Non-linear capacitors

[partzman]

@partzman

I just tried with a very large 9x14 cm rectangular ferrite core with a 3.3x3.3 cm square section, which was used in a KW power supply.
I have no anomalies.
The difference compared to my other tests is that here the coils are strongly coupled, distributed on both long sides of the rectangar core, and wide to the point of touching in the middle, so that it seems impossible to me that the opposite flux loop between the two coils.

I also tried with another toroidal ferrite, smaller than the other three already tested. I have the anomaly, with always Vout about equal to Vin/10.

What is your winding geometry on this core with the anomaly?

Could it be that the anomaly only occurs when the 2 flux can loop back between the two coils?

Actually, this is all I've tested that is, two identical counter-wound coils operating in a bucking mode.  My original intention was to experiment with the H field in the center of the toroid that is the result of the opposite poles at the start and finish gaps between the windings.

Have you some ferrite toroidal core with no anomaly, and if so of which geometry?

Out of twelve identical (supposedly) ferrite toroid cores as described in the paper, only one exhibits this anomalous behavior!  That's why I suspect I somehow changed the molecular structure in some previous tests.  Otherwise, I have not seen this type of anomaly in any other cores I've tested or used in the past.

Pm   

[Smudge]

A magnetic simulation such as FEMM can show that flux will escape through the full length coils when in bucking mode.  It is interesting that a transformer where the two coils are wound on opposite sides of a ring core will exhibit significant leakage flux similar to that of the coils in bucking mode.  This is of course because the secondary mmf from the load current bucks the primary load mmf from the input current (they are in antiphase and don't contribute to the magnetization, only the small 90 degrees magnetizing current does that).  It is easy to see that leakage flux across the centre of the ring core using a sampling coil there, and it is proportional to load current.

I like to think of the two bucking coils as two bent solenoids, C shaped electromagnets that have been brought together.  Bringing each pole close to the opposite one of the same polarity just  makes the ring core into an electromagnet with N and S poles diametrically opposite to each other.
Smudge   

[F6FLT]

Hi Smudge,

I agree with you on the physical effect. The question now remains as to why we do not measure half of the voltage at the mid-point.
I performed realistic simulations with LTspice, including coil capacitance, coil resistance, capacitance in between, oscilloscope probe capacitance. Anyway, LTspice shows that we must have half the voltage at the mid-point, and we don't have it.
I made new measurements this morning with an HP3468A instead of the scope, and same observation in agreement with the scope.

[F6FLT]

...
Out of twelve identical (supposedly) ferrite toroid cores as described in the paper, only one exhibits this anomalous behavior!
...

How is it possible?
On my side, all the four toroids that I tried show the anomaly. Only the big one doesn't (see photo).

Note that the anomaly does not occur at frequencies too low. The impedance must be significantly higher than the internal resistance of 50 ohms of the FG, i.e. the working frequency must be high enough that the voltage drop due to the impedance of the coils does not exceed 10 to 20% of the no-load voltage of the generator (with the cores I have used, frequency always higher than 800 KHz)

[Vasik041]

How is it possible?
On my side, all the four toroids that I tried show the anomaly. Only the big one doesn't (see photo).

Note that the anomaly does not occur at frequencies too low. The impedance must be significantly higher than the internal resistance of 50 ohms of the FG, i.e. the working frequency must be high enough that the voltage drop due to the impedance of the coils does not exceed 10 to 20% of the no-load voltage of the generator (with the cores I have used, frequency always higher than 800 KHz)

I think at 800KHz or higher you just have some resonance effect with parasitic capacitance.
Have you tried attaching some load resistors (to Vout on your schematic) ?

[partzman]

How is it possible?
On my side, all the four toroids that I tried show the anomaly. Only the big one doesn't (see photo).

Note that the anomaly does not occur at frequencies too low. The impedance must be significantly higher than the internal resistance of 50 ohms of the FG, i.e. the working frequency must be high enough that the voltage drop due to the impedance of the coils does not exceed 10 to 20% of the no-load voltage of the generator (with the cores I have used, frequency always higher than 800 KHz)

The major difference I see is in the operating frequency and generating source.  In my paper, I use a solid state power amp with a very high damping factor thus a very low source impedance and my test frequencies are in the range of 6-7kHz.

With these test conditions, none of the other identical toroids produced the anomaly.

However, this same anomalous toroid driven with a 50 ohm SG still exhibits a tap to input gain of 1:07 in one direction at 1kHz!  Reverse the input and output winding connections and the gain is 2.01.  The inductance of the primary with the secondary shorted is typically 760uH.  The individual winding inductance is 34mH.

Edit:  I might add that the measured inductance seen by the generator at the input of the series connected bucking coils is 648uH.

Regards,
Pm

[F6FLT]

I think at 800KHz or higher you just have some resonance effect with parasitic capacitance.
Have you tried attaching some load resistors (to Vout on your schematic) ?

No resonance. Same level from 800 KHz up to 3 MHz or 4 Mhz (resonances are above).

With a load resistance of 1 Kohm, Vout is divided by about three.

I have just improved the configuration by putting an isolation transformer (which is also a toroid with 2 coils) between the FG and the toroid to be tested. Then I can connect the oscilloscope ground on either side of the bi-coils.
Same measure! I have about Vin/10, which I measure from one side or the other. The configuration being now symmetrical, this means that it is the choice of the ground point that determines the coil on which only Vin/10 will be measured.
We face a measurement problem.

[F6FLT]

The major difference I see is in the operating frequency and generating source.  In my paper, I use a solid state power amp with a very high damping factor thus a very low source impedance and my test frequencies are in the range of 6-7kHz.

With these test conditions, none of the other identical toroids produced the anomaly.

However, this same anomalous toroid driven with a 50 ohm SG still exhibits a tap to input gain of 1:07 in one direction at 1kHz!  Reverse the input and output winding connections and the gain is 2.01.  The inductance of the primary with the secondary shorted is typically 760uH.  The individual winding inductance is 34mH.

Edit:  I might add that the measured inductance seen by the generator at the input of the series connected bucking coils is 648uH.

Regards,
Pm

For my configuration, I guess that it is a question of measurement method, probably that the impedance seen from the midpoint is high, and therefore the capacity of the scope probe collapses the voltage (nevertheless rather strange because not depending on the frequency). Not yet verified this idea.

On your side, I don't know if it's the same anomaly or another. What do you mean by "input gain"? Is it the voltage gain? How to connect the FG by inverting the input and output, do you keep the same ground point and connect the FG to the midpoint of the coils? (because if you double the voltage, everything happens as if your coils were in series but not in opposition).

This thread being dedicated to non-linear capacitors, I propose to open a new thread on your anomaly. Do you agree?

[Vasik041]

I think the bias voltage on the capacitor is a requisite for oscillation, which an LC alone does not provide. The proper point on the capacitor's non-linear curve must be found, I presume.
Reversing charge on the capacitor probably doesn't get you to the sweet spot on the non-linear curve.
As someone said: "things must be made as simple as possible, but no simpler". (don't remember the exact quote by Einstein?)
Regards
edit: Attached is a proposed test setup using a signal generator with LF offset output to induce a variable bias on C2.
       Look for small HF burst as variable V traverses sweet spot.

Hi ION,

Today I tried this test setup (see attached picture).
I use signal generator (with isolation transformer) and power supply with variable output voltage to create sine with DC offset.
I think it is close to setup you suggested.
With load resistors 1k and 100 ohm I don't see any non-linearity, but with 10 ohm I see sawtooth like distorted signal.
It is also interesting that amplitude depends on offset voltage, the higher offset the lower amplitude gets. I still not sure about FE, but at least it should be possible build a volume regulator

I tried simulate this setup and simulation looks close to reality. However I am observing some extra small peaks on top of sine (see attached photos).

This could be just isolation trafo non-linearity, or it also could be something...

Regards,
-V.

[partzman]

[snip]

This thread being dedicated to non-linear capacitors, I propose to open a new thread on your anomaly. Do you agree?

Yes, I agree.

Edit:  Actually IIRC, there was a thread where this was discussed so let me take a look to see if it would work.

Edit2:  Here it is, see if you can access the thread-  http://www.overunityresearch.com/index.php?topic=3278.msg55390#msg55390

Pm

[F6FLT]

I can't access. Message "An Error Has Occurred! The topic or board you are looking for appears to be either missing or off limits to you."?

[partzman]

I can't access. Message "An Error Has Occurred! The topic or board you are looking for appears to be either missing or off limits to you."?

Try the link again.

Pm

[F6FLT]

Try the link again.

Pm

It works! I will continue there.

[F6FLT]

I'm testing unknown recovery components of ma collection.
I get a huge effect with one of them: 1.9nF at room temperature down to less than 1.3nF when above the soldering iron. The increase of capacity is faster than the decrease.

Does anyone know what it is?

[Vasik041]

I'm testing unknown recovery components of ma collection.
I get a huge effect with one of them: 1.9nF at room temperature down to less than 1.3nF when above the soldering iron. The increase of capacity is faster than the decrease.

Does anyone know what it is?

May be this ? https://www.allaboutcircuits.com/technical-articles/safety-capacitor-class-x-and-class-y-capacitors/

Edit: see dielectric properties here https://www.vishay.com/docs/28536/geninfo-cd-rfi-safcaps.pdf

[F6FLT]

Thanks for the information, Vasik.

I hadn't classified it with the capacitors because I couldn't read a capacity value. Could it be the mention "2.22" (nF), and that the component has drifted (I measure 1.9nF)?

[Vasik041]

Thanks for the information, Vasik.
I hadn't classified it with the capacitors because I couldn't read a capacity value. Could it be the mention "2.22" (nF), and that the component has drifted (I measure 1.9nF)?

Yes, I think it is capacitor marked with 222 i.e. 2.2nF

Edit: it 10% or even 20% tolerance, so 1.9nF quite ok