"If we have a perfect coupling, voltage gain obeys turn ratio": it obeys, but as explained, we have in series another inductance that is not coupled, i.e. the inductance of the secondary can viewed as two inductances in series, one is 100% coupled like the secondary of an ideal transformer, and the other is 0% coupled.
The 0% coupled inductance L is resonant with the terminal capacitor. So the circuit is exactly as if we had a LC circuit directly powered by the output of a perfect transformer. If the perfect transformer has a 1:n ratio, the output voltage before the not coupled inductance L is n*V, but the final output voltage depends also and mainly on the Q factor of LC (in the same manner that we can obtain an overvoltage with a simple LC circuit without transformer).


I'm not sure to understand what you are refering to. In big coils may be we are outside of the quasi-stationary states approximation in which the above explanation stands. If so the current can't be considered constant along the coil due to phenomena of propagation (wire length not negligible compared to the wave length of the signals), and nodes can appear.
In this case the model is much more complex than a ideal transformer with an inductance in series. A modelization by transmission lines would be needed.

In regards to the inclusion or exclusion of a spark gap:
In a High Q tank the [pump] can simply be high speed switching of cmos.

You make a very good point.  The tightly wound helical "antenna"
is not an efficient radiator of Radio Frequency waves for long
distance communication since the Electric and Magnetic fields
of the "squeezed" standing wave are not properly oriented.
It does, however, produce a very intense "near field" which can
be used to produce very interesting effects by induction.

On the other hand, a very loosely wound helical antenna which is
only slightly shorter than the natural free space 1/4 wavelength
does make an efficient radiator.  Such antennae were very popular
during the heyday of CB Radio back in the '60s through the '80s.

yes that's very true Dumped, there will always be a little bit of radiation even from a Tesla coil.   Speaking of helical antennas,  at high frequencies, they can radiate in an "end-fire" mode with high directivity along the axial direction.  these are the prefered antennas for hand held satelite comm, or even mobile apps that do not track.  the beam is circularly polarized so orientation is not a problem and the beam width wide enough depending on the design that no special aligning is required, just point the antenna roughtly in the vertical direction.    I was contemplating making a directional power link with this helical design, it would work great I think for short distances, maybe across the yard type of application.

EM

Ah yes, you've jogged a memory.

Those kinds of antennae were very important at
numerous Naval Installations for the transmission
of telemetry data.

An example is here.

Thanks for comments.   I wanted to re-emphasize this excellent post by EMdevices:


Feeding a Tesla coil secondary from a signal generator -- can you use this then to find the principle resonant frequency?  Evidently so -- but would you be specific on how to do this?

Hi Steven,

The Corums have studied Tesla coils thorougly and you can find a practical test procedure in one of their papers, see mainly Page 2 of their PDF file here:
http://www.tuks.nl/pdf/Reference_Material/Corum/Corum-Tesla_Coils_and_the_Failure_of_Lumped-Element_Circuit_Theory.pdf

If you have access to a signal generator (or function generator with preferable a sine wave output) with a wideband frequency coverage then you could make a single turn wireloop of a diameter about the same or similar to that of the air core coil to be measured and connect this loop to the output of the generator via a series 50 Ohm resistor, you can use a small length of 50 Ohm coaxial cable between the generator output and the wire loop.  Lay this single turn loop onto a (wooden) table surface and place the coil to be measured above it with its axis perpendicular to the plane of the one turn loop (so the cylinder coil stands on one of its base on the table).  

To indicate the voltage levels around the coil maybe you could use a dual channel oscilloscope, Channel 1 would show the voltage at the top of the coil, fix its probe near to the upper end of the coil and Channel 2 would show the voltage at the middle part of the coi, fix the Channel 2 probe near to the middle part of the coil. On fixing I mean an isolated, loose coupling and no direct electrical contact.  Then you can start sweeping the generator frequency, starting from the lower frequencies and going upwards while watching the scope.  Biggest amplitude is received at the the quarter wave frequency of the coil at the top part of it (Channel 1) and you may sense this when coming up from low to higher frequencies gradually and you can see an increasing amplitude on both channels but then passing the 1/4 frequency the channel 1 voltage decreases fast as you pass it (after a definite voltage maximum and beautiful sinewave shape) while channel 2 amplitude still increases a little. Do not change probe couplings near the coil  during the measurements because it affects the amplitudes received.

To make it easier, if you happen to have access to a Grid Dip meter (GDO)  you could check the self resonance frequency of the coil first, then divide it by four to be in the ballpark (due to the coils velocity factor, you get an approximate 1/4 frequency).  GDOs are still available, maybe you can find one in one of labs of a university or borrow one from amateur radio friends. See this link on dipper usage:
http://www.qsl.net/iz7ath/web/02_brew/15_lab/02_dipper/english/pag02_eng.htm#USO%20DEL%20DIPPER  

This is in a nutshell of course but hope it helps.

Gyula

Edited for name correction,  sorry,

Very helpful, thanks, Gyula!

--Steven Jones

   

Observing the thread at EF evolve has been rather fascinating.

Bipolar transistors enable the construction of some very
unusual oscillators and power converters.  It would be very
interesting to see examples of signal waveshapes in both
the primary and secondary circuits.  Those could shed
considerable light on certain operational parameters.

How long will it take for one or more of the experimenters
to try a well driven MosFet in place of the bipolar transistor?
That will be the "secret" to maximum efficiency.  But the
challenge will be optimizing the driver pulse width and
frequency.  Although, in reality, this is far easier and much
more predicatable than "tuning" base current of a bipolar.

In addition to matching transformer characteristics to the
load that is...  Or the load to the transformer as the case
may be.

It's a bit like seeing the evolution of switching circuits in
a time capsule.

Dumped,

I hope you don't think a FET is always an improvement over a Bipolar. FETs can be used for about any switching or amplification work but don't work well for anything but ON/OFF unless the user understands the application between ON and OFF.
As an example: The circuit under construction on my home bench is an opto-isolated 8-channel analog adapter for a PIC project. The outputs are each pairs of Mosfets being used as variable resistors.

Don't laugh  I'm building it from what I have, not the latest single chip solution 

--------

Gyula,

Great documents on tuning a Tesla coil 

I always love to see someone else understands such coils ARE NOT conventional transformers. 

You are correct of course.  Bipolars are just fine for many
applications.  Even in switching circuits they are capable
or reasonable efficiencies with effective base drive.

But in low voltage switching circuits their extremely low
ON resistance coupled with their extremely fast Turn Off
capabilities gives the MosFet considerable advantage over
the bipolar option.  With the caveat that they must be well
driven.

I suppose the real "secret" to getting the best efficiency out
of any electronic switch is knowing how to "drive it" to get the
most out of what it has to offer.

Many experimenters who are broadening their experience
level to include MosFets do not yet fully understand the need
to drive the Gate adequately.  Unless a good Driver Chip is used
their performance can be very disappointing.


Fantastic!!  I'm a very great advocate of using what we have.

me too,  in fact,  I built an airplane once out of nothing but .......  PAPER!     

 
   As I am also an advocate of using what is on hand...  a scavenger at heart, and do also like to recycle components when ever possible.  I have quite a few fets to choose from to incorporate into the Joule Ringer, or the Exciter circuits that I'm currently working on. So, How can these MosFets be used in place of the pnp or npn transistors, without any further complications or additional drivers. Is this not as simple as is sounds? What would be the advantage, other than possibly not as prone to overheating and burnouts? Of which, I've suffered my share of, using the smaller 2n2222 and such, even the bigger and stronger TIP31 just blew up in my face the other day, while using up to 12 volt input on the Exciter circuits.  So, are they really worth looking into? Or not?
 
  By the way, does anyone know what has happened to Dr. Stiffler, or if there is any additional news from his work???

  Dumped:  Good to see you here... as some of the other forums that we've been in, seam to be going by the wayside.
                                                                                                                           
  NickZ
                               

Good questions and commentary NickZ.

Bipolar Transistors are well suited for self oscillating
circuits which employ direct feedback.  And they are
capable of operating at low voltages without many
complications.  Their circuits can be very simple.

The difficulty with bipolar transistors is getting the
base drive "just right" in order to attain most efficient
switching in power circuits without either underdriving
it (runs hot) or overdriving it (also runs hot or self destructs.)

"Tuning" the base drive entails providing enough base
drive current to get the transistor just into hard saturation
during the conduction pulse and to follow this at turn-off
with a momentary polarity reversal of the base signal to
clear the base region of residual current carriers forcing
a speedy cut-off.  This is best accomplished with a
transformer "feedback winding" which develops the
necessary pulse current and voltage.

Once "tuned" the transistor circuit will best operate over
a narrow range of input voltage.  If the applied voltage is
increased measures will have to be taken to reduce base
drive to prevent "overdriving."  This is why many experimenter
oscillator circuits have a variable resistance, usually bypassed
with a capacitor, in the base connection to enable either
increasing or decreasing base drive.  The bypass capacitor
functions as a "speed up" to assure that turn-on and turn-off
of the transistor are as fast as possible.

MosFets, as you have noted, demand a bit more complexity
because the Gate Drive signal must be from 5 to 10 Volts
magnitude.  Also, unless oscillating at high frequencies of
100 KHz or preferably higher, a Driver Chip is best used
rather than a transformer to provide Gate Drive.

To take advantage of the superior switching characteristics
of the MosFet it is best to use a Variable Pulse Generator
circuit to generate the gate drive pulses at the desired
frequency and pulse width.  This can be a low power, low
voltage circuit made up of CMOS chips or a CMOS version
of the 555 Timer Chip at from 3 to 5 Volts.  It can be powered
by 12 Volts too if desired.

For low frequency switching from about 50 Hz to 5 KHz the
MosFet Gate can be driven with a CMOS chip if desired, though
a Gate Driver Chip is best.  Most Gate Driver chips need from
5 to 10 Volts to function so the MosFet circuit is best for the
6 to 12 Volt range for battery powered devices.

Yes, the MosFet implementation of the LED or CFL Lamp Driver
circuit would have some additional parts but it would be easily
adaptable to virtually any transformer;  it would only require
adjustment of the pulse width to avoid saturation of the transformer
primary winding.  Then adjustment of the pulse frequency for
optimum power transfer.  Once the "hot frequency" region is
located power can be easily decreased by decreasing the drive
pulse width.

Keeping the drive pulse width short enough to prevent core
saturation of the transformer will eliminate the principal cause
of wasted power which is characteristic of bipolar transistor
power oscillators.  The fast turn-off of the MosFet when properly
driven will eliminate the second main cause of wasted power in
the bipolar power oscillator circuit;  bipolars are slow turning off
from saturation so transformer flyback at the collector always
results in some "leakage" during the turn-off transition.

MosFets are more sensitive to excessive flyback pulse magnitude
so, when used, they must never be driven into an "unloaded"
transformer.  The excessively high flyback pulses when exceeding
the voltage rating of the MosFet can take it into avalanche mode
which can be destructive if repetitive.

Once the MosFet is "mastered" for low voltage high current switching
you'll find yourself wanting to use it all the time.  When properly driven
it will run cool to slightly warm, even when switching tens of Amperes.

Yes, Forums tend to run Hot or Cold.  Some of them are presently running
Cold at the moment.  Experimenters occasionally get "burned out" and
take a recuperative rest from toil and trouble.

They'll be Hot again when a new challenge comes up to re-kindle the
fires of curiosity.  Unless the "economy" takes a hard turn downward.
Some are already feeling the hard times...

   As usual, you simply blew me away with your slightly more than overwhelming knowledge, of almost any subject.
Must have been the salt air, still running through your veins. You are better than a wikipedia, and always available.
  I will keep what you've said in mind concerning the MosFets, and their driving circuits. As it seam you are also very right about the cause of overheating or over driving of the bi-polars, as well as the solutions to be implemented.  Face masks, and such.
  As you may know I'm now just a retired, although not yet retarded kitchen table experimenter, but I will do my best to follow through with the above mentioned advice.
       Thanks, once again,
                                      NickZ

  "Some are already feeling the hard times"...  Aye Captain... the tank is empty. Plastic money, no credit, no fun...
   In time all may be feeling more than just hard times. As this may be an specially hard year to overcome, or even to plan ahead, if such a thing exist anymore. That is also why this Lasersaber circuit is especially important now, even before the lights go out, if they do,  as we've been repeatedly warned.
  
   I wanted to relate how I am feeling that this Joule Ringer 2.0 circuit is looking more and more like something related to the illusive Dr. Stifflers SEC devices, and how they may not be as easy to "tune" as we might think.
  We were hoping that the hand wound E-core ferrite transformer and single 2n3055 transistor would be a cinch to make, but, so far nobody has been able to replicate it, even using the same parts. 10 led bulbs = 200mAs draw.  Why???

  The chicken or the egg...
   Kacher- anyone 
   http://www.youtube.com/watch?v=xMiNnHzFsRQ&feature=youtu.be

The circuit configuration is unusual.  It's an example of
a "curiosity" than can be made to work with a lot of
tinkering but which in reality offers no advantage over
the more conventional circuitry.

Placing the "load" in the Base Drive portion of the circuit
is an approach that will be extremely difficult to control
and/or to adapt to various loads.  There are far better, easier
and more reliable ways to develop the properly shaped and
polarized base drive signal for most efficient switching.

As an example, the Horizontal Output Transformer of the old
style TV set is a pretty good approximation of the Joule Ringer
circuit for high power output.  If it is possible to obtain a
schematic diagram for an old transistor TV set zero in on the
Horizontal Output Transistor and how it is driven at its Base.

For maximum efficiency the transformer used must have
three windings:

A Primary driven by the Transistor,

A Secondary to supply power to the Load,

and

A low voltage secondary which is dedicated to
providing feedback and drive to the Base of the
Switching Transistor.

Capacitors are often used across the primary and/or
secondary windings to reduce flyback transient magnitude
and this will often significantly decrease heating of the power
transistor.  The capacitors are small so that the resonance
of the windings will be a frequency much higher than the
operating frequency of the oscillator. 

To make an oscillator for powering lamps with just an
ordinary two winding transformer perhaps the "Stingo"
circuit of Sucahyo's would be the best way to go.

  It looks like Slayer007 is one of the people that still continue to run with the Stiffler type circuits, which are currently of much interest to me. As the normal or regular circuits that can best be simply bought ready made don't offer much incentive or excitement, even though they make work very well for what they do. But, the Kacher circuit, now is also of interest, and especially if regular incandescent bulbs can be used. As they are what I've grown used to and still prefer. As I love their warm and pleasent glow, similar to candle light, or firelight. Leds, daylight rated 5000 to 6500 ones, can even be dangerous, and CFLs contain mercury.
  So, for the moment until the fine points of the Ringer circuits are "mastered", I will continue to dwell into the depths of the Exciter circuits, or Katcher circuits, which still offer much mistery, and are more economical to explore, and use.
 Led bulbs at $12 each, compared to a buck or so, for equivalent incandescent don't really excite me yet... but, neither does does paying for the higher electric bills that are constantly rising. Solar is on my A.S.A.P. list. even though it is still a rip off.
The Ringer circuit if it can be made to run on 200mAs for the 10 bulbs, as shown in the videos, is still quite amazing though, all in all.
  NickZ

Agreed, 10 bulbs at reasonable brightness on less than 3 Watts.

If what we see on the videos seems impossible - then perhaps
it really is impossible.  Some videos are very deceptive.

But then again, perhaps it is possible...  

Before electricity the brightest light was the Limelight.
Maybe we'll see it make a comeback.  It can be done with
an alcohol flame.

Really, really bright with an Oxy-Hydrogen flame...

Wow that was interesting, i had heard the phrase Limelight but never realized the process, i presume the Lime gets consumed a bit like burning Magnesium, if so surely this could be an extremely cheap fuel additive for rockets.