD theory that is extensively used as DC power converter for the AC electric motor in a hybrid car.

It is clearly shown what is the relationship between the train pulse wave-form and the actual sine wave-form that drives the AC motor to power the vehicle.

Hi thanks for your reply... but I am sorry, I fail to understand what you have tried to convey... We are not able to explain the exact phenomenon occurring in the Motor when a 50% duty cycle switching is done on the three coils with 0 phase shift with switching frequency of 16KHz. May be the title I have chosen is not very correlative... Could you assist me in understanding what happens in the coils when all three terminals are simultaneously connected to the 310 VDC bus and at the next instant to the 0VDC bus in a repetitive cycle with the On and off time being 32us.

Ok, this (generation and distribution) is the analog way to explain how the phases are out by 120 degrees and alternate for AC power.

In the previous example I did not realise it until later that they only show the pulse wave-forms. I assume here things become clearer for you to understand how the relationship between the pulses as they are width modulated (PWM) affect the sine wave shape applied to the AC motor.

If there's no phase shift the quasi arrangements will cause short between themselves and blow........something

Hi Isti, Sorry I don't know your name... Well I am quite clear on the working and generation of the sine - wave from the pwm sequence. The query is not about when the system is in normal running. In a VF control system, during no running, that is when 0V is supplied to the system, actually a 50% duty cycle is applied to the system. This does not short the system as all coils are connected to the same state of either 310V or to 0 VDC. During this time no current flow occurs in the system. But interestingly during that time a voltage is available across each coil. around 155 VDC. This 155 V DC is the DC offset applied to the system. The sinusoidal wave is from 0V to 310 V with the system being shifted to the off set results in -155V to +155V , i.e. a sinusoidal wave of 155V AC. The query is only with respect to how the voltage is 155V and not 0V across the coil when the 50% duty cycle, 0 Phase shift, 16KHz PWM is applied to all three coils simultaneously, shifting all coils from 310 VDC to 0 VDC in 32ms, cyclic.

I think I have a problem to understand your question. I am in the middle of trying to interpret it but I think your title is incorrect also.

Stephen

Yes I guess it should be charging of coils of a 3 phase induction motor under a special case. I know what is happening as it is evident through the practical measurements and observation but we are interested in the why... as we are not able to explain what is happening when the previously mentioned 50% dutycycle frequency is applied to the system with no phase difference. Thanks for your time Isti...

I wanted to help but I did not understand your questions and I still don't for example:-

Which point are you referring to measure the 150v on nd wrt to what?

How are those 3 coils wired up? (Just give me a proper explanation!)

(You can even forget about the freq for a moment since we're talking about voltage devision across the network that's only made up of inductors (coils). The term 'charge' for a coil is also inappropriate to use!)

We could be going in circles due to our misunderstandings here, as I see it, rather being able to get anywhere.

It's up to you if you wanna give it another shot.

Hi Isti,

forget about the 150 V,

What would happen if the three coils are altogether held at 310VDC line and later at 0VDC line with no return path? The frequency of connecting to 310 V and to 0V is 16KHz with each state being held for less than 32 micro second.

a. No Voltage and Current

b. The central node of the three coils acts as virtual ground (node N) and an AC voltage is induced with 150V average Voltage in each coil.

Let us forget about a motor here and yes the three coils are identical.

If you switch a coil to 310V dc (energise) and when switch it off then its magnetic field will collapse (de-energise). You are going to waste this energy unless have another inductor to capture it (like a secondary coil does in a transformer).

What I gather you have 3 inductors connected in parallel to 310v dc then (32micro sec later) they get switched off, am I right? Meaning that they all get the same voltage and same current flowing through them while they're on. During turn off however, they will generate emfs which will oppose the supply voltage as the magnetic field collapses in each coil.

I do not know your configurations of the coils but, if they are in star mode then the center point will be half i.e. at 155v due to the coils divide the supply voltage even if they are switched on at the same time.

Thank you Isti.

I understand that part. What I am not able to understand and consequently explain is as below.

Given:

Only three points of input exist in the system - A, B and C.

Problem:

When all points are switched together to either the 310V terminal or to the 0V terminal, how is the return path formed?

Is there a mathematical explanation for what is happening in the circuit due to the inductive property of the coil.

Is there a Virtual ground/reference terminal created at the Star node - N.

Thanks again Isti...

I'm sure there's a formula some sort for it and perhaps wikipedia has some also unless somebody later join in and help you out with it.

However, the zero potential of the 'nod' is achieved when all three phase line load currents are equal, which I doubt you can ever perfectly achieve due to even tolerance values let alone other factors (this is why you need, theoretically, a lighter gauged return line as the phases will never be perfectly balanced).

Nonetherless, for all practical purposes you can say that the nod is at zero potential when the currents in the phase lines are equal.

I can only say this providing I understood your question right.

Hi vish,

look at my last post: the return path is the stray capacitance from coil to zero.

All three coils will have a positive current after rising the voltage from 0 to 310V, flowing through the coils and leaving over the distributed capacitance to ground.

If you do the time-resolved measurement with an oscilloscope you will be able to have a sound understanding of the effects.

RHABE

Sorry vish, after I read your posting again I have started to come to grips with your question and realise how badly I misunderstood it.

Can you help me out here... Is there a mathematical explaination or equations that would relate what happens...???? Thanks!!

As I understand it now you do have a considerable capacitive reactance (XC) between the coils at 16khz, am I right?

Therefore, I thought the best for you would be is to get a proper capacitance & inductance measuring equipment and work it out like that. As far as I'm concerned using the normal formulas to calculate either (C) and (H) is never gonna give you a good enough result let alone an accurate one to work out your Xc or XL for that matter, especially couple these values with the motor's temperature.

Bear in mind if both these exist (Xc&XL), as they do, it means your motor's performance is severely affected by the the pulse rate, theoretically the larger the motor the worse it can get.

If you would once more reconvene your configuration a little more clearly nd concisely could help me to better understand. Nonetheless, I do believe my aforementioned assumptions are near enough nd providing that you're not over looking at some issue/s (whatever that may be) in which case we could end up in circles.

Also what's the core material of the motor?

Hello Isti,

Well we actually do not have a problem with running the motor using sinewave pwm control from about 2 Hz to 90 Hz. It is a 200W motor.

While explaining the sequence rendered for 0Hz, a question was raised by one of the new batch of trainees.

He was of the opinion that as during this 0Hz switching there actually is no proper return path. Then what is th purpose of switching. Can't we just go ahead and have all three coils and connect them to the 0V potential.

We could not explain clearly to him what was happening with solid base or any papers or documented equations.

regards,

Vishal...

Hi vishal,

Look, the only reason I must give up on all this is because I went back trying to understand your original question but I've got lost.

I mean what is 0Hz anyway? In my understanding it is constant dc voltage. (Some manufacturers do specify how much dc power to apply to their motors to break/lock them to hold the load without the need for any extra mechanical break to be used.)

In my definition when you have a switched off power supply it cannot be regarded as 0Hz simply coz there's no power applied to the motor.

Further, if your frequency is between 2Hz-90Hz the capacitive effect, for all practical purposes, must be zilch between the coils. If there's a capacitive effect however, then it must come down to the boot strap caps you mentioned about before linked to the motor circuit. They would have more to do with your capacitive reactances (or call it whatever) than all the other things we have talked about so far.

I can only help you further if you're willing to present a true (not hypothetic) configuration between the controller unit and all the associated components with the motor connected.

rgds,

Hi Isti,

Thank you for your replies and for your patience.

We have designed a drive based on AT90PWM3, AVR.

We are successfully running it for nearly all practical situations required. 2 - 90 Hz.

When normal running, the PWM carrier wave is incident on a frequency component of our desired requirement to provide a sinusoidal pulse of that frequency with the PWM waveform, each coil charged at 120 degrees offset electrically. The effective DC applied on the coil is based on the Volt-Hertz curve of the motor.

The situation of applyin the earlier mentioned pwm with all coils at 0 degree phase shift occurs when we are decelerating from the operating speed towards 0.

As the frequency approaches 0, the system operates normally with 120 degree phase shift. Upon reducing the applied frequency to 0V, instead of switching all the switches off we continue to operate them at 16KHz frequency with phase difference 0.

We would like to know why we should provide the 0 degree phase shift 16KHz pulse to the motor during stop rather than just switching off the motor by taking its supply off.

Hi vishal,

"We would like to know why we should provide the 0 degree phase shift 16KHz pulse to the motor during stop rather than just switching off the motor by taking its supply off."

Indeed, I don't know myself what's the purpose of it either other than to keep the motor locked, which does not make much sense to waste power for unless it is designed for special applications.

Should you be lucky to find out more about it would you let me know? I'm curious myself.

rgds,

Stephen

Thanks for your time Isti. I will when I find that out. Take care...

That's fine.

No, No, No!!!

Hi,

this is totally different.

Make a plot of ideal or real three-phase voltages.

As there is a phase difference between these voltages, at any moment with a closed top-switch there is a closed bottom-switch - and the current is flowing through coils and both switches.

If any of these switches is going off, then there is a short time until the next switch closes - to prevent short circuit and immediate damage to the switching transistors.

Get an oscilloscope and a current probe to monitor currents and voltages and their time dependence.

RHABE

Hi Rhabe, Thank you for your reply. We are clear on the 120 degree phaseshift pulse being applied during normal run for a frequency application of 1 Hz to 90 Hz. That is how we are running the Variable Frequency Drive Control. No questions about its working. When we want to apply no running, i.e., 0 Hz, a pulse of 50% duty cycle with 0 phase shift is applied at a frequency of 16 KHz. It causes all the coils to switch to the 310 VDC bus together and after 32micro seconds to shift to 0 VDC. Then all the coils are again subjected to 310 VDC after 32 micro seconds and the cycle gets repeated. The question is how does a voltage of 155 VDC register across each coil in such a situation when no actual current passes in the system. For more clarity please read the steps mentioned in the query. These steps only explain about the 50%, 0 phase shift condition and not the normal running sequence. Please help us understand what happens at such a condition. Also kindly go through my reply posted to "Itis80". Thank you for your time.

Sorry Rhabe but that was not the question.

Sorry Vish, I did not understand your question.

To make things clear: you put (with a triple H-switch) the A,B,C inputs either to 310V or to 0V and measure the N that is the star-point where the (non-A,B,C)-sides of the coils are connected but floating and not grounded.

More information needed: how did you measure the 155V, you did state that you find on N?

Oscilloscope with 1MHz bandwidth or more? or any other method?

How fast is your switch? Usually these switches can just operate at the existing frequency and not very much above so any form of voltage may be severely distorted by the switch behaviour.

What I would expect: you have a coil inductance and its distributed (winding to winding) capacitance: this is a parallel L-C-oscillator that is connected to ground by the stray capacitance of the coil to ground.

If your switch is fast then you have in the very first microsecond (the inductance has a near infinite resistance) a pure capacitive voltage-divider. After some time the current through the inductance grows until fully charging the coil to ground capacitance. So within some time the voltage should rise from the value of voltage division up to full 310V.

If you try to measure voltage and current with µs resolution - may be you see this at switching to supply voltage: a fast jump followed by an exponential rise to reach the supply voltage.

And at switching to zero: a first fast decline followed by an exponential decay to zero.

Zero and supply voltage may not be reached if the time constant of your L, (C1 parallel C2) is not short compared to a half-cycle or 35µs.

Oscillatory response to switching steps may be existing depending on resistive loss. If so, this would be superimposed onto the above described situation.

Please take a measurement.

RHABE