[Prev][Next][Index][Thread]

Re: Tesla Coil Power Factor

Hi John,

At 10:30 AM 9/25/98, you wrote:
>
>  Terry -
>
>  I have been studying the TC waveform photos you made and have come to the
>conclusion that to properly analyze the waveforms the primary and secondary
>currents instead of the voltages should be monitored. The following are my
>suggestions of how the Tesla coil should be tested to determine if the TC
>is operating properly. Note that the following checks the TC input
>(VA,watts,PF), the TC internal operation (waveforms), and the TC output
>(streamers). Other coilers may disagree and are welcomed to make their own
>suggestions.

OK,  I don't quite see the advantage of measuring currents but the current
is what seems to be feeding the streamers.  There is probably much more
data in the current waveform than the voltage.  Since Tesla coils have so
much capacitance in their output sections, The output voltage tends to stay
rather constant during arcs and other high speed events.  The current,
however, is rich in high frequency information which tells much more about
what is going on. 

>  1. The power factor will tell you how much real power vs. reactive power
>the coil is using. This means that a greater amount of real power will
>indicate a greater amount of energy being used by the sparks. The power
>factor can be found by connecting a voltmeter, ammeter, and wattmeter at
>the input of the HPF power transformer (NST). The power factor would be: 
>   Power factor = watts/volt amps
>  The PF should be as close to one as possible. A hot wire ammeter to take
>care of the distortion is preferred but a standard ammeter will give a
>close enough reading at higher PFs. 

Since this is such an important measurement, I would prefer to get a "black
box" power transducer that could measure the voltage, current, and power
without having to do much thinking or worrying about distortion and all
that.  These cost a few hundred dollars and there accuracy is easily within
1% even on nasty loads like TCs.  Not super cheap, but the hook up and use
is very simple and reliable.  If anyone has one of these laying around that
they could part with, contact me and we can make a deal.  (150VAC, 15A,
2000W single phase with external CT if needed).  If anyone knows a good
place for me to buy one I would also be interested.  Please e-mail me
off-list at terrf-at-verinet-dot-com . 
  
>  2. To check the waveforms a current transformer would be required in the
>ground wire of the TC primary coil. A current transformer would also be
>required in the ground wire of the secondary coil. The scope channel #1
>would be connected to the primary CT and the channel #2 connected to the
>secondary CT.

Current transducers tend to have limited bandwidth and can distort phase
information in systems like TCs that have important information at 60Hz and
200kHz.  My new probe system will use high speed isolated resistive current
transducers that will go from 60Hz to tens of MHz with accurate current and
phase information preserved.  The new revision transducers should also
eliminate the saturation problems that have been a problem in the past for
me.  The measurement of the waveforms you mention should be trivial with
this new equipment.

>  3. The ideal TC waveforms would be as follows. The primary current trace
>should show a ringdown to almost zero to indicate most of the energy in the
>primary circuit is transferred to the secondary. This is the time the
>operating gap should open. The secondary current trace should show an
>increasing amplitude as the energy is transferred from the primary to the
>secondary circuit. The secondary trace should then show a decreasing
>amplitude to almost zero as the secondary energy is dissipated in the
>output streamers from the toroid to the AIR. The streamer to air should be
>used because sparks to ground collapse the waveform. The toroid should be
>sized so the streamers begin at the maximum secondary trace amplitude. 
>
>  The overall efficiency of a Tesla coil is difficult to determine because
>the energy of the sparks (energy out) cannot be easily found.
>   Efficiency = energy out/energy in
>  Where the "energy in" is at the input and the "energy out" is at the
output.
>  Another possibility to indicate the performance of a TC is to use 
>   TC COP = input watts per foot of spark
>   Input watts = .5 Cp Vp^2 BPS/eff
>  Where COP is coefficient of performance and the sparks can be streamers
>or horizontal continuous sparks to a ground point.
>
>  John Couture
>

I am glad you want to measure streamers instead of grounded target hits.
When one sees the dramatic difference in the waveforms, it is obvious that
ground hits are not a typical situation.  Hopefully, my new rotary gap can
discharge the primary system much better that a static gap can and
eliminate that major source of power loss.

I would suggest a different test scenario since my equipment should be
capable of doing much more than you suggest.  Although, all the
measurements you mention would be made in this test anyway.  It would go
something like this:

1.	Set up an operating condition that is stable and repeatable.  This would
include streamer operation and such so that the coil would be close to a
real situation.  Since probes and such would need to be moved around, it is
necessary to turn the system off and on several times.  Thus, a constant
stable condition is needed.  All the usual stuff would be known to good
accuracy (dimensions, coil data, coupling.....).

2.	Using a transducer, measure the power, voltage, and current entering the
variac (I think the variac loss is a part of the efficiency equation).

3.	Measure the power, voltage, and current leaving the variac and entering
the transformer.

4.	Measure the voltage, current, and power entering the protection network.
 This requires a digital scope that can download data to a computer that
can do the calculations to arrive at RMS voltage, current, and power of a
complex waveform.  The data is provided by fiber-optic probes.

5.	Measure the voltage, current, and power entering the primary system
(gap, cap, primary coil).  Latter, I will be needed to go back and see how
much power is being lost by each of these too.

6. 	Now it may get funky.  I can easily measure the secondary ground
current and the secondary terminal voltage.  However, I can't use these to
arrive at power because the streamers are removing energy from the
secondary system.  However, If the probes work they way they are supposed
to, I will also be able to measure the current entering the top terminal
and leaving a point on the terminal that could be set up so that the
streamers are emitted from it.  I can also measure field current.  The
field voltage and current can be measured and analyzed to 60MHz which is
the bandwidth of my digital scope (the antennas go to 100MHz).  The power
can then be found by computer as before.  The current transducers in the
terminal will probably go to 20MHz.  Hopefully, the terminal current at the
field current will be identical and the high-bandwidth field antennas can
be used to determine streamer power.  Only the real test will determine if
all this works well enough to measure streamer power accurately enough.
The real problem is that the streamers exist in the nano-second region
where very high speed measurements are needed to determine the real power
of these high speed events (there is also a delay from the light passing
through the 30 feet of fiber that would need to be compensated for).  Even
if accurate voltage and current information is available the phase
information may be too distorted to do the power calculation.  The phase
information is ten times as sensitive to bandwidth and the voltage and
current.  So if you need 10MHz bandwidth to measure the voltage and
current, you would still need 100MHz bandwidth to preserve the phase
relationship information.  There is also a possible problem that I can only
capture 2500 points on each channel in a given frame.  If very high
bandwidth and low bandwidth information are mixed, I may have trouble
preserving all the information.  Perhaps grabbing many small frames in
sections and adding them all up would solve this.

	I also have temperature transducers in my bag of tricks.  These can
measure temperature rise over time.  This can be compared to static tests
to give a power vs. temperature measurement of say the secondary coil
temperature and other power dissipating components.

	Perhaps, with a long well documented test like this, with all kinds of
data, we could stand back and see what is important and what isn't.  Then
we could concentrate on the important areas for study and improvement.
Also, it should then be possible to say what parameters are really
important in TC operation.  I am sure there are some treasured beliefs that
such a test will disprove.  However, we may replace these with new
parameters we may have never realized were important.  I don't think we
care much about winding our coils with 1/4 wavelengths of wire anymore but
having static gaps loosing 1/4 of the system's power is a new and real
concern.  And, why does a rotary gap work so much better than a static gap?
 Well, we now have the tools to find out. 

	Only the real experiments will tell if this all works out.  If not I'll
just have to whip up some better test stuff and try again :-)

	Terry
References: