Hey Guys,
Look at the grid circuit of the Collins carefully, and look at a few
things.
The theory presented (Orr also presents the same theory in his
"Super Cathode Driven" circuit article) is that the high grid to
ground impedance caused by the GK capacitance forms a
divider between the cathode and grid (let's call that C1) and the
grid and chassis (let's call that C2). The theory is offered that the
normal grid to cathode capacitance forms C1.
The theory is the capacitive divider elevates the grid for RF,
allowing the grid to follow the cathode. This is supposed to reduce
the grid to cathode voltage and add negative feedback.
In order to form a linear divider, the impedance of C1 has to be much
larger than the lowest expected value of grid to cathode impedance.
If the impedance of C1 is many times less than the grid
impedance, any changes in G/K impedance do not affect the voltage
division.
The problem is the ~6 pF grid to cathode capacitance (I'm
using Svetlana's published value) is about 7500 ohms on 80 meters,
and decreases to just under 1000 ohms on ten meters. If the
changing grid to cathode impedance of the tube even approaches this
value, the divider falls apart. When the input signal is positive the
grid draws no current, and the grid to cathode impedance is nearly
infinite. The feedback system sees only the value of G/K capacitance.
When the cathode swings negative, the G/K impedance falls "through
the floor". If the peak grid current is 100 mA (this is peak, not
indicated current) and the G/K voltage is 50 volts, the G/K impedance
is 500 ohms. Even on ten meters the grid would not have a linear
voltage division with the 220 pF shunt capacitance (C2 in this text,
about 200 ohms on 80) the 30L1 uses.
That means grid to cathode voltage and RF feedback is non-linear
with drive and over the RF cycle.
This system works fine in the 30S1, because the tube never draws grid
current. If it does the grid to cathode impedance drops and drive is
suddenly reduced. But in ANY PA that draws grid current, this system
is seriously flawed in operation.
A second problem is the grid is the ONLY shield from the output to
input in the triode. Allowing the grid to partially "float" with RF
increases feedback through the tube. This de-stabilizes the tube and
makes low frequency oscillations more likely. This is especially true
when an inductance is in the path, forming a parallel resonant grid
circuit. A TPTG, TPTK, or TGTK oscillator is just one resonance
away! That's why Collins had to swamp out the grid choke.
SB-220's have this problem also.
In the 30S1, shielding loss by floating the grid was NOT a problem.
The screen was grounded, and it prevented feedback. So when
engineers took the great idea of the 30S1 and applied it to grounded
grid sub-class 2 (ie AB2 not AB1) triodes, they misapplied a good
circuit.
As for arc protection, putting arc protection in the grid is not very
smart. The grid to chassis resistance allows the grid to pull up to
anode voltage during an arc. The next thing in line is the cathode,
and the exciter is connected right to the cathode through a few low
voltage capacitors. An anode to grid fault (the normal discharge
path) can be passed right along to the cathode system, and straight
out to the $$$ exciter.
Arc protection belongs in the anode's dc supply line, not in the
grid.
I'm not saying the "Super Cathode Driven" circuits always cause
problems, but in PA's I've measured gain flatness and IMD on I've
found that system to INCREASE IMD unless the PA was very carefully
tuned and drive properly adjusted. Removing the "Super Cathode
Driven" system and grounding the grids produced much better gain
flatness and generally better IMD performance, while INCREASING
stability and improving damage protection.
Draw it out on paper and plug in some values, and remember the GK
impedance is time-varying with drive. You'll see the SKD system does
almost exactly what we don't want it to do. If the 220 Heath
continued, we were planning on pulling the grid chokes and small mica
caps and grounding the grids.
If you have one working, fine. But unless it's in a class A or AB1
tetrode that feedback system probably isn't helping things.
73, Tom W8JI
--
FAQ on WWW: http://www.contesting.com/ampfaq.html
Submissions: amps@contesting.com
Administrative requests: amps-REQUEST@contesting.com
Problems: owner-amps@contesting.com
|