PI holes

[Qiaozhi]

I'll try to figure out how to post the XL files.

Compress the xls file to zip format, and then click on "Go Advanced" to get access to the "Manage Attachments" button.

[green]

Zero time is fet gate off. Let me know if there any questions.

[Qiaozhi]

I don't know if any of you are already aware of this, but LTSpice has support for a behavioural inductance model. You simply need to remove the inductance value, and add this to Spice Line 1: flux=L*Isat*tanh(x/Isat)
where L=the inductance, Isat=the magnetic saturation flux density, and x is a built-in variable that equals the current through the coil.
For example: If you define Spice Line 1 as: flux=300u*10*tanh(x/10), it will give the same results as a standard 300uH inductor. Remember to also add the parallel capacitance and series resistance.

You can try lowering Isat, and note how this affects the results as the core starts to saturate.

I was wondering whether this might be useful in target modelling, since a metal target essentially acts as a core to the detector coil. However, it seems that LTSpice assumes a tight coupling to the core, so I suspect this approach will be of little use.

Finally, LTSpice also has a built-in core model (John Chan, et al, 1991) which allows the addition of several parameters:
Hc - coercive force (Am-turns/meter)
Br - remnant flux density (Tesla)
Bs - saturation flux density (Tesla)
Lm - magnetic length [excluding gap] (meter)
Lg - length of gap (meter)
A - cross-sectional area (meter^2)
N - number of turns (-)

Might be of some use.

[Davor]

At a time I devised two more useful models, varicoil and a vari-K-xformer. Varicoil is a device that responds as an inductor with voltage controlled inductance, while vari-K-xformer has voltage controlled coupling - quite useful for modelling moving targets. Neither is built with inductors that can be coupled with other inductors. So far I did not make a model that does both, but there is a need for it when modelling ferrous targets.

Want them?

[green]

Wondering if anyone came up with a slope of -1 from the zip data in reply #47. The data has more noise than I would like, some of it due to scope resolution. I have some clay from the yard and lava rock from a volcano from somewhere that plot the same slope. Thinking about plotting the output of a integrator (gain controlled by percent time on). I'm thinking if delay time = sample time the integrator output should be the same at any delay setting if the slope is -1. Trying to get better numbers at longer delay times. Is the circuit causing the slope difference or my measurement. Any suggestions on how I should try?

[Silver Dollar]

I remember reading some posts by Eric Foster on his web site where he found the number
was not -1 too. I thought he got -1.2 but it might have been -1.3. He had some explantion
for it too. I'll have to poke around later and see if I can find that info...

[green]

Was looking at integrator output today, I have two different ferrite beads, 1 inch OD x 5/8 inch ID x 1/2 inch wide and 11 mm OD x 1 mm ID x 5 mm long. When plotting amplifier out for the ground curves with the scope I couldn't see the ferrite. Today looking at the integrator out I could. clay_-1600, lava rock_-550, large ferrite_80, small ferrite_8.The ferrite changed the opposite polarity of ground and other targets. Does that make sense? The coil is wrapped on a 1.5 inch card board tube. I used 35mm film containers for the clay and crushed lava rock. The delay and sample time was about 10 usec. They weren't the same due to some delays. Have to make another circuit to plot the ground response with = delay and sample times.

[Davor]

Your explanation is a bit confusing. Regarding noise, I've put the values in excel, but not as a scatter, and some oscillation emerged. XY scatter is nice because of log log scale, but otherwise it may prove a bit stealthy on simple oscillation.

[green]

Was looking at integrator output today, I have two different ferrite beads, 1 inch OD x 5/8 inch ID x 1/2 inch wide and 11 mm OD x 1 mm ID x 5 mm long. When plotting amplifier out for the ground curves with the scope I couldn't see the ferrite. Today looking at the integrator out I could. clay_-1600, lava rock_-550, large ferrite_80, small ferrite_8.The ferrite changed the opposite polarity of ground and other targets. Does that make sense? The coil is wrapped on a 1.5 inch card board tube. I used 35mm film containers for the clay and crushed lava rock. The delay and sample time was about 10 usec. They weren't the same due to some delays. Have to make another circuit to plot the ground response with = delay and sample times.

Tried the ferrite again with sample and delay close to the same at 5, 10, 20, 40 usec. Got the same result as above. The voltage change in integrator out is the opposite polarity of other targets with ferrite. Any idea why? I read a reply the other day stating a target signal would only increase with a mono coil.

[Davor]

Proximity to a high mu half plain (a big chunk of ferrite) should alter (increase) the coil inductance and should affect coil damping. Is it possible that the response is overshooting like underdamped coil, so that the response is coming from below? I'd expect a faster recovery though.
I've seen some discrimination attempts that (ab)used this effect before.

[Qiaozhi]

Tried the ferrite again with sample and delay close to the same at 5, 10, 20, 40 usec. Got the same result as above. The voltage change in integrator out is the opposite polarity of other targets with ferrite. Any idea why? I read a reply the other day stating a target signal would only increase with a mono coil.

If your ferrite target is close to the coil, it's possible [depending on the target's physical size and material) for it to become magnetically saturated, which would cause the coil inductance to be reduced. The result would be a faster decay.