Ham Radio, June 1986, has a regulated screen grid PS article by Rich, AG6K.
Uses string of zeners and a TMOS FET.
>Small tetrodes have the somewhat unpleasant habit of generating
>potentially-fatal reverse screen
Not only small but big 'handled' tetrodes too. The Burle (RCA) 4616 tetrode
almost always exhibits bipolar screen current for me, depending on the
operating point
(plate and screen voltages, and power output). We have a 1000 Ohm shunt
resistor
on the output of the PS, where it feeds the screen. This is the screen to
cathode
path for the reverse current. We also put a 1800 Volt spark gap across
screen to
cathode, in case something happens (loose resistor or a plate to screen
arc) and the screen voltage rises. Normal operating screen voltage is 1400.
There is a small power resistor in series between the PS and the gap, so
that if it ionizes and shorts out, the power supply will blow this fusable
resistor open. Saved a lot of tubes this way.
Take a look at the constant current curves for the 4CX250B series. Even
with no grid current flowing, in linear operation, there are 'islands' or
families of current curves for screen current that are negative. And it
appears that as the plate voltage swing gets closer to 500 volts, (near
screen potential) the screen current starts turning around. This is the
current in the screen connection of the tube. Obviously, internally there
is some secondary electron emission from the screen. These electrons are
then being attracted to the plate.
A lot of the Eimac tubes are well behaved, according to the DC curves, from
the 4CX3000A up to the 10,000. This doesn't guarantee that they won't have
reveresed current. The tube tested for the curves, for the data points
recorded, did not show it. Another tube of the same type, maybe aged with a
little cathode material bombarded on the screen, or with a warped grid, or
at different voltages than recorded for the curves, or made slightly
differently (production variations) might have it. It's best to assume that
your screen power supply will have to handle it someday as you adjust an
amplifier, so plan on it.
In the 4CX12,000-20,000 families, curious curves show where the screen
current isn't exactly monotonic, and decreases for a moment, then increases
again. These discontinuities cause the fluctuations of the screen current
direction as well. The net current into the screen is positive, but drops
for a moment in the direction of generally increasing screen current (lower
plate voltage).
4CW50,000-100,000-150,000E family: With plate voltages from 8 to 15 KV,
negative screen current is possible. It must be dealt with in the power
supply as before. In higher power tubes, pyrolytic graphite grids have
eliminated secondary emission from the screen. These tubes have monotonic
curves, that show no signs of degenerative behavior due to the screen
current direction. Both Eimac and Thomson manufacture new tubes with PG
instead of wire grids.
However, a series pass regulator with a
>sufficient bleeder resistor will do essentially the same job.
>- If the zener-string/chain is powered from the anode supply through an
>appropriate resistance, the screen is automatically protected against
>loss of anode potential -- and there is no chance of damaging the screen
>by too light loading or by sudden loss of the antenna during a tornado.
Also can use a sensitive adjustable relay in the plate voltage, at the
bottom of a resistor divider, to determine that, yes, it is actually
present at the amplifier, (not at the PS, but at the PA side of the HVDC
connector!). This relay will allow screen power enabled only if the plate
is up to normal voltage.
john
K5PRO
--
FAQ on WWW: http://www.contesting.com/ampfaq.html
Submissions: amps@contesting.com
Administrative requests: amps-REQUEST@contesting.com
Problems: owner-amps@contesting.com
Search: http://www.contesting.com/km9p/search.htm
|