Kapanadze Cousin - DALLY FREE ENERGY

[verpies]

As for measuring Mosfet current a shunt in the source line is a much better solution than a shunt resistor in the drain line.

It is more convenient but it is not better because CSR increases the source inductance (sth that Ixys paper doesn't mention) and most importantly it allows the Source to raise above Gnd. The voltage developed between Gnd and Src subtracts from the Gate voltage (V_G-S - CSR*I_S) creating another feedback loop that is prone to oscillations and decreasing the switching speed of the MOSFET.

However it needs to be in the range of 5 milliohms or even lower for measuring above 10A. Read http://ixdev.ixys.com/DataSheet/99061.pdf pages 7 and 8 for details. 10 amps then give 50mV, which can be displayed on a scope without problems.

Indeed, such low resistance CSR is needed below the source to minimize the effects mentioned above. Such shunt resistors usually have special coaxial design to minimize the inductance of a straight wire.

[verpies]

While tuning for a max nano-pulse after rebuilding my Drain voltage filters to get rid of oscillations, my little toroid went up in flames.
I was working with 42V on the drain only, so probably there was a short on the magnet wire isolation on the MOSFET side of this toroid.

That is unusual.
I don't think that a failed insulation on the magnet wire could cause the core to break because ferrite cores are not electrically conductive.

Also, in 500ns and <0.1% DU.C, the damaged insulation could not have heated up the core sufficiently to damage it.
Note that significant current cannot flow between the turns of unbroken wire with damaged insulation since statically, no significant voltage can develop between unbroken turns wound over that core. 
Dynamically, significant voltage between unbroken turns can exists only while the current rate (di/dt) and the inductance are high. This means - before the core saturates and in your setup, the unsaturated time lasted only for tens of nanoseconds. 
That's not enough to locally heat up the core via I²R heating at such low duty cycle. 
Besides you measured the temperature of T1 many times and it was not heating up.

The core could have been heated up only through magnetic hysteresis losses.
Also, it could have been mechanically stressed due to magnetostriction. 
The core has an intrinsic acoustic/mechanical resonance frequency - we do not know what it is.

P.S.
An excessive current flowing for an excessively long time (>1μs) would have damaged your MOSFET and CSR.  Did it?

[itsu]

That is unusual.
I don't think that a failed insulation on the magnet wire could cause the core to break because ferrite cores are not electrically conductive.

P.S.
An excessive current flowing for an excessively long time (<1μs) would have damaged your MOSFET and CSR.  Did it?

I presently have no csr installed, but no, the MOSFET was fine, i just wound another toroid with this teflon wire and shot the video i showed above.

This "went up in flames" happened when i removed the ground lead (ground lead from the HV probe) after measuring the 1KV peaks.
I have seen this earlier where i removed a ground lead on the DSRD and something blew up, can't remember what.

The toroid was very hot, but i thought that would be caused by the high current as a result of the short, not sure.
I think with this shorted coax cable instead of the 50 Ohm resistor the DSRD and the toroid do get hot, i will test that.

Regards Itsu 

[verpies]

As I mentioned earlier now you need to mix two signals coming from L1 and from coax cable and find proper alignment

And by this T-1000 most likely means proper phase alignment between L1 and L3.
Practically that means that the frequency of U1 and U2 must be very close to each other (or close to the multiple of frequency).

Since it is impossible to synchronize two free running oscillators without some kind of phase locked loop, then even when U1 and U2 are tuned to the same frequency, their phase difference will slowly drift... allegedly hitting the "sweet spot" once in a while and powering the FWBR made of of D10-D13 (should be made out of fast diodes), and recharging C35, C36, C37, etc...

[verpies]

The toroid was very hot, but i thought that would be caused by the high current as a result of the short, not sure.

Think about it:  Intermittent 40A at 0.1% Duty Cycle (DUC) is equivalent to 40mA continuously flowing through this winding.
If the winding has a resistance of 0.01Ω (a gross overestimate) then the average resistive heating of this winding would be
0.040A² *  0.01Ω = 0.000016W = 16μW of Joule heating.
Any doubts remaining?

I think with this shorted coax cable instead of the 50 Ohm resistor the DSRD and the toroid do get hot, i will test that.

Perhaps that's due to those pulse reflections of inverted polarity.
As you have seen in your TDR experiment, shorting the coax at the far end, reverses the reflection's polarity,
It would be interesting to see if they also heat up when the coax is opened at the far end.