DEEPER PI DETECTION DEPTH

[Carl-NC]

Simon, what other result could you have gotten? If a higher voltage appeared in one of the RX loads then I would be questioning the whole concept of transformers! The V*I product in a (lossless) transformer has to remain constant, so as you increase I (by making R smaller) then the developed V is smaller. A constant L/R ratio in the target should always produce the same result in the RX coil.

Davor, target inductivity is mostly controlled by skin effect. Target R is largely determined by metal conductivity. Target size & shape affect both. However, it is true that, for modeling purposes, you only need to get the tau right.

- Carl

[Davor]

Got it, eddies take care of it. Skin effect is due to current flow within a target. Thanks.

Anyway, I made a small conceptual drawing of a high L -> high R and low L -> low R with a constant L/R.

BTW, same goes for search coils as well.

[Davor]

I previously mentioned a weighted Rx option for PI frontend. It could be implemented with no sweat by Tayloe type chopper with weighting resistors of appropriate values. Such weighted output could have much more manageable features, and I guess even a possibility to directly synthesize a tone representation of a target.

Here is a representation of what's on my mind, yet with a x^2 weighting function. You may play with R2 (target) to see a full picture.

Again, patent trolls - keep off!!!

[Aziz]

Another Brain Food

Hi all,

here is an another brain food for you tinkerers.

Continuous single frequency mode (VLF) vs. interrupted (paused) single frequency mode.

The first mode do not need a FFT display. All the spectral energy is focused in the fundamental frequency.

But what happens, when I interrupt the resonant mode? I.e. first half-sine current pulse, pause, next but negative half-sine current pulse and pause again?

Look at the bottom picture. One wouldn't probably guess, that the FD response is a wide band response. (odd harmonics 1,3,5,7,9,...)

The wide band characteristics disappear, when I set:
half-sine, negative half-sine, pause, pause =
full wave sine, pause, pause

So what's next?
-> interrupt the continuous single frequency resonant mode.


Aziz

[simonbaker]

Simon, what other result could you have gotten? If a higher voltage appeared in one of the RX loads then I would be questioning the whole concept of transformers! The V*I product in a (lossless) transformer has to remain constant, so as you increase I (by making R smaller) then the developed V is smaller. A constant L/R ratio in the target should always produce the same result in the RX coil.

Davor, target inductivity is mostly controlled by skin effect. Target R is largely determined by metal conductivity. Target size & shape affect both. However, it is true that, for modeling purposes, you only need to get the tau right.

- Carl

Well, a point I thought I was getting at is that it seemed we are forcing a certain physical interpretation by specifying the same mutual inductance for each target in our simulation -- however, that may be justified, I need more thinking.

What still seems lacking is some data showing that "tau" correlates with target conductivity. Looking at our target models, the 200 uH / 8 ohm target seems like a lump of coal, and the 200 pH / .00001 ohm target seems like a little piece of silver. But maybe my instincts are wrong, and just scaling a piece of silver will give the characteristics of the "lump of coal" target; it all boils down to non-obvious physics such as skin effect, etc.

-SB

[simonbaker]

Hi all,

here is an another brain food for you tinkerers.

Continuous single frequency mode (VLF) vs. interrupted (paused) single frequency mode.

The first mode do not need a FFT display. All the spectral energy is focused in the fundamental frequency.

But what happens, when I interrupt the resonant mode? I.e. first half-sine current pulse, pause, next but negative half-sine current pulse and pause again?

Look at the bottom picture. One wouldn't probably guess, that the FD response is a wide band response. (odd harmonics 1,3,5,7,9,...)

The wide band characteristics disappear, when I set:
half-sine, negative half-sine, pause, pause =
full wave sine, pause, pause

So what's next?
-> interrupt the continuous single frequency resonant mode.


Aziz

Those discontinuities look wide-band to me... like part square wave, which is broadband.

Can you show the last one you described that is supposedly narrow band?

-SB

[Carl-NC]

What still seems lacking is some data showing that "tau" correlates with target conductivity.

But tau doesn't (necessarily) correlate with conductivity... a US silver dollar has exactly the same conductivity as a US half-dime, but the taus are radically different. And, interestingly, I have a Mardi Gras token that's almost exactly the same size/thickness of a US silver dollar and almost exactly the same tau... yet it is made of aluminum.

[simonbaker]

But tau doesn't (necessarily) correlate with conductivity... a US silver dollar has exactly the same conductivity as a US half-dime, but the taus are radically different. And, interestingly, I have a Mardi Gras token that's almost exactly the same size/thickness of a US silver dollar and almost exactly the same tau... yet it is made of aluminum.

Ok, now we're getting somewhere. That's the drum I've been beating for a while, sorry to nag.

So how are we planning to use tau for discriminating with PI detectors? Or did I misunderstand?

-SB

[Carl-NC]

Target tau is to time-domain what target phase is to frequency domain. They are interchangeable. Mathematically,



In the time domain you really need to be looking at a driven response to determine tau. In a normal PI, we look at the decay way too late for this; we need to look at the first 1 or 2 us, right in the peak of the flyback, or look during the TX period. Either of these require an IB coil.

- Carl

[Tinkerer]

Target tau is to time-domain what target phase is to frequency domain. They are interchangeable. Mathematically,



In the time domain you really need to be looking at a driven response to determine tau. In a normal PI, we look at the decay way too late for this; we need to look at the first 1 or 2 us, right in the peak of the flyback, or look during the TX period. Either of these require an IB coil.

- Carl

We keep coming back to the IB coils. I have been working so long with IB coils, that I have forgotten where the problems with IB coils are supposed to be.

Could somebody refresh my mind what advantage a mono coil is supposed to have? Or what disadvantage the IB coils are supposed to have?

Tinkerer

[simonbaker]

Target tau is to time-domain what target phase is to frequency domain. They are interchangeable. Mathematically,



In the time domain you really need to be looking at a driven response to determine tau. In a normal PI, we look at the decay way too late for this; we need to look at the first 1 or 2 us, right in the peak of the flyback, or look during the TX period. Either of these require an IB coil.

- Carl

Ok, but how do you plan to use it to discriminate a US dollar from a Mardi Gras token?

-SB

[Aziz]

Those discontinuities look wide-band to me... like part square wave, which is broadband.

Can you show the last one you described that is supposedly narrow band?

-SB

Ok,

I admit, I was wrong. (Good mistake, which keeps learning.)
The full wave sine and pause pulse has a wide band spectrum as the sine wave abruptly discontinues in the time-domain. So the spectral energy isn't totally focused to the fundamental frequency.

See below.
Cheers,
Aziz

[Carl-NC]

Ok, but how do you plan to use it to discriminate a US dollar from a Mardi Gras token?

-SB

Can't do it. At least, not with induction phase/tau.

[simonbaker]

Can't do it. At least, not with induction phase/tau.

To follow up:

Aziz showed a PI simulation above with targets whose taus had 4 orders of magnitude difference.

Typical metals we are interested in are within 1 order of magnitude in conductivity.

To do discrimination, it appears we need to map 4 orders of magnitude of tau onto 1 order of magnitude of conductivity.

That would not be surprising if tau had some exponential relationship to conductivity that spread it over a greater range.

However, if I understand correctly, it might be inferred that targets with the same shape and size mainly differ in tau due to conductivity alone, and that the "R" of L/R is based on the metal conductivity (since the target shape and size is identical).

In other words, to me it looks like there is a disconnect between the statements that a) targets have 3 to 4 orders of magnitude tau range and b) that tau roughly corresponds to metal conductivity and can be used for discrimination.

Perhaps (b) holds if targets are typically the same size and shape; however, Aziz's simulation would not be realistic representation of that assumption.

Forgive me for beating this, but it seems important to clarify how targets behave to better design our MDs.

-SB

[simonbaker]

Davor, target inductivity is mostly controlled by skin effect. Target R is largely determined by metal conductivity. Target size & shape affect both. However, it is true that, for modeling purposes, you only need to get the tau right.

- Carl

I'm very interested in how the "skin" effect comes into play and how it varies for different targets. Are we also talking about a "shielding" effect where it might be difficult to stimulate much current in a thick target due to surface currents?

What can we say about targets that are identical size and shape but different metal conductivities, especially their taus?

-SB