[Amps] Direct rectification of AC mains to drive the amp,

[Manfred Mornhinweg]

> ## Has anybody actually measured the PF on a transformerless supply, with a 
> big C filter ???

I have done that. The result depends a lot on the impedance of the power 
grid!

Assuming that the grid has a negligible impedance, which is usually the 
case when we are considering very small power supplies of a few watts, 
the power factor changes from perfect to very poor, as we increase the 
filter capacitance from zero to a high value that provides low ripple at 
the output. A power factor as bad as 0.2 is typical!

Typically small off-grid supplies, such as those built into compact 
fluorescent lamps, intentionally use small filter capacitors that 
produce strong ripple at the output. This results in improved input 
power factor, typically between 0.5 and 0.6. Many CFLs are rated for 
0.55 power factor, and my measurements confirm that. I have measured 
0.56 and 0.57 on many of them.

But with larger power supplies, the impedance of the grid is never 
negligible. This impedance acts in series with out power supply, 
improving the power factor! I can imagine that a 3kW power supply, 
consisting just of a rectifier bridge and a big electrolytic cap, will 
be no worse than 0.5 at full load, when powered from a typical home 
outlet. I haven't measured that situation, though.

There are simple measures that improve the power factor: Lowering the 
filter capacitance is one. That creates more ripple at the output, which 
in many situations is acceptable. Adding a series impedance is another. 
This can be a resistor (cheap but lossy), or an inductor (more 
expensive, large and heavy, but better). Such series inductors are very 
often used in industrial power supplies (motor controllers) up into the 
100kW range. I have measured the power factor of such big things in 
industrial settings, with very low impedance power lines. With the 
proper chokes, about 0.7 to 0.8 at full load is typical. Without the 
chokes, forget it! Under 0.3. It can bring down pretty large circuit 
breakers.

A small series inductor along with a modest series resistor will produce 
the exact same power factor as a transformer input power supply. This is 
because a transformer improves power factor just because of its internal 
loss resistance and leakage inductance!

> ## Your typ commercial 1.5 kw CCS  SS  amp is huge $$ these days.  The 
> switching supply is the
> cheapest part.   $475.00  for a 3.3 kw CCS  switcher...... and unity/perfect 
> PF to boot ! 

Yes, sure... but those $475 exceed the total parts cost of my entire 
amplifier!!!  In this forum I usually talk about HAM amplifiers, not 
commercial ones. For a ham, the amplifier is just an accessory to 
practice his hobby, and has to fit a hobby budget. It has to be bought 
or built with whatever allowance his wife lets him have! It's very 
different from a commercial setting, where reliability is prime, 
performance is desirable, and cost is secondary.

Actually I can build an isolated, power factor corrected 3.3kW switching 
supply for less than $475 in parts. But not for very much less. 
Interestingly, the biggest cost is in electrolytic caps, inductive 
parts, and heatsinking, rather than the semiconductors!


> So the real savings
> in any transformerless scheme is minimal at best......

That depends! For example, it depends on the voltage ratio between input 
and output. When you want to get a regulated 13.8VDC from a 230VAC line, 
a switching supply with a transformer is actually the cheapest solution. 
  A transformerless one would cost MORE! But when you need roughly 
300VDC, needing no regulation and with low ripple not necessary, a 
simple rectifier/filter can't be beaten for cost! And when you need a 
regulated 150VDC from a 230VAC input, a non-isolated buck regulator is 
typically the cheapest solution, and will cost only about half as much 
as an isolated scheme, achieving better efficiency too.

Engineering is all about providing the cheapest solutions that will just 
fulfill all required specs. Using any solutions that are more expensive 
than the cheapest ones, providing either no advantages, or producing 
better specs than required, is bad engineering!

Manfred

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