Kapanadze Cousin - DALLY FREE ENERGY

[itsu]

Hoppy,

i tried that, but the basic 50V square wave pulses stay, but the ringing now is much worse, peaking 110V 

So it seems it (with the UF4007's and C24) is doing its work as a snubber ok.

I changed the transformer into 2 resistors (190 Ohm each) fet in the middle, and now the MOSFETs do what they suppose to do, see screenshot
So its something with the transformer used, both input coils are about 55uH


Regards Itsu

[Tomtech29]

Hi.
Itsu you have a lot of facilities to hold on to so tightly scheme This is a very good and reliable the devil is in these shoes (KIT)
keeping in such and no other although it was to prove that the plan is with errors that you can see it on a plate
as they go work on a push pull system I do not know what you are missing almost already you have the full set?

I would like to see how the control system operates with the possibility of phase shift on one channel or will be any difference whether the transistor to heat up more?
Here is a simplified version of 50/50% push-pull if someone wanted to try based on the same driver but that's another story
- you believe that this transformer step-up must be necessary for the elevated frequency up to 85Khz I use PWM to limit the voltage at the input where you specify 24Volt the frequency of 35kHz it probably does not matter that much as 130V DC power supply to power Kacher
This single pulses give the frequency and length of the bell

[Dog-One]

So it seems it (with the UF4007's and C24) is doing its work as a snubber ok.

I changed the transformer into 2 resistors (190 Ohm each) fet in the middle, and now the MOSFETs do what they suppose to do, see screenshot
So its something with the transformer used, both input coils are about 55uH

Welcome to the jungle Itsu.  Smart detective work using a resistive load to verify proper circuit operation.  I had the same trouble with my PWM board I posted a while back.

Two things are going on here.  You are getting BEMF and the snubber is doing its job protecting the MOSFETs.  The resistor is probably dissipating a fair amount of heat.  The ringing on the square wave is a clue.  That ringing tells me you are likely saturating the core and there are at least two factors involved:  The amperage switched to the primary and the dead-time between switching--either one will make the core misbehave.  If you have big capacitors holding the rail voltage up, you will have big current traversing your MOSFETs when they turn on.  Think of the primary and core as a simple inductor acting like a water tank--you fill it too fast and it will pressurize itself.  Your rail caps are responsible for this rapid filling.  The second thing is your frequency and dead time.  You have to allow enough time for the magnetic field to transfer into the secondary; if you don't, that energy will push back out the primary, hence the snubbers.  Also keep in mind if no current is exiting the secondary, as would be in the unloaded case, that energy has to go somewhere, which would be back out the primary.  I recommend making sure you have a lamp or something connected to the secondary that simulates the approximate power transfer you actually need from this inverter.

Now, what Nick mentioned about the Mazilli for this application is indeed brilliant because it automatically allows the transformer to run at its optimal frequency.  This is probably what I will switch my inverter to.  Just be sure to either limit the current or bias the transistors so they barely turn on.  Reducing the size of the rail capacitors is a first start or use some resistors between them and the MOSFETs.  The running frequency of the inverter isn't important because we are going to filter and rectify this output back to DC anyway.  Once you have this working, the only adjustment you need to make is the turns ratio, particularly on the secondary, to get the voltage in the range you want.  Again, I would run this circuit under a mild load to verify power transfer.  I think you'll find the circuit to be way oversized for what we actually need from this inverter, which with filter caps and such is probably only a few milliamps of continuous power.

[NickZ]

   Hey Tomtech:
  I like your 50/50 PWM push-pull. Can that set up control frequency? Or, only the duty cycle?
  Do you have a pic of your "simpler" version that you've built?
 
   IF I can get close enough to the right combination of coil windings turns and tuning caps. And by adding ferrite pieces inside the grenade core, and inside of the Kacher core, and the magnetite magnet in the yoke, etz... I can pretty much not only control the frequency, but add GAIN to the output, as well.

   I'm about to re-connect my Mazilli/yoke back up to the grenade/Kacher circuits.  And fire it up, again.
Preliminary tests are complete. If I could only remember them.... Ha!
   

[skywalker66]

All,

question about the DC2DC converter (24V to 130V) as in the red box, see below.

I have found a suitable transformer (SPI 804-2007-35), i think, and hooked it up, but the output voltage was way to high (600V pp).
Measuring the input from the both IRF540 MOSFET drains (see yellow / blue lines in the diagram), shows
a square wave signal, but at about 74V!!, see screenshot.

Input into the circuit is 24V from 2 batteries in series.


Why do i have 74V square wave signal on the drains?  How to get it back to the required 24V? 


Thanks,  regards Itsu

Hi Itsu!

Your setup work just fine as is supposed  to work.
If you check across each half of your primary you'll notice a square wave of about 24V amplitude which is correct.
As you have has check drain voltage, there is another story. When a mosfer are in conduction you'll get 0V reading,
when not in conduction you'll measure 24V(Power supply) + voltage across coresponding primary which is forward induced by
other half primary which conduct in that period = about 48V.
I hope I managed to clarify that aspect for you.
As for your 600Vpp output, try other pair of primary (coresponding to 12V out if your trafo is from a standard
PC PSU with +/-5V & +/- 12V at the secondary and a common connection for gnd).