Lenzless resonant transformer

[Jack Noskills]

What is the root cause?  The root cause is that silly nonsensical L3 coil that straddles the toroidal core.  I already discussed the magnetic circuit for that core when it is being driven by L3.  I said that the magnetic flux leaves the "top blue" travels through the air, and then enters the "bottom blue."

Here is what I did not say:  When L3 drives the toroidal core as a quasi-cylindrical core, the left half of the toroid has CLOCKWISE flux, and the right half of the toroid has COUNTER-CLOCKWISE flux at the same time.  It's totally nonsensical!

Therefore, when L1 and L2 are in "normal resonance" at 157 Hz, the counter flux generated by each winding is CONNECTED by the toroidal core and you get a near-perfect MAGNETIC FLUX SELF-CANCELLATION, a magnetic SHORT.  Hence, for ALL AC excitation frequencies, there is a near magnetic short-circuit and the effective inductance L is reduced to a very small value.  If L1 and L2 were a perfect match, the self-resonant frequency would be "infinity" (divide by zero.)

So this one was "revenge of the nonsensical L3 and associated magnetic circuit - explained."

MileHigh

I should have drawn a picture of how I think fluxes will go in this case in the pdf, but this is the effect nicely put in words by you: Near perfect magnetic flux self cancellation. I must add that those two L2s also feed each other and because of two opposing fluxes going on in the toroid L3 does not see if there is load or not. 


I disagree with you a bit though. The inductance that is left comes from the local inductance field which is low. Just compare same amount of turns in closed loop core and in a solenoid and effective L that they produce. In this case 4.8 kHz and 1100 nf gives 0.9 mH for L in that one LC. Did not found a calculator for rectangular 8*25 mm coil on 80000 core that would give sane results though. Maybe I did it wrong, I tried with this one: http://www.eeweb.com/toolbox/rectangle-loop-inductance/



Presence of local inductance field can possibly be proved easily. If same amount of turns is used but there is more space between turns then resonant frequency will go up as the influence of wire to a neighboring wire is reduced while 'looped inductance' stays the same. In itsu's case this could be done by moving the part of coils by hand that are visible, no need to touch anything that is under L3. When coils are parallel the spacing does not need to exact, it is the total L (and also C) that is left that counts. Maybe this can be done at a later time once we learn more about the basic setup from itsu.

[verpies]

Just compare same amount of turns in closed loop core and in a solenoid and effective L that they produce.

Yes, it is obvious that L1 & L2 exhibit higher free inductance than L3. 
The inductance of a winding increases with the number of its turns and decreases with the reluctance of the magnetic circuit.  It is given by the  formula;
Inductance = Turns² / Reluctance.

The difference in inductance between L3 and L1 (or L2) is mainly caused by the difference in the reluctance of the magnetic circuit.
For L1 & L2 the magnetic circuit has low reluctance through a high permeability toroidal core and for L3 the magnetic circuit has high reluctance through mostly air.

Low inductance (e.g. L3) presents low reactance to AC signal source and that causes high current to flow in it.  Colloquially this is called a "heavy load".
To decrease that load the inductance and reactance would need to be increased. L1 & L2 are examples of such loads.

Note that in Itsu's video, the mere presence of L3 increases the inductance of L1 & L2. 

@Itsu
What's the inductance of L1 & L2 with L3 shorted and open?

Did not found a calculator for rectangular 8*25 mm coil on 80000 core that would give sane results though. Maybe I did it wrong, I tried with this one: http://www.eeweb.com/toolbox/rectangle-loop-inductance/

This calculator is for a rectangular loop of wire without any core in that loop.
You will get the most sensible results with the formula:  Inductance = A_L * Turns².
You can find the A_L value in the manufacturer's datasheet for this toroidal core

This formula will not work for L3 because the magnetic circuit is not contained within the core for such winding and the reluctance of this path is affected by any permeable objects outside of that core (mostly air).

[verpies]

I feel like the guy in the drawing

That means it's the time to set your scope to XY mode and do frequency sweeps.
The exponential sweeps produce the least artifacts because they present constant number of cycles for each frequency (Rigol calls them "logarithmic" for some reason)

Those sweeps will make you like an octopus instead of a guy with only four limbs in that funny drawing.

[verpies]

If you had a true closed-loop toroidal core for L3 then it would look like much more of an AC short-circuit.  (In the past few minutes I am very confident I figured it out, and that last sentence is the big clue, but moving on....)

No because "AC short-circuit" means low inductive reactance, X_L = 2πfL.

This means, that the inductance of L3 must be low in order to present low reactance for the same frequency.
Since L = N²/R that means that either turn count (N) has to be small or reluctance (R) has to be large in order to minimize inductance and reactance X_L = 2πfN²/R.

A "true closed-loop toroidal core" has less reluctance than air, thus a winding over a closed permeable toroidal magnetic circuit will have higher inductance than the same winding over a less permeable magnetic circuit (e.g. one containing air gaps). 

In other words, a "true closed-loop toroidal core" offers less reluctance than air and causes more inductance, which means more reactance and less current, thus less of an "AC short-circuit"

[itsu]

Thanks all for you comments, it will take some time to digest.

Mags, i will keep those suggestions in mind as they involve some addons and/or modifications, nice info in that pdf.

MileHigh,  thanks for your reasoning, i need to get over them again when on the bench to check out some things if possible.

Jack, i am glad you can see something positive in all of this, i will check out what you said.

verpies, i will get back on your question about L1/L2 inductances while L3 is open/shortend.

I did notice that my LCR meter showed a correct value when the caps where removed (eg 710mH), but it went into a negative sign value (like -600) when i hooked up the cap on the other coil.
Normally a negative sign on this LCR meter means a too low value, a short or measuring a capacitor in the L position or vv.

Regards Itsu