Because the spinning motor stores significant energy. This is how RPC
are able to generate the missing phases.
Recall the long discussions of sizing RPC motors to handle the startup
surges of various machine tools, and the staged startup scenarios (where
one starts the smaller motors first, to have enough spinning capacity to
later start the larger motors).
It makes no difference in this picture if the first source of 3-phase is
a RPC, or a VFD.
To my knowledge, for an induction motor to act as a generator, and
that is in essence what you're proposing, it must be rotating over its
synchronous speed. Clearly this is not the case if the motor is
powered by a VFD, and it won't be the case if you switch in a second
motor while holding the frequency constant. I guess you could drop the
drive frequency at the same time the second motor is started and take
advantage of the idler's kinetic energy, but that hardly seems
It only has to be spinning a bit faster than what it's trying to drive.
It does not have to be at synchronous speed. If it did, no induction
motor would work.
An unloaded "idler" motor will by definition spin faster than anything
under load, or anything that is starting.
If you plot motor torque versus speed you get an "S"
curve which passes through zero at synchronous speed.
At all speeds below synchronism, torque is positive. The
motor draws sufficient current from the supply to provide the
mechanical output. In the case of an idler this is sufficient
torque to overcome windage and friction.
The machine CANNOT return power to the supply unless it is
externally driven to the negative torque region above synchronous
These comments apply to a 3 phase motor driven from a true
3 phase supply. This differs from the case of a 3 phase motor
connected to a single phase driven idler acting as a rotary
In this case, although the idler is mechanically running
below synchronous speed, the single phase excitation induces
rotor currents at slip speed which cause the magnetic field
rotation to occur at line frequency. It is THIS magnetic field
rotation that generates the output at the phantom phase (at the
other terminals it is of course the back EMF). The load that is
delivered by the phantom phase is provided by a corresponding
increase in single phase input power - it is still acting as a
motor - there is no net power generation.
Yes. The motor has to be slipping to develop any torque.
Umm. The situation is that there is a motor that is turning slower than
the idler. Why cannot the slower motor draw power from the faster
motor, regardless of the VFD?
I'm not so sure about that, as the idler stores energy in either case,
simply because the idler is spinning.
True for sure, but is it relevant? The situation under consideration is
an idler driven by a three-phase power from a VFD. The question is if
motors abruptly connected can draw power from the stored energy of one
or more large spinning idler motors. One reason to believe that it can
is the observation that it matters what order one starts motors with a
RPC. The game at each step is to ensure that what is currently spinning
is bigger than what one is now trying to start, the unspoken objective
being to ensure that the new motor isn't able to stall the current
collection by demanding too much too quickly.
On Tue, 10 Feb 2009 23:09:44 -0500, Joseph Gwinn
The mechanical energy stored in the idler rotor rotation is
largely useless. Useful energy can only be extracted from it if
the rotor speed falls below the supply source/VFD defined
The reason for sequential rotary converter startups is reduction
of phantom phase voltage drop due to the source impedance of the
phantom phase winding. This is mainly determined by the series
resistance and leakage inductance of that winding. The bigger the
machine the lower the source impedance.
In a rotary converter multiple system all machines are
directly connected in parallel. All machines that are spun up to
speed contribute - its the total spinning HP that counts.
This part I'm not following. Generators can motor, and motors can
generate, although they are usually best at one or the other role.
Let's simplify the picture a bit:
Consider a number of 3-phase induction motors of similar size connected
together, in parallel across a common power bus, all having been spun up
while that bus was connected to commercial 3-phase power.
The motors are not mechanically connected in any way.
Disconnect the commercial power, so now the motors are all coasting, but
still electrically connected to one another.
One of the motors drives an adjustable mechanical brake/dynamometer,
which is then applied. How much energy is dissipated in the
dynamometer? Is it only the mechanical energy of the one motor, the
energy of all connected motors, or some other intermediate value? How
do we know?
Said another way, can we stop one motor without affecting the connected
These two paragraphs seem to contradict one another. The first implies
that only the RPC motor matters, and does this by being a better
transformer between the single-phase power coming in and the three-phase
power going out, so the effect is that startup surges are supplied
entirely by a corresponding instantaneous surge in demand on the
This no doubt happens, but if it's the whole story, why does total
spinning HP matter, versus the HP of the RPC motor alone?
Yes, it does. This back EMF is precisely the motor acting as a
I've been thinking about the issue, and there are a number of datapoints
VFDs will trip (complaining of overvoltage on the DC filter capacitor)
if trying to stop a motor driving too much rotating mass. The classic
solution is to provide a braking resistor to absorb the energy causing
the overvoltage. The VFD manuals all say that the energy comes from the
kinetic energy of that spinning mass. The more mass, the more energy to
be handled, the hotter the braking resistor gets.
When starting, the rotor is stationary and so the motor draws the
locked-rotor current. The difference between current at normal load and
with the rotor locked is around six to one. As the rotor spins up, the
current drops. What's happening is that the motor is generating a back
EMF proportional to rotation speed, so at speed the motor is driven by
about 1/6 the supply, and the back EMF is the other 5/6 of the supply
voltage. This is true of almost all motors, not just polyphase
Induction generators do exist, and are usually based on stock induction
motors, most often three phase.
There are many articles on induction generators available. The best
article on the theory I've found so far is "Induction Generator Theory
and Application", J.E. Barkle and R.W. Ferguson, AIEE Transactions
(Power Apparatus and Systems), February 1954, pages 12-19.
The big problem with induction generators in practice is that it's too
hard to control them, and it's cheaper to use a wound rotor (allowing
control by varying the field current).
Right, and it occurs because at any point in time (while the VFD is
decelerating the motor) the motor's speed is greater than its
synchronous speed, as determined by the VFD's output frequency at that
Yes. The VFD provides the current to maintain the field the motor uses
to be a generator.
If one instead has a resistor across the motor and abruptly disconnects
the power, leaving resistor and connected motor to their own devices,
the field is maintained by regeneration, and too much load on the
"generator" will cause the regeneration effect to collapse.
Pentagrid has mentioned this regeneration and collapse in past postings,
and I've also seen it discussed in textbooks.
With a VFD driving multiple parallel motors, some of which are more
heavily loaded than others, all these effects will play out
simultaneously, and the net result will depend on the details.
I have VFDs for my tools, so no RPC to play with. I may have to get a
pair of three phase motors just to experiment upon. My local source of
cheap motors vanished a few years ago, so it may be a while.
Almost zero - as soon as the supply is removed the magnetic
field which it induces into the rotors collapses. Without this
magnetic field, although the rotors are still rotating, the
motors can no longer generate back EMF.
The first paragraph refers to the basic case of a single idler
motor starting a single load motor.
The second paragraph considers what happens if if there are
multiple motors in circuit. Each of the motors that are spun up
to speed are capable of supplying power to the phantom phase.
Because of this the parallell connection results in a
correspondingly lower phantom phase source impedance.
The customary distinction between an RPCidler motor and a
load motor is entirely artificial. The setup is simply two or
more motors all connected in parallel. Any of the motors that are
spun up to speed act as RPCs and can supply power to the phantom
phase. Equally, it is perfectly acceptable to place a mechanical
load on a motor nominaly designated as an idler
This is the key. Now, you don't mean almost zero, but I know what you
meant: the dynamometer will dissipate the mechanical energy of the
motor mechanically connected to it, but not the mechanical energy of the
I recall discussions of people using induction motors as generators.
Once started, they would generate their own field and keep on going. If
one opened the circuit, the field would collapse, and peobably would not
This needs to be tested experimentally. I would try this, but don't
have enough big three-phase motors to do it, even though it is an easy
As discussed above.
This experiment is the key to the rest of the discussion as well.
And all motors are doing an instantaneous phase transformation, but not
storing much mechanical energy?
Given that the single-phase supply is connected to one but not the
other, one would think that there has to be a distinction.
Although I understand in general how RPCs work, I've never read a real
theoretical analysis of this. One thing that nags me is that there has
to be energy buffering by the RPC, or there would be no way to get from
pulsating power (singlephase) to continuous power (polyphase). And the
rotation of the rotor is essential, as no static transformer or
collection of transformers can do this, while there are transformer
configurations to convert between N-phase and M-phase, where M and N are
all at least two. Something is smoothing those pulsations out, filling
the gaps in. Can you suggest some learned articles for me to read?
I think of a RPC as a three phase motor running on single phase and
connected to a three phase induction generator. Where the three phase
motor and the induction generator are one and the same motor. Try
thinking of it first as a single phase motor mechanically coupled to a
three phase motor. Once the single phase motor is running, there is
no capacitors in the system.
With a sine wave single phase power in, there are not what I would
consider " pulsations".
Joseph Gwinn sez:
". . . .Can you suggest some learned articles for me to read?"
You've already seen the best description of RPCs in Jim's comments. Go back and
read them again.
The key here is Jim's reference to the currents flowing in parallel. Try to get
notions re. "generator and load". A RPC is a very complex network consisting of
through 2 or more 3-phase motors connected in parallel with each other and with
1 set of those legs
in parallel with the input mains.
On Wed, 11 Feb 2009 23:18:32 -0500, Joseph Gwinn
All motors can phase transform. They also store a significant
amount of kinetic energy but this is not available as a useful
ALL motors are parallel connected so every motor receives
single phase excitation.
The torque pulsations of the single phase drive are
smoothed out by the rotor inertia.
Nothing really appropriate comes to mind but you might find
my own book "Electric Motors" ISBN 13 978 185486 246 4 reasonably
I suppose we could posit a 3-phase 220/440 motor, with single-phase
power supplied to the 220-volt inputs (center taps), but with the
440-volt terminals of the various motors connected together.
So the energy is stored in the rotor's mechanical energy. That's the
fundamental meaning of "inertia" in this context.
The electrical parallel of inertia is inductance, which is the ratio of
energy stored in the magnetic field to the current causing the field.
The higher the energy, the higher the inductance.
I'll get a copy. To be continued.
But I'm still interested in learned articles on how RPC work, at the
Polytechforum.com is a website by engineers for engineers. It is not affiliated with any of manufacturers or vendors discussed here.
All logos and trade names are the property of their respective owners.