Kapanadze Cousin - DALLY FREE ENERGY

[itsu]

@all,

thanks for the great suggestions and advices, i will start with using a more rugged MOSFET like a IRF840.
I can always build me a "Inductor Saturation Tester" like suggested by Сергей Ð'

Regards Itsu

[verpies]

blah; how could I kill a relay?
RS3âˆ'1D40âˆ'21*

The usual ways: overcurrent when closed or overvoltage when opened.

If you think of using these Solid State Relays in Dally's circuit then forget it.  They are too slow.
I'd be lucky if they can make and break in 1ms (1kHz).

[verpies]

but why is it then that i can vary the voltage on the drain between 0 and 200V without any problem on a fixed pulse width (still opening the MOSFET and current to flow every 10Khz or so), but when i change the pulse width a tiny fraction it blows the MOSFET.

Most likely because at 40ns gate pulse width the MOSFET is not conducting fully.

Take a look at the snapshot taken at 1m57s of your video.
You say that you are supplying 200V and your scope is at at 20ns/div and 500mV/div (with 0.1Ω CSR that translates to 5A/div).
(I artificially added the green trace showing an imaginary gate signal as I envision it)

You can see the current ramping up to 20A in 5ns (blue line) and then it stops increasing despite the gate pulse not being finished.
Why doesn't the current continue to increase further (along the yellow line) ? ? ?
After all, the R=CSR + R_DS-ON + R_WINDING with the 200V supply, establish the V/R limit ≈ 2000A.

There are several answers to that:
1) Your MOSFET is not fully conducting (gate signal is degraded at this pulse width).
2) Your MOSFET is reaching its current limit.
3) Your power supply (with the bypass capacitor) cannot supply that much current
4) Miller effect

Most likely, when you increase the pulse width above 50ns, then the gate signal stops being degraded and the MOSFET begins to conduct even more current and the drain amps continue to ramp up at the rate 4A/ns, or more. 
If your MOSFET can withstand that current, then the high voltage flyback pulse can still kill it when it opens later.
That voltage will reach a high value, because no flyback diode is connected across T1 and no snubbers provide a path for the current, when the MOSFET opens.

P.S
Since L= V*dt/di than the inductance of your T1's primary calculates to 200V * 5ns/20A = 50nH.  A very small inductance !

[verpies]

Read this and make one
Inductor Saturation Tester
http://www.vk2zay.net/article/200
Watch video
http://www.youtube.com/watch?v=L6J5FvqBJY0

I don't think this saturation tester can measure such a small inductance as Itsu has: 50nH.

But this is an excellent video!
On the attached snapshot it nicely shows how the current in an inductor ramps up linearly (the blue line) and how it starts to curve up when it enters saturation (red line).  If it was an air cored inductor of the same inductance then the current would continue to increase linearly along the yellow line (air cored inductors do not exhibit saturation effects).

P.S.
On another thought. Maybe Itsu's scope shows such low 50nH inductance because it is already the inductance AFTER the saturation point.
In other words, maybe the saturation happens so early that we can't see the knee between the two slopes (blue & red) on the scope.

[itsu]

verpies,

There are several answers to that:
1) Your MOSFET is not fully conducting (gate signal is degraded at this pulse width).
2) Your MOSFET is reaching its current limit.
3) Your power supply (with the bypass capacitor) cannot supply that much current
4) Miller effect

great analysis,

points 1), 2) and 3) are obvious to me as i can see the gate signal decrease at this small width, the IRFU320 spec's show pulsed drain current to be 12A and i know i have a 120VA isolation transformer for the 200V supply.

But concerning the low inductance, well that's the design of this toroid i guess.  Or should we deviate from this and use much more inductance?

Regards Itsu