Kapanadze Cousin - DALLY FREE ENERGY

[verpies]

I See two kinds of frequencies across your drains. The first 14Mhz oscillation happens when mosfet stops conducting, and a second lower freq. oscillation to the same drain, when the other mosfet starts conducting. 
Looks like that between the two oscillations we have a change to the values...

That's Itsu's scopeshot?
Think of an experiment to verify that the open winding rings at a different frequency than the closed winding.

parasitic capacitance and leakage inductance which determine the oscillation value. I hope Verpies will through some light to the shade.

Yes, it is the leakage inductance that can form an LC oscillator with parasitic and stray capacitances.
The full winding inductance cannot form it, because most it is consumed by the "transformer action" (mutual inductance).

Is there any way for someone to predict the capacity value of a diode when this works in combination with a coil? I see for example that the diode bridge at the output of grenade, gives specific resonant rise to a frequency value according diode's capacity and grenade's inductance. Can this be calculated from the begin?

Yes but it is not easy.  Also note that any added parallel capacitance should decrease the frequency.  The bridge at the output loads the secondary and that can decrease the apparent inductance of the primary.  Less inductance there results in higher frequency.

[Jeg]

As mentioned earlier, i build a new yoke consisting of a T-520 Toroid using the loss-less clamp design with the same number of turns on the primaries (2x 12) and secondaries (3 and 28),
see picture were i compare the new toroid with the first one (not loss-less) i had.

When firing up, the 24V battery stack hits its auto fuse of 10A and thats it.

It seems that this "toroid yoke" gives way less inductance to the primaries (5.8uH each compared to the 64uH on a "yoke yoke") so it presents itself as a short to the MOSFETs.

I had the same effect as when i by mistake had wound my present loss-less clamp design yoke as bucking primaries, also then the inductance was too low and the 10A auto fuse came out.


So how to remedy this?   Use more turns on the primaries, but that will influence the secondary turn ratios as well which could cause the toroid to run out of space for all those windings.

Or stick with the present "yoke yoke"?


Itsu

Hi Itsu.
I guess your new core has lower permeability which makes it able to oscillate at higher frequencies than the yoke material. If in was in your place I would try a smaller cap at the output of your choke filter. I haven't try it but normally it will limit the inrush current. 

[Jeg]

Yes but it is not easy.  Also note that any added parallel capacitance should decrease the frequency.  The bridge at the output loads the secondary and that can decrease the apparent inductance of the primary.  Less inductance there results in higher frequency.

Thanks Verpies.
I am in changing my yoke today to a lossless one. Will it be any advantage if I use it in combination with my dissipative snubber? Won't it cut this way all this ringing that we see in Itsu's setup?

[starcruiser]

@ITSU,

Did you look at the specifications of the materials used in the cores you are using? The TV Yoke is designed to operate at a much lower frequency (Khz) than the standard torrid (typically Mhz or High Khz). TV Yokes are optimized to operate around the 15Khz area as this is the typical sweep rate (horizontal), the vertical is usually 30 to 60hz. I believe this is something most are overlooking when substituting cores in this project.

The test instruments to check inductance are using a low frequency signal (100Khz) to calculate the inductance. Ideally we should be using a frequency close to the operating frequency of the circuit to make sure the measured inductance is accurate.

[itsu]

The drain to source capacitance is more that 2pF.  The additional inter-turn capacitance would make it even higher.
But it is not the ~60μH that is ringing with this capacitance.  It is the much smaller leakage inductance which forms an LC tank with this capacitance.  The leakage inductance should be measured with the secondaries shorted.

So with these numbers the LC resonance of the leakage inductance seems plausible.
Good. That confirms that the 14MHz ringing comes from the drain circuit.

However that does not explain the 35.25MHz peak and the 9.416MHz peak with its harmonics.
Please repeat this test with a 100pF capacitor and note whether these peaks are also affected.
This means that the magnet forces the alternating magnetic flux out of the core, because the core forms a circular magnetic path and affecting any part of it would affect the inductance of both windings equally.

So that would be at the splits were i also have the stacked magnets.