THREE COINS IN THE P.I. FOUNTAIN

[Carl-NC]

Eric, this is classic skin effect. The thickness of the metal determines the depth of penetration of the eddies, and the deeper they penetrate, the more "drag" there is in the decay. You can really see this with aluminum foil; I have a bunch of pieces cut to 1" square, with stacks of 1x, 2x, 4x, 8x, 16x, 24x, and 32x, each stack tightly bound in plastic. The tau (TD) and the phase (FD) get higher for thicker stacks. It's easy to use these stacks to emulate just about any coin response.

A very very thin gap between each layer (relative to the layer thickness, anyway) makes little difference as there are no vertical currents, assuming the target is centered in the coil field. You can test this by using a thin poly food wrap as an insulator. As the gap is widened, the tau increase drops because the magnetic field weakens across the gap. I've never thought much about it, but the boundary conditions across a gap may also give rise to effects which alter the exponential. Probably not.

The decrease in amplitude with thicker metal is interesting... did you increase the pulse width as you stacked the coins? I would expect thicker metal to suffer more from residual reverse turn-off current.

Mike, of course there is skin effect in the TD! Try the experiment, and you can see it for yourself. No, increasing the pulse width will not produce the same decay curves, and pulse frequency is largely irrelevant. The shift in tau is real, and due to skin effect.

- Carl

[mikebg]

How about a frequency domain rendition ?

Eric.

The frequency domain shows that participants of this thread use a term "skin effect" which is not attributable to time domain. There is no skin effect in TD. The suitable term for TD is "relative time" normalized to fundamental timeconstant of target.

The frequency domain shows that your decay curves depend on width of TX pulses and their repetition frequency.
If you increase the TX pulse width, you can obtain the same decay curve for three coins as for a single coin.

The frequency domain shows that each coin has series of several timeconstants inverse proportional to squares of natural odd numbers: TC₂ = TC₁/9 ; TC₃= TC₁/25 ; TC₄ = TC₁/49 etc.

Note that there is no emoticons to this post. All is serious.

[Ferric Toes]

A blast from the past
http://www.youtube.com/watch?v=8YxRNdgY5vg
http://en.wikipedia.org/wiki/Three_C...ain_%28song%29

How about a frequency domain rendition ?

Eric.

[mikebg]

A blast from the past
http://www.youtube.com/watch?v=8YxRNdgY5vg
http://en.wikipedia.org/wiki/Three_C...ain_%28song%29

[Qiaozhi]

Nice plots . I expect the results will be rather different for other coins, particularly coins that are made of a more conductive metal e.g. silver or copper. Each coin in turn is attenuating the high frequency components of the TX waveform, giving less excitation to the next coin. Lower frequencies still get through to give the enhancement at late times. Star = 1 coin, Circle = 2 coin, Triangle = 3 coin.

Here is a further plot for my coins where you see the initial amplitude reduced for the next coin in line, but the enhancement at late times. I am hoping that some brave person will try a Spice simulation - if it is possible.

Eric.

It looks to me like the plots show an interplay between permeability (absorption) and eddy current effects (redistribution). With the single coin, the cross-sectional area is relatively large when compared to its mass. As more coins are added, the overall target mass increases relative to the cross-sectional area, resulting in an amplitude decrease below 40us and an increase at later times. I would expect this effect to become increasingly prominent as the spacing between the coins in reduced.

[Ferric Toes]

Yes, nice plots. Thanks for sharing.

Don't forget, that the first coin is shielding the magnetic field lines so the successive stacked coins won't see much magnetic field strength. See also the Skineffect.

Cheers,
Aziz

As the first coin shows just a single exponential (straight line on a log lin plot) any individual skin effect must take place prior to 20uS, otherwise there would be an upward deviation at early times. When second coin is in place the first coin acts as a coupled LR damping network such that the magnetic field is that radiated by the first coin's eddy current. This results in a slower switch off and less initial emf in the second coin. This situation is repeated by the slowing of switch off by the second coin in energising the third. Seems as though the coins act as low pass filter sections for the magnetic field step. It's just like skin effect in that each coin is a skin for the next one. At least, that is how it appears to me.

Eric.

[Aziz]

Yes, nice plots. Thanks for sharing.

Don't forget, that the first coin is shielding the magnetic field lines so the successive stacked coins won't see much magnetic field strength. See also the Skineffect.

Cheers,
Aziz

[Ferric Toes]

4 inch coil, coins centered in coil, not stacked coins were touching, not stacked nickels about 1 inch above coil center, stacked coins were touching. Plotted difference of no coin reference and target. Noisy but still shows trend.

Nice plots . I expect the results will be rather different for other coins, particularly coins that are made of a more conductive metal e.g. silver or copper. Each coin in turn is attenuating the high frequency components of the TX waveform, giving less excitation to the next coin. Lower frequencies still get through to give the enhancement at late times. Star = 1 coin, Circle = 2 coin, Triangle = 3 coin.

Here is a further plot for my coins where you see the initial amplitude reduced for the next coin in line, but the enhancement at late times. I am hoping that some brave person will try a Spice simulation - if it is possible.

Eric.

[green]

4 inch coil, coins centered in coil, not stacked coins were touching, not stacked nickels about 1 inch above coil center, stacked coins were touching. Plotted difference of no coin reference and target. Noisy but still shows trend.

[WM6]

It demonstrates that as coins are added to the stack the TC increases even though they are insulated from one another. I am not clear as to why.

Probably you need to add some EM shielding from side of stacked coins. Could be that some of composed pulse spectrum penetrate insulation.

[ODM]

Mutual inductance rising more rapidly than resistance falls? The contacts between coins are likely to be a different magnitude of resistance compared to the skin of a single coin, too?

[Ferric Toes]

Here is a plot for the three coins. Lower curve is 1 coin and show a good single exponential. The starting amplitude is set to 1000 for each of the 3 curves, although the exponential fit for the second curve (2 coins) was at some variance, particularly at the start. For 3 coins (top curve) it is better. It demonstrates that as coins are added to the stack the TC increases even though they are insulated from one another. I am not clear as to why.

Tomorrow I will post a different graph set showing how the starting amplitude varies as coins are added.

Eric.

[bbsailor]

Eric,

After scratching my head a few times I came to the conclusion that the three coins must act like the links in a fine gold chain. The TC of the individual link (or coin) governs it's detectability not the mass of all the poorly electrically coupled links (or coins).

bbsailor

[Overtheedge]

Admittedly I've still got a lot to learn, but here is my thoughts.

Amplitude reduction with an extension in the time delay.

My reasoning is the eddy currents in each coin affects the other coin's emitted signal. You are stacking emitters whose phase delays are not in synch.

Effectively you get a smeared return.

But what do I know? Not all that much until I have tested and confirmed for myself.
eric

It's not what you know that gets you in trouble, it's what you know that ain't so.

[green]

I tried the test with 5 quarters. Got different results. The single coin had a slightly lower voltage at the start but a higher voltage at longer decay times. I still don't have my preamp done and am looking at coil voltage after the decay with a scope. Subtracting a no coin reference signal from the target signal in Excel. The coil is laying on the bench with the coins stacked on the bench in the middle of the coil. Should I expect result to be a lot different measuring the coins at a distance 1,2 and 3 times the coil radius with a preamp and integrator?