This is from a 50% duty cycle @ 1 kHz modulated RF signal
sent to the Rife/Bare tube. The oscilloscope picture shown here is the
resulting optical signal from the tube.
After some tests, it was determined that the length of the optical
spike was, for all practical purposes, very close to 6.5 microseconds. The
pulse duration did not seem to vary greatly with changes in RF power,
modulating frequency, or RF feed methods. In order to get the most RF power
into the pulse, but send as little RF power as possible into the tube during
the remainder of the operating cycle, I settled on an RF pulse length from the
transmitter / amplifier of approximately 10 microseconds in duration. I found
that with a 10 microsecond RF pulse, the optical pulse from the tube was
maximized in amplitude, and had a fairly uniform rise and fall time.
This picture
shows the optical signal from the Rife/Bare tube when driven at a modulation
frequency of 456 Hz at a peak RF power level of 1 Kw (1,000 watts).
The average RF power level is quite low, only about 4.56 watts. The
horizontal scale of the oscilloscope screen is 5 microseconds per division.
Note that the optical pulse averages about 8 microseconds in width.
If the modulating signal is changed to a 50% duty cycle square wave,
then the RF power to the tube increases to 500 watts average power, but the
peak power in the optical spike remains the same as is shown here. This
indicates that in order to obtain the best optical spike, it is not necessary
to send RF power to the tube for a longer time than the optical spike lasts, or
about 6-10 microseconds.
This is what
the Rife/Bare tube looks like when driven by the 10 microsecond RF pulse at 456
Hz and 1 Kw peak RF power. The color is actually a more deep Violet than this
photo shows, as the camera is not very red sensitive, and shifts the color
towards the blue end of the spectrum. The beam is very "tight," and drives
straight across the tube with no wavering or flutter. Even though the average
power is very low, the tube can be heard to faintly "sing" if you are within
two or three feet of the tube.
If the tube is driven with the "standard" 50% duty cycle square wave
RF signal, the discharge in the tube tends to be less well defined, and
generally fills the tube from wall to wall. The color also shifts, from a
blue-violet to more of a blue-grey color, depending on the RF power level and
the modulating frequency used.
The behavior of the tube is interesting, and a bit different from that
is observed when driving the tube with a 50% duty cycle RF signal. When viewed
in a darkened room, the discharge is fascinating, particularly at low
frequencies.
At frequencies below about 600 Hz, the discharge snakes from one end
of the tube to the other, in a "sheet" discharge along the inner wall of the
tube. This sheet is variegated, and has a spider web like appearance. It tends
to concentrate itself in areas of highest RF field strength.
As the frequency is raised to about the 1100 Hz range, the sheet
thickens and begins to lift away from the tube wall, gradually becoming a
violet fog in the center of the tube.
Raising the frequency to about 1500 Hz finally causes the discharge to
free itself from the tube walls entirely and concentrate itself into a pencil
thin brilliant violet beam down the center of the tube. The point at which it
will do this is determined by the average RF power into the tube, and not the
pulse width.
Increasing the frequency further causes the discharge to become
thicker and brighter, until finally it fills the tube from wall to wall at
about 20 kHz.
It may seem odd that since every RF pulse is the same in amplitude and
power level, that the visible discharge in the tube should vary so much as the
frequency of the applied pulses change. Consider that when the pulses are
infrequent, as they are at low modulating frequencies, the gas in the tube has
a chance to completely deionize, and in addition, heated gas in the tube (yep,
the gas heats up when we zap it with RF energy) will have a chance to diffuse
out of the ionization path before the next pulse arrives. This makes each pulse
behave as though it were the very first one the tube sees after having been
turned off for a while.
Since the cool, deionized gas is a very good insulator, the RF pulse
has a hard time getting the ionization started. This allows the RF field to
build up to a fairly high level. When it gets high enough, the gas nearest the
electrodes will begin to ionize. Since the electrodes are usually around the
outside of the tube and against the glass, the discharge starts at the inner
surface of the tube wall closest to the electrodes. Because of the high
dielectric constant of glass as compared to air, the RF electric field is
somewhat stronger at the tube wall. This helps to cause the gas to ionize near
the tube wall rather than in the center of the tube, and accounts for the
reason that the discharge "sticks" to the wall of the tube at low pulse
rates.
As soon as the gas ionizes, it becomes a conductor, and increases the
concentration of electric field at the end of the ionized gas the furthest
distance away from the electrode. As long as the pulse is present, the
ionization will progress across the length of the tube until it finally
connects both electrode areas. When the concentration of ionized gas along the
tube wall increases enough, any additional ionization will be forced to occur
further and further away from the tube wall. As soon as that happens, the
discharge begins to shift away from the tube wall and start to concentrate in
the center of the tube. As soon as that happens, the power density in the
discharge column increases to a critical level, and the discharge brightens
considerably. At this point, the load resistance of the tube falls rapidly and
so does the VSWR reading. For a more detailed explanation of this, please see
Understanding The Rife/Bare RF System.
If the pulses are close enough together, and the discharge is
concentrated in the center of the tube, the gas cannot completely deionize, nor
can the heated gas readily "get out of the way" of the next pulse. The tube is
now operating in what I consider the "normal" mode.
This
oscilloscope picture compares the Rife/Bare tube optical output against the RF
signal applied to the tube. Note that the optical signal is slightly delayed
with respect to the RF energy applied to the tube.
The RF signal was acquired using a 1 inch sample antenna placed about
4 inches from the feed electrode at one end of the tube. The optical signal was
acquired using my NanoMeter Receiver.
In this picture, the horizontal time scale is 2 microseconds per
division. The width of the RF pulse may be seen to be 10 microseconds, and the
optical pulse is slightly less, but delayed by about 4 microseconds.
Why is the RF pulse not a square wave? If the RF bandwidth of the
system were greater and if the Rife/Bare tube were not in the circuit, it would
be. What happens here is that as the RF voltage begins to rise at the beginning
of the pulse, the tube slowly starts to conduct. At first, the ionized gas has
not spread from one end of the tube to the other, and the tubes impedance is
fairly high. As more and more of the the gas ionizes, the RF field is loaded
down more heavily, and the rate of rise of the RF voltage begins to slow down,
and then actually decreases before the RF pulse reaches its maximum. In the
meantime, the light from the tube is increasing, and does so until the RF pulse
reached its maximum value. This change in impedance of the tube also causes
VSWR readings to fluctuate wildly during the operating cycle of the Rife/Bare
tube. For a more detailed explanation of this, please see
Understanding The Rife/Bare RF System.
RF System
Considerations for Pulse Drive of Rife/Bare Tubes
Using an RF amplifier for pulse service is quite a bit
different from using it for AM, FM, or CW service. In standard Rife/Bare use,
we are sending a 50% square wave to the modulator circuits of the transmitter.
The rise and fall times of the modulation waveform is usually the limiting
factor in how well the transmitter reproduces the waveform. In some cases, the
inherent distortion in the modulator circuits and the RF stages help, rather
than hinder, the generation of an acceptably fast rise time on the waveform.
In the case described here, a different approach is
required. The entire time of the pulse is close to the rise time generated by
some of the modified CB radio systems I have tested, making them unacceptable
for pulse service. A system designed for pulse work is needed. A much faster
rise time is required in the RF stages, as well as a wider bandwidth through
out the RF circuitry.
My system uses a GT-310 frequency generator card which is
used to generate the desired square wave modulating frequency. This is fed into
a variable pulse width generator, which is adjustable from 6.5 to 250
microseconds pulse width. Following the pulse circuit, is a 50 ohm impedance
coaxial cable line driver. The frequency response of the line driver is in
excess of 30 mHz. This feeds the RF modulator input on my
Yaesu FLdx-400 transmitter which I have modified
for direct RF modulation. The output of the Yaesu goes to my
Heath Warrior Amplifier, which is capable of
700 watts RF output on CW or 1 kW when operating in pulse mode. My
Heath Antenna Tuner and my
Twisted Transmission Line complete the RF
circuit to the Rife/Bare tube.
Tests have shown that the RF bandwidth of the system is
really a bit too low for what I feel is optimum performance, although it does
well enough for initial tests and measurements. Let's examine what's required
for better performance.
If we assume that about 10 microseconds is close to
optimum for the RF pulse width at the tube, then we want to have the rise time
very short. A pulse time of 10 microseconds equals a 50% duty cycle at a
frequency of 50 kHz. That is, at a frequency of 50 kHz, each half cycle would
be 10 microseconds in length.
To have a good quality square wave, we need a bandwidth
which will handle out to at least the 20th harmonic, or 1 mHz. This is a
minimum value, and preferably, should be about five times that, or close to 5
mHz in bandwidth. This value far exceeds any available CB radio transmitters or
Amateur band radio equipment commonly available. The one exception to this are
some of the solid state RF amplifiers made for export, and for Amateur 10 meter
band operation. These have bandwidths far in excess of the 5 mHz we require for
pulse service.
It is possible to broad-band some of the RF circuits
inside some of the CB radios to improve the RF response for pulse service, but
doing so will lower the power output somewhat.
This leaves the antenna tuner as the weak link in the RF
chain. As it turns out, the tuners commonly used are fairly broad banded at
27.12 mHz. The problem is that as you operate away from the exact center
frequency to which the tuner is adjusted, the tuners internal reactance changes
drastically from the optimum value required to match the amplifiers output to
the Rife/Bare tube. This mismatch causes distortion in the shape of the pulse
by effectively filtering out some of the harmonics of the square wave from the
signal. To some degree, this can be compensated for by deliberately
misadjusting the antenna tuner to obtain the best pulse shape. This what I have
done with my system for the oscilloscope pictures you have seen here.
It was necessary to adjust the various RF stages
throughout the RF chain to obtain the best pulse shape. In this respect, it is
just like tuning up a television transmitter. In fact, the required bandwidth
for an NTSC television transmitter is the same as we need for our pulse system,
5 mHz.
In pulse service, the amplifier must be capable of
providing the highest amount of RF power you need. The average power the
amplifier has to handle will depend on the pulse length times the pulse
frequency. For instance, with my unit, which is capable of running 1 kW peak
power, the average power through the tube will vary from less than 1/2 watt to
as much as 300 watts. The higher average power is obtained at higher modulation
frequencies. your amplifier must be able to handle at the same time, both the
peak and average power levels required.
Happy experimenting!
I thought it might be
interesting to take a look at what happens in a Rife/Bare tube when it is
connected to a standard "Neon Sign" transformer.
Interestingly enough,
there are remarkable similarities between the operation with a luminous tube
transformer and the RF system of the Rife/Bare device. Needless to say, there
are considerable differences when it comes to the method of of feeding power to
the tube. When using the internal electrodes for 60 Hz feed, the limiting
factor is the current carrying capacity of the electrodes themselves. With an
external RF feed on the outside of the tube, the only real limit is that of
overheating the walls of the tube and puncturing it.
This picture
shows the optical output from the Rife/Bare tube when it is connected to a 7.5
kV @ 60 mA luminous tube transformer.
The horizontal scale in this picture is 2 milliseconds per division.
Note that what is displayed here is NOT a 60 Hz waveform, (the North American
standard power line frequency) but, is instead, a 120 Hz optical signal. This
is so because the Rife/Bare tube is actually turning on and off with each
half-cycle of the AC power from the transformer. The raggedness of the waveform
is caused by the fact that the discharge in the tube is uneven, and visibly
"grainy" to the naked eye. The optical detector clearly shows the random
optical noise present in the signal. When this type of optical noise is present
on the light from an RF driven tube, it generally indicates the presence of
undesirable phase noise in the modulating signal or other unwanted amplitude
modulation of the RF carrier due to power supply problems or amplifier
oscillation.
In this picture, the trace is at the lowest point on the screen when
the tube is turned off. The higher the trace, the brighter the tube. Look at
the turn-on optical transients visible at each half-cycle. The first is visible
just to the right of the second vertical division, and the transient at the
other half-cycle point is visible just past the 6th vertical division.
Why does the spike seemingly occur AFTER the tube has started to light
on each half-cycle and not just as the half-cycle starts?
It turns out that some light is emitted from the tube as the gas
around each electrode first begins to ionize. This happens as the voltage from
the transformer is beginning to rise at the start of each half-cycle. As soon
as the gas ionization extends across the entire tube and connects the
electrodes together, a much higher current begins to flow between them, and the
discharge becomes much brighter, resulting in a pronounced light spike from the
tube.
This picture
is an expanded view of one of the half-cycle turn-on optical pulses. The
horizontal scale represents transformer voltage, with the lowest voltage being
at the left side of the picture, and the highest voltage at the right.
The first pulse occurs immediately after the half-cycle
starts. The initial pulse uses up the energy which is stored in the
inter-winding capacitance of the luminous tube transformer. As soon as this
stored energy is used up, the light level decreases to a very small value. In
fact, the discharge during the time interval between the first pulse and the
second pulse does not completely connect the electrodes together. The discharge
does not reconnect the electrodes until the transformer voltage increases
sufficiently at the time of the second pulse.
When the second pulse occurs, the transformer voltage is high enough
so that the tube stays on for the remainder of the half-cycle of AC power. When
the discharge connects the two electrodes at the time of the second pulse, it
begins drawing a heavier current from the transformer. The surge of current
causes a momentary "bump" of the the inductance and the distributed capacitance
of the secondary windings which form a resonant circuit, making the secondary
voltage "ring" or "bounce." This results in the dampened sinusoidal optical
waveform shown for the remainder of the trace. Note that for the remainder of
the trace, the general trend of the light signal is rising, as the transformer
voltage continues to rise.
Hmmm... well, if stray capacitance can
cause a pulse in the light output, what would happen if we add a lot of
capacitance to the circuit? Let's try
it!
In the picture below, I have added an additional capacitance of 0.08
microFarad across the secondary of the luminous tube transformer. This added
capacitance resulted in a severe change in the optical waveform, as is shown
here:
Wow!! What a
difference! What has happened here is that the tube produces NO light at all
during most of the time, but instead generates very brief, intense pulses of
light. In fact, the intensity of the pulses is more than 500 times that of the
previous pulse light level. It's not easy to see, but the pulses at about the 3
1/2 division mark extend completely off the top of the screen. The added
capacitance is storing up energy from the transformer and dumping it through
the discharge in the tube whenever it turns on. This also changes completely
the way the tube appears to the eye. Read on...
This picture
shows how the tube looks when it is driven from the transformer, but without
the added capacitor in the circuit. The discharge is dim, fuzzy, and tends to
fill the entire tube from wall to wall. The color is a pale blue with a touch
of violet in the core of the discharge. The visual brightness is quite low.
Zot!! Here we
have added the extra capacitor to the circuit. The tube appears quite bright to
the eye, and in fact, will readily cast a shadow whereas the tube running
without the capacitor will not do so. The discharge column is now pencil thin
and concentrated tightly down the bore of the tube. The color has changed to a
brilliant pink-violet, which exits from the inside of the electrodes. The
electrodes get quite hot, and the color of the discharge near the electrodes
begins to change to a blue-white after a minute or so of operation, indicating
possible electrode overheating may be taking place. The general appearance of
the discharge is quite similar to that of the tube when it is running with the
10 microsecond RF pulse. This is probably because in both cases, the tube is
operating in pulsed mode, rather than average power mode.
