Steve Thompson wrote:
>> Not so. The gain is determined by the total anode load impedance, not
>> just the resistance in the network. A given length of resistance wire in
>> a coil has a certain reactance and a certain resistance, when
>> measured as series components. As frequency goes up, so does the total
>> impedance when considered as series components, and so does the total
>> impedance when considered as parallel components - at the same frequency.
>>
>> An inductor in shunt with a resistor in parallel shows an impedance
>> change as frequency goes up that is asymptotic to the impedance of the
>> resistor. That says the network impedance has a maximum value, and thus
>> the amplifier gain has a maximum value.
>Only at frequencies where the rest of the network between the
>suppressor and ground looks close to a short circuit. Any inductance in
>that path means that the anode load is not the Rp of the suppressor
>network. Outside the valve, there's several things in series from the
>anode to ground - you've got to look at the whole lot of them in their
>series equivalent form, get the overall Rs and Xs and only then work
>out the parallel equivalent if you want to know the parallel load that
>the anode is seeing.
>
There is a spreadsheet on my home page (below) which does just that. It
was originally for designing Pi and Pi-L networks, but after a previous
round of this discussion in 1999-2000 I extended it to analyse the
effects of LC suppressors.
It calculates the parasitic frequency Fp, and also the parallel load
resistance that the suppressor creates from anode to ground at Fp.
The complete solution requires a long sequence of series/parallel
transformations, and it involves a number of components in the tank
circuit. In other words, the effect of the suppressor will depend on the
amplifier it's used in.
>Worrying about the vhf Q of the suppressor in isolation isn't quite
>irrelevant, but is misdirected.
When you crank out the complete solution, the VHF Q of the suppressor
turns out to be totally irrelevant. The solution needs R and X
separately, and their ratio Q is no use at all.
--
73 from Ian GM3SEK
http://www.ifwtech.co.uk/g3sek
_______________________________________________
Amps mailing list
Amps@contesting.com
http://lists.contesting.com/mailman/listinfo/amps
|