View Single Post
  #1   Report Post  
Posted to sci.electronics.repair
Robbie Hatley Robbie Hatley is offline
external usenet poster
 
Posts: 25
Default How do variable-speed 3-phase motor drives actually work?

Greetings, group. Where I work, we repair several different types of variable-speed
3-phase motor drives. All of them I've seen so far have very similar circuitry.
The main power-control module generally contains:
6 hi-current hi-voltage silicon diodes in 3-phase bridge (to generate +320VDC)
6 hi-current hi-voltage IGBTs in 3-phase bridge (to power the motor)
thermistor (for temp sense)
braking IGBT (in some units)

The IGBTs are always wired as 3 push-pull pairs, with the 3 center points
connected to the 3 phases of the motor. (The down side of this technology
being that if both transistors on the same phase turn on at same time, the
IGBTs explode with a deafening "BANG!!!" and you're out $65 in 65ns.)

I had assumed that the IGBTs would be switched so as to generate 3 AC waveforms,
120deg out-of-phase with each other, and that the frequency would be variable
from about 2Hz to about 60Hz, with motor speed proportional to frequency.

But no such thing is true!!! When I actually looked at the waveforms at the gates of
the IGBTs, with the motor at lowest speed (about 0.5 RPM), the frequencies of
all 6 IGBTs are all about 400 Hz, about 50% duty cycle. When I turn increase
commanded speed, the frequencies all stay at 400 Hz, but the duty cycles
begin to fluctuate, with rate of fluctuation of duty cycle being equal to motor
speed in revs/second, and amplitude of fluctuation also increasing with speed.
(Eg, if motor is turning at about 10 revs per second, the duty cycles of the signals
are fluctuating from about 40% to about 60% at about 10 fluctuation per second.)

So how the hell does this bizarre technology work? If the frequency is always
stuck at 400Hz, why isn't the motor always turning at exactly 24000 RPM?

(My guess: at minimum speed, the three signals are almost perfectly in-phase;
but at higher speeds, the phases fluctuate slightly away from 0deg, with the
fluctuations "cycling" around the 3 phases at a rate equal to motor revs/sec,
but the "carrier" frequency still staying at 400Hz all the while. But I'm
just guessing.)

So, does anyone here know what this bizarre motor-control technique is
called? (Hard to google something one doesn't know the name of.)
And, roughly, how it works?

--
Curious,
Robbie Hatley
lonewolf [at] well [dot] com