Detect a target with a TC less than mono coil TC

[Carl-NC]

Coils with separate Tx and Rx and mono coil with Rin going to -input should have one slope.

True, a mono coil with an inverting-mode preamp has one effective damping R all the way down.

[green]

If Mr. Greens latest simulation were changed to R_d = 430 R, and input protection resistor R = 1k0 , then he would have critical damping ( 300 R ) until the protection diode stopped conduction, and a modestly under-damped response thereafter.

Tried it with spice. Added R2 and V4 to cancel leakage current thru M1 so coil would decay closer to zero. Oscillates with 430R damping resistor. Probably wouldn't matter using an integrator with a few us average, wondering if there could be a problem taking a sample of amplifier out with an A-D during the oscillation.

[Skippy]

Looks like 'modestly under-damped' is not the way to go ! So presumably you'll have to work with: over-damped until the protection diode stops conducting, then critically damped?

Re: leakage: I don't know how it works in a Spice model, but in the real world, taking the gate of an enhancement-mode N-FET negative turns it off more. I assume this is easy to model, setting gate voltage source to ( -5 10 -5 etc.. ) would do it ?

[green]

Looks like 'modestly under-damped' is not the way to go ! So presumably you'll have to work with: over-damped until the protection diode stops conducting, then critically damped?

Re: leakage: I don't know how it works in a Spice model, but in the real world, taking the gate of an enhancement-mode N-FET negative turns it off more. I assume this is easy to model, setting gate voltage source to ( -5 10 -5 etc.. ) would do it ?

The decay signal looks a lot better critical damped with spice, a little underdamped with a real circuit doesn't seem to matter with what I've tried. Haven't tried sampling with an A-D and was wondering if oscillation could cause a problem.

Leakage in the simulation is actually thru D2 and V3(added for fet avalanche). Leakage in real circuit would be thru D1 to the voltage stored on the drain after coil decay.

[bbsailor]

This is an important discussion in order to form a good mental model for the relationship between coil TC and target TC. A different but related question to consider is the relationship between target TC and coil delay. The eddy currents in a target with a 2uS TC will decline to almost zero in 5 TCs or 10uS. Thus, a PI machine that operates at a 10uS delay will miss this 2uS TC target. Getting this coil down to about 7.5 uS delay will allow you to pick up the tail end of the 2uS target decaying eddy currents.

Here are the things to consider when planning a low delay PI system and coil.

Any capacitance seen by the mono coil will cause the coil to have more oscillations, and need a lower resistance damping resistor (Rd) but at the consequence of lengthening the coil discharge TC and delay. In your mental model, think about all the ways you can lower the capacitance as seen by the coil.
1. Coil turn to turn wire capacitance due to insulation dielectric and thickness.
2. Coil to shield capacitance due to spacing between the coil bundle and the shield, the area of the shield and the dielectric of the spacer materiel.
3. Preamp saturation time and how quickly it comes out of saturation.
4. MOSFET COSS capacitance and your ability to isolate the coil from it by using a series diode between the MOSFET and coil.
5. The coil wire size and type and minimizing the coil wire from holding any eddy currents at the minimal delay time.
6. Coil power and ability to maintain critical damping even if Rd gets a little warm (slightly raising the R value) due to using higher coil current.
7. When using a PI design that integrates many RX signals, coil size and sweep speed are important operational search considerations.
8. Coil style and size based on the search area soil type, minerals and moisture either fresh water or salt water.

As you get your PI design and coils operating below about 8uS to 10uS many things that do not interact at higher delays will start to interact more at lower delays.

Summary, lower TC targets will need PI circuit designs and coil designs that are capable of sensing the eddy currents in the targets before they fully decay.

Happy New Year of hunting low TC targets

Joseph J. Rogowski

[Carl-NC]

Getting back to the question of detecting a target TC lower than the coil TC... ain't gonna happen. Consider a 300uH coil that has a high enough SRF to use a 1k damping resistor. The coil tau is therefore 150ns. You might think that the natural coil response will decay in 5*tau but that's not nearly enough.

First, it is not a simple RL decay but rather a critically damped RLC decay, which has the form t*e^(-t/tau). Second, it drives a usually high-gain amplifier which must settle far enough to be well within its linear range. It would not be unusual that the coil voltage must settle from several hundred volts to a few mV. This will require more like 10-15 taus to settle, on the order of 2us for the example.

So a coil tau of 150ns isn't usable until 2us. And this doesn't include the opamp recovery delay. A 150ns target is long gone.

[green]

Getting back to the question of detecting a target TC lower than the coil TC... ain't gonna happen. Consider a 300uH coil that has a high enough SRF to use a 1k damping resistor. The coil tau is therefore 150ns. You might think that the natural coil response will decay in 5*tau but that's not nearly enough.

First, it is not a simple RL decay but rather a critically damped RLC decay, which has the form t*e^(-t/tau). Second, it drives a usually high-gain amplifier which must settle far enough to be well within its linear range. It would not be unusual that the coil voltage must settle from several hundred volts to a few mV. This will require more like 10-15 taus to settle, on the order of 2us for the example.

So a coil tau of 150ns isn't usable until 2us. And this doesn't include the opamp recovery delay. A 150ns target is long gone.

Thanks