Best target simulation.

[moodz]

Hi all ... what is the best target simulation for LTspice.
I use a an inductance looped with a small resistance but does anyone have any good / realistic models or this OK ????
My simulation detects a 0.1 nH inductance with 1 ohm loop resistor. The main coil is 300 uH. Any comments ...

Regards,

Moodz.

[moodz]

Hi all ... what is the best target simulation for LTspice.
I use a an inductance looped with a small resistance but does anyone have any good / realistic models or this OK ????
My simulation detects a 0.1 nH inductance with 1 ohm loop resistor. The main coil is 300 uH. Any comments ...

Regards,

Moodz.

...anyone ?? ...

The target inductance is now 0.01 nH ( 10 pico Henry ) and 1 ohm resistor. Provides a 10 mV unamplified response AND some descrimination ...using complex direct pulse balancing .... I will test this on the weekend for real hopefully.

Here is a nice paper to read ( study and academic use only ) ...

tn7.pdf

Aziz and TInkerer will appreciate it

moodz

[moodz]

...anyone ?? ...

The target inductance is now 0.01 nH ( 10 pico Henry ) and 1 ohm resistor. Provides a 10 mV unamplified response AND some descrimination ...using complex direct pulse balancing .... I will test this on the weekend for real hopefully.

Here is a nice paper to read ( study and academic use only ) ...

[ATTACH]12641[/ATTACH]

Aziz and TInkerer will appreciate it

moodz

The maths for your musket balls and alum foil lies within ...

[Aziz]

Model Accuracy

Hi Moodz,

make a following parameter sweep in your spice model:

.step param list Ltgt 1n 10n 100n 1µ

Ltgt is the inductance of your target coil.

If your target response follows the following relationship:

Uindtgt direct proportional to -Ltgt*dI/dt,
where Uindtgt = Induced voltage of target coil
then the results might be accurate enough.

But generally a fine spice model takes too much time and hard disk space to calculate it. Also the gmin is generally at 1e-12 and your Ltgt=0.01 nH = 1e-11 H, so it doesn't make sense to go lower then 1 nH as the calculating errors will be higher (cumulating error).

Best to see the calculation error result is to make a copy of the schematics without the target response and compare both response outputs (in the same project).

Aziz

[simonbaker]

...anyone ?? ...

The target inductance is now 0.01 nH ( 10 pico Henry ) and 1 ohm resistor. Provides a 10 mV unamplified response AND some descrimination ...using complex direct pulse balancing .... I will test this on the weekend for real hopefully.

Here is a nice paper to read ( study and academic use only ) ...

[ATTACH]12641[/ATTACH]

Aziz and TInkerer will appreciate it

moodz

This statement on page 3 doesn't seem believable to me -- regardless of resistance, the initial current is the same? How does he get that? It almost seems like he left a term out of the solution. But I'm beyond rusty with E&M equations.

-SB

[mikebg]

SERIES OF TIMECONSTANTS

Hi all ... what is the best target simulation for LTspice.
I use a an inductance looped with a small resistance but does anyone have any good / realistic models or this OK ????
My simulation detects a 0.1 nH inductance with 1 ohm loop resistor. The main coil is 300 uH. Any comments ...

Hi Moodz,
Here I propose not the best SPICE model of the target, but it predicts experimental results with a coin at error below 40% in HF region. The coin is presented as two time constants with ratio 9:1, ie the second time constant is 9 times smaller than the first (fundamental) time constant. The mutual inductance between search coil and the second eddy loop is also 9 times smaller than to fundamental loop. For example if the fundamental time constant is TC1 = 3,6E-5s, then the second TC2 = 4E-6s and if linkage coefficient k1 = 9E-3, then the k2 = 1E-3. However, if the target is like gold nugget, the k2 should be larger than 1/9 of k1, for example one eighth to one seventh and appears influence of the third timeconstant.
For more accurate modeling, even at coin should be introduced third time constant, which is 25 times smaller, as shown in the attached equivalent diagram, but no simulation benefit from such complexity.

[moodz]

Thanks guys .... I will give all your tips a whirl. I may have been chasing an rainbow in the simulation error margin .. however the information is very useful for scaling the targets realistically and doing simulation runs.

Thanks again .... moodz