Designing for high Power Factor

["Tesla list" <tesla-at-pupman-dot-com>]

Original poster: "R.E.Burnett by way of Terry Fritz <twftesla-at-qwest-dot-net>" <R.E.Burnett-at-newcastle.ac.uk>


Hi guys,

Original Poster: "Mike Doyle" wrote :

>  How would one go about designing a high PF system?

Maybe I should have explained this power factor thing better,  but I
didn't want to make my original reply too lengthy.  I have been looking at
AC resonant charging behaviour for some time,  but have been delaying
presenting my findings here until I was sure they were valid for both 50Hz
and 60Hz countries.

In a system using a separate power transformer and external ballast, it is
possible to achieve a power factor of 0.85 or better for a wide range of
rotary speeds by careful choice of the BALLAST and TANK CAPACITOR
components.

Rather than considering the ballast and tank cap values separately, I
believe we should really work with them as a pair.  It is not the actual
ballast and capacitor values which are important.  Instead it is the
resonant frequency,  and characteristic impedance of the charging circuit
which define its behaviour.

The resonant frequency of the charging circuit essentially defines the
charging profile of the tank cap after each bang.  It controls all of the
timing related stuff like Power Factor,  and response to various break
rates.  For any chosen rotary firing rate there is a natural frequency for
the resonant charging circuit which gives best power factor.

The characteristic impedance of the charging circuit defines the power
throughput aspect of its behaviour only.  Resonant frequency changes with
the product LC,  and the impedance is proportional to L/C,  so it is
possible for us to control each of these things separately by manipulating
the ballast and tank cap values together.

Eg. Double L, double C,  to  halve Fres and leave Z unchanged,
    Double L,  halve C,  to  double Z and leave Fres unchanged.

If the resonant charging circuit is designed such that it has the correct
natural frequency,  then the values of the ballast and tank capacitor can
be tweaked together to set the characteristic impedance and achieve the
desired power throughput.

Therefore I propose the following design approach:

1. Choose the desired rotary BPS,

2. Find the optimum resonant frequency for the charging circuit,
    (this fixes the LC product and guarantees good power factor at
    the chosen rotary speed.)

3. Adjust characteristic impedance to get desired power throughput.
    (this fixes L/C,  so values for ballast L and tank C can now be
    calculated.)

Using these L and C values will provide the desired power throughput at
the best achievable power factor. (Like biggest spark for minimum
current.)  (It's difficult to get a PF better than about 0.93 because of
the current waveform is not a pure sinewave.)

There is more information about this at my web site:

www.staff.ncl.ac.uk/r.e.burnett/async.html

Be sure to click the link at the bottom for externally ballasted supplies.

This stuff is kind of heavy going,  but I am planning to develop a simple
computer program to do all of the clever stuff.  I was inspired by Bart's
excellent MMC Designer,  and realised that there really is not much
information about designing AC resonant charging circuits.

The program will ask the user to input the chosen rotary speed in BPS  and
the desired power throughput in Watts.  It will then calculate the ballast
inductor and tank capacitor values necessary to achieve the specified
power throughput at the chosen BPS,  with good power factor.  I'm still
working on this,  and will be asking for some people to volunteer as Beta
Testers shortly ;-)

If there is much interest on this list,  I will try to go through a 
"Design Example".  Something like "What ballast and cap values would give
10kW at 371BPS with a PF of 0.9 ?"  I think that correct design of the
charging circuit is very important,  however I don't want to bore the
socks off those who don't really care about this stuff, and get along just
fine without it !

							Cheers,

							-Richie,
							(Newcastle, UK)