skin effect

[green]

As Carl has mentioned before:
Response conductivity is due to target metal type.
Response inductivity is due to target thickness.
Response strength is due to target surface area.

Which means that the inductance is mostly controlled by the skin effect (which depends on target thickness), and resistance is mainly controlled by the conductivity. Both target size and shape have an effect on the inductance and resistance. Therefore, although you only need to have the correct value of tau for modelling purposes, the problem is essentially a matter of figuring out what the tau needs to be. Which is why you'll not see any hard-and-fast numbers published anywhere.
(posted in another thread)

I've been thinking different:
Response conductivity is target metal type and target thickness
Response inductivity is size, width and length mainly width(maybe only width)
Response strength is due to surface area

Have read in some of the replies skin effect effects thin targets not thick targets. When I search skin effect, current isn't the same thru out target thickness(higher on the outside). Been thinking if target charts straight line decay linear log chart, target not effected by skin effect. If decay isn't straight line decay, faster decay in the beginning its because of skin effect.

Including some charts with examples.
TRT_29: Stacked foil targets chart straight line decay, not effected by skin effect. Slopes not the same so would effect detection depth.
Quarter and stacked quarters, not straight line decay, effected by skin effect. Chart straight line decay at longer delay times.>2 or three target TC's? Chart close to same at normal sample times so wouldn't effect detection depth.
TRT_32:second graph, copper clad board and lead sheet chart same(TC=(L/R)). 1oz copper clad board 1.4mils thick, for same resistance lead would need to be 16.7mils(measured 16mils, close enough)

Probably been discussed in other threads but I couldn't find what I was looking for so I'm starting a thread labeled skin effect. Maybe I'm thinking wrong or it's a matter of definition.

[bbsailor]

As Carl has mentioned before:
Response conductivity is due to target metal type.
Response inductivity is due to target thickness.
Response strength is due to target surface area.

Which means that the inductance is mostly controlled by the skin effect (which depends on target thickness), and resistance is mainly controlled by the conductivity. Both target size and shape have an effect on the inductance and resistance. Therefore, although you only need to have the correct value of tau for modelling purposes, the problem is essentially a matter of figuring out what the tau needs to be. Which is why you'll not see any hard-and-fast numbers published anywhere.
(posted in another thread)

I've been thinking different:
Response conductivity is target metal type and target thickness
Response inductivity is size, width and length mainly width(maybe only width)
Response strength is due to surface area

Have read in some of the replies skin effect effects thin targets not thick targets. When I search skin effect, current isn't the same thru out target thickness(higher on the outside). Been thinking if target charts straight line decay linear log chart, target not effected by skin effect. If decay isn't straight line decay, faster decay in the beginning its because of skin effect.

Including some charts with examples.
TRT_29: Stacked foil targets chart straight line decay, not effected by skin effect. Slopes not the same so would effect detection depth.
Quarter and stacked quarters, not straight line decay, effected by skin effect. Chart straight line decay at longer delay times.>2 or three target TC's? Chart close to same at normal sample times so wouldn't effect detection depth.
TRT_32:second graph, copper clad board and lead sheet chart same(TC=(L/R)). 1oz copper clad board 1.4mils thick, for same resistance lead would need to be 16.7mils(measured 16mils, close enough)

Probably been discussed in other threads but I couldn't find what I was looking for so I'm starting a thread labeled skin effect. Maybe I'm thinking wrong or it's a matter of definition.

green,

Your analysis needs to include another variable: Full stimulation of the target you are testing.
According to theory, full stimulation of a target occurs when the coil discharge TC is 5 times faster than the target TC. The coil discharge TC is calculated by dividing the damping resistor value into the coil inductance. A 300uH coil with a 600 ohm damping resistor creates a coil discharge TC of .5 uS. This would fully stimulate a target of 2.5 uS. Since the damping resistor value is largely controlled by coil and coil circuit capacitance, reducing this capacitance allows higher values of damping resistance, thus being able to fully stimulate smaller or lower TC targets.

A fully stimulated target signal will fall into the noise level in about 5 TCs. That is why early sampling of smaller targets is desired for low TC targets. When designing a PI metal detector you must make trade offs in TX power, pulse length, TX frequency, delay, RX gain, noise, fast time to come out of RX op amp saturation, and RX signal integration to focus your design on a small range of desired targets. Once you get into trying to optimize all of these factors, you will come to realize that it is a balancing act to ensure that optimizing each of these variables moves your design in the same "more sensitive to your selected targets" direction.

Look up how lock in amplifiers work to sense very low signal levels. Integrating many PI RX signals uses the concept of lock in amplifiers to improve RX sensitivity but only as long as the target is within the coil area and coil moving speed.

I hope this helps?

Joseph J. Rogowski

[green]

green,

Your analysis needs to include another variable: Full stimulation of the target you are testing.
According to theory, full stimulation of a target occurs when the coil discharge TC is 5 times faster than the target TC. The coil discharge TC is calculated by dividing the damping resistor value into the coil inductance. A 300uH coil with a 600 ohm damping resistor creates a coil discharge TC of .5 uS. This would fully stimulate a target of 2.5 uS. Since the damping resistor value is largely controlled by coil and coil circuit capacitance, reducing this capacitance allows higher values of damping resistance, thus being able to fully stimulate smaller or lower TC targets.

A fully stimulated target signal will fall into the noise level in about 5 TCs. That is why early sampling of smaller targets is desired for low TC targets. When designing a PI metal detector you must make trade offs in TX power, pulse length, TX frequency, delay, RX gain, noise, fast time to come out of RX op amp saturation, and RX signal integration to focus your design on a small range of desired targets. Once you get into trying to optimize all of these factors, you will come to realize that it is a balancing act to ensure that optimizing each of these variables moves your design in the same "more sensitive to your selected targets" direction.

Look up how lock in amplifiers work to sense very low signal levels. Integrating many PI RX signals uses the concept of lock in amplifiers to improve RX sensitivity but only as long as the target is within the coil area and coil moving speed.

I hope this helps?

Joseph J. Rogowski

Trying to understand 5 times faster TC. Tried a simulation in spice. Left side, Tx TC=target TC. Right side, Tx TC=1/5 target TC. The target signal(b) and Rx signal(c) are a little higher after 5usec delay when Tx TC=target TC. Signals are lower when Tx TC is 1/5 of target TC. Tx(a)mono coil maybe not usable when Tx TC=target TC. Maybe doing something wrong?

[bbsailor]

Trying to understand 5 times faster TC. Tried a simulation in spice. Left side, Tx TC=target TC. Right side, Tx TC=1/5 target TC. The target signal(b) and Rx signal(c) are a little higher after 5usec delay when Tx TC=target TC. Signals are lower when Tx TC is 1/5 of target TC. Tx(a)mono coil maybe not usable when Tx TC=target TC. Maybe doing something wrong?

Green,

Check out this web link at Eric Foster's forum.https://www.findmall.com/read.php?34...51#msg-1119751

Eric had posted a research paper (somewhere?) on his forum that stated that a coil discharge TC of one fifth the target TC would stimulate that target just as good as if the coil discharge TC was faster.

Your models may not incorporate this coil discharge TC data.

Another thing to consider is where on the TX current rise graph, the current is being turned off. If you turn off the current at 3 current rise TCs then the current is on the more horizontal portion of the current rise graph at about 95% and penetrates the target more. However, if you turn off the current at around 1 TC the current rise graph is more vertical and less efficiently stimulates the target due to cancellation of some of the charge current.

Joseph J. Rogowski

[green]

T-C = 0.072 x D x T x %IACS

Where T-C was in microsecs
D was edge length in mm
T was thickness in mm
%IACS is the conductivity relative to a figure of 100% for pure copper. from gold nugget simulation, Skippys reply#165

Target has L and R. I've been thinking D is L and T is R. Maybe not correct, any thoughts?

[bbsailor]

Trying to understand 5 times faster TC. Tried a simulation in spice. Left side, Tx TC=target TC. Right side, Tx TC=1/5 target TC. The target signal(b) and Rx signal(c) are a little higher after 5usec delay when Tx TC=target TC. Signals are lower when Tx TC is 1/5 of target TC. Tx(a)mono coil maybe not usable when Tx TC=target TC. Maybe doing something wrong?

Green,

The mono coil presents a problem with the discharge TC due to the fact that the op amp input resistor and clamping diodes put the input resistor (typically 1K ohm) in parallel with the damping resistor until the voltage falls below the diode on voltage (typically about 0.6V). Even a 1K ohm damping resistor would look like a 500 ohm resistor while the clamping diodes are conducting. A DD coil would not show this problem and allow some earlier sampling compared to a mono coil.

Joseph J. Rogowski

[waltr]

Green,

The mono coil presents a problem with the discharge TC due to the fact that the op amp input resistor and clamping diodes put the input resistor (typically 1K ohm) in parallel with the damping resistor until the voltage falls below the diode on voltage (typically about 0.6V). Even a 1K ohm damping resistor would look like a 500 ohm resistor while the clamping diodes are conducting. A DD coil would not show this problem and allow some earlier sampling compared to a mono coil.

Joseph J. Rogowski

Very good point that I had not thought of. Thanks.
I did find that concentric and DD coils do allow earlier first sample than a mono coil.