Kapanadze Cousin - DALLY FREE ENERGY

[itsu]

Please increase the horizontal ns/div on your scope and compare the rise/fall times with and without the MOSFETs connected.

Ok,  i made some tests with my MOSFETs and (old) toroid just to see how they behave.

It shows that there is way to much current running through the 2 turn primaries, so i need to start building another one.

Video of this testing here: http://www.youtube.com/watch?v=jwwLWSdm4pM&feature=youtu.be

I also started a breadboard build for the nanopulser using the LT1073, but realized that a breadboard is not a good platform for such a delicate pulser, but i will see what comes out of it.

Regards Itsu.

[Vortex1]

Itsu:

Running at around 4.6kHz, you have not nearly enough inductance with two turns primary on your core. You should see a nice ramp, but your waveform  shows nearly instant high current draw across your 0.1 ohm current sense resistor.

This core might be fine if you were switching at 50 kHz or more, but that would raise a host of other issues. I realize you want to maintain a certain frequency and turns ratio, therefore use a larger core and a few more turns on the primary until you see a nice sawtooth current ramp up.

Also remember what verpies pointed out earlier,  that a load on the secondary is reflected to the primary, and a short circuit on the secondary will kill primary inductance (except for leakage inductance).

Tungsten filaments are extremely non-linear. A cold light bulb is a fraction of an ohm, a temporary near short on your secondary which is killing any primary inductance until it heats up. Right now it seems you need a lot more inductance to run at 4.6kHz.

I suggest using a resistor load of a thousand ohms or whatever you deem appropriate for the power into the load you expect to see in place of the bulb.

Right now your FET,s are your heater load, as you can see with so much of your power dissipated there and so little in the load.

You can switch to a light bulb later when the core inductance issue is resolved.

Verpies: Thank you for the chart posted earlier. It should be carefully studied by anyone designing blocking oscillators or any switchmode converters.

[d3x0r]

On the  http://realstrannik.ru/forum/44-freeenergylt/81742-dally-ustanovka-free-energy.html forum they did a spectral analysis of the sound in the videos ... They did it in a line graph, this is a full graphic view, figured I'd post it.  When he first turns on the machine  0:30 (  http://www.liveinternet.ru/users/edward_lee/post235237573/play ) the frequency is higher, then at the dotted line the frequency shifts when he turns on the nanopulser.


At some point he switches on the light at the power output , but there is no detectable change so I'm not sure where that is.

[verpies]

Verpies: Thank you for the chart posted earlier. It should be carefully studied by anyone designing blocking oscillators or any switchmode converters.

Actually I have made this chart for the Orbo motor people.
In this motor a stationary coil is charged up with current, next a magnetic rotor pole passes by the coil mechanically and when it is far away, the current in the coil is discharged (energy recovered).

I was trying to show these people that it makes no sense to energize the coil longer than 0.575 Tau because after twice that time (1.15 Tau) needed for charging and discharging the coil, the energy dissipated in the resistance as heat is greater than the energy stored in the coil as magnetic field.

That is true about any energy recovery from a constant inductance coil.

[verpies]

@Itsu
I basically agree with the statements made by Vortex1 so I will not repeat them here.

I can add these points:
1) The coiled filament of a light bulb has inductance.  For kHz loads it's better to use a bulb with straight filament (or a resistor),
2) The TL494 is inadequately bypassed. Solder several 0.1uF monolithic capacitors directly to the power supply pins of the TL494 using very short capacitor leads. An IC with a capacitor soldered from the top can still be plugged into the breadboard.
3) Measure your MOSFET gate waveforms at the two MOSFET pins (between gate and source) - not at the TL494 output pins.
4) Scope your drain current with an open secondary winding at first. You must see the current waveform similar to the one shown on this graph as i_L.
5) Keep your wires short and use plenty of low ESR capacitors for power supply bypassing, closely placed between the sources of your MOSFETs and your ground.
6) Take care what happens during the "dead time" when both MOSFETs are off.  The inductor (primary winding) "does not like" to have its current interrupted and will defend against it by producing a high voltage to keep this current flowing in the same direction and at the same level. This creates a HV spike between the source and the drain of the MOSFET that is turning off (before the other MOSFET starts conducting). This spike is bad. You should minimize the "dead time" during which both of your MOSFETs are off and try to kill the spike with a zener diode or at least a 1kΩ resistor.
7) Watch out for the conduction of the parasitic diodes inside your MOSFETs.

Your current waveforms are so wrong that I will not even attempt to analyze them at this stage.
Please read this.