[AMRadio] Re: capacitors for tank circuits

John coleman wa5bxo at gmail.com
Wed Apr 16 19:17:17 EDT 2008

```Yep, very good points Don.  What I was mostly trying to say to Barrie was
that tubes should be chosen first so as to acquire a ratio of plate current
to plate voltage that will be used and then design the tank circuit around
that.  400 Watts DC input can be achieved by 8 - 6146s in push pull parallel
but because of the low voltage and high current the tank circuit would have
a much larger amount of capacitance and less coil than a tank that is for a
pair of 75ths at high voltage and less current.  Also Barrie, when using the
charts, you should be sure to choose the one for class C not the chart for
linear operation.  The tank Q is all together different.

I used the example of the 8 - 6146s only as an example.  If you were to try
that you would have a difficult time finding a modulation XFMR to match the
900 ohm load that those tubes would represent to a modulator.

John, WA5BXO

-----Original Message-----
[mailto:amradio-bounces at mailman.qth.net] On Behalf Of D. Chester
Sent: Wednesday, April 16, 2008 11:42 AM
Subject: [AMRadio] Re: capacitors for tank circuits

> From: "John Coleman" <jc at pctechref.com>

> The size of the spilt stator capacitor that you are looking for (will)
> need to
> be determined by the voltage and current that you intend to run, The
> ratio of I:E determines the capacitance  There are charts in the handbooks
> for this purpose.  Be sure to consider that you are doing push pull with a
> balanced tank circuit as the required capacitance is only a quarter of
> that
> required for non-balanced. The voltage will determine the needed spacing
> of
> the plates.  If you raise the tank capacitor above the chassis on
> insulators
> and connect the rotor to B+ you can get by with less spacing but you will
> need to use and insulated coupler to the knob shaft and be sure the knob
> shaft is grounded for safety.  You can also accomplished the same thing by
> using two large RF chokes, one per tube, and capacitive coupling the RF to
> the tank circuit as is done in most PI net circuits.  This way the plate
> tank cap can be mounted to chassis and the RF choke at the center of the
> tank coil can be grounded.  The idea is to not have a DC + modulation
> voltage across the plates of the tank capacitor but just RF voltage.
> Capacitive coupling, as was just described, is a neat way to do this but
> it
> requires the very large long RF chokes and good coupling capacitors.

For a balanced tank circuit, the capacitance is one fourth, but the voltage
rating must be double that of single-ended.  The reason for this is, that
for a single-ended final, the tube is working into only half the balanced
tank circuit, but the coil acts as a step-up autotransformer and the induced

voltage across the other half of the coil is approximately equal to the
voltage on the half that the tube works into, giving an rf voltage
end-to-end that is twice that which is actually generated by the tube.  It
is exactly the same in the case of push-pull, since in class-C or even
class-B service, only one tube is working into the tank circuit at a time,
and each tube is working into one half of the tank circuit.  Or you could
think of it as each tube generating equal rf voltage, but the outputs of the

two tubes are in series as they work into the tank circuit.

But a capacitor with twice the voltage rating and one fourth the capacitance

is equivalent to taking the original single-ended capacitor, splitting it in

half, and wiring each of those halves in series.  That is exactly what we
mean by a split-stator capacitor.  For example, take the BC-610, which runs
the final at 2000 volts @ 250 mills.  The tank capacitor is split stator
with 150 pf per section.  If the final were changed to unbalanced output,
for example by substituting a 4-250 for the 250TH and converting to a
pi-network, or by changing the grid tank to balanced and running the plate
tank unbalanced,  the final tank capacitor would need to be 300 pf in order
to keep the tank circuit Q the same.  That could easily be accomplished by
wiring the two sections of the 150/150 split stator capacitor in parallel.
The parallel connection gives 300 pf at approximately 7 kv rating.  The
series connection as used in the balanced tank, renders 75 pf at 14 kv
rating.  A single ended capacitor rated 75 pf @ 14 kv would be about the
same size as the dual 150 @ 7 kv, and when new, the cost would have been

The point is, for a given power level and plate voltage/plate current ratio,

the physical size and cost of the tank capacitor is approximately the same,
whether the tank is balanced or unbalanced.  If a split stator capacitor is
used, it can easily be connected up as a balanced or unbalanced tank.  Of
course, the number of turns in the balanced tank will be double that of the
single ended one, since 4 times the inductance is required to maintain
resonance at the same frequency.    OTOH, the wire size in the unbalanced
tank will need to be heavier, since the circulating rf current is higher
with the higher L to C ratio.

I always raise my entire pushpull tank above ground and put the full HV on
the whole capacitor, and series feed through the tank coil.  The rf choke
has to do much less work than with parallel feed, because with series feed
the rf choke connects to a zero rf voltage point on the tank coil, while
with parallel feed, the choke goes to the plate of the tube and thus the
full rf voltage appears across the choke.  Theoretically,  with series feed
the B+  lead could be connected directly to the cold spot on the coil with
no rf choke at all, but in a practical circuit, the B+ leads needs to be
isolated from the tank coil with a choke since it is virtually impossible to

maintain perfect balance in a nominally balanced tank circuit.

Don k4kyv

______________________________________________________________
Our Main Website: http://www.amfone.net