John Coleman jc at pctechref.com
Tue Jan 27 10:43:52 EST 2009

```Don:
It's so nice to see a good tech post from you.  You have such way
with words it is difficult for me to see how anyone would not understand.
Some times I get to writing and make assumptions that folks know what I mean
such as the statement about audio needing to be twice the DC.  (You will
notice I corrected that in a later email).

I was explaining to one of the other members that the tubes do not
produce the audio or RF energy.  They simply control the current flow
through a load.  In respect, putting in a bigger switch, alone, will not
make the light brighter.

John, WA5BXO

-----Original Message-----
[mailto:amradio-bounces at mailman.qth.net] On Behalf Of D. Chester
Sent: Monday, January 26, 2009 11:07 PM

> The output voltage of a modulator is determined by its plate supply
> voltage
> and the modulation XFMR turns ratio.

> You need to think in terms of voltage transformation.  If you are using a
> common power supply on the final and modulators, or more exact, the
> modulator plate supply and the final plate supply have the same voltage,
> then the ratio that you use is what determines the maximum modulation.

That is true irrespective of the nominal impedance of the transformer, the
p-p load impedance of the modulator tubes or the modulating impedance of the
final.  Of course, the tubes have an optimum p-p load impedance at any given
plate voltage, and a transformer has an optimum primary and secondary
impedance at which it works best.  But the actual transformation is based on
turns ratio.  The impedance ratio stamped on the nameplate of the
transformer is a nominal value, and a good transformer should be able to
work at up to twice the nominal value and as low as half the nominal value
with little degradation in performance.  The impedance transformation ratio
is the square of the turns ratio.  For example, a transformer with a 2:1
turns ratio has a 4:1 impedance ratio.

> 100% modulation occurs when the audio voltage from the modulation XFMR is
> 2
> X the plate supply.

Actually it's when the peak output voltage, the combined audio and DC
voltage from the winding adds up to 2 X the plate supply voltage.  The peak
a.c. output voltage from the transformer is equal to 1 X the DC plate supply
voltage.  When it is in the same polarity as the DC voltage, the two
voltages add together to produce a sum that is 2 X the plate voltage at the
positive modulation peak.  At the opposite peak of the audio cycle the
polarity is reversed and the two voltages cancel, leaving zero volts on the
plate of the final.  This is the negative modulation peak.

>
> At maximum drive the modulator tubes conduction (assuming they or big
> enough) takes the plate voltage close to 0 Volts at the peak of the audio
> for that conduction cycle.  Nothing you can do will take the voltage lower
> than zero.  As one tube hits the Zero volt peak then the other tube will
> hit
> the 2 X plate voltage point.

That is theoretical.  In actual practice, there is nothing you can do to
pull the instantaneous plate voltage below about 20% of the power supply
voltage.  In the case of screen grid modulators, the plate voltage can never
be pulled to a lower voltage than the DC screen voltage.  As one tube
reaches maximum conduction, the instantaneous voltage on the plate of the
other tube will reach about 1.8 X the DC plate supply voltage.

> Something between 2:1 and 1:1 is what is needed.  You need a little extra
> to
> make up for the fact that the modulators will use some power in plate
> dissipation and you will want a little head room for voice lopsidedness
> (everything is not a perfect sine wave).
>
> Experience information from Don, K4KYV, indicates that between 1.2:1 and
> 1:4:1 is generally a good choice.  1.2:1 will give you more head room but
> will require more modulator current perhaps larger tubes.  1:4:1 will
> probably just be enough audio with very little head room, but will require
> less modulator current and lighter demand on the modulator tubes.  If you
> chose 1.2:1 for plenty of head room then choose modulators with a little
> more current capability or double up (push pull parallel).

Using a higher step-down ratio of 1.6:1 will just barely allow you to reach
close to 100% modulation with no headroom whatever, but the modulator tubes
will run more efficiently.  Somewhere between 1.2 and 1.4 will allow more
headroom at the expense of efficiency.  But that extra headroom is needed
for minimum distortion and splatter, since driving a modulator or linear
amplifier (exactly the same thing except the modulator amplifies audio while
the linear amplifies rf) right to the saturation point results in more
distortion.  But watch the modulator plate current and make sure you don't
exceed the tube ratings.  If so, double up to use a pushpull parallel
modulator.  However, this may increase the audio driver requirements.

>You may want to
> consider a modulation reactor even if your XFMR says it can handle the
> secondary current.  Keeping the current out of the secondary will greatly
> improve the low frequency capability of the XFMR.  You want regret it.

Taking the DC off the secondary greatly reduced the talk-back when I was
using a UTC VM-5 modulation transformer.  With the DC going through the
secondary, the thing sounded like a small speaker inside the transmitter
cabinet.  With the reactor, it was totally quiet.  This will also reduce the
distortion from the modulator, since it will reduce the magnetic saturation
of the core over the audio cycle.

A modulation transformer designed to run DC through the secondary will
usually have a gap in the core, filled with paper or some composition
material.  A transformer designed to be used with a reactor will usually
have a core stacked like a power transformer, with no gap.  But this is not
strictly true.  I have seen transformers designed for DC with an extra large
core and no gap, while I have seen ones designed for use with a reactor that
had a very narrow gap, not much more than .001 inch.

Don k4kyv

```