>> much lower working Q and so offer higher efficiency.
>I'm wide open to
> correction here, but I think If you are matching between two
> resitances, the L match has the lowest loaded Q. To get lower, you
> have to cascade L matches.
Looking at a typical inductor Q=300 and capacitor Q=5000
matching 50 ohms to 1000 ohms with 1500 watts on 1.83 MHz:
L network Qeff=4.36 C=380pF L=18.96uH and loss = 25.7watts
T network Qeff=4.7 C=389(2) L=19.45uH and loss=24.5w
Pi Qeff=4.7 C=389(2) L=19.45 and loss=24.5w
While the effective operating Q is slightly higher in the T and Pi,
loss is lower.
The main difference between all of these is the T would have the
widest impedance matching range of all three choices, and the L
the poorest impedance matching range for a given maximum value
of components. Both the pi and T are very restricted in matching
range when reasonable size components are used, and neither
have lower loss.
That's the reason almost everyone uses a T network.
I suppose some people might think a pi or L is somehow better,
but for a given maximum value of component size the T actually
works out best by far.
Another interesting thing is some people assume moving the balun
to the 50 ohm side of a T or pi makes the balun work better. That
isn't true either, so far as flux density in the core or overall balance
is concerned.
The balun sees the same common mode problems, no matter what
side of the network it is on. The exception is when a grounded
center true-balanced network is used.
I sometimes see L's and Pi's or even "balanced" networks with a
floating center point, and wonder why they bothered. All my big
tuners are T's, with choke baluns on the outputs.
73, Tom W8JI
W8JI@contesting.com
--
FAQ on WWW: http://www.contesting.com/FAQ/amps
Submissions: amps@contesting.com
Administrative requests: amps-REQUEST@contesting.com
Problems: owner-amps@contesting.com
|