Re: VFD's

--- In CAD_CAM_EDM_DRO@y..., Jon Elson <elson@p...> wrote:

>
>
> jmkasunich wrote:
>
> >Something isn't right with the currents you see on your
> >mill motor. I've never seen a motor that draws more than
> >about 50% no-load current, in 10 years of working with
> >VFDs and induction motors from 1HP to 1000HP.
> >
> Well, your work experience is a LOT broader than my
> meager experience, but I have 2 data points from recent
> observations. As I said, the current readout on my
> mill's VFD shows idling current of 3.0 A, and the
> motor's full load rating is 3.3A, and that shows up
> on the VFD at a reasonable load. Now, maybe the VFD's
> display is way off!

I don't doubt your observations - a couple of real world
measurements are sometimes worth much more than lots of
general experience. However I am inclined to consider
your results the exception and 25-35% no-load current
the norm.

Addressing your specific measurements - is 3.3A is the
motor nameplate amps? What is the nameplate volts
and Hz? I assume the Hz is 60 for a conventional motor,
but the volts might be 200 (for 208 nominal line), or
230 (for 240 nominal line). Is it a dual voltage motor?
What voltage is the drive set up to produce at 60Hz?
Is the drive set for straight volts per hertz control?
Is the line voltage close to nominal? Some drives will
adjust the output to correct for high or low line, others
will not. Was the drive running at or near 60Hz when you
saw the 3.0A no-load current? Currents can vary widely
at low speed, especially under 5Hz, but from half speed
up they should be pretty stable.

Maybe the motor you have is a real exception to the
rule. If so, it's a pretty important exception, since
Bridgeport pancake motors are a lot more common than
600HP ones in the home shop!

>
> I measured current on a capacitor-run single-phase
> motor on my air compressor. This compressor has an
> unloader, which is not the same as no load, but
> is pretty low, maybe 1/4 Hp. The motor has a 2 Hp
> rating. Current is about 9.5 A unloaded, and rises
> to 11.4 A just before the compressor goes to unload.
> I know that single-phase motors have worse power
> factor than 3-phase. I used a clamp-on for this
> measurement.

What I know about single phase motors would fit in a
shot glass with room to spare for a nice shot. I'm
not too surprised to see high idling current with a
single phase motor, but I don't know whether it means
anything with regard to three phase motors or not.

>
> <snip>
>
> >That and other experience lead me to say that no-load
> >current (magnetizing current) over 50% almost certainly
> >means something is wrong. Either there is actually a
> >load on the motor (belt friction, bad bearings, etc.), or
> >the drive is misconfigured and is overfluxing the motor.
> >For example, applying 240 volts to a motor rated for 208
> >volts can drive the flux high enough to cause partial
> >saturation, and the magnetizing current can rise to
> >70-80% of nameplate full load current.
> >
> >
> This is quite possible. I will mention that the motor
> runs remarkably cool, even after running for many hours.
> Motor and most head bearings were replaced when I rebuilt
> the head. The belts, even though it is a step-pulley
> version, still draw a bit of power. I will drop the
> belt off it and see what I get, and also try a clamp-on
> meter. I don't know how reliable that reading will be,
> though.

Cool motor is good - that would tend to dismiss my theory
about a 208V motor running at 230V. If you have a v-belt
head, I don't think there would be that much friction.
My comments about belts were directed more at ajdustable
speed sheaves. I am interested to see what results you
get with the clamp-on.

>
> I will also fire up my lathe motor that has a really
> nice Toshiba VFD, and see what it reads, no load. I
> don't have an easy way to load it, as the headstock is
> off the bed at the moment.

Load isn't needed. The 25-35% "rule" relates no-load
current to nameplate full load current. Actual current
under any particular load in a machine doesn't matter.
I doubt most machine motors reach rated load more than
a few percent of the time. Full load represents a pretty
heavy cut on any machine. Often rigidity or belt slippage
limits the cut long before the motor runs out of steam.

>
> The point I was originally trying to make, way back in
> the beginning, is that using a VERY large motor on a
> small VFD has one other side effect that is quite serious.
> The VFD is tuned for the expected inductance range of
> the motor. If the inductance is too low, like a big
> motor, then the on-times of the transistors can saturate
> the motor magnetics, because current will rise much faster
> than expected.

When you start talking about on-times, it seems you are
concerned about the PWM carrier frequency, and the current
ripple at that frequency. We refer to the 60Hz (or whatever)
current as the "fundamental" current, and the PWM frequency
current as the "carrier frequency ripple". On a scope,
the ripple is a triangle/sawtooth wave superimposed on top
of the fundamental frequency sinewave.

I wouldn't use the word "tuned" here. The carrier frequency,
and therefore the on-time, are not tuned to the motor, and
can often be changed by the customer. On the drives I am
familiar with, the carrier can be adjusted from 2KHz up to
10 or 12KHz. The larger drives only allow 4 or 6KHz, because
the switching losses get out of hand. Lower carrier frequency
means more carrier frequency ripple current, and more audible
noise from the motor, but a cooler drive. High carrier freq
means less ripple, less motor heating, and a quieter motor,
but the drive runs hotter. However, even at 2KHz, the ripple
current is usually less than 10% of the fundamental current.
At higher carriers, ripple is even less. At 10KHz, the ripple
is just a narrow band of fuzz on the fundamental waveform.

As long as the ripple is only 5-10% of the fundamental,
it isn't a major problem. With a "typical" 3.3A motor, I
would expect the magnetizing current to be about 30%, or
1.0A. This is the fundamental current at no-load. If the
ripple is 10% of the full load fundamental, or 0.33A, the
total no-load current is 1.33A. I would argue that a
drive rated at half the motor current (1.6A) should be
able to handle it. We both agree that a drive rated less
than half shouldn't be used.

> This is what blew up my first VFD, when I tried to run
> a 1 Hp motor with a 1/2 Hp VFD. It was ok at full speed,
> but passing through a low speed range, it would trip, and
> eventually popped a transistor.

The on-times and PWM are not much different between part
speed and full speed. There is a known instability of
induction motors that takes place near 15-20Hz (for a
conventional 60Hz motor). I don't know the details of
the theory (I work mostly on the power side, not the
control side), but perhaps instability set in and the
drive couldn't deal with it. I sure hope it wasn't one
of our drives - they are designed to safely shut down
with any overload, from a steady 10% overload, to an
oscillating motor, to a direct short circuit, without
blowing transistors.

> Later, I got help from Magnetek on using my air-bearing
> drilling spindle motor on their VFD. The solution was
> to put inductors in series with all the motor output
> lines. Even still, there are certain combinations of
> V/Hz that create long on-times of the transistors that
> make the inductors literally jump off the table and
> fault the VFD. The optimal setting seemed to be
> 75 V@400 Hz, but I had to set it for 80 V@400 Hz to
> avoid the jumping inductors. The air-bearing motor was
> designed for very low inductance, as it is really
> designed for operation in the range of 400 - 800 Hz.
>
> Jon

That high speed motor is definitely an exception to my
earlier mention of 5-10% ripple current. You would want
to use as high a carrier frequency as you can get from
your drive to keep the ripple down. As you found, even
with a high carrier, extra inductance might be needed.

If your inductors are "jumping off the table", something
isn't right. If I had that problem here, and I could
replicate it, the first thing I would do is put a current
probe on the motor (or inductor) leads and trigger a
storage scope on the fault. 100-200uS per division to
start with. Without a storage scope, I would be up
the proverbial creek without a paddle. When the scope
triggers, I would expect to see several PWM cycles of
normal sawtooth current, followed by the current heading
for the ceiling when the fault occurs. The current slope
is more-or-less fixed by the inductance and DC bus voltage.
If the inductors are large enough to limit the di/dt during
a normal PWM cycle, then the transistors would have to
remain on for several consecutive cycles to do what you
describe. Straight V/Hz control shouldn't do that. Vector
control, or sensorless vector, or some of the other fancy
schemes that some drives have, might do that, and the first
thing I would do is try to disable all that crap.

We might be getting off topic, but I'd be happy to
continue this by email. It's very interesting.

jmkasunich at ra.rockwell.com

John Kasunich