Targets frequency response

[Aziz]

Hi B^C,

have a nice trip. I am looking forward to your progress. You are enviable.

Aziz

[amtech2005]

Targets frequency response

There is quite a bit posted on the web "www" about coin characteristics and also archaeomagnetic coins. You just have to be carefull how you type your search!! If not You will get 50,000 "metal detector for sale" sites!!! . I started reseaching today on Magnetic fields of coins and I found this on a public website!! Eugene

[B^C]

Hi All,

I have intruments that i can collect some good data on.
Is there any particular data that will be useful to others?
If so make a list & in the coming weeks i will set about doing some tests & give the results. I am moving to a new location tomorrow & will start to set some things up for testing.
I have a lot of things to test out for myself but if i can help out just let me know.

[Aziz]

Hi B^C,

I would appreciate the following data (if possible):
- noise characteristics on the signal decay curve (e.g. after the pre-amp)
- ground signal (+noise) on the decay curve (without target) on most critical mineralized grounds (I live in Germany, so I have no idea about Oz ground).

This would help, to analyse and eliminate the noise problem. As the lock-in amp can not be used easy and conventional on the PI's, I have to look for other solutions. So using lock-in amp sucks a little bit here.

Regards,
Aziz

[B^C]

Gday Aziz,

I will set things up & get some data for you, just leaving now.
I'll be in touch.

[Aziz]

I have found a method, to compute the target frequency response.
The way is to setup two differential equations of the second order for the primary coil and for the target (coil). As the system is inductively coupled, these two differential equations are also coupled together.

For the analytical solution, it is not trivial and often it can not be done. Using numerical methods (like finite elements, numerical derivations) gives a good approximation.
The mutual inductance between the primary coil and the target (coil) must be taken into account. Also any capacitance of the primary and target.

I will search for more infos..
Aziz

[wyndham]

induction test

Bruce, hope the move went well.
If I understand the different screens you posted that for a given induction most of the decay, regardless of density or surface area(except the ring) of the material, decayed almost the same. Does that mean that the targets producing the eddy currents did this or the RX only saw this amt of decay?
Maybe I'm reading the charts wrong, but shouldn't diferent target show more of a difference in the decay(except for the ring)?
I am very inexperienced(none) in the nuts and volts of circuits and the math that goes with it, so I'm limited in many areas here.
It would seem that non magnetic metal objects might have a min decay time(that doesn't seem right) and though you didn't post it yet, is the decay time in linear or exponential relationship to the induction. I would guess exponential since that's what I seem to see on other screens and post that indicate the exponential amt of tx energy to add a modest amt of depth.
Also would it be possible to use a 24 bit sound card in a pc for the input of a induced signal and sound software to see the transient signal? That's about all I have that might be useful.
Thanks for your time when you have some for a wannabe
Wyndham

[Tinkerer]

Different target response

Bruce, hope the move went well.
If I understand the different screens you posted that for a given induction most of the decay, regardless of density or surface area(except the ring) of the material, decayed almost the same. Does that mean that the targets producing the eddy currents did this or the RX only saw this amt of decay?
Maybe I'm reading the charts wrong, but shouldn't diferent target show more of a difference in the decay(except for the ring)?
I am very inexperienced(none) in the nuts and volts of circuits and the math that goes with it, so I'm limited in many areas here.
It would seem that non magnetic metal objects might have a min decay time(that doesn't seem right) and though you didn't post it yet, is the decay time in linear or exponential relationship to the induction. I would guess exponential since that's what I seem to see on other screens and post that indicate the exponential amt of tx energy to add a modest amt of depth.
Also would it be possible to use a 24 bit sound card in a pc for the input of a induced signal and sound software to see the transient signal? That's about all I have that might be useful.
Thanks for your time when you have some for a wannabe
Wyndham

The target response is composed of several different components.

The resistive component
The reactive component
The skin effect eddy currents
The core eddy currents
And maybe a few more.

To visualize or read a specific component, the processing of the signal has to be specifically designed for that specific component.

For example: to read the response of the skin effect, the sample has to be taken very early, meaning after 5 to 10 us after the transient.
The skin effect relates more to the surface area presented to the field.

The effect of the core eddy currents last much longer, they are related to the TX time and represent mostly the TC of the target.

If You want to see the reactive response of FE targets, you need to know where and how to look for them.

A tip from Eric Foster: If you present a thin, FE disk the flat way to the coil, you see mostly the resistive component of the target.
If you turn the same disk on edge, you see mostly the reactive component of the target.
Try this and compare by doing the same test with a thin disk of similar dimensions of non magnetic material. The results will prove without doubt.

Starting with this good piece of advise, I developed a fully discriminating PI platform. Thanks Eric.

Tinkerer

[Monolith]

Fully Discriminating PI

The target response is composed of several different components.

The resistive component
The reactive component
The skin effect eddy currents
The core eddy currents
And maybe a few more.

To visualize or read a specific component, the processing of the signal has to be specifically designed for that specific component.

For example: to read the response of the skin effect, the sample has to be taken very early, meaning after 5 to 10 us after the transient.
The skin effect relates more to the surface area presented to the field.

The effect of the core eddy currents last much longer, they are related to the TX time and represent mostly the TC of the target.

If You want to see the reactive response of FE targets, you need to know where and how to look for them.

A tip from Eric Foster: If you present a thin, FE disk the flat way to the coil, you see mostly the resistive component of the target.
If you turn the same disk on edge, you see mostly the reactive component of the target.
Try this and compare by doing the same test with a thin disk of similar dimensions of non magnetic material. The results will prove without doubt.

Starting with this good piece of advise, I developed a fully discriminating PI platform. Thanks Eric.

Tinkerer

Tinkerer,

A fully discriminating PI has been everybody's dream for some years.
Would you care to tell us how you accomplish that?

Monolith

[Tinkerer]

Fully discriminating PI

Tinkerer,

A fully discriminating PI has been everybody's dream for some years.
Would you care to tell us how you accomplish that?

Monolith

Monolith,

I have been posting bits and pieces in different posts on the Forum. If people show interest, I will start a dedicated thread and explain the methods and show the circuit.
Tinkerer

[mikebg]

Resonance Frequency of soil and ferrous target

"Resonance" is term of frequency domain. It means that there is a frequency, at which the phase lag of response is zero because the imaginery part of complex transfer function is zero. Since EMI sensor excitates environment with magnetic field, induced EMF in RX coil has the same phase angle as TX current at resonance frequency.
Visual analysis in complex plane shows that there is resonance frequency of soil or target if it possesses both conductivity and magnetizm. This is shown in attached figure. Magnetizm pulls working point upwards in ferro quadrant, but conductivity pulls it downwards in conductive quadrant. In this case, the curved line of transfer impedance intersects Re axis at resonance frequency.
In reality, the phase lag of induced in RX coil voltage can not be zero at resonance frequency. It is 360 deg phase delay, because the response is causal. Conductivity generates signal ON at resonance frequency. The phase angle of point N is negative. Ferromagnetizm generates signal OF at resonance frequency. The phase lag of this signal is more than 270 deg because is shown lossy ferrite.
The vector sum of both generated signals OF and ON is the resultant received signal OQ. In frequency domain, point Q coincidents visual with the axis of excitation Re. In time domain, resonance point Q can be achieved by appropriate sample delay.
There is an comprehensive explanation of term "Electrical resonance" in Wikipedia:
"Electrical resonance occurs in an electric circuit at a particular resonance frequency when the impedance between the input and output of the circuit is at a minimum (or when the transfer function is at a maximum). Often this happens when the impedance between the input and output of the circuit is zero and when the transfer function equals one."

[Tinkerer]

Mikebg wrote:
<In time domain, resonance point Q can be achieved by appropriate sample delay.>

Thanks for the input.
Your figure explains a lot. I wish I knew how to do the mathematical analysis of what I do with the circuit.
In general the discriminating PI, (Time Domain) has to overcome the problem as you mention above. The steel disk test mentioned above is an easy way to try to coax a specific response.
A much more difficult target to differentiate is a clad coin. That is a coin with a FE core and a copper, silver or other non magnetic alloy outer cover.
The problem with the decay times is that different size\mass targets will produce different decay times even if they are of the same metal or alloy. Talking about core eddy currents.
Now, when you look at skin effect eddy currents, it is rather the surface area that is important. Example, a thin aluminum candy wrapper, gives a large response in the skin effect time domain.
If we lump all different responses into one single output, we will know very little about the target's identity.

Therefore we need to spread the responses into several times so that we can sample each window that can give us the desired information.

At present I use 6 samples and I can differentiate quite a few different metals and alloys. But, there is still a lot to learn. For instance, I still don't quite understand what causes the different readings, why the readings change when the target is very near the coil or why a gold leaf shows a negative response like FE, but a gold coin or ring is of opposite polarity.

Oh, by the way, I can differentiate a clad coin from a FE disk or a copper disk.

Tinkerer

[mikebg]

Combination magnetizm &amp; conductivity

Transfer impedance of conductive permeability

Curved line 3 represents transfer impedance form in common case when soil or metal object possesses conductivity as well as permeability. The half-cardioide 3 is obtained by summing pairs points – one of line 1 (impedance of ferrite) and another of line 2 (impedance of nonferrous conductivity) having equal frequency.
The line 3 can't be normalized exactly because the line of ferromagnetic impedance 1 can't be normalized at all. Besides, the line 1 is not exactly straight and represents nonlinear properties – hysteresis, saturation and magnetic viscosity. Complex plane can represent correct only linear properties. Despite all we can show shape of impedance form and make several conclusions:
1. If a soil or a metal object have both conductivity and permeability, transfer impedance seems like half-cardioide line 3 laying in first and fourth quadrant intersecting Re axis. The intersecting point Q represents resonance frequency. The shape of half-cardioide can differ from shown form because depends on permeability/conductivity ratio. After rain conductivity of soil increases and changes impedance form of ground signal.
2. We can define three characteristic frequencies on half-cardioide 3:
- Resonance frequency (point Q), at which phase lag is zero (the actual phase angle is minus 360 deg),
- Quasi (pseudo) cutoff frequency (point C), at which the phase lag is 45 deg (the actual phase angle is minus 405 deg) and
- Mirror cutoff frequency (point D), at which phase lead is 45 deg (the actual phase angle is minus 315 deg).
3. If an object is made of ferrous metal and is excited by relative high TX frequencies having impedance points in 4-th (conductive) quadrant, we can't determine whether it is ferrous. Only at relative low frequencies we can identify ferrous objects by impedance points situated in first (ferrous) quadrant.

[Tinkerer]

Mikebg,

thanks for the detailed explanation. Could you explain to me how to correlate frequency to transient signals?
Can I use the slew rate as an approximation?

Tinkerer

[mikebg]

Step-down response in frequency domain

The attached figure is suitable for visual analysis only, but it shows methods for elimination signals of mineralized conductive soil and ferrous metal objects.
There is difference between Pulse Induction and Step-down response. At PI the TX pulse has width, but at Step down the TX width is infinitely.