Many mods being circulated for Ranger and Valiant audio systems seem to be poorly thought-out. It appears mod creators didn't think through the problem and take logical steps to ensure the mods being created are truly beneficial. Good modifications include changing the troublesome 18K resistor in the VFO compartment. Solid state rectifiers also can be worthwhile, and so can replacing old electrolytics.
There also are bad modifications. Some audio changes remove nearly all upper frequency rolloff. The modulator section shows increasing phase shift at very high audio frequencies. When the rolloff is eliminated at frequencies we can't use anyway, unwanted positive feedback occurs in transmitters like the Ranger. The modulation transformers also are subjected to high-energy low bass response that does nothing to improve a signals readability. The increased bass and treble removes power available for the 300-3000Hz range so important for communications, and taxes the transformers so much that the user runs a serious risk of damaging expensive transformers. At the very least, he winds up ruining QSO's up or down the band from splatter or irritating SSB ops, who sometimes make it a point to return the "favor" and retaliate against all AM'ers.
If you ever listen to the bassy AM ops from a distance, you'll see they are very difficult to copy compared to normal communications audio. Collins worked this all out many years ago and settled on an optimum passband for readability. It wasn't 10Hz to 10,000Hz as some would like us to think!
Finally, I don't believe the people who designed your Ranger were so stupid or careless that they made a mess of the total audio system. It would be normal engineering practice to design a good system, and then adjust one component to tailor slope. My bet is they intentionally sized C52 to increase audio punch at the expense of making people sound like Ted Baxter.
Useless Mods
The Diode Super modulation Mod
One popular change is the addition of a silicon rectifier diode in series with the modulation transformer secondary. The thought is this diode limits negative peaks and prevents splatter caused by negative peaks. This is sideways thinking. Factually the diode does nothing at all.
While reference materials sometimes say excessive negative peaks cause splatter, that is a gross oversimplification. In a AM system, the device being modulated actually must behave as a linear response mixer. The audio, in effect, is the signal and the carrier the local oscillator. It isn't the fact the PA receives negative voltage that causes the splatter, it is the fact the "local oscillator" or mixer is shut off abruptly with a rapid transition in wave slope as the PA reaches zero output. We can go as negative as we like with anode voltage and the energy in splatter does not increase UNLESS we change the rate of slope at the zero crossing transition.
Just like with a CW envelope (CW is really AM), the rate of change or slope angle at any point sets the bandwidth. Moving the diode from inside the tube (the cathode to anode path is the diode) to the outside of the tube doesn't modify the slope at all. It does not allow us to reach 100% negative peaks without generating splatter. The PA still reaches zero volts with an abrupt waveform slope transition, and splatter is just as bad.
What the author or authors of that mod intended to do was produce a negative peak limiter, but they missed. Properly designed negative peak "hard-limiters" must always be followed by a suitable low-pass filter. The low-pass filter would round off transitions, and cause the modulator waveform to gently slope to zero. This limits the maximum frequency of distortion products and the transmitter's bandwidth, although it does nothing for in-band distortion. If the low-pass audio filter is omitted, the diode does nothing at all to constrain bandwidth. It's a totally useless mod.
I actually spent considerable time with my spectrum analyzer (it makes direct measurements of adjacent channel power) and my Viking Valiant, and found no matter what I did with the diode there was absolutely no improvement in adjacent channel energy for a given level of audio. The diode, no matter how configured, did not make things better.
What did allow me to run asymmetrical peaks was a simple mod to the 6AL5 clipper in the Valiant. I disabled one section of the dual diode, leaving the clipping effective only on one audio polarity. Since this "hard-limiter" is followed by a good 3.5-4kHz low-pass filter, the transition is rounded and off-frequency energy caused by clipping is attenuated.
The Weak Interstage Transformer Myth
Another myth is the audio driver transformer is "weak". What I found through careful experimentation is no matter what you do with the driver transformer, when the modulator tube control grids are driven positive the grid voltage flattens off and clips. This is actually caused by the entire audio driver being unable to deliver the power required to force the modulator control grids positive. The grid waveform remained exactly the same in my Valiant regardless of the size of transformer I used, but the impedance ratio made a difference. A lower impedance ratio reduced grid voltage, but at the same time it reduced clipping more. Loading the secondary with a resistor had much the same effect.
The real problem is the modulator isn't capable of producing clean low-distortion class AB2 operation no matter what you do with the existing driver. Even if you overcome the grid problems, the tubes themselves cannot deliver low distortion audio in class AB2. The modulator tubes begin to saturate and produce harmonic and IM distortion, and that distortion is passed on right to the PA stage and out on the airwaves.
I have a simple solution to this problem. Run less power. If you reduce the plate current you won't have to push the modulator into class AB2 to obtain 100% modulation. You can run it AB1, and distortion will be considerably less.
The Audio Choke in the PA Screen Myth
Another myth is the PA screen grid requires an audio choke to modulate linearly. What the screen actually requires is a moderate percentage of modulated voltage that is in-phase with the anode. This is because a tetrode, even in hard class C operation, is a very poor mixer with a very non-linear transfer function when only the anode is injected with the modulation signal. The screen must also be modulated.
The engineers at Johnson were not as dumb as people might have us believe. The Ranger and Valiant already came very close to optimum screen audio voltage. Johnson did this by obtaining screen voltage from the modulated plate voltage via a dropping resistor.
If you want to optimize this voltage, you can do so by adding a 5 to 20uF, 900 volt (two 450's in series) capacitor from the positive end to the screen end of the screen dropping resistor. By adding a resistance in series with the new capacitor, you can adjust the amount of screen modulation for maximum peak linearity. I found there was very little room for improvement over what Johnson did from the factory when the PA was operated to specification!
Save your money and don't drill up that old rig, the choke isn't needed and actually can be harmful.
What about
removing
C-53?
C53 is 200 pF in
parallel with three
resistances. Those
resistances are the
plate resistance of
V7A, R19, and
R23.
200pF @ 3kHz is
265k ohms. That 265k
ohm impedance is in
parallel with about
50k ohms, so
removing C53 would
have a negligible
effect. Let it be.
Remove C-56?
Again about C56 is
470pF or 113k @
3000Hz in parallel
with the plate
resistance of V7B,
R23, and R27. That's
113k in
parallel with about
45K ohms. Again, not
a big change in
level although very
slightly more effect
than removing C53.
Let C53 alone.
Remove
C-60?
Removing this cap
could destabilize
the audio system! It
is in the negative
feedback loop. The
internal Ranger
audio feedback loop
has negative
feedback around this
component. Removal
of this cap won't
affect response when
feedback phase is
negative!!
Removing this
capacitor impacts
frequency response
and gain on
frequencies where
feedback is
positive, such as
those far above
normal audio
frequency ranges!
Any advice to remove
this part is very
bad advice!!!
Change
C-52 from 500 pfd to
.02 mFd
OK, this is about
the only major
worthwhile effect,
since it
will bring low
frequencies up
several dB.
Change C-51
and C-55 from .1 to
20 mFd
This is a wasteful
change. This change
won't do anything
except reduce very
low sub-audible bass
slightly. The reason
why is very easy to
see. The impedance
at the point where
C51 and 55 are
attached is very
high compared to the
reactance of the
capacitor. C51
at .1mF is 5.3k ohms
reactance. The
impedance at that
point in the circuit
to the audio path is
470k ohms. Obviously
any change in
voltage across C51
or C55 caused by the
time-varying anode
current of the 12AX7
is so miniscule the
20uF is a wasted
effort. I measured
only a few nanovolts
of AC across C51 at
300Hz.
Change
C-57 from .02 mFd to
.1 mFd
C57 is 26.5K @
300Hz. It's in
series with about
100k ohms total R
(counting feedback).
This reactance
causes about 1.5 dB
rolloff at 300Hz.
Changing C57 to .1uF
will increase lower
end gain, but it
will also unbalance
overall response by
about three dB or so
when low end is
compared to high
end.
The Ranger (and
Valiant) really
don't require a bass
gain increase beyond
increasing C52!
Audio response is
very flat with only
a change in C-52. As
a matter of fact
changing C-57 to a
larger value creates
a problem.
Increasing the value
of C57 causes the
lows to be
emphasized more than
highs, and then this
change forces you to
go back through the
rest of the circuit
to boost highs just
because you
unbalanced the
response!
Change
C-71 from .1
mfd to 20 mfd
I'm not sure of the
effect of this, but
my instincts tell me
it has no effect.
Any effect would
depend heavily on
peak screen
current. In my
Ranger, there was no
measurable effect.
By the way, you can use disc capacitors. It makes absolutely no difference in sound or performance if you use an orange drop or a ceramic disc in audio circuits. Save your money and time, and use whatever is handy to you.
Pictures of my Valiant and its waveshape.
SPICE simulation of a minimal modification improvement to Ranger and Valiant audio
Original response at anode of V7A:
Note the rolloff is only 1dB at 3000 Hz.
C52 produces the following slope:
Below is frequency response at anode of V7B.
Note the large low-frequency rolloff. Rolloff is from -252dBv to -267dBv or a bass loss of 15dB. This was done intentionally to peak the QRM-and-noise-cutting high frequencies for DX work.
Note how very little C57 affects the audio curve in the sweep below.
Frequency response slope is almost exactly the same on either side of C57. C57 is large enough.
Changing C52 Only
Overall response when C52 is changed:
*
The graph above is 1/2dB per cross line. Overall response actually is quite good when only C52 is changed. Response falls within about 1/2dB from 200Hz to 5000Hz. Expanded view of C52 mod result:
Between 400Hz and 3000Hz there is less than .2dB variation! All this from changing only one capacitor in your Ranger.
-244.720dB @ 400Hz
-244.545dB @ 1150Hz
-244.705dB @ 3000Hz
This will give excellent sounding full audio without excessively taxing audio transformers. At the same time your Ranger is preserved, it isn't all hacked up with dozens of changes that produce insignificant results.
The old timers at Johnson weren't that stupid after all, were they? I'll bet they did the math when they picked parts.
as
of 1800Z on 2005 Dec
26