Kapanadze Cousin - DALLY FREE ENERGY

[NickZ]

   Itsu:
   Have you tried to make the non dissipative snubber, yet?
   I have not found a combination of RC or Oleg snubber components that work well without heating up the resistors.
   But, I don't really fully understand just how to make the non dissipative snubber.
   Any chance that you could make a small diagram, showing the correct components and how to connect it all up to our induction system?

[itsu]

Nick,

no, i did not make the non dissipative snubbers yet, but i can do that.

The combination 0.1uF and 1mH will result in a resonance frequency of about 16Khz which is similar as being used by the push pull.
So that will not account for the ringing in Voids screenshot, which looks way higher (MHz).

Anyway, give me some days, i need to find the correct components etc.

Itsu 

[Void]

Hi Itsu. Yes, the ringing frequency seems higher than just from the L and C values alone.
Good observation. The original article that I referenced states:
"The resonant circuit formed by L2 and C1 when Q1 is turned on rings at a frequency of:
F=1/(2*PI*sqrt(L2*C1))"
From what they said, the ringing is supposed to be determined by the capacitor and inductor values used,
but, yes, that doesn't appear to match my test results.

There are also two diodes in there and the MOSFET, and it may actually be that the ringing
frequency is determined by the L value and the capacitance of the diodes and MOSFET in
series/parallel with the capacitor to get that higher ringing frequency. Changing the capacitor value
does seem to influence the ringing frequency however. I think I tried it with some larger capacitor values
between 0.1 uF and 1 uF and the ringing was at a lower frequency.

From my testing, an L value of around 450uH does not give good results, so I would
suggest trying at least 1 mH or higher for the L value.


Hi Nick. If you have any questions on it, feel free to ask me.
I am attaching a schematic drawing which shows how the non-dissipative
snubber, consisting of one capacitor, one inductor, and two diodes, connects
between the MOSFET Drain and Source and V+.  'V+' is the power supply plus terminal
which connects to the center tap of the yoke primary winding.
You would need two non-dissipative snubbers. One for each MOSFET in the PWM driver.

I am going to try using a couple of commercially wound 1 mH toroids for the inductors.
They are wound on an approximate one inch (25.4mm) outer diameter toroid.
I plan to solder my snubbers to my MOSFETs on a perf board this evening if I
have enough time, so it might be better to wait before ordering any parts for me to solder
together my snubber circuits and connect to my PWM driver and retest it to make sure all is OK.
Once I have it all soldered together using the 1mH toroids, I will restest my PWM driver board and
let you know if it is still working about the same as in my previous test. If everything is still
working about the same, then I will have more confidence that the non-dissipative snubber circuit
works reliably with the component values I am using.

P.S. I had accidentally omitted labelling the other two diodes in my previous drawing of the snubber circuits,
so I re-uploaded the drawing with the other two diodes labelled. I am using UF4007 diodes for all the diodes
in the snubbers.

[Void]

I soldered my snubber board together and have started retesting. 
I am getting different results now, but I did make a couple of changes from
my previous setup. I don't have the two varistors connected in from the Drain to Source
of each MOSFET yet, and I added a 10uF cap from V+ (at the yoke primary center tap)
to ground.

I am getting quite different results now, but it may be due to the 10uF cap I added at V+.
Now I don't get the high frequency ringing any more, but just a single pulse that is limited
at a lower peak voltage. This is causing diodes D3 and D4 to get hot with V+ set only
to 12V. With V+ at 24V it looks like they could well burn. I am going to need to do
more testing yet, but my guess is the addition of the 10uF cap at V+ may be the reason why
things are working differently now. I will try removing the 10uF cap when I get a chance and
see if it starts working the way it did previously. If I leave the 10uF cap in, it seems I would have
to go with a different type of diode which can handle a fair bit more current.

See the attached scope shot of what I am getting at the moment with the 10uF connected.

Nick, it seems it might be an idea to hold off on the non-dissipative snubbers for the time being,
unless you don't mind spending some time experimenting with different values to try to find
the best configuration for your setup.

[itsu]

Void,

thanks for the info; that spike is rather "thick", so probably contains a fair amount of energy, but on the other hand it will stay within
your MOSFETs max voltage (200V) range when on 24V probably.

I do notice your V+ is sagging during the off time (when the other MOSFET is on), so are you sure your V+ ps is up to the job?

Furthermore, the diodes (D1 - D4) in the original paper are shown as schottky diodes, perhaps they are better suited for this situation.


Itsu