> At higher frequencies, where given
a typical tube the plate C
> represents a significant part of the
plate tune cap, the effect is to
> significantly raise the plate
impedance (as seen by the Pi-Net input)
> and the balance of the pi-net would
need to be adjusted to maintain
> the desired Q & impedance ratio.
Not true. It sounds "nice", but it is
not factual. In truth, the
impedance barely changes.
Let's assume the blocking cap, at the
lowest frequency, would be
set a reasonable value of 10% of the
impedance looking into the
tank, or approximately equal to the
value of the tuning cap with a Q
of ten (using the simple but not
absolutely correct value of Rp/Xc.
Using an Xc equal to 10% of the plate
operating impedance, and
assuming a choke somewhat larger than
the plate impedance, the
change in impedance looking into the
tank is:
300pF blocking (~300ohms 1.8 MHz)
2878ohms
1200pF blocking (~75 ohms 1.8MHz)
2964ohms
The tank input impedance, with no other
changes except a
readjustment of the tuning cap by 2 pF
to compensate for the
reactance change, changes less than 100
ohms out of 3000
ohms....a totally insignificant change.
As I said, I learned my lesson many
years ago when I laughed at
someone for reading a blocking cap wrong
by a factor of ten in a
160 meter amp. After poking fun at him,
I was amazed when we
changed the capacitor to a "traditional"
value of 1000pF and
nothing we could see changed.
People waste far too much time worrying
about enough blocking
capacitance, and spend far too little
time considering current
ratings of the capacitors and the plate
choke design.....which has a
MUCH larger effect than the blocking
capacitor on system Q and
other problems.
And the small effect above is on 160
meters, where the problem is
at its worse point. On higher bands, the
effect would be less
because Xc would decrease with
increasing frequency.
Hi Tom
At the higher frequencies, (more
significantly), doesn't the blocking cap
form the top half of a tapped C
transformer with the
tube's Cout making up the bottom half??
Your example of 10% RL=blocking XC @ 50
mhz for RL=2500 ohms is 127pf blocking.
With tube Cout = 20pf
this works out to a new RL of 3349 ohms
(as seen by the PI network), using the
tapped C equation, (choke not a factor).
The same parameters but @ 29mhz
transforms the RL to 2977 ohms. At the
higher frequencies there normally is no
more room for re-adjustment of the
tuning cap
so the loaded Q has to rise with the new
higher RL.
This is how it plays on the bench when I
use a lower value of C-block.
I use a low value to capitalize on a
better current rating @ 50mhz but
it can get too small.
Paul
PAUL HEWITT
WD7S PRODUCTIONS
QRO HOMEBREW COMPONENTS
http://wd7s.home.att.net
--
FAQ on WWW: http://www.contesting.com/FAQ/amps
Submissions: amps@contesting.com
Administrative requests: amps-REQUEST@contesting.com
Problems: owner-amps@contesting.com
|