Body
Hi Mick,
"Yes your right it could be an avalanche punching through the gate would look the same as the ferroresonance, using an IGBT but same potential problem."
The avalanche happens through the drain-source n- or p-channel. If the gate insulation is punctured the device is destroyed.
"I would like to use a snubber without rounding off the corners of my square wave but this may also stop the coil ringing which is what I want."
The parasitic body diode of MOSFETs will prevent the oscillations from going below the ground voltage if switching to ground or above the positive voltage if switching to that voltage during magnetization. An AC switch is needed for exceeding the rail voltages like two back to back MOSFETs. Else the oscillations must stay below the voltage between the voltage rails. Or at least when the flyback pulse returns to the zero level is mustn't go below the zero level as this forward biases the parasitic body diode of the MOSFET. This will make the voltage bounce with new flyback pulses instead of a true decaying sine wave oscillation.
"Is a snubber cap any different than the same rated polycap?"
A dedicated snubber capacitor may have a series resistor integrated for dissipating the energy. If there is a load at the secondary the resistor isn't necessary. If there is nothing to dissipate the energy the oscillations aren't damped and thus goes on forever. This will not happen in a real circuit as real components have some parasitic resistance which will dissipate the energy. Ideal components only exist as simplified mathematical models. Thus a real oscillator must have a little amount of power supplied to keep it oscillating. For extracting energy from the ambient some parametric change has to be made along with the oscillation.
In a normal flyback transformer the capacitor in parallel with the switch is chosen to limit the peak voltage and control the flyback time which is one half of an oscillation. The type of capacitor must be able to withstand the repeating voltage peaks and current. A good capacitor has low loss and thus won't heat at the job it is chosen for. The flyback pulse is a parabola or half sine wave. If a square wave is wanted a clipping circuit is needed. A coil and capacitor will make sine wave oscillations. Clipping is done with diodes or some active switching. The energy then has to go into something that can absorb it fast enough.
This image is taken from here 'Flyback Converters for Dummies:' http://www.dos4ever.com/flyback/flyback.html http://www.dos4ever.com/flyback/flyback.html.
At point a the switch is on. Between point b and c the flyback pulse is generated but is clamped during c to d. This is done by the output load. At point d the voltage drops below the output voltage and oscillates freely until reaching slightly below the ground potential (point e) where the parasitic drain-source diode of the switch is forward biased. Between e and f the oscillation is continued. At point f the switch is turned on again and continues during g where the current increases. The points a and g are the on period where the primary coil is magnetizing the core. Between c and d energy is taken out of the coil. Only a small amount of energy is dissipated during e. This is because of the small resistance of the wire and the voltage drop across the parasitic diode of the MOSFET. If the switch wasn't turned on at f there would have been a damped oscillation until all the energy was dissipated.
Regards
Ole
---In [email protected], <mkjekyll@...> wrote :
Hi Ole,
All great points, still thinking about ways of harvest as calcs will get close but there is always the gnats ass final nudge. Have not decided on the caps being filled quickly through loose coupling like a spark gap or power corection into a step down transformer,
Just the capacitance coupling of me being nearby changes things so I can see what Hector said about tuning is a bitch. Nice info on the flyback time transfer function.
I am trying to use distance as my answer the the capacitive coupling.
Yes your right it could be an avalanche punching through the gate would look the same as the ferroresonance, using an IGBT but same potential problem.
I would like to use a snubber without rounding off the corners of my square wave but this may also stop the coil ringing which is what I want.
Is a snubber cap any different than the same rated polycap?
Working in the major investigative stage now so all your points a very welcome!
On 5/6/2018 10:32 AM, onielsen@... mailto:onielsen@... [EVGRAY] wrote:
Hi Mick,
The resonant frequency is given by the circuit parameters. If the iron core saturates most of the inductance vanishes which increases the resonant frequency. How fast the coil is demagnetized determines the peak voltage of the flyback pulse at a given current and inductance. The product of the voltage and time is the same for the input and output pulses when the circuit parameters aren't changed. Thus an output pulse of one tenth the length of the input pulse gives ten times as high peak voltage of the flyback pulse. The secondary then transforms that voltage by the ratio of the number of turns between the primary and secondary coils. For a tesla coil this ratio has to be calculated like for a transmission line transformer. I don't think transmission line theory is strictly necessary for calculating the output voltage of an iron cored induction coil. Use the formulas for normal transformers. A capacitor is placed across the switch to limit the flyback voltage peak to a safe value for the switch. If using a power MOSFETs as switch the MOSFET probably goes into avalanche mode without this snubber capacitor. A mechanical switch will just arc across the contact points without the capacitor. What is seen as ferro resonance could be an avalanche (breakdown) through the MOSFET. This happens very fast just like ferro resonance. It may degrade the MOSFET to give it a shorter life.
How fast the coil can be demagnetized depends on its parasitic capacitance and the load of the coil. A totally open circuited coil without any capacitance will demagnetize immediately and thus induce an infinitely high voltage. Such ideal coils doesn't exist as even vacuum has a dielectric constant. Thus parasitic capacitance is formed between neighboring turns which dampens the voltage. I.e. the energy is converted from magnetic field energy to electric field energy.
The induced high voltage of the secondary coil creates an electric field between the output terminals. If close to the earth or any metal parts the field will charge those parts even through a vacuum. If the distance isn't great enough the field will punch through and discharge the field. Else the energy is stored as an electric field around the coil and in the dielectric material between the windings. When the windings share a common connection between the primary and secondary coils the electric field has to be stored on the outside of the induction coil. Metal parts close by may get charged and discharge to the air by ionization or to persons touching them. Thus be careful when experimenting with high voltages. Everything is coupled by the capacitance between them.
Regards
Ole
---In [email protected] mailto:[email protected], <mkjekyll@...> mailto:mkjekyll@... wrote :
Warren,
The earth spike seems to work weather or not there is a ferroresonance signature on the coil waveforme, just a whole lot better when there is.
Therefore the frequency become speculative as to what resonance one is working with, be in the spark gap or the modulation or the beat frequencies.
I think a wideband approach of filling some caps from a spark gap or plasma tube as fast as they can be discharged is the ticket. This is where the experiment becomes costly as the high voltage multi mF caps are not cheap.
In the Colorado notes Tesla has an extra coil in his Magnifying transmitter that no one seems to understand what he was doing with it. I propose this is the ferroresonant modulator.
A less costly way to harves is by PF correcting the energy to work in a transformer so the field requires much research and moving slow in order not to burn up parts