Streamers with current transformer feedback, half wave rectified.
I thought it was about time I made a SSTC. This is actually the product
of several weeks of trying different designs and secondaries. Instead
of going with a fail-proof concept first I tried various discrete
designs with little success, which I later see was most likely due to
the uber high fres secondary I was testing with. Anyway I decided I
just wanted it to work, so I made a low frequency secondary, and
settled on Steve Conner's SSTC PLL driver
from his Dirty Weekend SSTC.
It has the advantage of easy startup thanks to
a constantly running oscillator, and with an antenna providing feedback
it's (nearly) always in tune. The best of both worlds- in one chip.
SSTCs function much like SGTCs. A large secondary coil with much
inductance also has capacitance in the form of parasitic
capacitance between turns and capacitance from to the topload ground.
These reactive components form a resonant circuit, and if a
waveform is applied at the right frequency the AC impedance of the
setup is reduced to zero. By then coupling energy into this setup
through transformer
action we step-up the voltage by a ratio at the base, and this voltage
is stepped up further by the resonant rise. The result is an incredible
increase in
voltage. For a much more thorough explanation see
Richie Burnette's page on SSTC driving.
Coupling is important for getting power into the secondary
system, also explained in Richie's page. Quite simply put though, try
to get the coupling factor
as high as possible as this seems to work best for SSTCs. (K=1 implies
that the primary occupies the same space as the secondary, which is
impossible. High K means that the
primary coil is wrapped tightly and over as much of the secondary as
possible.) The limiting factor is
flashover, when sparks occur between the primary and secondary. Due to
insulating pipe size restrictions I was unable to tune for optimal
coupling, and thus the mosfets heat quite a bit. Ideally I would have
the primary reaching half-way up the secondary and with a few more
turns. This would reduce magnetizing current and couple more current
into the secondary.
A phase locked loop (PLL) is a combination of variable frequency source, and a phase
comparator, with a feedback loop and filter. The variable frequency source is typically
in the form of a voltage controlled oscillator (VCO). The CD4046 is a PLL integrated circuit,
containing two different phase comparators of type 1 and 2, along with a VCO. The type 1
phase comparator can be thought of as an XOR gate, whereas the type 2 variant is a digital
circuit with memory elements. In the PLL driver here the timing components set the rough
frequency range, and the potentiometer sets a DC bias (positive or negative) on the VCO input.
The type 1 phase comparator is used, and will output Vdd/2 when at 90 degrees phase difference,
and hence in tune, and 0 to Vdd as the phase difference moves from 0 to 180 degrees. Closed loop
operation works because the phase comparator output is set to lower the frequency when above
resonance and raise it when under resonance. Tuning the phase bias voltage with the
potentiometer will shift the filtered phase output voltage, so even though the phase is locked
and the comparator output is at Vdd/2, when mixed with the potentiometer output, the voltage
sum will be whatever is required to place the VCO frequency at the Tesla coil resonant frequency.
Mounted on a board for quick and easy demonstrations. Safety is paramount.
In this SSTC driver, the CD4046 is configured to use the type 1 comparator,
along with a low pass filter to control the voltage controlled oscillator on pin 9.
The one input of the phase comparator is fed directly from the VCO output, and the other from an antenna
sensing the voltage from the Tesla coil topload.
The VCO upper frequency is determined by the capacitor on pins 6 and 7, along with the resistor
on pin 11. The resistor on pin 12 sets the lower frequency the VCO can operate at. Feedback in
the form of square waves is provided by the antenna, which is clamped by the signal diodes.
The potentiometer and resistor feeding into pin 9 are intended to provide a voltage bias to the
VCO input, allowing for tuning of operating frequency. When operated in this manner the PLL will
lock it's output to the input with a phase difference of 90 degrees at the VCO center frequency,
and close to 0 and 180 degrees at the ends of the frequency range. The center frequency is given by
(((F_upper - F_lower) / 2) + F_lower). Where F_upper and F_lower are the frequencies when the VCO
input is at the supply voltage or ground respectively.
With this understanding of the circuit, the Tesla coil resonance frequency must first be
determined. Then, the timing components need to be picked such that the resonance frequency
of the coil lands near the VCO center frequency. I'm unsure of how close the F_upper and
F_lower frequencies should be to the Tesla coil resonant frequency, but far enough away that
changes in resonant frequency can be tracked. Too large a range seems to cause problems
for the PLL in achieving a lock. The potentiometer can be used to bias the VCO, in order to
allow for tuning of the phase offset. The fact that the PLL can run without an input signal
from the antenna means it is easy to set it up offline, and get things or more less in tune
before applying power to the system. Once setup, it's just a matter of powering up, and tuning
the phase potentiometer until break out occurs!
Audio modulation is also possible by further biasing of the VCO voltage, but
I had to run my coil from half-wave rectified mains to keep it from burning up, so it was
never implemented. Check Steve Conner's page
to see how. Later on I built a continous wave SSTC, the
Audio Modulated Tesla Coil, which features audio modulation.
Tuning tips
A sharp breakout point is needed. It is not enough that it is sharp, but it must be far away from the electric field
generated by the topload as well. A large topload is ok, but the breakout point needs to be moved far away from it
in order for breakout to occur. A poor breakout point seems to make the system work unreliably. In some of my other coils
with toploads, it wouldn't work at all until making a longer breakout point. Performance wasn't acceptable until the
breakout point was placed far from the topload.
If using a current transformer for feedback rather than an antenna, the polarity is important. Change by switching
which direction the ground wire enters the current transformer.
Tuning the PLL is somewhat tricky. Unsure if breakout point issues were the main culprit or not. If building this,
consider using potentiometers for the timing resistors, in addition to the phase adjustment potentiometer.
Left: Nice thick streamers. This was before I changed
the logic power supply to an auxiliary smps
which altered the streamer appearance. The other three picture show the result, much longer,
but thinner streamers. The pic on the top right shows how salt on the breakout point colors
the streamer. This works well with other salts too, just look up flametests on wikipedia
for other colors.
After taking the PLL SSTC 1 out of the closet for fun one day, it
suddenly
didn't work! The windings on the secondary had expanded while in
storage, causing some turns to overlap or get tangled, and the PLL was
somehow out of tune. So I refurbished the secondary, and set the PLL
range again. Still the results weren't as impressive as I remembered,
so I gave the coil a complete overhaul. First I put a ducting topload
on, which reduced the secondary fres from 500kHz to 250kHz. The only
changes in the driver circuit was changing the timing resistors to 3.3k
and 10k from 12k and 16k respectively, and increasing the timing
capacitor size from 330pF to 1nF. The 12V SMPS was also changed to
a good old fashioned linear supply,
and this stabilized the streamers greatly. Since the GDT was far from
ideal to begin with and now obsolete due to the low frequency, I made a
new one with 5-windings to support a full-bridge. With a lower drive
frequency and more primary voltage, I had to make a new primary as
well. Tuning it took a great deal of time as flashovers occurred
frequently. Eventually I found that simply dumping all of the required
turns at the base of the coil was enough to get the best coupling and
avoid flashovers. With all the new jazz I hardly thought a mere board
was enough to contain my new coil, so I modified an enclosure intended
for an old flyback SGTC project. In the end,
the refurbished SSTC 1 turned out the be a whole new coil, the PLL SSTC 1,5.
Disclaimer:
I do not take responsibility for any injury, death, hurt ego, or other
forms of personal damage which may result from recreating these
experiments. Projects are merely presented as a source of inspiration,
and should only be conducted by responsible individuals, or under the
supervision of responsible individuals. It is your own life, so proceed
at your own risk! All projects are for noncommercial use only.
This work is licensed under a
Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
