Re: Power Factor Correction

To: tesla-at-pupman-dot-com
Subject: Re: Power Factor Correction
From: "Jim Lux" <jimlux-at-jpl.nasa.gov> (by way of Terry Fritz <twftesla-at-uswest-dot-net>)
Date: Fri, 30 Jul 1999 16:51:38 -0600
Approved: twftesla-at-uswest-dot-net
Delivered-To: fixup-tesla-at-pupman-dot-com-at-fixme


> 
> >
> >I notice that same AC capacitors (like ones from GE) are significantly
more
> >expensive than others.  They are all full-time motor run caps made with
> >polypropylene (I bet some or oil filled) but I was wondering why the
price
> >difference for what appears to be the same cap?  the 60UuF were $15 for
one
> >manufacturer and $65 for GE...  Digi-Key sells only 20uF but the price
is
> >faily good for the value I need even if I do need to use 7 of them...
> >
> >Are there any hidden things I should know about like inrush currents,
cap
> >explosions, openload or short problems?  I ran some models with various
> >faults and all looked very good but...

Motor run (and PFC) caps are designed for 100% duty cycle.. Remember that
the RMS current through the cap is fairly high ( 100 uF = 1/377E-4 ohms or
about 25 ohms, so the current is going to be 4-5 Amps (off a 120V line)).
As you know, caps aren't the best heat dissipators in the world, so if you
use a lossy dielectric (say a loss tangent of .01), you're still going to
be dissipating 5 watts..  Fortunately, it's at 60Hz, so it's not all that
tough to find a suitable dielectric.

Motor Start caps are designed for very short duty cycle operation (a few
seconds as the motor starts). Same C magnitude, but they can be built
cheaper, because they don't get as hot, etc.... There are probably 100's of
times more Start caps made than Run, so they are a lot more common, more
cheap, etc.

> >
> >Is there any easy way to calculate the cap value for a real coil with
all
> >of it's dynamics?  MicroSim could easily find the value but I was
wondering
> >if there was an easy equation or something like that.  The value is not
> >real critical.

Empiricism reigns, I suspect...
> >