The following has been updated to improve construction information
and to note a follow-up experiment using graphite powder lubrication
in the stainless steel bearings:
http://www.mtaonline.net/~hheffner/HullMotor.pdf
For convenience, the sections updated follow:
Experiment Construction
Some nonmagnetic stainless bearings were obtained from KMS Bearings:
http://www.thomasnet.com/catalognavigator.html?cov=NA&what=nonmagnetic
+ball+bearings&heading=3920402&cid=270891&CNID=&cnurl=http%3A%2F%
2Fkmsbearings.thomasnet.com%2FCategory%2Fradial-ball-bearings-3
http://tinyurl.com/mk3o4d
Two SSR8A-6-1/2B bearings with 1/2” ID, 1 1/8” OD, 3/8” width were
ordered. They are single row, 316 Stainless Raceways Radial Ball
Bearings fitted with SS316 Balls.
Their lubricant washes out with soap and water. I didn't, but
probably should have ordered the model with a high temperature cage,
but it only has to run (or not run) for a few seconds to test the
thermal expansion hypothesis.
In addition, two Z99R8 1/2” ID, 1 1/8” General Bearing Corp. ordinary
steel bearings were obtained locally. Fig. 5 is a photo of the
four bearings obtained.
The stainless bearings were washed with hot water and Goo-gone soap.
Both sets of bearings were then given washes in gasoline, followed by
24 hour washes in mineral spirits, and then acetone. The extra washes
were due to initially poor and intermittent conductivity through the
bearings.
FOr use as flywheels, two all (nonmagnetic) zinc 4 inch OD die cast V-
belt pulleys, at $6.70 each, were obtained from:
http://www.mcmaster.com
The steel setscrews were removed, and plastic electrician's tape was
used to adhere the pulleys to the shafts. The shafts were made of
1/2” solid aluminum bar obtained from Home Depot.
Dimples were made on the shafts in order to limit the range of motion
of the bearings on the shafts, and to increase the firmness of
electrical contact.
Fig. 6 shows a side view of the experiment board under construction,
including the 12 V motorcycle battery purchased at Wal-Mart.
The Experiments
Fig. 7 shows the experiment board wired with the two motors in
series. The nichrome resistor was measured at about 0.3 ohms.
Channel 1 of the scope is across the battery, Channel 2 is across the
current resistor.
This experiment did not produce continuous rotation of either type
motor, due to insufficient current. The oscilloscope trace did show
good series conduction through all 4 bearings however.
It was decided to next try a single motor at a time, to be sure
enough current could be obtained. Single motor mode was achieved by
shunting across the bearing mounts on the left and moving the
magnetic bearings over to the right side. No luck first try. The
resistor was cut in half, to 0.15 ohms roughly. Measured 10 V across
the resistor, and a 1.6 reduction in potential across the battery,
givnig 67 amps for the run, but still no luck getting sustained
rotation. The magnetc bearings did show signs of an increased spin
down time.
Then the resistor was cut roughly in half again, giving the
configuration shown in Fig. 8 below. The resistance was thus about
0.075 ohms, the voltage drop across the resistor 9.20 V, and the
voltage reduction across the battery was 2.00, down to 10 V. The
current was therefore 123 A. Fig. 9 shows the scope post run. This
current estimate seems high, so if actual current should be important
the resistance of the nichrome can actually be determined by bridge
method. The motor ran, not very fast, but sustained a faster rpm.
Here is a video of this control run.
http://www.youtube.com/watch?v=A2XBPzxXtJk
Note the resistor slowly picking up an orange glow as the run
continues, and dropping it when the power is cycled off.
Next, for the live run, stainless bearings were exchanged for the
magnetic steel bearings. The voltage drop across the resistor was
8.80 V, the voltage reduction across the battery was 2.4 V, to 9.6
V. The current was 117 amps. The scope trace is shown in Fig. 10.
It proves that current was flowing uniformly through the stainless
bearings. It was 117/123 = 95 percent of that flowing through the
magnetic bearins, so is definitely enough to show a positive effect
on spin down time, rather than the negative effect observed. Here is
the video of the test run:
http://www.youtube.com/watch?v=3cllaQFkxQQ
Notice how quickly applying current puts the breaks on. This is just
the opposite of the magnetic bearings, which speed up. The same
grinding noise does appear when the current is on. This may be due
to arcing.
The nonmagnetic bearings were removed from their shaft. One of them
had a rough feel to its rotation, confirming that arcing and possible
intermittent welding had roughened up either the balls or the races.
Powdered graphite lubricant was added and the bearings work a while,
which improved spin down time, but the bad bearing still made noise.
When reassembled and a run made, the griding noise when power was
applied had disappeared. The voltage drop across the 0.075 ohm
resistor was 8.5 volts, giving 113 amps, only a slight current
rduction. The voltage reduction across the battery was 2.4 V. The
breaking effect and noise from applying current disappeared. There
was still no sign of increased spin down time.
Conclusions
It is discouraging that the stainless bearings could not be made as
friction free as the magnetic ones. Their plastic spacer rings are
the likely source of the friction problem. Use of powdered graphite
is recomended because it reduces friction, maintains good electrical
contact, and avoids the problem of arcing quickly damaging the
stainless steel bearings.
It would have been an improvement to run these tests with more
current, but the nichrome resistor is about at its limit. The
battery probably is too. However, the videos demonstrate clearly
that the nonmagnetic stainless steel bearings slow when current is
applied, and the ordinary magntic steel bearings speed up to a
sustained motion. This result eliminates thermal expansion as an
explanation for the performance of the Marinov Ball Bearing Motor.
Best regards,
Horace Heffner
http://www.mtaonline.net/~hheffner/
