PopularFX
Home Help Search Login Register
Welcome,Guest. Please login or register.
2019-12-12, 13:26:51
News: Forum TIP:
The SHOUT BOX deletes messages after 3 hours. It is NOT meant to have lengthy conversations in. Use the Chat feature instead.

Pages: 1 2 3 [4] 5 6
Author Topic: Magnetic Delay Transformer  (Read 3894 times)
Group: Moderator
Hero Member
*****

Posts: 1289
Well that 8.8nS is about 8 degrees at 2.5MHz.  Take that off your measured phase delay and the phase is then less than 90 degrees, not more than 90 degrees that is displayed and used in the math channel.  That means the negative power is an artifact.  So we are back to the situation that we arrived at with Graham's work.  Input resistance drops to a minimum at a frequency above resonance but doesn't go negative.  That drop in input resistance results from the time delay across the transformer, if there were no time delay it wouldn't happen.  Which is where we started on this thread and I wanted to deliberately increase the time delay across the transformer to see if we can get that input resistance to actually go negative.  I have already suggested means for doing this.
Smudge
   

Group: Elite Experimentalist
Hero Member
*****

Posts: 1866

Thanks Smudge,

had some family matters to attent, but let me try with a 100 Ohm resistor as csr instead of the current
probe, both with and without your input circuit to see of all negative wattage/resistance has gone.

Itsu
   
Group: Moderator
Hero Member
*****

Posts: 1289
The csr route should solve the probe time delay problem.  It would be helpful if you could do one simple thing, and that is feed your FG onto the primary, load the secondary with your 100 ohm non inductive R and then measure the time delay across the MDT.  This will be about 35nS and it would be good to get this verified.
Smudge
   

Group: Elite Experimentalist
Hero Member
*****

Posts: 1866
Thanks Smudge,

had some family matters to attent, but let me try with a 100 Ohm resistor as csr instead of the current
probe, both with and without your input circuit to see of all negative wattage/resistance has gone.

Itsu

I used a 100 Ohm csr in the return line and measured the circuit as proposed in below diagram.

First using the direct attached FG to the primary, then the Smudge input circuit.

Strange to me is that in the 1st case, the voltage across the 100 Ohm resistor is NOT in phase with
the voltage across primary + R when at resonance.

Going up in frequency it gets in phase.
Not sure the formula mentioned in the diagram below is valid for non "in phase" signals.

For the 2th case the signals are in phase almost all the time and at 2.7Mhz the voltage peaks.

Video here:   https://www.youtube.com/watch?v=nTCiTh56DRI


Itsu
   

Group: Elite Experimentalist
Hero Member
*****

Posts: 1866
The csr route should solve the probe time delay problem.  It would be helpful if you could do one simple thing, and that is feed your FG onto the primary, load the secondary with your 100 ohm non inductive R and then measure the time delay across the MDT.  This will be about 35nS and it would be good to get this verified.
Smudge


Using a FG (10vpp sine wave @ 2.5MHz) directly attached to the primary, and a 100 Ohm resistor only
across the secondary, i measure a 72nS time delay, see screenshot.

Yellow FG input signal,
blue signal across the 100 Ohm.

Itsu
   
Group: Moderator
Hero Member
*****

Posts: 1289
That's quite a significant time delay.  Looking at the 3F4 ferrite data I see it is quite lossy at 2.5MHz with its mu" at about one third of its mu'.  I think maybe those losses may account for why the input resistance remains above zero.  I think it would be worth using a lower frequency, say 500KHz and see what happens.  That means loading the secondary with a 6.5nF capacitor.  My input circuit values no longer apply so just connect Fg directly to primary + csr. Measure input voltage and current via csr and tune for minimum input watts on the math channel as before.  There should still be a minimum point above the resonant frequency.
Smudge
   

Group: Elite Experimentalist
Hero Member
*****

Posts: 1866
ok, i will set that up time permitting.

I also have a T520-2 toroid (iron powder) which is designed for 2-30MHz which i might try,  see:
https://www.overunityresearch.com/index.php?topic=3847.msg78156#msg78156


Itsu
   
Group: Experimentalist
Hero Member
*****

Posts: 949
Smudge and Itsu,

I have these powdered iron cores on hand so I tested one and the results are below.

It is a T225-26B core with 10 turns each for primary and secondary.  The load is a non-inductive 100 ohm film and with the frequency at 500kHz, the delay appears to be ~ 210ns.

Pm

Edit: Added single pulse scope pix.
   

Group: Elite Experimentalist
Hero Member
*****

Posts: 1866
Thanks PM,  my T520-2 (iron powder) with 2x 6 turns shows a delay at 500KHz, of 36ns,
At 2.5MHz i measure 34ns.


Smudge,
 
Measuring propagation delay across the T107 transformer using a 100 Ohm resistor shows that at 420KHz there
is a delay of 117ns, so worse then at 2.5MHz (72ns), see screenshot.


Having 3x 2.2nF in parallel gives a measured 2.35nF cap at the secondary.
F-res seems to be 420KHz then.

The current probe delay is still 9ns at this 420KHz, but represent about 1.4° phase @ 420Khz.

Using the FG directly at the primary and using the yellow probe and current probe for power measurements,
i never see any negative power value's, see video: https://www.youtube.com/watch?v=2pOoidqNaHM



Itsu
« Last Edit: 2019-10-19, 09:54:34 by Itsu »
   
Group: Moderator
Hero Member
*****

Posts: 1289
Hi Itsu,

I wasn't expecting negative power, but looking for minimum input power.  Since there is no actual resistive load your input power hovers around a few milliwatts.  I did note that at the highest frequency you went to (630KHz or thereabout) the secondary voltage is almost at 180 degrees to the primary voltage.  As voltage is directly related to flux and since you are using sine waves we can reasonably assume that the flux in the secondary is at 180 degrees to that in the primary.  That is most interesting and it means that it is equivalent to the two coils being in bucking mode.  There must then be flux driven outside the core, see FEMM image below.  You can check this by placing a coil in the centre of the ring core to sense the flux there.  With the secondary flux opposing the primary flux, it makes me wonder how the transformer will then respond with a load across that secondary, keeping the capacitor there of course.  If the phase delay for the signal return from that load is also 180 degrees, does the primary see that load as a positive one.  And if so, if the frequency were chosen slightly lower where the phase delay across the transformer is 90 degrees (i.e. the two way delay is 180 degrees), will a secondary load then reflect back as a negative one?

If there is flux outside the ring core as shown below, can a coil placed there be driven so as to reduce that flux?  What effect would that have on the transformer efficiency?  You can see there is much to learn about this MDT.

Smudge
 
   
Group: Moderator
Hero Member
*****

Posts: 1289
Hi Itsu,
That talk of bucking coils reminds me of a paper I wrote a couple of years ago, and here it is again.  As you now have a large toroidal core that has a significant phase delay across it perhaps you could quickly connect your two coils in series opposing and drive them directly from your FG, measuring voltage and current to see whether you achieve this magical negative input resistance, as suggested in figure 5 of that report.
Smudge.
   

Group: Elite Experimentalist
Hero Member
*****

Posts: 1866
Hi Itsu,

I wasn't expecting negative power, but looking for minimum input power.  Since there is no actual resistive load your input power hovers around a few milliwatts.  I did note that at the highest frequency you went to (630KHz or thereabout) the secondary voltage is almost at 180 degrees to the primary voltage.  As voltage is directly related to flux and since you are using sine waves we can reasonably assume that the flux in the secondary is at 180 degrees to that in the primary.  That is most interesting and it means that it is equivalent to the two coils being in bucking mode.  There must then be flux driven outside the core, see FEMM image below.  You can check this by placing a coil in the centre of the ring core to sense the flux there.  With the secondary flux opposing the primary flux, it makes me wonder how the transformer will then respond with a load across that secondary, keeping the capacitor there of course.  If the phase delay for the signal return from that load is also 180 degrees, does the primary see that load as a positive one.  And if so, if the frequency were chosen slightly lower where the phase delay across the transformer is 90 degrees (i.e. the two way delay is 180 degrees), will a secondary load then reflect back as a negative one?

If there is flux outside the ring core as shown below, can a coil placed there be driven so as to reduce that flux?  What effect would that have on the transformer efficiency?  You can see there is much to learn about this MDT.

Smudge


Due to some problems here, my time is still limited.
 

Quote
I wasn't expecting negative power, but looking for minimum input power.

Concerning this, here an input / output sweep from 1KHz to 1MHz for both input power (white) and output power (red).

The signals are NOT at scale, so look for the numbers to compare.
Not sure how to present those figures (mean or pp or rms), but it shows the sweep shape and that we have more input then output all the way (measuring inbetween the vertical lines).

I was using the current probe for current measurements as i guess the propagation delay concerns both measurements equal.

Output was measured across/through a 1M resistor parallel to the output LC.

Itsu
   

Group: Elite Experimentalist
Hero Member
*****

Posts: 1866
Hi Itsu,

I wasn't expecting negative power, but looking for minimum input power.  Since there is no actual resistive load your input power hovers around a few milliwatts.  I did note that at the highest frequency you went to (630KHz or thereabout) the secondary voltage is almost at 180 degrees to the primary voltage.  As voltage is directly related to flux and since you are using sine waves we can reasonably assume that the flux in the secondary is at 180 degrees to that in the primary.  That is most interesting and it means that it is equivalent to the two coils being in bucking mode.  There must then be flux driven outside the core, see FEMM image below.  You can check this by placing a coil in the centre of the ring core to sense the flux there.  With the secondary flux opposing the primary flux, it makes me wonder how the transformer will then respond with a load across that secondary, keeping the capacitor there of course.  If the phase delay for the signal return from that load is also 180 degrees, does the primary see that load as a positive one.  And if so, if the frequency were chosen slightly lower where the phase delay across the transformer is 90 degrees (i.e. the two way delay is 180 degrees), will a secondary load then reflect back as a negative one?

If there is flux outside the ring core as shown below, can a coil placed there be driven so as to reduce that flux?  What effect would that have on the transformer efficiency?  You can see there is much to learn about this MDT.

Smudge


Concerning the second part of your post, i made the below video showing what happens when inserting
a pickup coil in the middle of the toroid.

There is some flux being picked up, but it is fairly stable across a broad frequency range so does not
appear suddenly when input and output are 180° off.

https://www.youtube.com/watch?v=lZtPKcVc4pc

Itsu

   
Group: Moderator
Hero Member
*****

Posts: 1289
Hi Itsu,

Thanks for doing that.  That flux across the centre of the core is to be expected when you have primary and secondary on opposite sides of the core and the secondary is loaded so that it supplies current.  Under normal conditions (resistive load) the secondary current is opposite to the primary load current.  The mmf's are equal and opposite, thus acting like bucking coils to the load current component (not the magnetizing current component in the primary).  If primary and secondary are wound over each other the load currents do not create any flux at all, only the magnetizing current creates flux which remains inside the core.  But with the coil separation across the core the load currents drive that leakage flux outside the core.  In our case we have the phase due to propagation delay across the core and the phase caused by the capacitive loading, so I need to examine your video carefully to see what conclusions I can draw.

Smudge
   

Group: Elite Experimentalist
Hero Member
*****

Posts: 1866
Hi Itsu,
That talk of bucking coils reminds me of a paper I wrote a couple of years ago, and here it is again.  As you now have a large toroidal core that has a significant phase delay across it perhaps you could quickly connect your two coils in series opposing and drive them directly from your FG, measuring voltage and current to see whether you achieve this magical negative input resistance, as suggested in figure 5 of that report.
Smudge.

Smudge,

using the Fig. 5 in the pdf and measuring voltage and current as shown in the diagram below shows no negative input power across the 100KHz to 5MHz range.
The input power is fairly stable around 2.5mW, see screenshot taken at 2MHz.

i was using my battery operated FG here.

Itsu


« Last Edit: 2019-10-25, 09:27:59 by Itsu »
   
Group: Moderator
Hero Member
*****

Posts: 1289
Smudge,

using the Fig. 5 in the pdf and measuring voltage and current as shown in the diagram below shows no negative input power across the 100KHz to 5MHz range.
The input power is fairly stable around 2.5mW, see screenshot taken at 2MHz.

i was using my battery operated FG here.

Itsu
Hi Itsu,
Thanks for doing that.  I was interested in the variable peaks in the math waveform and I noticed that the current waveform had a DC offset.  Wasn't sure whether this was just a scope setting.  I downloaded your screenshot, added my own cursors to take reading of the peak math power to see whether the DC offset was the cause of the variable peaks.  I took two peaks, one positive and one negative and for each took the displayed voltage and current values to show that these were indeed used by the math channel, see first image below.

Next I measured the phase between voltage and current, not using the zero crossings but using the DC offset crossings to see whether this made any difference to the phase, see second image.  The phase agreed with the scope's measurement.  So although it appears your FG puts out a sine wave with a small DC offset, this doesn't materially affect the scope readings.

If you are willing to proceed I think the next step is to introduce some deliberate phase delay along the magnetic core to see whether this can be beneficial.  A series of separate windings of a few turns each along the top and bottom of the core, with each coil shunted by a low value capacitor, will make a lumped constant magnetic delay line that will increase the time delay from primary to secondary.  Increasing those shunt capacitor values will increase the time delay.  Perhaps you could try this and see what time delays you get, initially using a resistive load on the secondary.

Smudge 
   

Group: Elite Experimentalist
Hero Member
*****

Posts: 1866

Smudge,

i measured my used FG to have no offset, so the shown blue current offset is somehow caused by the
setup used, see my circuit diagram in post #89.


Concerning:

Quote
A series of separate windings of a few turns each along the top and bottom of the core, with each coil
shunted by a low value capacitor, will make a lumped constant magnetic delay line that will increase
the time delay from primary to secondary,

i am not sure i understand what you mean.

Should i keep the 6 turns primary and secondary intact and add these "A series of separate windings
of a few turns" next to them?
These are stand alone coils right?


Perhaps a simple drawing would explain, thanks.

Itsu
   

Group: Professor
Hero Member
*****

Posts: 1479
A series of separate windings of a few turns each along the top and bottom of the core, with each coil shunted by a low value capacitor, ...

Perhaps a simple drawing would explain, thanks.
« Last Edit: 2019-11-01, 07:49:02 by verpies »
   

Group: Elite Experimentalist
Hero Member
*****

Posts: 1866

Thanks vepies,

so these extra coils/caps turn the ferrite inbetween the prim/sec into a magnetic delay line which
should increase the time delay between prim/sec (without them now 107ns @ 420KHz and 72ns @ 2.5MHz).

Itsu
   
Group: Moderator
Hero Member
*****

Posts: 1289
Hi Itsu

Been busy with family so couldn't reply earlier.  It is important to get the winding directions right for the version that Verpies posted, the first image below shows this.  An alternative version is shown in the second image.

As regards that 107nS @ 420KHz and 72nS @ 2.5MHz, I think that variation is down to measurement technique, IMO there really shouldn't be any variation with frequency.  I think it best to input a continuous sine wave and measure the delay as phase (if the math does that automatically) or as time between zero crossings.  Of course that is with a resistive load, the addition of a load capacitor will affect the measured phase.

That DC offset on the input current, if that is truly being caused by the item under test then that could be an important finding, it must not be dismissed as a curiosity.   I say this because electron spins and orbits that create magnetic fields, can be considered as perpetual motion quantum dynamos, they can both deliver power and absorb power.  They are miniature atomic perpetual-current loops.  Under normal cyclic conditions the flow of energy back and forth from or to the quantum domain averages to zero.  It is thus hidden from view and is not taught in mainstream physics.  However if it is possible to induce a DC voltage into those atomic current loops, they will deliver power continuously.  If you look into my Marinov Generator work you will see that I believe it is possible to create such an effect that explains where the energy comes from.  Maybe this MDT is also a means for doing this.

Smudge
   

Group: Elite Experimentalist
Hero Member
*****

Posts: 1866

Thanks Smudge,

i will use the setup as mentioned by verpies, so all the extra coils sides the same direction, but connected
to their opposite side with the caps.
I will use 22pF smd caps.


Concerning the "107ns @ 420KHz and 72ns @ 2.5MHz" time delays, it seems that the phase between input
and output is changing by frequency and thus the time delay.

It follows the below table (see screenshot 1 and 2 for min. and max.):

100KHz: 118ns / 4°   (so input and output almost in phase).
500KHz: 115ns / 21°  (our 117ns @ 420KHz)
1MHz:    105ns / 37°
2MHz:    81ns  / 58°
2.5Mhz: 72ns  / 65°  (our 72ns @ 2.5MHz).


     
Concerning the DC offset on the current via the csr, its still there when using a 10 Ohm csr instead
of 1 Ohm when using by battery operated FG.

But when using my Rigol FG (and swapping the scope groundleads to the FG black lead side, it is almost
completely gone, so i think its caused by this cheap FG, see screenshot 3.

Itsu
   

Group: Professor
Hero Member
*****

Posts: 1479
An alternative version is shown in the second image.
Please elaborate on the differences between the two versions of the magnetic delay line, that you've posted.
   
Group: Moderator
Hero Member
*****

Posts: 1289
Please elaborate on the differences between the two versions of the magnetic delay line, that you've posted.
I think they would both behave in the same manner.  However the first version requires the coil placings to be geometrically identical top to bottom.  If not, the magnetic delay along the core is different for each top coil relative to its partner at the bottom and this will affect the performance.  In the second version there is no connection between top and bottom coils so this problem does not arise.

Smudge 
   
Group: Moderator
Hero Member
*****

Posts: 1289
Concerning the "107ns @ 420KHz and 72ns @ 2.5MHz" time delays, it seems that the phase between input
and output is changing by frequency and thus the time delay.

It follows the below table (see screenshot 1 and 2 for min. and max.):

100KHz: 118ns / 4°   (so input and output almost in phase).
500KHz: 115ns / 21°  (our 117ns @ 420KHz)
1MHz:    105ns / 37°
2MHz:    81ns  / 58°
2.5Mhz: 72ns  / 65°  (our 72ns @ 2.5MHz).

Hmm, I need to think more about that.  Maybe the L/R time constant at the secondary is having an effect (50uH and 100 ohms).
Quote
Concerning the DC offset on the current via the csr, its still there when using a 10 Ohm csr instead
of 1 Ohm when using by battery operated FG.

But when using my Rigol FG (and swapping the scope groundleads to the FG black lead side, it is almost
completely gone, so i think its caused by this cheap FG, see screenshot 3.

OK, we can ignore that strange effect.  Thanks for that check

Smudge
   

Group: Elite Experimentalist
Hero Member
*****

Posts: 1866

Still working on the magnetic delay line on my toroid so now and then.
I plan to have 40 turns on each half with a capacitor (22pF) on every 2th turn.

Getting it nice and tidy is the problem.

Itsu
   
Pages: 1 2 3 [4] 5 6
« previous next »


 

Home Help Search Login Register
Theme © PopularFX | Based on PFX Ideas! | Scripts from iScript4u 2019-12-12, 13:26:51