DEEPER PI DETECTION DEPTH

[Tinkerer]

I haven't thought this through, but there seems to be a fundamental difference between this "half-cycle" current pulse, and a traditional PI pulse which starts with a steady-state current and then suddenly switches off.

If you look at the di/dt of these two pulses, they are very different, are they not? I am wondering if the "half-cycle" works against itself by "unringing the bell" right after it strikes it. In other words, the di/dt has a quick "counter-pulse" -- could this partially cancel its effect on slower targets?

Just wondering... I guess we need to put it through the simulator.

-SB

Using the usual resistive 200us and 15us TC targets.

[simonbaker]

Using the usual resistive 200us and 15us TC targets.

Nice - can we compare this with a traditional pulse on the same graph?

-SB

[Aziz]

I haven't thought this through, but there seems to be a fundamental difference between this "half-cycle" current pulse, and a traditional PI pulse which starts with a steady-state current and then suddenly switches off.

If you look at the di/dt of these two pulses, they are very different, are they not? I am wondering if the "half-cycle" works against itself by "unringing the bell" right after it strikes it. In other words, the di/dt has a quick "counter-pulse" -- could this partially cancel its effect on slower targets?

Just wondering... I guess we need to put it through the simulator.

-SB

What is a "slower target" mate?

Again: only dI/dt matters!
The half sine current pulse has it's maximum dI/dt at the beginning and at the end.
Both generate a wide band pulse. The only difference is the frequency response.

But the half sine current pulse has more advantages over the traditional PI pulse.

Aziz

[simonbaker]

Using the usual resistive 200us and 15us TC targets.

Hey Tinkerer, can you post the LTSpice file?

Regards,

-SB

[Tinkerer]

Hey Tinkerer, can you post the LTSpice file?

Regards,

-SB

Here is the zip with the traditional PI and the test_PI. I think that to put the result on the same graph, you need to re-draw one of the asc files onto the other one.

[Davor]

What is a "slower target" mate?

Again: only dI/dt matters!

I'm afraid he just might be right. Responses to cosine edges (dI/dt) do tend to cancel each other for taus much longer than the stimulus duration, much like a gradient microphone does. So this idea could use some more thinking over. I think there is a potential in it, heck, gradient mikes are not bad at all. Maybe it will be perfect for gold - who knows?

I am still much more in favour of voltage step, but somehow got carried away with this. I've even put some flash around the bones, and not just one but two versions. Instead of a H-bridge which could prove better but much more complicated, I put just another coil to do the same thing in opposite direction with another diode. Again no stress to any of the components, and only plain vanilla parts. Simple. As MOSFET gate floats with voltage change, it required some exotic approach, and I assembled both opto coupler, and gate transformer. Since good transformation is obtained with somewhat too large inductivity to be found in normal stores, I made a simple solution with optos. Works as the concept or better, but I'll have to convince myself that there is any merit to it. Single coil is a tempting possibility though.
Here it goes...

[Aziz]

Half Sine/Cosine Pulse FFT

Hi all,

so what's the difference between a half sine and half cosine current pulse?
See below in the order half sine and half cosine:
Color cyan represents the FFT magnitude response.

Cheers,
Aziz

[Davor]

There is a distinct phase traversal at the middle which will have some saying in choice of a pulse duration in case you are into discrimination.

I guess we'll have to stick to a single convention here as to what we observe and why. I assumed a voltage as a more relevant (dI/dt and constant ampere-turns), but for all practical purposes - it is the same. I missed to include a library of a high voltage pnp in the schematics, so here it goes...

[Carl-NC]

so what's the difference between a half sine and half cosine current pulse?

The half-cosine has discontinuities which results in higher broadband FD response. Interestingly, a half-cosine current has a half-sine voltage response, and vice-versa, so when you look at the FFT are you looking at the voltage or current? Does it matter?

In the far more interesting time domain, half-cosine has the advantage of a PI-like discontinuity; half-sine has the advantage of a di/dt=0 that is useful for GB. Personally, I would combine them.

- Carl

[Aziz]

The half-cosine has discontinuities which results in higher broadband FD response. Interestingly, a half-cosine current has a half-sine voltage response, and vice-versa, so when you look at the FFT are you looking at the voltage or current? Does it matter?

In the far more interesting time domain, half-cosine has the advantage of a PI-like discontinuity; half-sine has the advantage of a di/dt=0 that is useful for GB. Personally, I would combine them.

- Carl

Hi Carl,

it doesn't matter, what you apply: current or voltage. The FFT isn't complaining at all. The current pulse is meant of course.

Well, the half-cosine pulse is DC free and has +6 dB more FD response due to +Max .. -Max range compared to the half-sine response (in the example +Max .. 0). If I would make a half-sine pulse over +Max .. -Max range, the whole half-sine FD response is shifted by +6 dB only.
But I like the wide discontinues of the half-cosine current pulse (wide band characteristics better than a half-sine pulse).

"half-sine has the advantage of a di/dt=0 that is useful for GB"

This requires an induction balance (IB) coil configuration (obviously). Unfortunately, the mono coil can't sample this in the half-sine pulse variant.

Cheers,
Aziz

[Davor]

I think it is just a matter of how you want to interpret the result, and by what means. In case of monocoil, your options are very limited.

So, being in a PI end of the world, let’s pretend that it is a monocoil we wish to cherish - no matter what. In such a case we need a strong di/dt pulse (voltage spike) followed by instantaneous recovery, and almost immediate Rx sampling. If humanly possible, we also wish to have all transitions free from stress against the switching electronics as they never bode well in pursuit of faint signals e.g. deep treasure.

I gave it a bit thought and it may all be achieved by stretching this hammer-and-mallet energy preserving rig a little bit. If I make a first half up to the current maximum 10 times slower than the rest of it, the resulting di/dt at the very end of it will result in 10 times di/dt of the starting one, and there will be no need for high supply voltage. A whole rig will become a neatly behaving flyback transformer with precisely defined voltage peak, no stress against the switching electronics, no stray charges, no ringing and no fuss. And low impedance all over. And no front against back signal cancelling. And energy conservation. And cold transistors too... hope no one ordered parts for the previous circuit - this becomes even more interesting.

Regarding the Rx part... Can anyone confirm to me if there was any attempt on chopper-type Rx sampling with PI before? I constantly bump to various patents (which are mostly useless if you really want to learn something due to poor technical descriptions). I think I am onto something. I believe there is a quite simple way of converting a PI Rx sample to an arbitrary VLF representation and use some of the tau->phase conversion to obtain a simple rig with multi frequency VLF advantages, and also a monocoil operation. That would be a three course lunch in a single pill. A true PI-VLF-Frankenmetaldetector.

[Carl-NC]

it doesn't matter, what you apply: current or voltage. The FFT isn't complaining at all. The current pulse is meant of course.

Well, the half-cosine pulse is DC free and ... I like the wide discontinues of the half-cosine current pulse (wide band characteristics better than a half-sine pulse).

And this is where it matters. The half-cosine current results in a... half-sine voltage! So it is a half-sine voltage you will be looking at in the receiver, if you were to choose to look at it at all.

This requires an induction balance (IB) coil configuration (obviously). Unfortunately, the mono coil can't sample this in the half-sine pulse variant.

The mono coil can't sample the half-cosine pulse either. Ergo, I don't know why you would bother with either one if you are just going to toss the extra information in the garbage. Just stick with plain ol' PI pulses.

Personally, I don't understand why anyone would be averse to using an IB coil if it gives you more information. And earlier sampling. And easier power supply design. And better sensitivity.

- Carl

[simonbaker]

Personally, I don't understand why anyone would be averse to using an IB coil if it gives you more information. And earlier sampling. And easier power supply design. And better sensitivity.

- Carl

Mono coil is just so much easier to build that if we can somehow get it to work well, great for DIY'ers!

Just another challenge to keep working on...

-SB

[simonbaker]

puzzle

Here is a not-so-hard puzzle, because I know some people here like riddles...

I think it also illustrates some pitfalls to modeling physical reality with spice circuits, and maybe reopens some discussion concerning targets and previous simulations.

The attached LTSpice circuit is an idealized PI pulse hitting two targets. Both targets happen to have the same "tau" (L/R). However, the current waveform in one target is 1000 times bigger than in the other, otherwise the same shape.

Also we have a two identical simulated Rx coils, one for each target. But... the responses in the Rx coils are identical even though one target has 1000 times the amplitude response than the other.

Holy cow... what gives?

The answer should make us go back and question some recent simulations for realism, or at least discuss them again.

Cheers!

-SB

[Davor]

That is the whole idea of discrimination - should work the same for objects of identical tau.

BTW, it is irrelevant how much inductivity your targets are, because it is always a short circuit. There is a circular current flow wether you separate it into many short-circuited wires or a single ring - what matters is ampere turns and tau. Tau is a related to L/R and additionally complicated by skin effect, but that's it. With L/R=const. you may scale L however you want. The targets you depicted in a puzzle are basically the same. The amount of ampere turns in spice is related only to coupling.

I personally prefer "normal" R values - makes things less awkward