Barkhausen effect

[exnihiloest]

Working on the possible elementary effects involved in the Coler device, I needed to wind a small coil of 110 turns around an iron wire 7cm long, insulated with green plastic. The iron wire diameter is 1 mm and the outer diameter of the coil is 2 mm. It's very small (see below).

The goal was to test possible anomalies when there is a current both in the coil and in the iron wire. I noticed nothing abnormal. Then I decided to use the coil as a detector of possible signals when a current passes through the iron wire, therefore I connected the coil to a sound card. There was shot noise of unknown origin. After many tests not yet ended, it seems that the problem came either from the sound card (a bit surprising because I had the same problem with two sound cards), or from the long cable connecting the coil to the sound card. So to avoid this problem, I decided to add a microphone preamp at the end of the cable, and then to connect the coil with a shorter cable to the audio preamp. This way the preamp adds a sufficient gain to make the shot noise negligible in comparison with possible coils signals.

Here are the coil and the preamp:


After that, I led miscellaneous tests, including one with a neodynium magnet. At this step I noticed strong noises when I moved the magnet near the coil, at about 10 cm, even very slowly. Obviously it was the Barkhausen effect. I was surprised that I could get a so strong effect with such a little coil.
Then I tried with a big coil with a big iron core and got only very poor results. It seems that using a thin iron core is much better than a big one. So this post is to share this information. It can be useful to know it, because if the Barkhausen effect is improved this way, it's possibly due to a best "cooperation" of magnetic domains when the volume is thin and/or to a best coupling with the tightly winded coil. Thus other devices with magnetic coils could perhaps be improved by the same way (for example by replacing a coil with a large core by the same volume of small coils with thin core).

Here is the Adobe Audition screen shot of the Barkhausen noise:

There are disturbances from the mains (50Hz and harmonics), visible in the horizontal lines. The Barhausen noise occupies all the spectrum that I had limited to 3 Khz in this test.

The MP3 is in the attached file.

[ion]

Interesting test Ex.

Cook claims thin rods operate his device better than a  large solid core.

It might be possible to use the pickup coil also as an excitation coil if suitably decoupled from a driving signal generator (low frequency triangle wave.) Might need a lot more turns to get the same strength as a Neo PM.

How many microvolts are you getting from 110 turns?

The Coler device is on my list of 5 most interesting for future research.

[EMdevices]

@EX

Then I tried with a big coil with a big iron core and got only very poor results. It seems that using a thin iron core is much better than a big one. So this post is to share this information.

thanks for sharing EX,  its good to know.   I imidiatelly think of of the TPUs, and the iron wires inside, perhaps Barkhausen is at work.

EM

[giantkiller]

This is profound. It is a part of the mechanism of manipulating the weak field. I am trying to state enough here.

[Peterae]

EX nice work

Isn't this exactly how SM said to wind his coils 
Why does the noise seem to be ordered at very low frequency, IE there are lots of consistant dots at the bottom of the spectrum that seem to make a perforated sheet sort of pattern, seems to be a low Hz oscillation?

[exnihiloest]

...
IE there are lots of consistant dots at the bottom of the spectrum that seem to make a perforated sheet sort of pattern, seems to be a low Hz oscillation?

I noticed them also. I don't know their cause. The gain of the amplifier is important, maybe it's a 1/f noise.

[exnihiloest]

...
How many microvolts are you getting from 110 turns?
...

I don't know yet, probably less than 100µV. The amplifier gain is important (the noise floor is very visible). When the coil is connected to a scope, the signal is not visible even at the scale 2mV/div.

The Coler device is on my list of 5 most interesting for future research.

His claim seems credible. The problem is to duplicate the device without clear description. For example, the composition of the iron used in the coil cores must be known if impurities play a role. This implies to make research in the history of the technology of Coler's time.
An interesting paper was published at Chavasciences. They say that the Coler's device is strongly related to a patent from R Norrby (FR512005A, in French). It can be a start point.

[exnihiloest]

A thinner iron wire is better. Here is a new test with a soft iron wire which should be of about half the diameter of the previous one, or less, meaning a volume divided by 4.
The copper wire is also thiner, so that the outer diameter including the winding is less than 1 mm.
On the photo the previous wire and the new coil are both shown:


Here is the signal record. In spite that the pass band has been widened up to 20 Khz, we see that the noise floor is weaker than previously:


I have also tested the first iron wire as core of a larger coil having much more turns but an inner diameter of around 1cm, thus 10 times larger than the outer diameter of the iron wire. Nevertheless I got very good signals. The important point is therefore less the coil than the thinness of the coil core.

[stprue]

Hello Ex,

Just a quick question, have you tried a longer piece of wire?

[exnihiloest]

Hello Ex,

Just a quick question, have you tried a longer piece of wire?

I tried with a 20cm long wire, but I kept the same 5-6 cm coil at the center. No improvement.
The only thing that I could conclude from this test was that a change of the magnetic domains by moving a magnet near the wire end, doesn't progate far from the end. The noise was rather proportional to the distance coil<->magnet, not to the distance iron wire<->magnet.

[Peterae]

So maybe how about a number of 1cm-2cm long pieces of iron wire all laid side by side and connect each copper coil that is wound over each iron in series

[BEP]

A circle of thin iron pegs with a copper basket weave coil covering them ?

[exnihiloest]

A circle of thin iron pegs with a copper basket weave coil covering them ?

It would be interesting to test it. I have had the similar idea to use an thin toroid made from an iron wire but I've not yet built it, my hesitation being due to the fact that the junction of the two ends will not be good. Perhaps I will use an iron ring if I find one of sufficient thinness.

[exnihiloest]

So maybe how about a number of 1cm-2cm long pieces of iron wire all laid side by side and connect each copper coil that is wound over each iron in series

This should give good results. Nevertheless it goes in the opposite direction that I would like to go, which is to have a maximum of synchronized magnetic domains. With many pieces, due to weak coupling, it's doubtful that the magnetic domains of different pieces synchronize with each other.
Given uncorrelated voltage noise sources, the total noise is given by the square root of the sum of each noise squared. If they were correlated, it would be the simple arithmetic sum which is much more. For instance, two uncorrelated noise sources of 1v each give √2 volts by adding, while two correlated sources give 2v.

[BEP]

I may try something like the attached image bent into a circle.

You are correct about uncorrelated noise. That would be a problem. Perhaps this winding method would allow for maximum results?

[exnihiloest]

I may try something like the attached image bent into a circle.

You are correct about uncorrelated noise. That would be a problem. Perhaps this winding method would allow for maximum results?

The winding doesn't change the way that the magnetic domains "feel" each other. The air (or copper) gap between the iron wires breaks the way for the magnetic flux.

[Peterae]

This should give good results. Nevertheless it goes in the opposite direction that I would like to go, which is to have a maximum of synchronized magnetic domains. With many pieces, due to weak coupling, it's doubtful that the magnetic domains of different pieces synchronize with each other.
Given uncorrelated voltage noise sources, the total noise is given by the square root of the sum of each noise squared. If they were correlated, it would be the simple arithmetic sum which is much more. For instance, two uncorrelated noise sources of 1v each give √2 volts by adding, while two correlated sources give 2v.

I see thankyou for explaining.

Maybe flat soft iron foil or strip.

[exnihiloest]

I have connected my last coil to the antenna input of a sensitive radio receiver (some µV are enough). The Barkhausen noise is clearly heared at 100Khz but not strong, the S-meter doesn't move. It is much weaker at 500 Khz. At 1 Mhz there is almost nothing audible. I have also tried at the coil resonant frequency (about 4 Mhz) and up to 30 Mhz, there was nothing. The power spectrum is obvioulsy in the low frequencies, in the audio range.

[ion]

Thought problem:

If the moving magnet were attached to a small motor assembly such that it moved linearly towards the end of the wire and back, the cost to the motor assembly (except for friction) would be close to zero since the force of attraction would be a "gain", and the force required to pull the magnet back an equal "loss". Since no field lines are being cut, Lenz's law would not apply, but domains would flip producing microvolts of noise in the coil.

The question then becomes this: Does the magnet need to be flipped to reset the domains or do they auto-reset at a reduced field.

Perhaps the change of field strength in the iron wire is the same as cutting lines of flux and Lenz's law does apply, so no gain here.

[exnihiloest]

@Ion
From what I understand of the Barhausen effect, the domains auto-reset at a reduced field. Whatever the way that the magnet moves, when a certain level of magnetic field is reached inside the iron, domains flip. They align more or less along the field lines of the magnet, and therefore they attract the magnet more strongly (it's not noticeable because the effect is too weak). This implies that the reverse motion of the magnet would need more energy to fight the attraction. In fact it is the same, because when the magnet moves away, there is a threshold below which the domains spin again, restoring their initial position and energy.

We can see the elementary magnetic dipoles in a magnetic domain as rings carrying a current and aligned with each other. The rings tends to align along the field lines of the resultant of the applied external field and of the neighboring domains field but there is a ratchet effect needing that a threshold to be reached for the effect to occur.

After a back and forth motion of the magnet, what I don't know is whether each domain is restored strictly identical or if it is the macroscopic mean of the fields of each domain that is conserved.