Magnetic Delay Transformer

[Smudge]

As this subject has recently been broached in another thread I have decided to create its own thread.  As the name implies this is looking at the effect of magnetic propagation delay along a transformer core to see whether it offers any benefits for OU operation.  Some considerable amount of work has already been carried out but unfortunately that came to an abrupt end when the company funding the work changed course.  As both the theoretical work and the experimental work showed some promise the hope here is that it can be continued to arrive at a proper conclusion.

Here is the first paper, more to follow.

Smudge

[WhatIsIt]

It is very interesting read!
Where ever there is propagation delay, there are 2 points of different potential.

Do you have some more papers on this topic? Or info?

[Smudge]

It is very interesting read!
Where ever there is propagation delay, there are 2 points of different potential.

Do you have some more papers on this topic? Or info?

I started another thread here https://www.overunityresearch.com/index.php?topic=3742.msg73171#msg73171 for a magnetic motor using magnetic delay, but it didn't get anywhere.  But here on the MDT I am on more solid ground as I do have some actual measurement data.  More later.

Smudge

[Smudge]

Here is the next installment.
Smudge

[Smudge]

Now here are some results.  Graham Gunderson built an automatic test rig that stepped through the frequency range so these are not swept frequency measurements, but are spot frequency measurements.  Lots of them!!  For different values of load resistance and capacitance.  Originally we were looking for the real part of the input impedance going negative at a lowish fequency (few MHz but above the LC resonant frequency) as predicted by the theoretical model.  That did not occur for the higher load capacitor values but it did for the lowest value of 10pF (scope probe) at around 12MHz.  This was maybe a measurement artifact but the fact that is was predicted suggests the opposite.
Smudge

[Smudge]

And here is another paper showing comparison between measurement and theory for other parameters at the 500pF 10K load condition.  We were still looking for that lowish frequency negative input resistance effect.  It was thought that the leakage flux (which forms an important part of the magnetic transmission line impedance) was introducing radiation losses that accounted for the fact that the effect did not show up in the measurements.  That led to a shielded version of the transformer also being measured.   More results are in the pdf posted here https://www.overunityresearch.com/index.php?action=dlattach;topic=3844.0;attach=33127

Smudge

[Smudge]

Just for completeness here is another viewpoint on the capacitive loading effect.
Smudge

[WhatIsIt]

Thank you,

I have to read all of this and then will come back.

[Smudge]

Ignoring the 14MHz anomaly for the moment, the whole exercise was set up to look for the input resistance going negative at a point somewhat above the the LC resonance, as predicted by the theory.  Although the measurements did not show that, it is interesting to observe what the device did at that theoretical point.  The pdf below shows the LC resonance clearly on the plots of input resistance and input reactance, both measured and theory.  Also shown is the ratio of transformer output voltage to input voltage which peaks at exactly the frequency where theory predicts that input resistance going negative.  That looks like a resonance, but it is not the LC resonance.  This suggests to me that we were very nearly there.  Of course if the input resistance did pass through zero there the chart would show an infinite spike there.  So I am reasonably convinced that with a bit of tweaking we would have got there, but unfortunately the work was stopped.  Hopefully someone here will take up the challenge and continue this line of investigation.
Smudge

[Smudge]

Here is another paper that I found on my computer.  I would have been better if I had presented this earlier as it tells more of the story, but my fling system is so chaotic that I only just found it.
Smudge

[Smudge]

Here are some further considerations on the 14MHz anomalous behaviour.  Never did get resolved.
Smudge

[Itsu]

An impressive amount of data in all these pdf's, but i miss some practical info for replicators like the used 3F4 toroid f.i.

Is this specific 3F4 type ferrite (u = 900) needed or can we use some other material?

I can only find a "T107/65/18-3F4" here: 
https://www.acalbfi.com/se/Magnetic-components/Cores/Ferrite/p/Ferrite-toroids---Ring-core/000000013F
but its not the one shown in the PDF.

What about how to drive the setup, is a FG used and if so, how is it set up?
How are these real and imaginary impedance values measured/calculated?

So to me there are much unknowns which prevent me to start a replication, so is there some more
practical info available on how these tests where done?

Thanks,   itsu

[Smudge]

Itsu,
The measurements were done by Graham Gunderson, I think he was in Spokane USA and I am in the UK.  I did not get sight of his rig, I only got his results which were obtained on a multi channel digital scope.  I do have a note of his circuit, see below.  I used his data to calculate various things such as input impedance real and imaginary. etc.  He provided phase information presumably done by the scope.  All the data analysis was done by me.  Below you will find the 3F4 data and the toroidal core data.  Use whatever you can get hold of that comes close to this.
Smudge

[Smudge]

Here is one of my earlier papers before the MDT program got started.  I modeled a  1 turn primary and a 10 turn secondary.  That could be the basis for a simple experiment.
Smudge.

[Itsu]

Thanks Smudge,

so V2, V3 and V4 are the scopes probe outputs?

Here again the scope probe grounds tie the primary and secondary together.

Anyway, i will see what i can find.

Regards Itsu

[Itsu]

I was able to obtain a T107/65/18-3F4 toroid which is on its way.

Looking at the connection diagram as proposed by Smudge (see below diagram 1), i was trying something with
another toroid, but seem to have problems to get the FG signal delivered to the primary (L5).
See diagram 2 for my LTspice setup from real life components (signal shown is from top L2).

It looks like that the RF signal (1MHz) is being choked big time and only mV are left over.
Its probably my component choice, but as they are from real life components here are the measured value's:

L1 / L2  toroid (green from PC PS) isolation transformer (9mH each)
L3 / L4  common mode choke 37uH each
L5 / L6  test toroid (T520-2) 6 turns each (24uH).
R2 / C1 = scope probe load (10MOhm / 8pF).

Using 1Mhz as frequency as the proposed 1Khz seems somewhat low to me.

Any suggestions on what i am doing wrong here?
I tried 8uH for L1 / L2 on LTspice to get a 50 Ohm impedance match with the FG (added a 50 ohm series R) on 1Mhz but this does not improve.

Thanks,  Itsu 

[Smudge]

Hi Itsu,

Is the signal being lost through the common mode choke or balanced choke?  If so it may be how the choke is made.  I have used baluns in the past and they used a twisted pair of magnet wires wound onto a ring core.  It so happens that a twisted pair of enameled coated wire creates a transmission line of about 50 ohms impedance.  Thus when fed at one end from a 50 ohm source the signal travels along the twisted pair and is matched to the source.  If that input from the source is unbalanced (grounded), with that twisted pair wound onto the core the signal at the other end becomes balanced (off ground).  That creates a wide bandwidth balun.  I don't know whether Graham's balun was like this but I guess it must have been, and it is unfortunate that the circuit diagram shows it like a transformer.  I am not sure that the isolation transformer is absolutely necessary, but I don't see that creating much signal loss.

Smudge

[Itsu]

Smudge,

the signal is already lost at the secondary (L2) of the isolation transformer.
The green signal in my simulator was taken at top of L2 (1Vpp in on L1, 1.7mV out at L2).

In my real life setup it showed the same.

So my isolation transformer is doing that, so i need a balun like setup there like you mention.
Will fiddle around with a balun tonight while waiting for the T107 toroid.

My common mode choke L3 / L4 (your balanced choke) should not influence the signal strength or
impedance much, only suppress some transients as i undestand.

Itsu

[partzman]

Itsu,

IMO, your coupling factors are too low on both transformers.  Realistically K=.9-.95 would be more in line but even at that, the iso transformer has way too much inductance for efficient transfer at 1MHz.

Also, I don't mean to throw cold water on your sim attempt, but in order to "see" any likeness to GG's bench results, the sim models will become quite complex.  The internal models for inductors do allow the various parameters to be added and there are two methods of creating non-linearity, but again IMO, these will not allow true replication of the device.

Regards,
Pm

[Itsu]

Thanks PM,

i did play around with the coupling factors and tried both higher (0.9) and lower values, also as
mentioned above, i tried for 8uH for the both iso transformer legs (50 Ohm reactance @ 1Mhz) but this also
did not improve much.

I do realize that the sim used is very basic, but i use it to test my real life setup and they confirm
up till now what i see on my scope (very low output on the iso transformer secondary).

Hopefully the 1:1 balun will change that, i also can do without the iso transformer by using a
battery operated FG.

Itsu

[Itsu]

I was playing with the twisted magnet wire on a toroid as iso xformer as suggested by Smudge and
that seems to work out ok.

It does not matter how to connect the FG to this iso xformer (series or parallel sort off), so i
decided to do it parallel so we then have a galvanic isolation of the FG ground wire, see picture.

The screenshot shows the signals and it now shows that the problem is the common mode choke.

White is the FG signal when measured stand alone (so no other probes attached), so 1Vpp @ 1MHz
Yellow is the iso xformer secondary, so nearly the same at 1Vpp @ 1MHz
Blue is the output from the common mode choke, so here we now loose the RF.
Purple is the test toroid output and load (10MOhm @ 8pF).

Itsu

[gyula]

Hi Itsu,

I think it would be informative to terminate the output of the common choke by a 50 Ohm resistor (remove the test toroid coil) and see the waveform across the 50 Ohm first.
If there is already an issue, then the ferrite core material of the choke "produces the issue", it could be lossy at 1 MHz and together with the inductive load from the test toroid coil it 'misbehaves'. 

Another test could be to terminate the output of the test toroid coil also by a 50 Ohm in the present setup you showed in the picture and see the waveform (now the output is practically open due to the scope probe).

Gyula

[partzman]

Itsu,

Yes, one coil or side of the CMC needs to have it's connections reversed.  The purpose of the CMC is to reject any common mode signal while having the ability to pass the differential signals.  The connections shown will pass the common mode and attempt to reject the differential.

Regards,
Pm

[Itsu]

Thanks Guys,

PM, i tried to reverse one side of the CMC (bottom side in above picture), but still no signal (or similar as before) out of it.

So i tried Gyula his suggestion and put a 50 Ohm resistor at the load of the test toroid, but still no
signal there or at the output of the CMC.

So i used a 50 Ohm resistor at the output of the CMC which showed some better signal now, but still
tens of mV's only.

Using some different CMC i have one now (green PC PS CMC) which gives about 160mVpp out with 1Vpp in,
see screenshot.
Reversing one side of the CMC (bottom again), decreases the output to 137mVpp.

Guess its the CMC makeup that causes this loss.

Itsu

[Smudge]

Itsu,
Are you measuring the output from the CMC using two probes and taking the difference?  If not the ground connection to the probe could be putting a short via ground onto the CMC.  The output from the CMC should be a twin wire balanced line that cannot sustain a ground connection on one wire.
Smudge