Lenzless resonant transformer

[verpies]

Opening the secondary and scoping the primary LC (while injecting a signal with an extra few (7) turns) shows a more peak like
resonance

Don't think that those 7 extra turns are loosely coupled.  They are not.
As such they will not allow the resonance amplitude to build up higher than SG's amplitude (measured on those extra turns)

Connecting my bulb to the secondary only, removes any resonance completely, at least i see no more resonance peak.
...
Q:Why is the resonance disappearing when connecting the bulb only to the secondary (load is to much)?

Due to the high magnetic coupling coefficient, connecting a bulb across the secondary is equivalent to connecting the same bulb across the primary (and leaving the secondary open).  You can try it and see the equivalence for yourself.  The resistance of this bulb significantly decreases the Q of the primary LC tank, decreases its amplitude and flattens its frequency response.

I then rewound my coil to have 2 complete turns on the circumference of the core, but after 1 revolution i turned back, so have now a forth and back kind of winding (this is where the next video starts).

By keeping the turn direction constant for 60 turns and reversing the circumferential advancement after 1 revolution and even number of revolutions (2), you have made a transformer that is close to an ideal.  It has almost not flux leakage, no leakage inductance and magnetic coupling coefficient close to 1.

That is very good for an IMT or a power supply transformer, but for the purpose of an oscillator it has no leakage inductance that could oscillate.

I redid the above tests, pointing to a somewhat similar resonance peak of around 22KHz (7nF cap prim, open end sec), while when attaching the bulb to the secondary again, the resonance disappears!

The same statement below is even more true because the coefficient of mutual coupling is even closer to 1 than before.

"Due to the high magnetic coupling coefficient, connecting a bulb across the secondary is equivalent to connecting the same bulb across the primary (and leaving the secondary open).  You can try it and see the equivalence for yourself.  The resistance of this bulb significantly decreases the Q of the primary LC tank, decreases its amplitude and flattens its frequency response."

Now the frequency response is so flattened that it appears to disappear.

Then i set up again the first test with this rewounded coil, meaning primary LC (7nF cap) feeding through the input bulb, and output bulb inbetween the secondary series LC.
Hunting for resonance now reveals a total different picture as yesterday, as the primary peaks (bulb off) around 22KHZ and the secondary peaks (output bulb dimly on) around 550KHz!!??
Q: Why is after rewinding the coil the both resonances way off (prim. res at 22KHz / sec res. at 550KHz)?

Again, that's due to the high magnetic coupling coefficient of an almost ideal transformer.  The free secondary inductance almost does not exist - it is converted to a voltage source by the transformer action.  Whatever remains resonates at 550kHz.  Also, all loads connected to the secondary are reflected into the primary almost without any losses.

It shows that the input current around primary resonance was leading or trailing the input voltage by 90° depending on which side of the resonance we are.

As it should, as it should

But it also shows that when the secondary is in resonance, the input voltage and current are in phase, meaning a pure resistive load?

Yes...or that any remainng inductive reactance has been cancelled by capacitive reactance (or vice versa).
It should be investigated

Video here:  http://www.youtube.com/watch?v=o2HzTHopA_Y&feature=youtu.be   

For OU hunting purposes the interesting situation would occur when the phase shift between the input voltage and current was 90º and on the output side - 0º.

[itsu]

Thanks for this great info.

Don't think that those 7 extra turns are loosely coupled.  They are not.

Ok,  what i meant was, it was not directly coupled into one of the both coils dampening them by the 50 Ohm impedance of the SG.

By keeping the turn direction constant for 60 turns and reversing the circumferential advancement after 1 revolution and even number of revolutions (2), you have made a transformer that is close to an ideal.  It has almost not flux leakage, no leakage inductance and magnetic coupling coefficient close to 1.

That is very good for an IMT or a power supply transformer, but for the purpose of an oscillator it has no leakage inductance that could oscillate.

Ok,  so i will keep this ideal transformer for some project where i need an IMT.
A good Impedance Matching Transformer is hard to come by.

Yes...or that any remainng inductive reactance has been cancelled by capacitive reactance (or vice versa).
It should be investigated

Ok,  i will do some more testing when the secondary is in resonance around 550KHz like also mentioned by Jack above.


At the moment i am winding a new toriod consisting of 2x 200 turns (cw and ccw) of 0.4mm magnet wire on opposite sides of the core like Jack seems to have and do some similar tests.

I can say now already that it does NOT take just half an hour to make it like suggested by Jack earlier 

Regards Itsu

[verpies]

Ok,  what i meant was, it was not directly coupled into one of the both coils dampening them by the 50 Ohm impedance of the SG.

I estimate that the 50Ω of SG appears as ~300Ω to the primry LC tank after impedance reflection and magnetic coupling coefficient of these 7 turns slightly less than 1.

Ok,  so i will keep this ideal transformer for some project where i need an IMT.
A good Impedance Matching Transformer is hard to come by.

A 1:1 IMT is not very useful.  To make a 100:1 IMT you'll need to rewind
What is the core's material anyway?

At the moment i am winding a new toriod consisting of 2x 200 turns (cw and ccw) of 0.4mm magnet wire on opposite sides of the core like Jack seems to have and do some similar tests.

That will make Jack happy.
The leakage inductance of this transformer will be higher.  This leakage inductance will form an LC tank with a capacitor.
Note, that the same effect can be accomplished with an ideal transformer by connecting external inductors in series with its windings.

I can say now already that it does NOT take just half an hour to make it like suggested by Jack earlier 

Maybe he has a Jovil machine
I hate winding toroids by hand - my spools are always too big to fit through the core's hole 

[itsu]

A 1:1 IMT is not very useful.  To make a 100:1 IMT you'll need to rewind
What is the core's material anyway?

Well, my plan was to put the both coils in series creating one primary using a new coil of a few turns as new secondary.

The core's material is unknown to me, but knowing the od (65mm), the id (40mm), the thickness (9mm) and the inductance with a 5 turns coil (20uH)
i calculate the permeability to be around 1000 using the methode given by Vasiliy Buslaev (where did he go?) here

Maybe he has a Jovil machine
I hate winding toroids by hand - my spools are always too big to fit through the core's hole 

It just fits 

Regards Itsu

[verpies]

Well, my plan was to put the both coils in series creating one primary using a new coil of a few turns as new secondary.

I hate to split hairs but my transformers always work better when the secondary is under the primary winding (...or interleaved together).