Re: capacitor

Jon,

OK. "When in doubt, run an experiment" is my motto, so I did. Let's
start with the stepper first.

The stepper experiment:

(1) Take two identical step motors and couple one to the other via a
flexible coupler. In this case the motors are double-stack 34s, 4A
per phase, 4-wire.

(2) Connect a drive (Gecko of course) to one motor and use a 24VDC
power supply.

(3) Connect a 'scope to a winding on the undriven motor.

(4) I run the combo up to a speed just short of 5,000 full steps per
second at 24 VDC. The reason I stop is I don't want to bust my fine
Tektronix scope; the voltage across the undriven motor reads 400
volts peak-to-peak at 1.25 kHz!!!!

(5) I switch over to a sturdier (600VAC max) multimeter and get 800
volts peak-to-peak, 400V peak, or 283 VRMS at 10,000 full steps per
second where the whole mess stalls. All this from a 24VDC supply.

How is this possible? Hint: it's the phase relationship between the
motor's lagging shaft position and the driving voltage at high speeds.



The DC servomotor experiment:

(1) Take a large size 42 servomotor and hook it up to a drive (G320).

(2) Toggle the direction input at a 1Hz rate.

(3) Connect a DC power supply (60VDC) thru a rectifier to the drive
so no reverse current can flow.

(4) Put a 'scope across the supply pins at the drive to monitor the
voltage. Remember, no current can flow back into the supply, so this
should a cause tremendous voltage spike during deceleration.

(5) Increase the speed on the motor until it is pulling 20A during
decel/accel on each direction reversal.

Rather than a voltage spike, the scope shows a 5VDC dip in the 60VDC
line on each reversal! This is due to the voltage drop as 20A passes
thru the 18 guage wire I am using for this lash-up.

Where did the energy go that is returned during deceleration? It was
dissipated in the motor's armature resistance and the drive's
MOSFET "on" resistance. Exactly the same way as it was delivered to
the motor during acceleration in the first place. It is symmetrical.

Mariss

--- In CAD_CAM_EDM_DRO@y..., Jon Elson <elson@p...> wrote:
> mariss92705 wrote:
>
> > Leslie,
> >
> > That is a good "worst-case" approximation but it is overly
> > pessimistic. It neglects to take into account the rate of energy
> > return.
> >
> > A burning stick of dynamite releases the same amount of energy as
one
> > set off by a blasting cap; the difference is the span of time that
> > energy is delivered.
> >
> > It's the same here. Friction and electrical losses in the motor
and
> > drive absorb the returned energy unless the rate of return is so
> > large (very rapid deceleration of a large mass from a high speed)
> > that it results in a net negative power supply current.
> >
> > This phenomena is of more concern with step motor because
their "back
> > EMF" can be many times the supply voltage.
>
> What? How is this possible? The back EMF, strictly, can't be any
> greater than the voltage required to MOVE the motor at that speed.
> This doesn't count inductive effects, so if you short the winding
while
> it is moving, and then remove the short, high voltages could be
> developed.
>
> > The DC servodrive does not
> > exhibit this effect at all. Deceleration at any rate or load
> > generates no returned power, ever.
>
> Now I KNOW you stayed up too late last night! This, too, is false.
> How can the energy inherent in motion (inertia) be dissipated
without
> going anywhere? Most certainly, energy CAN be reterned to the power
> supply. Your drives may have topology that prevents this, mine
> CLEARLY return energy to the supply, as the voltage on the cap bank
> rises when the motors are decelerated.
>
> If anyone doubts this, connect a light bulb to a servo motor and
turn the
> shaft. The bulb (if appropriate voltage and current) will light
up, and get
> brighter the faster you turn the motor. How is this different from
decelerating
> a machine tool?
>
> If you mean that the back EMF of the motor must always be less than
the DC
> power supply, then you are ignoring the inductive effects of the PWM
> techniques. Linear servo amps are not generally able to produce
regenerative
> braking, and disspate the full energy as heat in their power
transistors. But,
> switching-type servo amps definitely can send the energy back to the
> power source.
>
> Jon