On Aug 16, 2009, at 1:13 PM, Stephen A. Lawrence wrote:
I looked this over, and two questions occurred to me.
1) What's the power factor when it's running?
It's DC. Power factor is 1.
It's got a back EMF, due
apparently to a hysteresis effect,
The hysteresis provides a field through which the current runs. This
provides the standard i L B force of the motor. The back emf is
provided by Faraday induction, as with most motors. The i L vector
that generates the torque(motive force) moves sideways through the B
(or vice versa) at velocity v, creating back emf epsilon, which by
Faraday is:
epsilon = d phi/dt = d (L * v * B)/dt
If there is no motion there is no back emf. If there is motion v
then the back emf should be proportional to v.
It makes the resistance *look* larger as a function of rpms because
the voltage drop across the motor gets bigger with rpms.
However, with this motor, B varies with v, and not linearly. B drops
if v is too high or too low. It likely is the case the motor is
running (especially when v is stabilized) in a range where B drops
with increased rpms, thus as v grows B drops nearly proportionately.
This, if solidly confirmed, is good confirmation of my hysteresis
model of the motor's performance.
This is my mental model of the thing anyway.
and on the face of it that word
"hysteresis" suggests the EMF and current may not be in phase.
It's DC, voltage and current have to be in phase because there is
only one phase.
In
particular, if the power factor varies as the RPM changes that could
have odd effects. (You may have mentioned that somewhere; if so,
apologies, I overlooked it.)
The main thing that varies is the resistance of the nichrome
resistors as they heat up.
2) Is there any way to measure the *resistance* of the running motor?
On page 12 of:
http://www.mtaonline.net/~hheffner/HullMotor.pdf
You can see the voltage drop of the single nichrome resistor
configuration was 0.063 ohms.
More recent experiments are documented at:
http://www.mtaonline.net/%7Ehheffner/HullMotor2.pdf
Note especially the experiment on page 6, and the wiring diagram for
it in Fig. 6. This experiment still used the 0.063 ohm cold resistor
at that time.
Here are the traces for the motor being stopped:
http://www.mtaonline.net/%7Ehheffner/HullVAmotorStop.jpg
The current through the resistor starts out at 5 V. The current is
thus (5 V)/(0.063 ohms) = 79 A, The voltage drop through the motor
starts out at about 2.8 V, so the resistance Rmotor = (2.8 V)/(79 A)
= 0.036 ohms.
I added some shunts to the nichrome resistor to increase current, at
the risk of blowing the battery. Here are traces from a regular run,
and here are traces from the stopped run using the new resistance:
http://www.mtaonline.net/~hheffner/HullShuntRun1.jpg
http://www.mtaonline.net/~hheffner/HullShuntStop1.jpg
I then jacked the starting speed way up and got this:
http://www.mtaonline.net/~hheffner/HullShuntHighRPM2.jpg
I'm wondering if it's the same as when it's stopped,
I don't think the resistance goes up very much because the motor
doesn't heat up a lot like the nichrome wires. In any case, when the
run starts the motor is cold, and thus the resistance is known. The
lack of back-emf change with rpms is clear right from the cold
start. Assuming the battery is fully charged, calculating the back
emf is easy, it is just the running voltage drop minus the stopped
voltage drop across the motor. Since the voltage drop across the
motor doesn't seem to change much despite rpms or run time, it seems
to be a bit of a mystery, at least to me.
but off hand I
don't see how to tease apart the effects of the back EMF from the
effects of the (possibly varying) running resistance.
Not sure why
resistance would be a function of RPM, of course, but on the other
hand
I'm not sure it wouldn't be, either!
The resistance is not a function of rpm. It is a function of
temperature, but the back emf effects start out when the motor is
cold and continue when it is hot.
Horace Heffner wrote:
Some typos corrected.
EXPERIMENT REPORTS
The Fig. 1 resistance R1 was increased by adding 4 nichrome
shunts, as
shown in Photo 1 below. Using the Fig. 1 circuit the motor moving
(Photo2) and stopped (Photo3) runs were made again, with a few
minutes
cooling time in between.
CH1
o
|
------(-)battery(+)--o---SW----Motor----
| |
| LED 4.7 k ohms |
| ----|<|---R2-------------- |
| | | |
-----o-----------R1-----------o---------
| ?? ohms |
o o
CH2 Ground
Fig. 1 - circuit to measure motor voltage drop
The results show a clear back emf effect. The resistors reach a
resistance plateau in 2-3 seconds when and as the motor runs (See
Photo2), and not when the motor is stopped (See Photo3). Two of the
filaments glowed, the old large blackened one, third filament from
the
top in Photo1, and the new one with fewest turns in it, second
filament
from the top in Photo1.
The stopped motor current stabilizes at 1.5 V across it or less, the
running motor stabilizes at about 2.7 V, giving a back emf of 1.2
V when
running.
I don't know why the back emf isn't higher than for the prior run,
which
had a stopped voltage across the motor of 0.7 V (due to lower
current),
and running 2.1 V, giving a back emf of 1.4 V. Perhaps the reason
is in
the prior run the manual start put the motor at a higher rpm than
where
it stabilizes, but the motor didn't get a chance to stabilize speed
because I had to cut it off due to the filament overheating. I
don't
see how it might have affected this, but I recharged the battery
before
taking this last set of data.
I'm pretty happy with the performance of the little motorcycle
battery.
Photo1: New probe configuration and shunts added:
http://www.mtaonline.net/~hheffner/HullShunt1.jpg
Photo2: Traces with motor running:
http://www.mtaonline.net/~hheffner/HullShuntRun1.jpg
Photo3: Traces with motor stopped:
http://www.mtaonline.net/~hheffner/HullShuntStop1.jpg
I thought one way to validate a back emf is to drive the motor to a
higher rpm and look for an increase in the back emf measured. I
stuck
a half inch buffing pad on my Dremel tool and stuck it into the
partly
exposed 1/2" shaft hole in the pulley and revved the thing up to at
least twice normal speed. I expected the back emf to double and
that
trurning on the power would slow down the motor. It didn't slow down
when power was turned on. If anything it just ran faster when I
threw
the switch than where the Dremel tool took the rpm. It appeared
to take
much longer for the filaments to heat up though, and the Channel 2
trace
in Photo4 below bears this out, showing the voltage across the
current
resistor R1 is almost flat at -7 V throughout the run. The voltage
drop
across the motor, shown in Channel 1 is nearly flat also at about
2.8 V.
The prior run stabilized at about 2.7 V, with the stopped motor
voltage
drop at 1.5 V. This means the back emf only increased by about 0.1 V
over the run in Photo2, even though the rpm doubled, and the motor
power
output apparently doubled with no increase in overall current.
From my hysteresis model, I expected torque to increase with RPMs
to an
optimum point where the magnetized material migrates into the
current i
such that i * M is at peak strength, and then to decline as RPMs
increase beyond that point because the material doesn't have time
to be
magnetized. What I would not expect is that the back emf would not
change significantly at all even though the RPMs doubled. It also
appears *superficially* that the motor power doubled and the
heating of
the current resistor dropped significantly, even though the voltage
across the resistor is measured at pk-pk 7.20 V, not too different
from
the 8.8 V for the stopped motor.
Weird. By starting at a higher RPM, the motor runs faster, system
current is less, yet back emf is unchanged. If the motor were not so
darned inefficient this would be a monumental discovery. The
inefficiency and quirky behavior of the hysteresis effect make
quantifying individual variables difficult.
Photo4: High rpm current start:
http://www.mtaonline.net/~hheffner/HullShuntHighRPM2.jpg
Best regards,
Horace Heffner
http://www.mtaonline.net/~hheffner/
Best regards,
Horace Heffner
http://www.mtaonline.net/~hheffner/
