Pete:
Yes, restricting the transformer to <4MHZ makes a big difference in
its design. It can make a HUGE difference. Unfortunately, as happens way too
frequently here, people who post the original information are not entirely
accurate. I was simply responding to a posted opinion that someone decided it
to be strategic to redesign the PINET for 200 ohms out, and then step it down
with a broadband transformer. Well, those of us who design and build such
transformers, especially high-power ones, consider this to be a serious
challenge and not something to be taken on to satisfy some whim. If it has a
payoff, it better be a BIG one.
This is the type of 'argument' that is very difficult to resolve,
because the outcome is not a clearcut victory for either school of thought.
But I think most readers of this would agree with me on this: I would rather
live with a couple of high-power fixed padders on Cout (a small amount of
space required, close to 0 loss) rather than a a pair of Dunkin Donut size
cores (large space and moderate to high loss). Admittedly, there is not a
huge difference, unless overall size and weight are important. But I don't
see a single advantage for the 2 stage impedance matching scheme.
I have made it a point to read anything that crosses my radar screen
from Collins Radio/Rockwell. Particularly info on the design of RF
communications gear and power amplifiers. It is interesting that within those
archives you encounter just about every concept in RF engineering that has
ever surfaced (I just today read an article on multi-stage neutralizing
techniques in BIG amps). Well, I don't recall ever seeing this concept being
massaged by that group, but I have seen many other schemes related to plate
circuit matching systems instead. In general, they seem to have avoided the
use of ferrite in RF power circuits, except where absolutely necessary.
Probably because they also knew the other dirty little secret of these cores:
they produce distortion, in copious amounts as the power level increases
towards saturation.
73
Eric K8LV
|