Confusion on Waveform

This is a guess on my part because I have to know a variety of other things to be sure but I suspect that that you are seeing the effect of the back EMF of the motor.

This is AC current waveform and high on transition points where current switching is very sharp. If you magnify it then you will also see DC part in DC current recording.

To Shyam....

dear sir,

will you please elaborate on your point...

with regards,

Amith Raj.N

what you are seeing is AC transition either directly picked up by the sensor or AC output of the sensor filtered for high frequency by capacitive couping. These do not look like waveform of actual current. Voltage waveform looks OK to me.

Hi redfred...,

I can provide the information as you need..

In this case here, i am running the motor by using the drive.. I kept the Oscilloscope to output of drive which is the input supply for Motor...

As motor is running it can produce back EMF but at the same time input supply is present... We can't measure back emf seperately right..

And We can,t say It is Open circuit... because motor winding is connected as load.....

I want to know that any one else faced like this situation....

With regards,

Amith Raj.N

Well let me ask you this question first: When the back EMF voltage matches the drive voltage, what will be the current?

Is this an AC motor or DC?

You have connected motor and oscilloscope in parallel. Osc is very high input impedance so it is workiing like a voltmeter. it is not drawing any current. it is only showing input voltage.

You say that you are measuring current...please let us know how you are doing this using an oscilloscope?

I'd use a clamp on current transformer probe but I have several nice toys at work.

You are not providing the information needed. You must specify the power of your drive (it ranges from 0.1 - 1,000 HP) and the Type of drive - Yaskawa makes about 10 different types. The waveform you show are totally useless as the scales are inappropriate. Your waveform shows the "carrier wave" - about 4,000 Hz, . . . not your 50 Hz power components. The current spikes you see are not current - as current can not "JUMP" so much on an inductive load (motor) - they are "capacitive pick-up" caused by the inadequacy of your connections and measurement.

Amith,

It appears that the voltage output is approximating a square wave instead of a sine wave. Also, it appears that the current scale is too large, so you only see the spikes when the voltage reverses its polarity. I'm not sure about interpreting the time scale. If each major division is 50 μsec, then you are seeing the individual IGBT pulses at a carrier frequency of about 7700 Hz.

Come back with further data regarding the time base, if you can. Lacking that, I would say that your 'scope data were with too small of a time base, and were taken at a time near when the output voltage sine wave was crossing zero. The essentially zero current is consistent with this.

Try using a time base in the range of 0.02 - 0.1 sec.

--JMM

Hi Good day to all CR4 members...

To redfred question....

If back emf is equal to drive voltage, then no current flows but measured voltage in oscilloscope should not equal to voltage at standstil condition... but in this case voltage at standstill is 370V and while running its 350V...

To Baxterm@kos.net question...

It is Permanent magnet synchronous servo motor.....

For daffy...

I measured current waveform I used clamp on meter made by FLUKE i310s meter...

Below are the detailed waveforms...

this waveform shows the current[Blue colour] & voltage[Red colour] made of 4KHz PWM....

in this waveform we can observe PWM of 4KHz...

In this waveform i decresed the time base

here further decresed time base...

Here further decresed... here we can observe the detailed current waveform when voltage PWM switches from +ve to -ve & vice versa... Remaining time its zero...

With regards.

Amithraj.N

Now you have the waveform, . . . , but you measurements are suffering from excessive amount of interference from the 4kHz carrier of the controller. You need superior isolation and common mode rejection to see the waveform you are after.

So many things to discuss in this collection of scope traces and so little time. First, thank you very much for answering my question. The way that you answered tells me how much you understand how a motor works in the first place so I know how to present my answer. (That was supposed to be a compliment, but it actually sounds condescending to me now. I apologize for the tone.)

I'll start my analysis here with the first waveform that is actually the most informative. I suspect from this waveform that you have some form of a mechanical feed back (velocity likely) into your motor control. The variations of the PWM drive voltages (red) above and below ground are because the drive controller is sensing that the motor is not maintaining a constant velocity. The drive is trying to gently nudge the motor faster and then slower than it is actually running at a rate of about 50 Hz. I suspect the feedback sensing circuitry is coupling your power distribution frequency in some fashion into your circuitry. The driver's power supply maybe coupling into the feedback path, too. It may also be because you actually have some mechanical resonance that happens to coincide with the 50 Hz power distribution that you have. Those three possibilities are presented in sequential order of most to least likely. To properly read the average voltage you are providing into the motor you should filter out the 4kHz PWM edges. My personal preference for that would be to fabricate a very high pole low pass filter with a corner frequency no higher than 2kHz. (Maxim MAX291 is a single chip 8th-order switched capacitor low pass filter. For a completely analog design I prefer an Ackerberg Mossberg bi-quad.) This will also remove aliasing problems due to the very high frequency voltage noises from switching an inductive load. This filter network though should only be part of your diagnostic signal path (into your oscilloscope) not part of your feedback path. Capriciously adding extra poles into a feedback path can quickly become a recipe for disaster.

Now onto your current trace. The very high frequency spikes that you are getting are clearly related to the edges of the PWM voltage drive signals. I suspect that these are actually coupled in your oscilloscope but they could be coming anywhere. If you fabricate the filter I described above for measuring the drive voltages recheck the current measurement first before modifying your current pick up circuitry. I suspect that most if not all of the spikes will be gone. If not then you will have to some how either add a similar filter network for your current detecting circuitry or identify how the voltage spikes are capacitively coupling into your current detecting circuitry. I say this because I agree with others here that the current through your motor is not changing that fast and far. Once these spikes are removed and before you correct the 50 Hz velocity wobble I believe you will see a tiny 50 Hz corresponding wobble above a small DC current. The DC current is the power required to fight drag and the 50 Hz wobble is the readjusting feedback signal.

Once you get these anticipated steady state signals then you can produce the most useful data acquisition traces for an analysis of your motorized system. This will take multiple attempts to get this trace. Set your oscilloscope up for a single sequence recording to record the start up of your motor. Ideally you want to do this without any feedback. This will allow you to characterise the combination of motor, drive and mechanical load. Changing the mechanical load will show you a very different startup time constant. Now you can do a proper design analysis of your optimal feedback characteristics so that your servo system can in a stabile fashion either maintain a constant angular velocity or position.

Looking at you first waveform, it is actually a sine wave at about 50Hz, produced by "chopping" the DC supply in the drive electronics. It looks perfectly normal. I think that somehow the switching frequency producing this waveform is being coupled to your current probe. If you could bandwidth limit this to show only signals below about 20KHz I suspect you could increase the sensitivity on the current channel to about 0.5A/Div and see a nice 50Hz waveform lagging the applied voltage.

I think this is your problem:-

It's not quite clear but I think that says 50 Amps per division.

Thanks Randall..., you gave one point to think...

Here my PWM frequency is 4KHz.... so inductive reactance(Xl) is 2*pi*freq*Inductance...

f=4000, L=28milli henry so Xl=2*pi*4000*0.028.....

as I= V/Z... Z=Impedence...

V=720V [Peak to Peak]...

R=6.5 ohms...

I=V/(R+(J*Xl))....

I= 0.00945 Amps...

As I kept the Scope ampere range too high, i am not able to see the current wave...

Is my thinking way correct..??

I will come back by tomorrow with real time testing result..

With regards,

Amith Raj.N

just <10mA for your motor? I seriously doubt it. What are the ratings for your motor?

Shyam, he calculated the ripple current from the 4KHz carrier. The 60Hz current would be much higher (66x = ~600mA). Still small for a decent motor but this could be a very small one. In this case, setting the probe to a lower scale would help

It looks like the current signal is drowned in switching noise. This may be from coupling from the voltage reading channel. Do you use an isolated probe? or is it a properly isolated voltage channel? Non isolated channel would pick up a lot of noise and potentially be dangerous if the probe ground is connected to the voltage source.

The motor might not be connected properly (open windings for star / delta connections) and only noise flows through capacitive coupling. Does it work when connected directly to the line?

Marcot,

There is a procedure for motor current measurement and it is all there on Maxim and Linear Technology web site. Isolation helps a great deal. Drop of voltage across resistor from motor current will also generate sufficient voltage output along with switching noise.

You will not be able to see the current wave because of interference from the 4kHz PWM controller . . .

Check your connections, tell us the scale of the current and voltage waveforms, show the measurement circuit...

...L