Kapanadze Cousin - DALLY FREE ENERGY

[TinselKoala]

Yellow and green dots indicate where the scope is.
Scope settings at 1V/1V  10uS  probes set at x10/x10.
Input voltage around 12V.
Choke; is about 10mH
Coils are about 908uH (0.9mH) 
Coil that is not connected is about 6mH


Resistor to ground is 10k  (9.8k measured)
resistor to power is 470


C1 is 500nF I think
M1/M2 IRFP4020H


Added another shot side by side with gate signal ; second scope 200mV/200mV  10uS; probes set at x10... But you can't see division lines...


Think slightly higher voltage on the second shot
----------
Edit; added scope (blue is on the choke before the coils, Yellow is on one mosfet base)

Thanks... I am puzzled by the low voltages indicated on the scope... you say your probes are set to 10x attenuation but the last photo shows the scope channel set to 1x... These settings must match! Scope channel and probe must both be set to either 10x or 1x for the numbers to read out properly. If your scope channel is set to 1x and the probes to 10x then I must mentally multiply the scope's indicated volts/division by 10, right? I mean there is no way a mosfet will turn on with eight hundred mV applied to the gate, like your scope shows, right?

Your circuit is basically the same circuit I use for wireless power transmission and flyback/yoke driving, without the gate protection Zeners I use, and with a different output coil arrangement (but very similar to the flyback driver). Those spikes that have shown up are the switching transients when the mosfets turn on and off quickly, which you are now seeing because of your superfast diodes, I think.

Your scoping scheme is "differential probe" style, with one probe on either side of the capacitor and both probe references to supply ground or negative rail. Inside the scope you can now select the trace "Add" function and add the traces together and display them as a single trace, which should look almost perfectly sinusoidal. Or, if your circuit is isolated from the scope ground you can actually scope directly across the cap: probe on one side and reference (ground) clip on the other side of the capacitor. This will give you the perfect sinus output on that trace. Be very careful about other grounded instruments connected to the circuit while you are doing this, even power supplies unless they are isolated.  For an example of scoping across the capacitor with a single probe you can look at the video I just coincidentally uploaded last night on a different topic.
http://www.youtube.com/watch?v=wI2r85CeCaw


The capacitor in these circuits is absolutely critical. It is under a lot of stress and should be well overrated in the voltage department and should be high quality poly-film type. It is better to build up the necessary capacitance with multiple smaller identical units in parallel, than to use one big cap, since this cap will tend to run hot. This cap is the tank capacitor for the LC tank formed by the coil set and this cap. You might like to try putting in some means of varying this capacitor's value. I have clips on my experimental testbed in parallel with the capacitor bank so that I can add more caps in parallel to "tune" the circuit. This is more convenient for me than varying the inductance because I use a single loop for wireless transmission and my flyback/yoke coils are immersed in mineral oil for HV safety and cooling.

Here is what your circuit can potentially do, without overheating mosfets, while driving a flyback transformer. This is with IRFP260N mosfets and a bit under 2 microFarad capacitance in the main tank:
http://www.youtube.com/watch?v=4XRwlNCF1PU

You can put a lot of power through this circuit and with the right mosfets, diodes and capacitors it will really "light your fire". The same basic circuit with a slight mod and using IRFZ44N mosfets and different capacitance makes my wireless power transmitters.

[d3x0r]

Thanks... I am puzzled by the low voltages indicated on the scope... then I must mentally multiply the scope's indicated voltsision by 10.

yes; I could change the multiplier in the scope....

Your circuit is basically the same circuit I use for wireless power transmission and flyback/yoke driving, without the gate protection Zeners I use, and with a different output coil arrangement (but very similar to the flyback driver). Those spikes that have shown up are the switching transients when the mosfets turn on and off quickly, which you are now seeing because of your superfast diodes, I think.

Hmm could be that they switch faster now; but previously I had leaky diodes that at this frequency really were acting as a wire (which I also tried at low voltage) and I didn't have spikes.... maybe they're DSRD'ing somewhat... The previous ones (1n4004) were allowing the voltage from the collector side to get to the gates, which would have required zeners since I get at least 2x supply voltage at the gate.
I also wonder if maybe it's got a small 'jump' when the coils switch that is causing a variable inducntanct... the U is just setting on the I, not even some tape holding them together....

Your scoping scheme is "differential probe" style, with one probe on either side of the capacitor and both probe references to supply ground or negative rail. Inside the scope you can now select the trace "Add" function and add the traces together and display them as a single trace, which should look almost perfectly sinusoidal.

That's an idea.... usually I just align them on the same baseline with ione inverted... probably should use a subtract instead of add though, or I'd get double humps...

Or, if your circuit is isolated from the scope ground you can actually scope directly across the cap: probe on one side and reference (ground) clip on the other side of the capacitor. This will give you the perfect sinus output on that trace. Be very careful about other grounded instruments connected to the circuit while you are doing this, even power supplies unless they are isolated.  For an example of scoping across the capacitor with a single probe you can look at the video I just coincidentally uploaded last night on a different topic.
http://www.youtube.com/watch?v=wI2r85CeCaw

Ahh yes I can do that too

The capacitor in these circuits is absolutely critical. It is under a lot of stress and should be well overrated in the voltage department and should be high quality poly-film type. It is better to build up the necessary capacitance with multiple smaller identical units in parallel, than to use one big cap, since this cap will tend to run hot. This cap is the tank capacitor for the LC tank formed by the coil set and this cap. You might like to try putting in some means of varying this capacitor's value. I have clips on my experimental testbed in parallel with the capacitor bank so that I can add more caps in parallel to "tune" the circuit. This is more convenient for me than varying the inductance because I use a single loop for wireless transmission and my flyback/yoke coils are immersed in mineral oil for HV safety and cooling.

Here is what your circuit can potentially do, without overheating mosfets, while driving a flyback transformer. This is with IRFP260N mosfets and a bit under 2 microFarad capacitance in the main tank:
http://www.youtube.com/watch?v=4XRwlNCF1PU

You can put a lot of power through this circuit and with the right mosfets, diodes and capacitors it will really "light your fire". The same basic circuit with a slight mod and using IRFZ44N mosfets and different capacitance makes my wireless power transmitters.

Regarding the cap, it's a 400V cap, I had at one point added a (smaller) 100V ceramic cap that was 2nF in parallen with a 1nF to have 3nF; this additional cap definatly heated up.  These red ones don't have that issue. (yet) 

Wow; that's a nice jacob's ladder   Didn't think you could push that much current though a flyback   and that's just arbitrary resonance?

My better mosfets should arrive tomorrow (200-500W instead of these 21W)...
Then the issue will be my power supply has a max of 3A output... so I'll have to move over to batteries, which I have; but trying to make sure I don't end up with a locked open mosfet...
Since I can go to 20V on the gate, I just removed the zeners and keep the power supply under 15V, although I have some 19V 5W zeners coming too which I will end up adding back. 


The coils are actually wound on a U/I core that is 1"x1" ferrite area and 4"x4" size... kinda big   really resembles a yoke core with more ferrite... I have a couple yokes I'll rewind in the future.. but yes, the coils are on the same core (like Nickz) where i see yours are entirely separate. The choke is wound on a similar core...


There is a small affect on the frequency due to voltage but only at the low end of voltage... after a certain amount it is fairly constant frequency




First is the one I'm using for a choke, connecting red/green in series.
Second is my drive coils (red/red) and the secondary (green)  Although I guess it needs to be just a few turns instead.

[TinselKoala]

It looks like you are handling things well. There was one other thing I thought of: in my experience, if the circuit is operated with bad connections to the coil this will destroy the mosfets in no time. Sometimes they can literally explode. So for best performance and reliability I would not use breadboard construction for the "tank" part of the circuit. The mosfet drain-source wiring to the capacitor and coil should be solid and symmetrical, with heavy conductors and soldered connections, or at least tightly screwed down connections to the coils.
Also, for people working with this circuit, don't slowly ramp up the supply voltage from zero. Sometimes the circuit won't oscillate and one mosfet will stay on and fry itself very quickly. Just supply at least 12 volts and turn it on with a switch or other fast connection. This will kick the circuit into oscillation immediately on power-up and save the mosfets.

Yes, invert one channel and add, or just subtract, that's right.   

The Jacob's Ladder thing uses a TV flyback transformer with my own primary wound around the outside part of the ferrite core: just 8 turns of heavy wire, center tapped.  This is the coil that is in the tank with the ~2uF capacitor, and is how all that energy is coupled into the flyback's secondary. The balance between the capacitor and the inductance of the coil is what determines the frequency of the oscillator and the power throughput.

[corry]

4 requirements for ou

1. hf oscillator(AC)= kacher(first generator)
2. spark (electrostatics+magnetics) hf dc dioded kacher can be used(2nd generator) those who want to avoid spark can feed  the hf dc to coaxial cable shorted at other end
3. output coil design must be cw+ccw
4. ferrite rings or ferrite rod bundle core

00

The resonator 1/4 lambda of hf, it is not necessary?

[00]

The resonator 1/4 lambda of hf, it is not necessary?

frequency resonance can be adjusted by moving ferrite core. or using  a split copper  tube inside ferrite rings or if you are using ferrite rod bundle core then place the ferrite bundle core inside copper split tube

00