Kapanadze Cousin - DALLY FREE ENERGY

[itsu]

Void,

thanks for your testing, very nice, it is very similar with my own experiments results.

I think that if you can eliminate the spike completly, then there is no ringing, so no need for the ringing (RC) snubbers.


I too first was thinking of these active protection snubber setup PNgx1 and PNgx2 outputs that they would go to the gates, but i cannot see that that would work.
Its to me more obvious that the g stands for ground, but also not sure about that, guess i just have to try it.


Itsu

[AlienGrey]

Hi fellers I have been going through some of the old data base on this thread 2012 I believe it was and believe it or not high voltage overshoots were a problem back then and wast energy dissipation of stored energy in the output circuits and putting it to good use in the device as wast energy is not free energy, and this was some 5 years ago,

Verpies came up with a good circuit, but that was a little too involved for me, I used Arunus's  Idea of the biffilar wind witch works very well.

If you scope the drain voltage I think you will find the threshold voltage has rison from the supply voltage so just dumping the overshoot resistively to ground results in current draw therefor a medium to raise the clamp voltage is required in this instance a zener perhaps (not shown in this circuit) ?  works for me.

[Void]

I just tried a quick test with a 'rate of rise' RCD snubber, and it seems
to work about the same as the 'voltage clamp' RCD snubber, but the Allega
snubber configuration seems to bring the overshoot spikes down a little more,
so that seems to be the best configuration for an RCD snubber. I might take a
look now at the passive non-dissipative snubber arrangement, but I don't
have exact design details so I will have to do some trial and error to see if I can
figure out suitable component values to make it work.

P.S. Wow! I just tried the non-dissipative snubber configuration, and it is quite simple to
implement and my initial testing seems to indicate that it works better than the Allega snubber, 
but produces virtually no heat because it does not use any lossy components. It also appears to
be lowering the input current draw from the power supply as it is recovering energy from the overshoot
spike and returning it back to the power supply V+. 
I will post more details later as I am still testing with this type of snubber at the moment.
It is looking promising.

P.P.S.
I have attached a scope shot showing the non-dissipative snubber on channel 1 (yellow).
There is no voltage spike snubber on channel 2 for comparison.
I am still experimenting with component values to see if I can get it better.

Nick, this appears to be a good way to go. No resistors are used, so no heating and
no worries about fried resistors. Also no power wastage. It almost sounds too good
to be true. I am still using the series RC snubbers to stop the ringing.

[Void]

Here is the info on the non-dissipative snubber circuit which I
have just been testing with to limit the switching overshoot spikes.
I show the component values I was using to create my previous scope
shot for channel 1. You can play around with the C1 and L2 values to change
the ringing frequency. You will need two of these snubbers, one for each MOSFET.

I found the info on this type of snubber here (see Method 3):
Snubber Circuit Design Calculators

From what I can tell so far:

Advantages
1 - No heavy component heating, so minimal power wasted.
The diodes seem to stay quite cool in my testing so far, but I am still testing.
2- Energy is returned back to the power supply.

Disadvantages
Adds high frequency ringing in the area of the limited overshoot spike.
This is due to the way this type of snubber works.
If your circuit requires to have a very clean squarewave with no high frequency
ringing, then this type of snubber may not be suitable for you.

Since this type of snubber is much more efficient than the RCD snubbers, I think
I will go with this non-dissipative snubber in my PWM driver for now.

Good luck with your experimenting.
All the best...

[NickZ]

 
  "Both of the preceding snubber circuits can dissipate large amounts of power through the corresponding resistors. Not only does this cause inefficiencies, but it creates heat transfer problems, and increases component sizes. The following non-dissipative snubber in figure 2 is overcomes these problems by storing the leakage voltage spike energy into a capacitor when the transistor shuts off. When the transistor turns on again, energy is poured into a parallel inductor, which resonates with the storage capacitor."
          end quote.

  Guys: I still feel that some ringing is needed for the HV signal to ride on. But, not over 200v, yet high enough for the "effect" to happen.

  These devices are NOT just like conventional electronics, where the top of the square wave should normally be flat, as possible.
  Look at Stalker's best drain signals, below.  Look at my signals, (below Stalkers), which I am working to emulate.

  But, for me, the heating issue on the fets, or snubber resistors persists, still.

   Do you see any Ruslan type self runner showing a flat top square wave?  Or that can self run on drain voltages signals showing less than 100v?  I don't. Most replication made by other guys have failed. In part because of these very tricky snubber circuits.

   Good luck with your tests Void, and Itsu, and anyone else working on this. This is the point where most guys toss the towel in.
   

                                                                                           NickZ