>
>> Is there a formula for calculating tuned input circuit Q. I see a
>> chart in Bill Orr,s Handbook for various input impedances with Q,s of
>> 2-3. Im probably missing something here. How would you calculate for a
>> tuned input Q of 5. Im trying to learn so if its a dumb question I
>> apologize for my lack of experience. Is it roughly Rin/xc1? Im curious
>> as how to arrive at the values for a tuned input pi net for a given Q.
>> 73,s Jeff
>
>Hi Jeff,
>
>Most Handbooks use the Q of one end of the network as the Q, but
>that is actually incorrect. The actual Q involves all parts of the
>system.
>
>Eimac almost certainly is using the entire network's Q, rather than
>the Q at one end of the network.
// Almost is undoubtedly right. In Care and Feeding, Eimac figures Q
based on Xc1.
>
>The nice thing is Q has very little effect on the system, it is fairly
>non-critical as long as the Q is more than 1+ the square root of the
>ratio of the impedances. Less than that and the network won't act
>like a pi. To behave like a pi, the network has to look like a step
>down and step up L network connected back to back. In other
>words, phase shift has to be somewhat more than 90 degrees and
>the impedance someplace in the network to ground has to be less
>than the lowest of the source and load impedances.
>
>In many cases I have looked at, the "simple Q" of two is nearly an
>overall Q of five. Orr and Eimac actually somewhat agree.
>
>Other than matching, which sets the minimum usable Q, the only
>other reason for a high input system Q is to provide a reasonably
>low shunt impedance for harmonics. That prevents the cathode of
>the tube from moving around at harmonics caused by the shorter
>than 360 degree conduction angle of the tube. Another thing is if
>any harmonics from the shorter than 360 degree conduction angle
>of the tube reach a solid state exciter, they can "fool" the SWR
>shutdown circuit into thinking the RF is being delivered to a
>mismatched load.
>
>Because of this, all input circuits should be a low pass C-L-C pi
>network or a parallel tuned network, and not any form of "T" or an
>L-C-L high-pass pi network.
>
>Fortunately Q is very non-critical in most applications,
// For g-g amplifiers with a tube-type Pi-net tuned output driver, this
is undoubtedly true. With a solid-state untuned output, this is hardly
the case.
>that is why
>so many approximations and rules-of-thumb work.
>73, Tom W8JI
- R. L. Measures, 805.386.3734, AG6K, www.vcnet.com/measures.
end
--
FAQ on WWW: http://www.contesting.com/FAQ/amps
Submissions: amps@contesting.com
Administrative requests: amps-REQUEST@contesting.com
Problems: owner-amps@contesting.com
|