LT spice

[green]

Have tried to make a target that decays straight line decay log-log. Got closer. Is there a better way? What do I need to do to make the slope less steep?

[Davor]

You are very close. Try my viscous model: https://www.geotech1.com/forums/show...276#post195276

[Davor]

Actually, scroll up to one post before where PI excitation and log-log scale results are presented.

[Skippy]

It seems that Mr. Green has chosen a constant multiplier for his R value ratio, namely square-root of 10, approximated as 3.3 in his spice model. It would be better to use a more accurate figure of 3.16 : 1 . Can Spice accept any value for resistors? Failing that, 47 : 15 is a pretty good approximation, so 47R, 15R, 4R7, 1R5 etc.
Davors model ( in the linked thread ) seems to aim for cube-root of 100 = 4.64 as its multiplier, but misses out some values and isn't consistent. Would it work better with the series 47R, 10R, 2.2R , 0.047R, 0.01R, etc.?
The more elements in the model, the more you can extend the 'length' of your straight-line, presumably ?

[waltr]

The more elements in the model, the more you can extend the 'length' of your straight-line, presumably ?

That was my thought also since mineralized ground is generally an infinite number of 'particles' that also extent way past the edges of the coil excitation.

[Davor]

The more elements in the model, the more you can extend the 'length' of your straight-line, presumably ?

Yes. The model is not very sensitive to component values, so standard ones from E6 series are too good. Picking only values in a series that are roughly 1:2 does the trick; 1 - 2.2 - 4.7 - 10 - etc.

[green]

Davor, thanks for the spice model. Recorded ground, US clad quarter, and 1oz copper coin decay with 200us and 5ms constant rate TX. Wanting to make a spice model for ground, clad quarter and 1oz copper coin to see if spice decay changes slope similar to what I measure.

[Davor]

You should be able to tune the model for the purpose. I was happy with the values already there, but I did not have a device like yours to check it against nature. Only some trade papers with mostly obscured details of the measurement rig.

You should also be able to see the effects of excitation duration to the response, in real sample and the model. When I played with this model, I concluded that in x/t response, as reported in literature, the natural value of x is 1 or very close to 1, but the variation of reported x values is chiefly due to the measurement apparatus, and mostly the excitation period. The difference in inclination of slopes for 5ms vs 200us should become even more dramatic if you shorten the excitation to less than the Rx duration time. You should see even soil response end up in ski-shape for short excitation periods.

[green]

Changed R values from multiples of .33 to .316 suggested by Skippy. Small change but better. Charted ground at some different Tx on times. Changed scale to see decay out to 200us, noisier but maybe ok. Looks like 100us Tx doesn't chart straight line decay log-log out to 200us. Like Davor replied, shorter the Tx time the steeper the decay curve. Changed spice Tx drive to a current source to match my bench circuit. Wondering how to chart the multiple Tx times on one chart in spice like the measured chart. Haven't figured how to change the target so the decay slope matches the quarter and copper coin yet.

Changing the Tx on time in spice seems to change the slope similar to the measured data. Wanting to chart the different Tx on times on one chart if possible.

[Davor]

Just add a directive (you may add a new line by pressing ctrl+M):
.param rise 3200u
.step param rise list 3200u 200u 100u

- note that the step param list always sorts the stepped parameters from the smallest to the largest, so the execution sequence is 100u, 200u, 3200u

And edit current source statement to:
PULSE(0 1 {3200u-rise} {rise} 0 10n 4200u 1)

I did not fix the effective coil voltage with this edit, but you'll manage

[green]

Just add a directive (you may add a new line by pressing ctrl+M):
.param rise 3200u
.step param rise list 3200u 200u 100u

- note that the step param list always sorts the stepped parameters from the smallest to the largest, so the execution sequence is 100u, 200u, 3200u

And edit current source statement to:
PULSE(0 1 {3200u-rise} {rise} 0 10n 4200u 1)

I did not fix the effective coil voltage with this edit, but you'll manage

Thanks for the help. Step rise works for me but I'm lost on how to step the start stop times to match the rise times. Should have changed V1 voltage to 8V in zip files reply #9.

[Davor]

But I already sent you the source statement that does it.
When you right-click the current source, you set the Tdelay[s] value to {3200u-rise}, what you get is that the current starts ramping delayed, so that the fall of all 3 rise times is synchronised. You may also right click the text statement below the current source, and paste the command I sent previously:
PULSE(0 1 {3200u-rise} {rise} 0 10n 4200u 1)

[Davor]

Like this:

[green]

Like this:

Thanks again, does just what I wanted. With spice and computers in general I can sometimes be as dumb as a rock. I got on this site to learn something new. I'll try to figure out what your suggestion does.

[green]

California ground chart and three different spice targets. All 100u, 800u and 3200u constant rate Tx. All fairly close. Any thoughts on a spice target for the quarter or copper coin reply #7? Thanks again Davor.