Kapanadze Cousin - DALLY FREE ENERGY

[verpies]

2) the ground clips are on the + power side of the resistor, and the probe on the other side, both on the resistor.
3) no

Those two answers above mean that you have your probe tips connected at points B & C and their ground clips are both connected at point A. (as illustrated below).
Such probe connections mean that when the falling edges of your CSR waveforms occur (marked in red on the attached snapshot of the scope), transistors Q3, Q6 are just starting to conduct between emitter and collector, which pulls down points D & E and this causes currents to flow through the primary windings (ideally they are current ramps). As these currents gradually build up, they pull down points B & C in the negative direction too, because the CSRs allow those points to move down a little away from Vcc.

Correct me if I am wrong here.  This is important because I do not see the signals at the bases of those Q3, Q6 transistors (points G & I) and I do not know when they are supposed to be conducting in relation to the CSR waveform.

Note that this red falling edge lasts for apx. 200ns and during this interval the absolute current in the primary increases from 0 to -1.5A (because -150mV measured across a 0.1Ω resistor means that -1.5A is flowing through it).
This determines the current increase rate {di/dt} to be 7.5mA/ns and because you are supplying the primaries with 12VDC, then it means that the inductance (L) of the primary is equal to V*dt/di which calculates to 12V * 0.0000002s / 1.5A = 0.0000016H, which is 1.6μH. A very small inductance!

After this red falling edge, the current in the primary starts to decrease (paradoxically it's a rising edge on the scope - marked with yellow color).
The transformer does not transfer any energy to the secondary windings while the current in the primary is decreasing. 

But why does the current start to decrease so early? There are two explanations:
1) The big transistor stops conducting (but why was it conducting only for 1% of the total period time?)
2) The current trace is somehow inverted and that red falling edge actually occurs when the transistor Q3 stops conducting and current in the primary decreases, and the yellow rising edge occurs when the transistor Q3 starts conducting and the current through the primary winding increases.

If pt.1 is true then, the energy is being transferred from the primary to the secondary only during the red falling edge. That's only 200ns out of the 18μs of the total waveform period. If we divide 200ns/18μS we get the ratio 0.0111...
This means that the energy flows from your primary to the secondary only 1% of the time !!! ...actually 2% when we account for two primary windings. 
In that case, it's no wonder that you have insuffucient power problems.

However it is still possible that pt.2 is somehow true and the current in the primary of T2 increases during the yellow rising edge.
Please look into it.

[verpies]

I wound another toroid with 20 gauge wire for secondaries, and wound 4 coils that are 70 turns in two layers; so I wound two 70 turn coils, then put a layer of tape and wound 2 more 70 turn coils on that; so I could wind 6 turns for the primary...

If pt.1 is true in my previous post, then 6 turns for the primary is not enough for this core permeability

[d3x0r]

If pt.1 is true in my previous post, then 6 turns for the primary is not enough for this core permeability

Right, I agree... but again, the original is said to be an AC transformer (220->9) (70:3) ( 3/70 * 220V = 8.57V)... then probably the 70 turns is going to be the side that would saturate such a core, and therefore I'd have to have a 70:1633 winding ratio....


I'm going to tinker with it a bit, and see if I even get saturation with 12V into the secondaries (hmm guess that needs to be 110 or 220V)


(I'm in america so we're 110)...


Was playing with the measuring on the new toroid, it has 0.216uh  ... the other one was 0.130 or something I think....didn't post a measure of it. 

[verpies]

Right, I agree... but again, the original is said to be an AC transformer (220->9) (70:3) ( 3/70 * 220V = 8.57V)... then probably the 70 turns is going to be the side that would saturate such a core, and therefore I'd have to have a 70:1633 winding ratio....

But the original did not specify the AL Value of the core (inductance per turn - related to permeability).
You can still stay true to the original by keeping the turn ratio and accomodate to the characteristics of your core.

[d3x0r]

But the original did not specify the AL Value of the core (inductance per turn - related to permeability).
You can still stay true to the original by keeping the turn ratio and accomodate to the characteristics of your core.

old toroid primary measures as 0.056mH (old one) new one is 0.226mH  ...


I reversed my connections, and 140 windings as primary driver.... and I still cannot saturate. 


And you asked why my pulses were so short...
only when the TL949 is running very slow do I get to adjust to a wide duty cycle... as I turn up the frequency, the duty cycle also seems to shrink... I probably have something miswired, right? 
But anwyay to get power out of this first transformer, running for long periods of time with current doesn't do anything.  You're really only going to get as much out as it first takes to charge, then you might as well turn it off the drive... driving DC-DC converters only work during the change in field...


I was thinking of connecting all turns (280) ... but then I get more resistance, and even though it is 20g magnet wire, it's not going to carry amps and amps, but I should be able to get a good spike from a charged cap...


Well I started to go ahead with this one, it has a much larger inductance because the 6 turns are turned around 1/2 the toroid each... the first all 3 are right next to each other.  I'm getting better power from this... it's a dirtier signal... I'm wondering if I have a bad connection somewhere now...


I damaged my nanopulser with a stray connection i think. 


But so then I'm thinking why not just another TL494?  since 200ns turns into 1.xus through the power transistor, unless the original russian part is just that good that it's high power and super high switching speed... and the tl949 at near 0% duty cycle is a 200ns pulse