Rich says:
>// Class AB2 is not linear with a grid driven 4-400. Class AB1 is
>linear.
This is not strictly true. AB2 can be more linear than AB1 - see the classic
work by Pappenfus et al. In order to get that performance, you need a well
(shunt) regulated bias supply, a low impedance RF source so the grid volts don't
get distorted when grid current flows, and well regulated screen grid supply.
Some negative feedback from unbypassed cathode resistors helps, too.
If you don't mind the Tbird becoming distinctly non original, I'd seriously
consider going to grounded grid. Turn the old grid input circuitry into switched
pi networks, go to a bridge rectifier with choke input and a stack of
electrolytics for the HV capacitor. 4800 volts AC will give about 3800 DC with a
choke input filter, although you'll need a bleeder resistor. You may need some
bias: probably something like a 10 volt 10watt zener, but you'll need to
experiment there. I'd go for about 70mA per tube standing current. The
4-250/4-400 family perform quite well at 3600 or 3800 volts. Although I use
passive grid input AB2 on the LF bands with 4-250s or 4-400s (the basic tube is
the same, the filament is the same so the peak cathode current is the same) I
also use one on 10m in g-g as above with good results. Incidentally, in g-g with
an 8.2v zener for bias, you could substitute 3-500Zs - same cathode structure.
Passive grid with a 50 ohm resistor will give you a <1.5:1 SWR with up to 40pF
of stray C across the swamping resistor.
If you want to stay with grounded cathode, the HV is far too low - for either
AB1 or AB2.
73
Peter G3RZP
--
FAQ on WWW: http://www.contesting.com/FAQ/amps
Submissions: amps@contesting.com
Administrative requests: amps-REQUEST@contesting.com
Problems: owner-amps@contesting.com
|