AC Drive IGBT question

OkiePC

Lifetime Supporting Member
Join Date
Mar 2005
Location
ENE of Nowhere Oklahoma
Posts
11,757
I am putting together a class for our technicians to teach them the basics of AC VFDs and stumbled on some conflicting information that surprised me and I am not sure it is accurate.

Today's inverters use IGBTs to switch the DC bus on and off at specific intervals. In doing so, the inverter actually creates a variable AC voltage and frequency output. As shown in Fig. 7, the output of the drive doesn't provide an exact replica of the AC input sine waveform. Instead, it provides voltage pulses that are at a constant magnitude.

IGBT_Waveform_000.JPG



It 1st says variable voltage...then says constant voltage...then the picture shows what looks like 4 different voltages, 2 magnitudes on each side of the zero line.

Don't IGBT drives just do PWM with full bus voltage like an SCR drive? With a single voltage level that produces a sinusoidal current?

Thanks in advance, Paul...

EDIT: sorry about the cruddy pic, but you get the idea...
 
sounds like your reading the same article I have here even the image is the same. I took that line to mean that due to its higher switching rates
as opposed to SCR's it will replicate Sinusoidal waveforms more accurately and its the average of the sinusoidal form that is of a constant magnitude. rather than the PWM
 
Yeah, reading further I see another error in the use of the word voltage, so I think their fuzzy little picture is in error as well. I believe I will teach the guys what I have always believed unless someone proves it wrong...

Thanks, Paul
 
Today's inverters use Insulated Gate Bipolar Transistors (IGBTs) to switch the DC bus on and off at specific intervals. In doing so, the inverter actually creates a variable AC voltage and frequency output. As shown in Fig. 7, the output of the drive doesn't provide an exact replica of the AC input sine waveform. Instead, it provides voltage pulses that are at a constant magnitude.

what-i7.gif

Figure 7, Drive Output Waveform
The drive's control board signals the power device's control circuits to turn "on" the waveform positive half or negative half of the power device. This alternating of positive and negative switches recreates the 3 phase output. The longer the power device remains on, the higher the output voltage. The less time the power device is on, the lower the output voltage (shown in Fig.8). Conversely, the longer the power device is off, the lower the output frequency.

what-i8.gif

Figure 8, Drive Output Waveform Components
The speed at which power devices switch on and off is the carrier frequency, also known as the switch frequency. The higher the switch frequency, the more resolution each PWM pulse contains. Typical switch frequencies are 3,000 to 4,000 times per second (3KHz to 4KHz). (With an older, SCR-based drive, switch frequencies are 250 to 500 times per second). As you can imagine, the higher the switch frequency, the smoother the output waveform and the higher the resolution. However, higher switch frequencies decrease the efficiency of the drive because of increased heat in the power devices.

Copied probably from the same site you were looking at WWW.joliettech.com. Although other articles typically say the same if not similar explaination

The other advantages of IGBT's is that due to its higher switching frequency it interferes less with other signals and those frequencies are outside the audible range for humans so do not hum like older drives
the circuit also would use two IGBT's one for the positive and one for the negative. Just some side notes for your training
 
Here is my understanding. Hopefully DickDV will hop on with the very best info.

I think the original article is right, and the source of confusion is the difference between voltage at the IGBT and voltage at the motor.

In essence a PWM drive switches constant DC voltage. I believe the switching is done at opposite polarities, which is why some of hte pulses are above the 0 line and some are below it. Because of the inductance in the motor windings, this looks like the ragged sinusoidal current waveform in your second graph. The square wave voltage at the IGBT "looks" like a sine wave current and voltage at the motor because the damping effect of motor inductance "integrates" the wave form, filling in the gaps in the square wave pulses.

At different frequency most motors want variable voltage so the Volts/Hertz ratio is always constant. By varying the width and number of pulses in the square wave the "effective" voltage at the motor is reduced as the frequency is reduced.

So, Dick, how close was I?
 
The output switching of PWM is 6 steps just like the VVI type inverters. When you turn on the igbt, you put full voltage thru to the motor lead.

At all times 3 transistors are conducting. 2 from one bus and 1 from the other. The PWM shown is actually to the MOTOR neutral. Most induction motors are Y type wound and you have 2 leads in parallel passing 50% amps each and 1 in series with them passing 100% of the amps.

The wave form you look at with a voltage probe will always be choppy. The fuzzy wave in the picture is the EFFECTIVE voltage wave. Some of you might remember PWM DC servos, Turn it on more you get higher volts. It usually takes me about an hour in a class room to get across how the PWM wave is actually generated.

Due to the fact that an IGBT is a voltage activated device, not current like the tripple darlingtions used in the earlier "BI-POLAR" it takes a lot less power to turn on and off. Switching times go up. The amount of pulses can be increased which allows for infinately better control of the output.

PWM's down fall is with long lead length, you get some hellish voltage spikes at the motor. That is when output reactors help.

Get a current probe for you scope and put it on a motor lead. It will some what resemble the fuzzy sine wave.
 
My stuff looka better: http://www.patchn.com/mtrwhtpaper.htm

Does not matter whether IGBT or SCR, the incoming AC is converted to DC then a sinsoidal type waveform is created for the output to a motor.

Regardless of which is used the waveform created may be PWM or more of a sawtooth, it depends.
 
Last edited:
Okie, you and others are being victimized by some material published by Yaskawa and which is applicable only to their recent drive designs.

All other inverter manufacturers create their output with a DC bus (at 460VAC the bus will be at about 660VDC) being chopped by six output IGBT's, three of them connected to the + bus and each of the three motor leads and three of them connected to the - bus and the three motor leads. That way, each motor lead can be switched to either the + or - bus.

The output voltage pulses are all rectangular and 660V high. When recreating the + part of the sine wave they are all from 0 to +660V. When recreating the - part of the sine wave they are all from 0 to -660V.

Only two IGBT's conduct at any time connecting the proper motor lead to either the + or - bus.

The rectangular pulses vary in width only. When forming the beginning of a sine wave, the pulses are very narrow and by the time the peak part of the sine wave is being formed the pulses are all at maximum width.

The inductance of the motor takes these voltage pulses and basically integrates them back to a sine wave in the current. Lacking the inductance of the motor, there would not be any sine wave, only pulses. So, the motor is a key piece of the process of recreating the sine wave.

A few years ago, Yaskawa came out with a 12 IGBT inverter which basically splits the DC bus into a full voltage 660V bus and a 330V bus. Six of these IGBT's are switching to the higher voltage and six are switching to the lower voltage. The output pulse pattern is with 330V pulses of varying width when recreating the lower portion of the sine wave and with 660V pulses of varying width when forming the highest voltage part of the sine wave. The effect resembles the old six step VVI voltage waveforms except that the pulse width varies.

When Yaskawa first came out with this needlessly high component count design, it was touted as the solution to third and seventh harmonic problems in the motor leads and the marketing people had a field day. After careful testing however, it was found independently that the claim was bogus and the motor insulation problems were just as bad as with ordinary PWM. You may have noticed that now, three years later, there is no more marketing hype about this. Basically, they are doing with twelve IGBT's what everybody else is doing with only six.

It takes a fairly fast scope to see the pulses clearly. It is very hard to sync on any one pulse because they are all of varying width. On the Fluke scopes that I use in my VFD seminars, you have to use the HOLD feature to "trap" a pulse for examination. Letting the scope run will produce too much jitter and hash to sync on anything useful.

Hope this clears this issue up for you, Okie. If you've got more questions, ask again.
 
Thanks everyone for clarifying this. I will find a picture that better illustrates the typical IGBT drive output to keep the confusion factor to a minimum.

Special thanks to DickDV for explaining the Yaskawa "dual voltage" bus setup which explains where the picture came from...

Paul
 

Similar Topics

All Respected experts I want to know to check igbt moduel of siemens simovert vc drive while input is connected to drive pls help me i...
Replies
5
Views
3,103
Kindly, has anyone tried to control Lenze servomotors with Siemens S120 drives ? Any special hints ? Have some concerns for the resolver and servo...
Replies
5
Views
178
Hi , I am trying to replace an old inverter with a new one and its giving me issues . Once the new one was installed , the S7-300 PLC started...
Replies
3
Views
98
I am trying motor startup with pf525 drive by using wizard start over ethernet communication but during direction test it's showing control error...
Replies
3
Views
159
Back
Top Bottom