SPICE Simulation Challenge

[Locator]

Thank you for your replies!

[waltr]

Does anyone know the Iron time constant for pulse induction metal detector?

Thank's in advance!

Like all targets it depends on the SIZE of the iron.
Smaller targets have lower TC. Example: A US silver quarter has a higher TC than a US silver Dime, even though both are the same metal.

Also, for non-round targets it depends on the orientation. A nail on end has a much different TC than on its side.

This is the main reason Discrimination can not be done by TC.

[Qiaozhi]

Does anyone know the Iron time constant for pulse induction metal detector?

Thank's in advance!

A metal target's time constant depends on more than what it's made of. The time constant is also affected by the target size, shape and thickness.
Asking for the time constant of iron is like asking "How long is a piece of string?".

[Locator]

Does anyone know the Iron time constant for pulse induction metal detector?

Thank's in advance!

[green]

The reason for offset to voltmeter, lower right of schematic https://www.geotech1.com/forums/atta...3&d=1556831504 adjust amplifier out to less than .5mV(no target) before recording data.

Unipolar might help if I was charting integrator out. Charting amplifier out and logout from amplifier. No EF sample.

Bipolar might help not Unipolar. Brain not working today.

[green]

I'm think it would. Log amplitude is very sensitive to offset. Unipolar pulses response has offset removed by means of a late EF sample, but in a case of ferrous targets that have less steep decay, their tail is knocked down by EF pulse, effectively introducing a negative offset.

The reason for offset to voltmeter, lower right of schematic https://www.geotech1.com/forums/atta...3&d=1556831504 adjust amplifier out to less than .5mV(no target) before recording data.

Unipolar might help if I was charting integrator out. Charting amplifier out and logout from amplifier. No EF sample.

[Davor]

no. don't know if bipolar Tx would improve charting target decay curve

I'm think it would. Log amplitude is very sensitive to offset. Unipolar pulses response has offset removed by means of a late EF sample, but in a case of ferrous targets that have less steep decay, their tail is knocked down by EF pulse, effectively introducing a negative offset.

[green]

Charted the 1oz copper coin and the screw with the smaller coil to get decay out to 600us.

TX has to be 5 times greater than the higher of the target tau and the coil tau to see no more pulse width effect.Carl's reply from https://www.geotech1.com/forums/show...977#post253977

Was thinking tau is infinite with constant rate Tx. Recorded the coin, screw and a US nickel at 100us, 1ms and 10ms Tx on times. Thinking the US nickel might chart three different lines but it didn't.
The deck screw didn't chart as straight a line log-log out to 300us with the smaller coil.
Target was placed on coil and recorded at each Tx time without moving target.

[green]

Green, where can I find the scheme for your test circuit?
Have you tried a bipolar tx?
How do you know that your targets are being fully saturated/ stimulated?
I assume that you are testing with various ranges of pulse widths and pps?

replies #113 and #115
no. don't know if bipolar Tx would improve charting target decay curve
I guess I don't know. Think if Rx doesn't change when increasing Tx on time it could be fully stimulated?
Tx: constant rate 1A peak or constant current .5A peak, on time adjustable from 50us to 50ms, about 1 pulse/sec

https://www.geotech1.com/forums/showthread.php?24294-Target-response-tester&p=239255#post239255

[dbanner]

Ah! Got it.
I'm still contemplating on more serious involvement in PI. Part of my reluctance is in inability to separate the influence of measurement from the phenomenon, so many odd things creep in and spoil the very reasoning of what is going on.

Maybe you can have a look at KingJl's tx front end. It looks very interesting as a platform for establishing a microprocessor based pulse machine.

[dbanner]

Green, where can I find the scheme for your test circuit?
Have you tried a bipolar tx?
How do you know that your targets are being fully saturated/ stimulated?
I assume that you are testing with various ranges of pulse widths and pps?

[green]

Ah! Got it.
I'm still contemplating on more serious involvement in PI. Part of my reluctance is in inability to separate the influence of measurement from the phenomenon, so many odd things creep in and spoil the very reasoning of what is going on.

One of the reasons I built the target response tester was to try and understand what effects and why. Recorded the coin and screw at 100 and 250us Tx time to see what effect Tx time has. Both targets have near straight line decay log-log Tx off. Shorter Tx time decayed faster. The two targets chart close to the same. Need to try with the smaller coil, higher signal to decay past 500us.

Tried the mono coil again since the ferrous and non ferrous decay looked the same(ferrous_non ferrous_3), wondered if I recorded the same target twice. No difference, think I've seen a thread with scope pictures showing a difference. Anyone know which thread or what I might try?

[Davor]

Ah! Got it.
I'm still contemplating on more serious involvement in PI. Part of my reluctance is in inability to separate the influence of measurement from the phenomenon, so many odd things creep in and spoil the very reasoning of what is going on.

[green]

Actually you need a Tx on time at least as long as the observed Rx period. Otherwise there is a steeper decay observed at Rx past the Tx on time (delay).
Perhaps you should also add a fast recovery diode (MUR...) to "hide" the MOSFET capacitance from Rx, and a series resistor of a few ohms to limit the current.
Unfortunately I did not follow the target response tester topic, but from what I can see, the schematic is nice Log amp and all.

Schematic is correct except I'm using the other log out off pin 12 Wrong. The schematic shows a MUR460 diode in series with Tx(erased). Tried both ways. Used without for awhile but has been included for a long time. The current is in a control loop, constant rate 1A peak or constant current .5A peak. Tx on time adjustable. To get a Rx signal during Tx, Tx current needs to be changing. Best if changing at a constant rate.

[Davor]

Actually you need a Tx on time at least as long as the observed Rx period. Otherwise there is a steeper decay observed at Rx past the Tx on time (delay).
Perhaps you should also add a fast recovery diode (MUR...) to "hide" the MOSFET capacitance from Rx, and a series resistor of a few ohms to limit the current.
Unfortunately I did not follow the target response tester topic, but from what I can see, the schematic is nice Log amp and all.