If it is a short winding, then we should just leave more space between each turn to make use of the entire toroid circumference?

That's one solution, but then you're left with uncancelled circumferential current.
A better solution is to parallel low-count windings.  This way even a one-turn winding can span the entire circumference of the core.

Thanks Verpies. Just to be sure I understand correctly, do you mean that for the purposes of reducing
leakage inductance, it is always better to use the whole toroid for windings than to wind multiple layers
on just one side of a toroid for a particular winding?

Below, I have listed the combinations from best to worst:

1) Spanning entire circumference, even number of forward&back layers,
2) Spanning entire circumference, odd number of layers,
3) Spanning part of the circumference, even number of forward&back layers,
4) Spanning part of the circumference, odd number of layers.

These winding rules apply any time a tightly coupled transformer with low leakage inductance and high coupling coefficient (k) is desired.
These rules are not specific to the lossless clamp windings, although they do help them to perform better, too.


See the illustration below.  The number of flux lines symbolizes the amount of magnetic flux.
Notice, that in Fig.1 and Fig.3, five lines of flux are leaking outside of the core and in Fig.4, six lines are leaking out.
This is what actually happens when narrow windings (not spanning the entire circumference of the core) are used.

BTW: The magnetic flux created by the circumferential current, is not shown on these illustrations because such flux is directed perpendicularly to the plane of the drawing ("out of the page").

In regards to  capacitance of the diode. T1000 is right.
In regards to   values of the resonance  circuit that would never be seen as capacitance  and inductance.
 You must think of it as components of impedance at given frequency.
In the Smith chart you can easily calculate- see( with no calculation) the values of impedance.


1.The same resonance circuit (or element acting as resonating component e.g single capacitor or single  inductor ) will have components of impedance  at the same very time  corresponding to
-Equivalent Circuit Series
- Equivalent Circuit Parallel
- Equivalent Circuit Series-Parallel ( mix)


2.The same resonance circuit (or element acting as resonating component e.g single capacitor or single  inductor ) will have components of impedance 
-with  total impedance showing capacitive character ( below horizontal line of Smith Chart)
characterized by high voltage low current ( at that given frequency)
or
-with  total impedance showing inductive character ( above horizontal line of Smith Chart)
characterized by low voltage high current ( at that given frequency)


3. With change of frequency character of the circuit will change.

4. instrument that uses Smith chart is Vector Network Analyzer (VNE) or  Network Analyzer (NE)[/size]
https://en.wikipedia.org/wiki/Smith_chart




5. some older pieces of  tools similar to VNE  or NE    are using Polar  or rectangular  graph
very helpful is to have (  Phase , Gain Analyzer HP 4194A (50 Ohm)


6. it does not matter how big is the voltage and current in resonance circuit, just bring it down to level of millivolts and now safely measure or calculate  its values


7. parameters of impedance will always be made of:
- inductive reactance
- capacitive reactance
- resistive component
However pure resistive components can have reactances close to zero.
Calibration resistors (  50Ohm) are made this way. They act as  50 Ohm resistors  in all  spectrum frequencies  of the bandwidth of the instrument from 30kHz to 6GHz   

http://www.arrl.org/files/file/Antenna%20Book%20Supplemental%20Files/22nd%20Edition/Smith%20Chart%20Supplement%20-%20Corrected%20Jan%202012.pdf

This feature permits the use of the Smith Chart for any impedance values,

Wesley

What a great info Verpies. Thanks

That's one solution, but then you're left with uncancelled circumferential current.
A better solution is to parallel low-count windings.  This way even a one-turn winding can span the entire circumference of the core.

Same question with Void. Do you mean to wind each turn perpendicular to the core axis? If yes, what about the secondary? Doesn't it need too to be wounded with the same manner so to have Lenz?

Below, I have listed the combinations from best to worst:

1) Spanning entire circumference, even number of forward&back layers,
2) Spanning entire circumference, odd number of layers,
3) Spanning part of the circumference, even number of forward&back layers,
4) Spanning part of the circumference, odd number of layers.

Can we safely assume that case 3 is better than case 2 when diameter of the core is higher than a critical value? Or case 2 is better than case 3 in any way?

...what about the secondary? Doesn't it need too to be wounded with the same manner so to have Lenz?

The same rules apply to secondary windings if the transformer needs to have a low leakage inductance and a high coupling coefficient (k) to minimize ringing and losses.

Can we safely assume that case 3 is better than case 2 when diameter of the core is higher than a critical value? Or case 2 is better than case 3 in any way?

The latter.
The ratio of toroidal to circumferential ampturns will almost always be in favor of the former.

Thanks Verpies!
I don't know when this replication can be considered as finished, but for sure we have learned how to make a decent push pull transformer for any occasion.

Below is a schematic of 1-turn distributed winding that could be spanning the entire circumference of a toroidal core.

Hi Verpies. Ok, I see what you meant by parallel windings now. Thanks.

Good day all:

Just wanted to post some very interesting reading material: If you want to really get into the basic mechanisms involved with coil resonance(s), I have attached an U.S. patent from March of 2010 dealing with *double* resonant systems, or as some say 'resonance within resonance'.
This is some of the *BEST* reading material I have come across with a plethora of information in one document. You might want to read it a number of times to actually grasp the scope of application........  I believe that this is what applies to the D.S. replication and probably has applications in relation to the Ruslan device as well.

I find very interesting the differentiation between L/C resonance and 1/4 wave resonance........  this patent give an iterative development cycle /method (actual steps involved) of combining the two resonances in one coil for the *preferred* results.

enjoy........

take care, peace
lost_bro

No problem,  here the same screenshots, now without (1) and with (2) a 100pF cap across 1 primary (drain to Vcc) with the 4 stacked ceramic magnets half way this primary on top of the yoke split:

Most of the peaks have moved in response to the added 100pF capacitance so they must have been a result of some LC resonance or the driving waveform harmonics.

However the dip marked with the purple asterisk did not move and that might be an indication of some energy absorption in the
core or a standing wave.

So please put two identical capacitors across 2 primary halves (drains to V_CC) and experiment with different pF values of these capacitors while observing whether that dip (trough) moves.

By the way, the spectrum analyzer is a DSA815-TG, so with the tracking generator option.

Oh, so you can do low power sinusoidal frequency sweeps of the yoke windings without worrying about harmonic distortions of the driving waveform.
This is worth doing when you have the gear for it.

EDIT, ok, yes, after also using a clamp-on toriod on the keyboard plug, no more lock-ups untill now.

So it seems that is it worth to choke or place ferrite beads on all "antennas" leading to the computer.