[Amps] Power factor

[Manfred Mornhinweg]

Jim,

I will take you message as a cue to write another of mine, but it's 
addressed to everyone really, and specially to those who are not up to 
date with power conversion technology.

Before some people here on the forum think that power factor and its 
correction is too big of a problem, I think it would be good to state 
that this is NOT so!

To improve the power factor of a rectifier/capacitor group, either a 
passive or an active system can be used.

The passive system consists simply of a series inductor. The idea behind 
this is the following: As has been said, a rectifier/capacitor group 
draw strong current pulses, near the peaks of the voltage waveform. To 
be exact, it draws the strong current from some time before the peak 
voltage, until just after the peak voltage. After that the current 
stops. So we have two problems: The waveform of these pulses is very 
different from a half sine, and it is somewhat early in phase.
A properly chosen series inductor will reduce the peak value of these 
current pulses, stretch them out, and will also phase shift them. If the 
inductance is properly chosen for the power level drawn, the phase of 
the current can be made to agree very well with the phase of the 
voltage, fixing one side of the power factor problem, and as a 
by-product the pulses are softened and broadened, making them assume a 
waveform closer to the half sine which would be optimal. The result is 
that such a simple inductor in series can improve the power factor very 
significantly, but it cannot make it perfect. Also, this correction is 
strictly optimized for one fixed power level. At any other power level, 
the correction will be less good. Usually the inductor is chosen to 
produce best power factor at full load. At partial load the power factor 
is then worse, but that's less bad than if it were poor at full power.

The series inductor is simple and reliable, but bulky, heavy, only 
partially effective, and somewhat costly.

The active system instead consists of a simple switching boost 
regulator, which is controlled in a special way. This boost regulator is 
spliced in between the rectifier and the filter capacitor, and consists 
basically of one MOSFET, one diode, and one inductor, one small filter 
capacitor, plus the control circuit.

The control circuit senses the voltage on the main filter capacitor, and 
produces an error signal that has a low frequency response - 
considerably lower than the line frequency. This error signal is 
multiplied with a sample of the rectified, unfiltered input voltage, and 
the resulting product is used to control the current flowing in the 
boost converter. As a result, the waveform of the current taken from the 
line accurately follows that of the line voltage, is very well 
phase-aligned, and the average magnitude of the input current varies to 
maintain the output DC voltage roughly constant.

This control system might sound complicated to make, but it comes all 
ready made inside an IC that costs less than one dollar! These ICs are 
manufactured by the hundreds of millions, and used in every conceivable 
electronic equipment.

To build such an active power factor correction circuit, one just needs 
the MOSFET, diode, inductor, IC, a capacitor of a few uF, and a very few 
small passive components. The hardest part for most people is the 
inductor. This uses a ferrite core, can be bought ready-made for the 
lower power levels, but for 3kW it probably needs to be wound at home. 
One of the largest ferrite cores typically available will be needed, and 
it will typically take something like 50 to 200 turns of wire.
Except for that inductor, whose cost is highly variable depending on 
where you get it, all the other parts for a power factor correction 
circuit for the 3kW level will be about 40 dollars. This will provide 
about 0.98 or better power factor, regardless of the actual voltage 
waveform in the mains, and across essentially the whole range from 
nearly zero to full power!

So, if any law should be passed that requires us to have excellent power 
factor even in our homebuilt equipment, this can be done simply and 
inexpensively even for a legal limit amplifier. For lower power 
equipment, it costs just a few bucks, as little as 5 dollars for small 
ham power supplies. So there is no reason to worry.

As long as there is no such law, we have the choice of including such 
power factor correction, or skipping it. Without PFC, we save a few 
bucks, about two percent loss, and some potential for RFI. Mainly the 
latter two are the reasons why I use active power factor correction only 
when it is truly required or clearly advantageous.

In high voltage power supplies for tube type amps, active power factor 
correction can be included too. In principle it could be done on the 
secondary side, after the transformer, at high voltage, but that would 
require very expensive MOSFETs and diodes. It would be stupid to use an 
expensive, heavy transformer, and then do PFC in an expensive way! So 
the obvious solution is to use a switching power supply to make the high 
voltage, and build the PFC circuit into it.

This switching supply might not even need regulation - after all, the 
PFC circuit itself delivers pretty well regulated DC, and tube 
amplifiers usually run from unregulated supplies. So, a very simple and 
effective, power factor corrected tube amp power supply could consist of 
an active power factor correction circuit followed by a simple, 
unregulated upconverter, operating somewhere in the 30 to 100 kHz range, 
using a MOSFET bridge with a free-running oscillator/driver chip, a 
compact ferrite core transformer, and a high voltage rectifier made with 
superfast silicon carbide Schottky rectifiers, followed by a small 
filter capacitor (less than 1uF is plenty).  This technology is 
available and inexpensive. Digikey, for example, has most of the parts 
in stock. Only the ferrite cores would have to be ordered elsewhere.

This power supply might weigh 3kg total, would produce much better 
voltage stability and less ripple than a conventional supply, have near 
to perfect power factor, and might cost around 150 dollars in parts, 
which is likely far less than a conventional transformer based supply.

Electronic engineering is the combination of money (for components) and 
brains.  The more you use of either, the less you need of the other.

Manfred

========================
Visit my hobby homepage!
http://ludens.cl
========================