Dear All.
It has been suggested that a new thread be started with the sole purpose of investigating the properties of Ferrite Cores in relation to the present round of experiments that various members are participating in.
Areas of investigation should be....
Composition.
Permeability.
Ferro resonance frequency.
I am sure that there many more to add to this list ?? So don't be shy, join in !! O0
Cheers Grum.
Is ferroresonance different from one of these?:
1) LCR resonance (what C ?)
2) Mechanical resonance between two core halves (a two masses and a spring system)
3) Electron Spin Resonance
4) Nuclear Magnetic or Nuclear Acoustic Resonance
5) Standing EM wave resonance inside the ferritee (related to its permeability and permittivity)
6) Resonance of acoustic standing waves (a.k.a. "dimensional resonance")
a) transverse
b) longitudinal
c) torsional
It would be prudent to have the subject of this thread deconflated (http://en.wikipedia.org/wiki/Conflation) before it starts.
Quote from: verpies on 2014.05.05, 16:53:15
Is ferroresonance different from one of these?:
1) LCR resonance (what C ?)
2) Electron Spin Resonance
3) Nuclear Magnetic or Nuclear Acoustic Resonance
4) Standing EM wave resonance inside the ferritee (related to its permeability and permittivity)
5) Resonance of acoustic standing waves (a.k.a. "dimensional resonance")
a) transverse
b) longitudinal
c) torsional
It would be prudent to have the subject of this thread deconflated (http://en.wikipedia.org/wiki/Conflation) before it starts.
Dear Verpies.
Thank you for getting the "Ball rolling" .
So. We need a "Bench mark", a basis from which our tests can be conducted. "The same hymn sheet" I shall wait for the more erudite members to contribute !! :)
Cheers Grum.
In the attached paper the author suggests winding at right angles and pulsing one of the windings with pulses around 1uSec,
He does not say which of the categories of resonance is excited.
I recommend experimenters do not confuse this with normal LC resonance.
verpies has suggested that a spring material be used between the core halves.
This would convert the inductor into a parametric device, activated at the acoustic or ultrasonic level or slightly beyond.
Maybe he had something different in mind and can explain it to us.
I welcome this initiative from Grum. It will hopefully divert us somewhat from our pre-occupation in dribbling over Akula's very unconvincing attempts to demonstrate self running devices.
Quote from: ION on 2014.05.05, 17:17:44
Verpies has suggested that a spring material be used between the core halves.
This would convert the inductor into a parametric device, activated at the acoustic or ultrasonic level or slightly beyond.
Maybe he had something different in mind and can explain it to us.
When core halves become attracted, they gain kinetic energy. This energy is lost if absorbed by a dampening spacer.
To return the core to its original position one of the three things below has to happen:
A) The halves collide and bounce back
B) A second asymmetrical winding periodically creates repulsion between the core halves.
C) A springy material pushes them back apart
I suggested a paramagnetic spring material between core halves in case the core is configured to experience resonance described in
pt.2 (http://www.overunityresearch.com/index.php?topic=2430.msg38324#msg38324) by the method from pt.C.
Dear All.
I have attached a picture of the Single pulse generator that ION posted in a PDF earlier in this thread.
Both ION and I are pretty sure that using this device into a single turn primary coil wrapped around the Ferrite under test. And a scope pick up coil also single turn and at right angles to the primary we should be able to see the ringdown. This signal would be un cluttered !!
I have tried a number of methods both acoustic and electromagnetic but have found in all cases that the received signal was very difficult to decipher !!
Thoughts ??
Cheers Grum.
Quote from: Grumage on 2014.05.09, 20:53:06
Dear All.
I have attached a picture of the Single pulse generator that ION posted in a PDF earlier in this thread.
Cheers Grum.
Good evening Grum.
Is a schematic available for the Single pulse generator shown in the attachment?
thanks
take care, peace
lost_bro
Concerning a pulser, in the days when i was replicating the Dally circuit/nano pulser, i was pointed by verpies to
an application of a LT1073 DC2DC converter chip named: "1.5V Powered 350ps Risetime Pulse Generator"
It can be found on page 14 of this PDF: http://cds.linear.com/docs/en/datasheet/1073fa.pdf
I did build it and in the crude version it was still able to create 3ns pulses @5V see this video:
https://www.youtube.com/watch?v=s-Alj7gm0RA
Lateron i used it to do some coax time domain reflection measurements with it which where successfull.
I will try to use this on my ferrites to see if it will show any ringing.
Regards Itsu
Ok, some tests with the pico-pulser on my E-core ferrite.
I have some solid ringing on 22MHz (45ns), not sure if its the 5 turns coil(s) or the ferrite, but my guess its the coils.
Sometimes however it shows a frequency around 1.1MHz which in my feeling would more match the ferrite resonance frequency.
Playing with the FFT function is not really helpfull, as the massive 22MHz signal obscures the rest.
Picture of the used setup, and a screenshot of the scope, yellow = output pulser, blue is ringing pickup.
Video here: https://www.youtube.com/watch?v=S2pBnRTGdIw&feature=youtu.be
Regards Itsu
I removed some windings on the both puls and pickup coils (went from 5 to 3 turns each) and as expected, the frequency response went up from 22MHz to 30MHz, see screenshot 1.
So the ringing is from the coils, not the ferrite.
I do however see, like mentioned before, still these occasional 1MHz responses and was able to capture one, see screenshot 2.
Here you see the frequency response of the scope again mentioning 1.043MHz which my feeling is the ferrite resonance frequency.
Need a way to verify this, any suggestions?
Regards Itsu
Hi Itsu,
I assume it is a mechanical resonance (=magneto-elastic resonance) for that ferrite E core around 1MHz. If it is, than perhaps suspending the E core on a piece of thread may increase the amplitude vs the amplitude you have found so far with laying the core on the table surface (damping may be less when the core is suspended). Instead of suspending, perhaps laying the core onto a soft material like a sponge may also reduce dampening the mechanical Q of the core.
Also, if you position the other E core to face the one with the coil i.e. close the open E core with the other E core, then perhaps the 1MHz response also changes (use also a soft base for the two). I know you cannot fully close the two E cores because of the output becomes positioned in the gap.
(Sorry that I am suggesting things I have not tested personally but this is my take on this.)
You may have noticed I uploaded a Philips data sheet on the ou forum for MnZn ferrite material 3R1 and it included a formula for calculating the mechanical resonant frequency of ring cores made from that material, and the formula includes OD and ID sizes only, besides a 5700 number in the nominator.
I attached the file here too, see pdf file page #8 for the remarks. I mention this because if you happen to have a toroidal core from MnZn material then you could also test it at the frequency the formula suggests. Maybe the formula works for other ring cores you have but with less precision. Seeing some output pulses for different toroidal cores in your setup may help identify the E core responses too.
Gyula
Quote from: Itsu on 2014.05.10, 09:34:20
Concerning a pulser, in the days when i was replicating the Dally circuit/nano pulser, i was pointed by verpies to
an application of a LT1073 DC2DC converter chip named: "1.5V Powered 350ps Risetime Pulse Generator"
It can be found on page 14 of this PDF: http://cds.linear.com/docs/en/datasheet/1073fa.pdf
Regards Itsu
Thanks Itsu; I've already ordered the LT1073 chip, will get to building it as soon as it arrives.....
take care, peace
lost_bro
Quote from: Itsu on 2014.05.10, 12:17:11
I removed some windings on the both puls and pickup coils (went from 5 to 3 turns each) and as expected, the frequency response went up from 22MHz to 30MHz, see screenshot 1.
So the ringing is from the coils, not the ferrite.
I do however see, like mentioned before, still these occasional 1MHz responses and was able to capture one, see screenshot 2.
Here you see the frequency response of the scope again mentioning 1.043MHz which my feeling is the ferrite resonance frequency.
Need a way to verify this, any suggestions?
Regards Itsu
Try make scope not 20 ns one division, but 50-100, then maybe you see after pulse lower frenquency.
Quote from: gyula on 2014.05.10, 14:14:20
Hi Itsu,
I assume it is a mechanical resonance (=magneto-elastic resonance) for that ferrite E core around 1MHz. If it is, than perhaps suspending the E core on a piece of thread may increase the amplitude vs the amplitude you have found so far with laying the core on the table surface (damping may be less when the core is suspended). Instead of suspending, perhaps laying the core onto a soft material like a sponge may also reduce dampening the mechanical Q of the core.
Also, if you position the other E core to face the one with the coil i.e. close the open E core with the other E core, then perhaps the 1MHz response also changes (use also a soft base for the two). I know you cannot fully close the two E cores because of the output becomes positioned in the gap.
(Sorry that I am suggesting things I have not tested personally but this is my take on this.)
You may have noticed I uploaded a Philips data sheet on the ou forum for MnZn ferrite material 3R1 and it included a formula for calculating the mechanical resonant frequency of ring cores made from that material, and the formula includes OD and ID sizes only, besides a 5700 number in the nominator.
I attached the file here too, see pdf file page #8 for the remarks. I mention this because if you happen to have a toroidal core from MnZn material then you could also test it at the frequency the formula suggests. Maybe the formula works for other ring cores you have but with less precision. Seeing some output pulses for different toroidal cores in your setup may help identify the E core responses too.
Gyula
Thanks Gyula,
so this specific 3R1 cores resonance point will be around 98KHz when using that formula.
This is lower then i would expect, so good to know.
Regards Itsu
Quote from: MenofFather on 2014.05.12, 08:02:41
Try make scope not 20 ns one division, but 50-100, then maybe you see after pulse lower frenquency.
Thanks MenoGather,
normally i do check the scope at various settings but i will double check for lower frequencies.
Regards Itsu
Gyula
QuoteYou may have noticed I uploaded a Philips data sheet on the ou forum for MnZn ferrite material 3R1 and it included a formula for calculating the mechanical resonant frequency of ring cores made from that material, and the formula includes OD and ID sizes only, besides a 5700 number in the nominator.
That is fantastic and important information.
Thankyou for finding this O0
Quote from: gyula on 2014.05.10, 14:14:20
You may have noticed I uploaded a Philips data sheet on the ou forum for MnZn ferrite material 3R1 and it included a formula for calculating the mechanical resonant frequency of ring cores made from that material, and the formula includes OD and ID sizes only, besides a 5700 number in the nominator.
...and that 5700 number is the speed of sound in the ferrite - 5700m/s
Quote from: verpies on 2014.05.12, 21:26:54
...and that 5700 number is the speed of sound in the ferrite - 5700m/s
So are we talking "shock wave" here? Maybe resonating the speed of propagation just above and just below the Mach speed in such a way as to flip/flop the magnetic dipoles?
If that happens to be true, maybe we need to investigate frequency modulation (http://www.youtube.com/watch?v=QnnlZ0yR6AE)...?
McFreey used the same formula to calculate resonance of his copper fuel ring, but instead of using the average of the internal and external diameter he just used the diameter as it was a thin copper wire ring, also the speed of sound in copper was lower
http://www.overunityresearch.com/index.php?topic=1526.msg33149#msg33149
Quotefr = VL/(pi*d )
A resonant LC is run at fr/2 to bring the ring into acoustic resonance at fr, the frequency needs to be pretty exact.
The frequency is finely adjusted using a microphone which listens to the ring, the ring calculated frequency may not be precise and therefore a sweep of the ring noting any resonant sines picked up with the microphone.
For copper the speed is between 4000 - 5000m/s
but there are problems with the McFreey device, the evil skin effect stops the electrical LC sinewave from permeating the copper because copper has a small skin depth even at low frequencies.
MnZn does not suffer this evil skin effect anywhere nearly as bad and this makes MnZn ferrite a fantastic fuel for NAR.
Ok, it became deafening silent overhere on this important subject, but fortunately our Russian friends still are active and via a PM i got some info on how to continue.
Not sure if he appreciate to be named, so i will call him Mr S and only forward info and links.
I got the following link to a PDF describing in Russian how to measure the resonance of ferrite, that link is here:
http://realstrannik.ru/media/kunena/attachments/6159/CE_ferrit.pdf
I have let google translate this and put it into a Word document which is attached and i will refer to this Word document in the video
This Word document info together with the following info should lead to a method on how to determine the resonance frequency of a ferrite core.
Received extra info via PM:
For exaple you want to find ferro-resonace of core from your HV flyback transformer.
1. Remove the core from Flyback
2. It has quadratic form. Now wind 2 test coils X and Y coil with smae number of turns and same wire diammeter. For example 10 turns in X axis than continue winding 10 turns in a Y-axis. As you will see it is in fact one coil electricaly but not magneticaly. 1/2 coil is positioning in X-axis and 1/2 in Y-axis.
3. Use generator from 50 Hz- up to 1..2 MHz for square wave impulses. You can try your generator on TL494.
4. Main goal here is time for opening MOSFET must be shortest as possible. Also time for closing them must be shortest as possible. You need 10ns to open output transistor and not higher 50ns for closing him. Both parameters for ezxample on 200kHz or upper (it's better).
5. Put 75 oms (non-inductive) resistor at output of MOSFET and measure turn-off time on your 3054C. For example MOSFET has gave 50ns turn-off time which is what you need.
6. Remove resistor and conect coil to MOSFET between VPlus supply rail and Drain of your MOSFET.
7. Wind 1/2 winding on your core for connecting your 3054C. Winding should be placed farther away from your excitation coil. So put it on the opposite corner of your quadratic core. Opposite side is maximum longer from your excitation coil.
8. One channel probe on Gate of MOSFET - (if you have signal ringing from TL494 generator put a several small ferrite beads in your gate terminal)/
9. Second channel signal from 1/2 winding.
10. Sweep frequency and stretch the time base to get nice big oscillogram. For example you get 400ns from excitation pulse-width 10ns....200ns. Now you have got impulse resonse of your ferrite core. As you carefully look at oscillogram you will see how ferrite is establishing self-oscillations on exactly one frequency but strictly at the moment after turn-off your MOSFET. Read pulse width response and check it on spectral analyser. You should get nice big peak eactly on Resonant frequency of your ferrite core. As shortest excitation impulse is better. If you have prfessional signal generator try excitation from 1nS up to 200 ns. Just turn on peak detector and sweep generator. Very simple if you have equipment.
I did follow the instructions in the Word document together with the above info, but the result is not as mentioned in the Word document.
The pickup signal is a nice copy of the input frequency, so a nice square wave.
See the video here: https://www.youtube.com/watch?v=muC-UlyDBcw&feature=youtu.be
Not sure what went wrong as i also sweeped a large frequency range (so not only at 173KHz) and varied the duty cycle between 10 and 90%.
Any comments are appreciated, regards Itsu
@Itsu
Seems to me the first procedure of using one coil then a pickup coil wound over the first coil is not right for finding ferrite resonance, since the pickup will be corrupted by the main coil impulses being imparted directly from wire to wire.
The second Russian procedure you just posted is a good start but again not complete.
Seems to me working with a single pick up is not the right method since the pickup coil will show two potential resonance types on the scope depending on how high your FG will go, you will see many resonance points, so how do you know if the resonance is engendered by the ferrite resonance or the pick up coils inherent resonance points.
Seems to me the only way to really find ferrite resonance by using a secondary medium like a pick up coil is to use several pick up coils all having a different number of turns located anywhere on the ferrite as long as they are not one over the other.
So you put one drive coil of let's say 5 turns, then you add let's say three pickup coils of 4, 7, 11 turns (the number is not important as long as each are different).
Then you drive the drive coil with the FG at 10% duty sweeping the frequency and scoping the first 4 turns pick up coil. Each time you see a rise in the waveform you jot down that frequency, then repeat the same procedure with the other two pick up coils.
Since each have different turns, their own internal resonance point should not be the same. By going over your three lists of frequencies, the frequency that is the same in all three will be the ferrite resonance since this is not dependant on any turn ratios. So ferrite resonance, since you cannot connect your scope to the ferrite itself, can only be done with a secondary pick up coil that itself may be misleading. So you need several pick ups to then arrive at a common resonance point.
wattsup
Quote from: wattsup on 2014.05.26, 11:39:17
@Itsu
Seems to me the first procedure of using one coil then a pickup coil wound over the first coil is not right for finding ferrite resonance, since the pickup will be corrupted by the main coil impulses being imparted directly from wire to wire.
I agree, it shows nicely the resonancepoints of the used coils, not of the ferrite, as also confirmed by Grumage.
Quote
The second Russian procedure you just posted is a good start but again not complete.
Seems to me working with a single pick up is not the right method since the pickup coil will show two potential resonance types on the scope depending on how high your FG will go, you will see many resonance points, so how do you know if the resonance is engendered by the ferrite resonance or the pick up coils inherent resonance points.
Seems to me the only way to really find ferrite resonance by using a secondary medium like a pick up coil is to use several pick up coils all having a different number of turns located anywhere on the ferrite as long as they are not one over the other.
So you put one drive coil of let's say 5 turns, then you add let's say three pickup coils of 4, 7, 11 turns (the number is not important as long as each are different).
Then you drive the drive coil with the FG at 10% duty sweeping the frequency and scoping the first 4 turns pick up coil. Each time you see a rise in the waveform you jot down that frequency, then repeat the same procedure with the other two pick up coils.
Since each have different turns, their own internal resonance point should not be the same. By going over your three lists of frequencies, the frequency that is the same in all three will be the ferrite resonance since this is not dependant on any turn ratios. So ferrite resonance, since you cannot connect your scope to the ferrite itself, can only be done with a secondary pick up coil that itself may be misleading. So you need several pick ups to then arrive at a common resonance point.
wattsup
Ok, i am trying to follow that second procedure first with comments coming in which i will post here.
If it still not works i will continue to follow up with your approach, thanks.
Regards Itsu
I got the following feedback on the above video (my answers between brackets):
The pickup coil needs to be at a right angle to the ferrite
(Which indeed was not 100 correct, so i changed this on 2 separate U-cores, but no improvement was seen)
Try to increase pulse width (average pulse power fro excitatio)!! Start from 1microsek and lower down to 500ns.
Amplitude of your Rigol DG4102 12V or higher 18-20V.
If this doesn't help try to wind bucking coil configuartion.
(pulse width was increased (duty cycle) from 1% to 99%, FG was driven at max (20V pp) amplitude, on a new bigger U-core,
the excitation coil was wound in bucking configuration, see also next comment).
Responded pulse width of ferrite shoud not depend any how from excitation pulse width not from number of truns.
Try simetrical configuartion on X axis 5 turns and Y axis also 5 truns (same wire diameter). Also try X axis 5 turns left winding - y Axis 5 turns right winding. Conect serially.
For oscillograph loop coil (output) try 2.5-5 turns on other side of core.
And don't put nothing in air gap. Both halves need to be fasten very hardly.
(I used a new bigger U-core (see picture) and used bucking mode (cc/ccw) symmetrical (2x 13 turns) configuration. U-cores have/had no air gap and are clamped together).
ps: One thinh more how to know you are finding Resonant frequency of ferrite. Both halves will be attracted so hard just like you using
neodimium magnets when you hit Main resonant peak. This is also confirmed by Avalon and some more people.
(never did i found a spot where there was ANY attraction (with core half's loosely together) noticeable)
Aleksandr and Vitaly were measured resonance on Russian Ferrite-Cores !!
Does the methnod work wit american ferrites i don't know !! It;s a different technolgy !!
(i have some Russian toroid cores which i could try, not sure yet how to setup the excitation and pickup coil on a toroid)
So i tested another evening on these U-cores, but never did i found a conclusive resonance point fixed to the ferrite.
It could be indeed that the Russian ferrite behaves differently in which case we in Europe/America have a problem.
I will try to test these Russian toroid's i have to see if any resonance can be found there.
Regards Itsu
As promised i tested one of my in Russia bought ferrite cores (no idea of this is really made from Russian ferrite).
I used the same setup as used on the earlier U-cores, meaning a coil of magnet wire consisting of 2 coils in series wound 90° apart of 15 turns each.
I used them in cw/cw and later on (video) in cw/ccw (bucking mode) setup.
Input signal via a bipolar cap (10uF) from by FG in pulse mode, 10V pp all positive pulse 50% duty cycle at about 173KHz leading edge of
the pulse set to 200ns, trailing edge to 10ns.
But i also swept across a broad range of frequencies, duty cycles and leading/trailing edge settings.
Pickup coil was half a turn squarely to the ferrite core.
The pickup waveform is a close copy of the input waveform, and never did i see that the leading edge of the pulse got missing like
showed in the ferrite Word document attached in one of my earlier posts.
Video here: https://www.youtube.com/watch?v=x9sdscgys7E&feature=youtu.be
So still no successful way to determine the used ferrite resonance frequency.
Regards Itsu
@itsu
The ferrite core seems to me that the russian pdf was then tested in a different way with a coil in the inner hole and another hole in the exterior.
Quote from: TutorialFE on 2014.05.28, 11:59:56
@itsu
The ferrite core seems to me that the russian pdf was then tested in a different way with a coil in the inner hole and another hole in the exterior.
Yes, you mean like in the picture below.
Well, its kind of hard to implement that way, but i could give it a try.
Regards Itsu
Perhaps a different approach is needed. How about applying electrodes directly to the ferrite material, 4 in all, 2 for excitation, 2 for pickup. These could be in the same plane or orthogonal but some distance apart.
When I get time I will do this experiment by using a silver epoxy or paint to fix the copper foil contacts to the ferrite.
Alternately a combination of electrodes for excitation and coil for pickup, or use the coil to excite and electrodes for pickup.
With the electrodes, measure the resistance and try to get a good impedance match from the driver.
Regards, ION
Thanks ION,
good idea, my ferrites Ohmic range differ enormous per ferrite so it should be tailored for each specific piece of ferrite i guess.
Regards Itsu
As suggested by ION, i tried to characterize one of my pot cores by inputting a sweeping frequency range (1KHz - 100MHz) directly into/onto the ferrite and measuring the response at the other side with my scope.
The pot core half's are separated by a piece of mica in the center.
It looks as if this specific pot core has a resonance peak around 42MHz which is kind of high i think.
Any comments are welcome if this could be a valid way to characterize a ferrite core.
Screenshot 1 shows the pot core frequency range starting at 1 KHz and ending at 100MHz with a strong peak around 42MHz.
I later on tried the same with a large U-core separated by 2 pieces of electrical tape:
Screenshot 2 shows a U-core with frequency range starting at 10Hz and ending at 100MHz with a peak around 25MHz.
Video here: https://www.youtube.com/watch?v=XVWFd6w0xI0&feature=youtu.be
Regards Itsu
Users Void, D3x0r and me are trying to replicate the shown ferromagnetic resonance of a flyback core demonstrated in a video by Akula.
The forum used is over on overunity.com in this thread:
http://www.overunity.com/14687/akula-eternal-lantern-4/msg412955/#new
Page 1 of this thread shows the complete diagram used, but we are concentrating on replicating the ferromagnetic resonance shown by using a very small part of this circuit as can be seen (and attached below) here:
http://www.overunity.com/14687/akula-eternal-lantern-4/msg412459/#msg412459
In this post:
http://www.overunity.com/14687/akula-eternal-lantern-4/msg412458/#msg412458
there are linked severall videos with translated text from Russia to English which gives good info what is trying to be done.
Up till now i have build this small circuit and found a pulse which seems to indicate this ferromagneticresonance of the core.
But making use of it like demonstrated by Akula is failing at the moment.
Screenshot of the pulse below, video here: https://www.youtube.com/watch?v=XJMx_BsUnms&feature=youtu.be
Screenshot data:
blue = collector transistor (MJE13005) signal
yellow = base transistor signal
purple = FG input signal (250Hz DC pulse, 2V amplitude, 4% duty cycle)
green = current probe signal from plus lead (3V dc), current probe controller set to 200mA/div.
What i find strange is the peak in the middle of the blue trace, so middle in the active transistor window.
When pushing on the core (gap by 2 pieces of electrical tape) the peak moves to the right.
Any idea what the cause could be of this peak?
Regards itsu
Itsu:
Sorry I missed this experiment in post #29 when you first posted it. I just now saw it.
Very nice experiment, however if I may, I would like to offer some suggestions for a more refined test.
Instead of a pot core, use a rod core or piece of 2 inch ferrite cut from a flyback transformer with square cross section.
The leads should be attached with conductive epoxy, (preferably silver) at two distant points on the rod on opposing faces.
The assembly should be put into a small aluminum box with BNC connectors at either end and 50 Ohm termination resistors.
Do not use probes, use BNC cables to connect to the generator and scope.
The test can also be repeated with a bar magnet instead of a ferrite rectangular rod.
I'll work up a sketch for you. Termination and shielding is important at these frequencies, and probe ground clips can pick up stray signals or be resonant loops at certain frequencies, so should be replaced with BNC cables terminated to 50 Ohms inside the shield box.
Thank you for doing this experiment. My generators lack the sweeping capability, and my 300 MHz scope is in repair with a triggering problem or I would try to duplicate this experiment. I am sorry that I cannot do the fine level of testing you are able to at this time because of equipment shortcomings. May be time for me to buy some new gear.
I got the idea from a video I saw where a fellow took a cut half section of a ring magnet (speaker magnet) and he attached leads to either end, four in all at distant points on the magnet section two connections at each end on the faces and was able to light a LED at a specific and very sharp frequency point. That video was later removed from youtube.
I thought this might also be used to characterize a ferrite resonance.
Thank you, Itsu for following up on this.
Kind Regards ION
Hi Itsu:
Regarding post #30 ,notice that the pulse coincides with the peak of the current waveform on the inductor.
To me this appears to be a parasitic turn off of the FET due to the high current drawn as the core tries to saturate or the breadboard layout is not stiff enough to support the current drawn at that point (some stray inductance or capacitance causing the turn off at the peak current.)
With additional stiffening of the power supply at the breadboard site by using quality capacitors and keeping all gate leads and drive leads as short as possible, the parasitic turn off of the FET should be able to be eliminated.
I cannot stress enough that clip leads and plug boards do not make for a good breadboard test at high frequency pulsing. Better to use a piece of copper clad as ground plane with devices soldered to the ground plane where required, and BNC input for signal from SG, with good ground to the ground plane. Keep all wires as short as possible, and local bypassing of the V+ to the ground plane is a important.
Hope this has been helpful, Kind regards, ION
http://www.eevblog.com/2014/01/15/eevblog-568-solderless-breadboard-capacitance/
http://forum.allaboutcircuits.com/showthread.php?t=38094
http://electronics.stackexchange.com/questions/2103/when-to-avoid-using-a-breadboard
Hi Itsu, For what it's worth I got the premature turn off when using a BJT with a low voltage circuit (runs down to under 1 Volt input voltage). I put it down to low base drive, due to the coil drawing current and dropping the voltage, then when the coil prematurely turns off the voltage rises and turns the BJT on again. I use mosfets rather than BJT's wherever possible or unless the application calls for it.
Just out of curiosity. What effect would exciting the ferrite to resonance have ? Would it vibrate the core and induce a current by way of a vibrating core and windings ?
What would be the mechanism for and nature of any extra energy to be harnessed ? What kind of resonance is desired ? mechanical, magnetic or other.
..
Here is a funny story from Bob Pease on the subject of breadboards.
Quote from: ION on 2014.08.04, 15:00:32
Itsu:
Sorry I missed this experiment in post #29 when you first posted it. I just now saw it.
Very nice experiment, however if I may, I would like to offer some suggestions for a more refined test.
.................
Thanks ION, good advice as always i have some silver based conductive glue, so might try it again with these suggestions.
These Russian Pdf's make it look all very easy :) , but obviously its not.
Regards Itsu
Quote from: ION on 2014.08.04, 15:40:31
Hi Itsu:
Regarding post #30 ,notice that the pulse coincides with the peak of the current waveform on the inductor.
To me this appears to be a parasitic turn off of the FET due to the high current drawn as the core tries to saturate or the breadboard layout is not stiff enough to support the current drawn at that point (some stray inductance or capacitance causing the turn off at the peak current.)
Guess you are right, as i found out that even with an air coil (68uH) i have this peak in the middle of the transistor opening, so its not coming from the ferrite.
I will use a copper groundplane setup to put this little circuit together and see if the pulse is still there.
Regards Itsu
Quote from: Farmhand on 2014.08.04, 16:58:07
Hi Itsu, For what it's worth I got the premature turn off when using a BJT with a low voltage circuit (runs down to under 1 Volt input voltage). I put it down to low base drive, due to the coil drawing current and dropping the voltage, then when the coil prematurely turns off the voltage rises and turns the BJT on again. I use mosfets rather than BJT's wherever possible or unless the application calls for it.
Just out of curiosity. What effect would exciting the ferrite to resonance have ? Would it vibrate the core and induce a current by way of a vibrating core and windings ?
What would be the mechanism for and nature of any extra energy to be harnessed ? What kind of resonance is desired ? mechanical, magnetic or other.
..
Hi Farmhand,
thanks for the info, i noticed that when varying the input pulse amplitude (2 - 4 v) on the base, the pulse is moving across the opening window of the collector.
Also this pulse is there when using an air coil only, so nothing to do with ferrite resonance :(
The effect we are after is as Akula demonstrated the eternal lightning of some leds :D
This ferromagnetic resonance should be able to supply power for a long time, but nobody probably knows how (transmutation, mechanical vibration, magnetic vibration?).
Regards itsu
Itsu:
Using a 2 inch x 1 inch x 1/4 inch bar magnet, I silvered copper electrodes 1/8 inch wide to the faces at the one inch ends.
The magnet is polarized on it's large faces, but I don't know what type it is. It was salvaged from an industrial "Printronix" printer
Driving with my signal generator at one end, I was able to light a LED two inches away connected at the other end and there was a peak at one frequency, but it was not very sharp, resembled a low Q LC resonant circuit.
My magnet was about 2000 Ohms on each end.
I will post more technical details and a photo when the glue is all dried and it is mounted in the shielding enclosure. (and when camera batteries are charged).
This is just a preliminary report. Next I will try with ferrites that have not been magnetized.
Thanks ION,
i have some 20/2cm (7.8/0.79 inch) ferrite rods which i could use, but not really sure about your setup, so i await the picture.
Also this "My magnet was about 2000 Ohms on each end" is not clear, do you mean the resistance measured with a multimeter?
I remember my ferrite resistance measurements were in the 80KOhm or so with the leads 1cm apart.
Regards Itsu
Attached is a picture of the magnet and box it will be put into for shielding and proper low noise connections.
There are two copper connections on each end, both left and right, a little hard to see in the picture.
2000 Ohms was measured on both the left two and the right two connections. I did not measure the resistance across the length, but will do so.
I drive input into the right two connections and scope the left two.
I'm having a little trouble with the conductive silver glue and may need to redo the contacts for lowest resistance.
Peak output was between 170kHz to 180 KHz.
I may also wrap a coil to around the magnet and look for output on it while driving various connections to the magnet (and vice versa).
Hope this helps.
OK, got it, do you plan to suspend the magnet in the air, only supported by the BNC connections, or will there be some support underneath?
Regards Itsu
Quote from: Itsu on 2014.08.06, 14:46:15
OK, got it, do you plan to suspend the magnet in the air, only supported by the BNC connections, or will there be some support underneath?
Regards Itsu
The magnet is rather heavy, I may use some small strips of double sided sponge tape at either end to fix it to the box or may float it from the connections if I get the glue to hold.
It does seem to be sensitive to damping with fingers.
BTW, the largest faces of the magnet is where the poles are located.
ION
I tried to do the same with a somewhat smaller ferrite magnet, but this silver conductive glue is not strong enough.
Guess i need to stiffen it up with some superglue over it.
I did not get any reading from the short time i was able to, so no 2000 Ohm or so.
Regards Itsu
Quote from: ION on 2014.08.04, 15:40:31
Hi Itsu:
Regarding post #30 ,notice that the pulse coincides with the peak of the current waveform on the inductor.
To me this appears to be a parasitic turn off of the FET due to the high current drawn as the core tries to saturate or the breadboard layout is not stiff enough to support the current drawn at that point (some stray inductance or capacitance causing the turn off at the peak current.)
With additional stiffening of the power supply at the breadboard site by using quality capacitors and keeping all gate leads and drive leads as short as possible, the parasitic turn off of the FET should be able to be eliminated.
I cannot stress enough that clip leads and plug boards do not make for a good breadboard test at high frequency pulsing. Better to use a piece of copper clad as ground plane with devices soldered to the ground plane where required, and BNC input for signal from SG, with good ground to the ground plane. Keep all wires as short as possible, and local bypassing of the V+ to the ground plane is a important.
Hope this has been helpful, Kind regards, ION
http://www.eevblog.com/2014/01/15/eevblog-568-solderless-breadboard-capacitance/
http://forum.allaboutcircuits.com/showthread.php?t=38094
http://electronics.stackexchange.com/questions/2103/when-to-avoid-using-a-breadboard
ION,
i made some changes like getting away from the breadboard and using a groundplane like PCB.
I did not go all the way by using BNC connectors for the input etc. or putting into a sealed box, as the frequency is very low (250Hz)
I also use a very stable 3V from 2 aaa batteries in series, but the mystery peak is still there.
So i guess its just the setup of this circuit that behaves this way and closes down the transistor early.
Video here: https://www.youtube.com/watch?v=OOPsny2pmps&feature=youtu.be
Regards Itsu
Itsu:
Try removing the 10nF capacitor and the 220pF cap one at a time to see what happens. It could be that the 10nF cap across the 10 ohm is providing a "leading current turn on" of the transistor which is then overcome when the inductor current exceeds the base current (x transistor gain) that the capacitor can supply to keep the transistor on, so the transistor turns off then on again when the SG drive "catches up" with the required base drive.
The copper clad breadboard looks good.
Thanks.
Nope, removing the 220pF variable or the 10nF or both does not make any difference on this signal.
Regards Itsu
Thanks for trying that, sorry it didn't work.
When I get a break I will build the identical circuit and test it.
This has me puzzled now.
Itsu:
What is your power source to the circuit? Is it set to current limit?. I notice the waveform drops back to some lower amplitude and holds to the "on" condition after the first peak of the waveform discharges something (a capacitor on the output of the power supply but internal to it)? or possibly the bypassing at the board. Something is supplying an initial high current then the current drops back to a lower level.
A current shunt resistor (0.1 ohm) in the emitter would give us a better picture of the instantaneous current drawn by the circuit.
ION,
in the latest video with the copper clad breadboard, i use 2x aaa batteries in series (3V).
In the earlier video with the breadboard i used my 10A bench PS set to 3V and current limiter all the way up (so to deliver max current).
I will put a 0.1 Ohm resistor in the emitter lead and measure the voltage across it using again the 2 aaa batteries as power source.
Regards Itsu
Ok, i used both the 2x aaa batteries in series (3V) again and lateron a 12V Battery with a 3.3V regulator as power source.
A 0.1 non-inductive resistor on the emitter lead shows the current through the transistor (mje13005).
The 220pF variable (base to ground) and the 10nF (input to base) capacitors are removed.
The screenshot shows again the peak in the middle of the base opening window with the 3.3V regulator as power source:
Blue trace = collector signal
yellow trace = base signal
purple trace = voltage across the 0.1 Ohm resistor in the emitter
Video here: https://www.youtube.com/watch?v=EnfXLH-w3ko&feature=youtu.be
Regards Itsu
Hi Itsu,
I am not a transistor expert, I learnt my trade before they existed so I can talk knowledgeably about vacuum tubes :).
Looking at your waveforms I can see the rising current charging the inductor up to about 400mA and that is the point where it all goes wrong. IMO it has got to be a characteristic of the transistor, i.e. the point where the emitter current has reached a value such that the base no longer drives sufficient current, the transistor switches off momentarily then recovers to a situation where the emitter current is just below that critical point. The inductor is fully charged to that now constant current, its voltage drop is then due to its DC resistance hence we see a further period of constant current draw while the collector voltage is near its maximum DC value. Then at the end of the base drive pulse the transistor switches off in the normal manner, the inductor discharges into the capacitance there giving the large resonant spike. The middle spike sees the inductor losing energy to the capacitance as its current drops from the peak of 400mA to something slightly smaller.
Smudge
Thanks Smudge, much appreciated.
I can agree with your explanation, but as i have this behaviour with severall different type of transistors and even with a MOSFET,
i wonder if this behaviour is specific to this type of circuit then.
If not, then all these type of (simple) circuits should show this behaviour, meaning that the transistor/MOSFET shuts down premature.
I don't think this is the case unless it is a known behaviour and somehow avoided by adding some components somewhere (snubber circuit).
Regards Itsu
For some reason I cannot view the video it is marked as private.
I agree to a large degree with Smudge's analysis.
If you still have your current probe handy, move it to the base drive and instead of looking at base voltage, look at base current.
I'm beginning to think the SG cannot supply enough current to keep the transistor saturated and at the peak point on the waveform, base current from the SG is insufficient to keep the ramp rising, so it turns off, then goes to the maximum current that the transistor can supply (it is now a current source limiting current to the level shown that is somewhat less than the peak current).
The current gain of the MJE13005 is a min of 10, max of 60. That is rather low, and if we consider that it may be as low as 10 then your SG must supply greater than 40 mA drive current to keep the transistor saturated. I suspect the SG is current limited to less than 40 mA output and the gain of the transistor may be a little higher than 10.
The transistor will certainly turn off or limit at some value when base current times transistor gain is exceeded by collector current.
At the low frequency you are driving, the core probably has insufficient inductance and is drawing a lot of current at the peak point on the current waveform.
A much higher gain transistor or higher inductance core would remove some limitations. If you must keep the present inductance, you could also go much higher with drive frequency, at least twice but three times f should remove the problem entirely.
I don't know if I have explained it clearly, but your transistor cannot supply any arbitrary current and stay saturated, it's collector current will be limited by base drive capability (multiplied by current gain) and exceeding that, it will become a current source and limit current at the collector.
Quote from: ION on 2014.08.10, 11:34:04
For some reason I cannot view the video it is marked as private.
If you still have your current probe handy, move it to the base drive and instead of looking at base voltage, look at base current.
Oops, sorry about that, changed it to hidden, so you should be able to see it now.
I will probe the base current with the current probe additional to the above other traces later on as i have some visitors now.
Thanks so far, regards Itsu
I don't think the transistor actually turns off, the current continues to flow. It switches to the new state as ION said. The sharp spike is merely the inductor having to respond to a change from rising current (where it produces a voltage that almost backs off the supply) to a small drop down to a new current plateau, hence it has to produce that spike in voltage followed by zero emf.
Smudge
Quote from: Smudge on 2014.08.10, 14:08:28
I don't think the transistor actually turns off, the current continues to flow. It switches to the new state as ION said. The sharp spike is merely the inductor having to respond to a change from rising current (where it produces a voltage that almost backs off the supply) to a small drop down to a new current plateau, hence it has to produce that spike in voltage followed by zero emf.
Smudge
Yes you are correct, the transistor never turns off but goes to the current limited state, and the spike seen is the change in current. Since it takes some time for the change in current in the inductor to take place, there is a slight overshoot of the holding current level.
The ramp down to the new plateau exactly matches the total width of the spike.
Ok, guess you are right, i put the collector signal (blue) in AC coupling and expanded the amplitude.
There is a difference of about 180mV between after the pulse and before opening the transistor, see first screenshot.
I am uploading a new short video with the base current included (green trace), see second screenshot.
https://www.youtube.com/watch?v=8vY3byFKsjg&feature=youtu.be
Regards Itsu
sorry for offtopic but I desparately need a solution how to cheaply cut ferrite Do you have experience with any hand tool able to do that ? what I need is to take yoke core from TV and cut the edge to obtain a ring
Quote from: forest on 2014.08.10, 16:06:11
sorry for offtopic but I desparately need a solution how to cheaply cut ferrite Do you have experience with any hand tool able to do that ? what I need is to take yoke core from TV and cut the edge to obtain a ring
I have also used small 1.25" diamond blades on my Dremel tool to successfully cut small cross sections of ferrite material. A very thin blade at high speed seems to work nicely and wastes less material. Save the powder waste for future casting with a binder or stuffing in a tube to make your own ferrites.
Quote from: ION on 2014.08.10, 17:26:16
I have also used small 1.25" diamond blades on my Dremel tool to successfully cut small cross sections of ferrite material. A very thin blade at high speed seems to work nicely and wastes less material. Save the powder waste for future casting with a binder or stuffing in a tube to make your own ferrites.
How much speed did you used ? Is there any range I should stick with ?
Regarding my post #31,re: the bar magnet excitation method, it seems the distributed capacitance and shunt resistance of the bar magnet creates a frequency filter that has a broad pass band around a particular frequency. This accounts for the transfer of voltage from one end to the other of the magnet.
The magnet with connections as shown is a complex network of resistors and capacitors, drawing the simplified network shows the reason it works the way it does.
So no special effect her, just an erroneous first assumption on my part of what I thought might be a possible new effect.
I invite others to try this as a backup check.
Forest:
QuoteHow much speed did you used ? Is there any range I should stick with ?
I don't know what the speed was, never measured it. My Dremel tool from Harbor Freight has three settings, I used the highest, but lower will also work, just takes longer.