Fw: [Boatanchors] rotor starting capacitor

[Brian Clarke]

Hi Peter,

A dc electrolytic capacitor has one of its [usually rolled] plates coated in
aluminium oxide to form the insulator and dielectric. This oxide coating is
formed
by an electrolytic process - hence the name; similar kind of thing to
plating metals,
in a way. Hence, it is polarised - ie, it works when the voltage is applied
in one
favoured direction. The oxide layer is very thin - the actual thickness
determines to
a large extent the voltage rating of the capacitor. When a reverse voltage
is applied
to the capacitor, the electrolytic process reverses, the oxide layer gets
eaten away
and eventually the plates meet => short circuit => the 'attendant
consequences'.

If you think that using two polarised capacitors back-to-back = depolarised,
I do
believe you are mistaken. You are using two polarised capacitors. If you
depolarise
them, they will eventually short.

Let's try a little science. Measure the capacitance and Equivalent Series
Resistance [ESR] of each capacitor at installation and then every, say, 25
hours of
operation. I suspect the reason you and many others claim to have had such a
trouble-free run is that the actual time during which the capacitors are
reverse
polarised is very small, and so, the loss of oxide coating is fairly small -
so far. And
none of you has measured the capacitance or ESR over time to detect when
danger
is approaching.

Let's really put our money where our mouths are. Get a dc electrolytic
capacitor
and connect it to a low impedance ac source whose voltage is near the
operating
voltage of the capacitor. Tell me how long it takes for the 'attendant
consequences'.
I'ld suggest a large Pyrex dish be inverted over the experiment to contain
the
'attendant consequences'. I've done it in teaching electrotechnology
students - time
to failure? - about a couple of seconds! You can put a current limiting
resistor in
series if you want to reduce the clean-up time.

OK, let's look at this 'use the same voltage rating' hypothesis. What
happens when
you put two capacitors in series? If they have the same capacitance, they
share the
applied voltage equally. If they are of different capacitance, the smaller
capacitor
gets the larger share of the voltage. So, it's the ACTUAL [read 'measured'
rather
than what's printed on the outside] capacitance that matters, rather than
the voltage
rating. Of course, it would be unwise to over-stress a capacitor - you'll
get
punch-through of the insulator/dielectric, local carbonisation, then gradual
spreading and failure.

One respondent to this group suggested running the capacitors at way below
their rated voltage.This may slow down the measurable loss of oxide. But how
many of you have measured the capacitance at different operating voltages?
Electrolytic capacitors achieve their specified [+100%, -50%] capacitance at
near their rated operating voltage. So, if you use a pair of capacitors
whose
characteristics you have not measured, at well below their rated voltage, in
a
motor start circuit, what happens to the phase of the current in the motor
windings? The motor has to work much harder to achieve starting torque,
and you may well burn out the running winding. Is this a wonderful idea??

Next time you inspect an ac motor start capacitor, have a close look at the
tolerance - you'll find it is much tighter than on the dc electrolytics we
amateurs are used to consuming. Why? So that the motor doesn't get its
starting current exceeding the windings' current carrying capability.

Now, go ahead and connect your dc electrolytic capacitors in series reverse
polarity - and the very best of luck to you. Most of the time, the laws of
physics are not based on luck. Albert Einstein was well aware of this, when
he said 'God does not play tricks.'

Another respondent to this group suggested using a pair of steering diodes
around your pair of reversed dc capacitors. This is close to a good
solution.
Philips, in their book on diodes, published about 30 years ago, suggested
putting a dc electrolytic capacitor inside a diode bridge. This way the
capacitor
is always polarised correctly, you only need one capacitor, and you can
probably afford 4 diodes. My only caveat would be to measure the
capacitance to be sure it's within cooee of the motor's requirement before
re-installing the whole lot at the top of your tower. Measure the
capacitance
OUTSIDE the diode bridge, ie, the way the motor and the ac mains will see
it.

73 de Brian, VK2GCE.