DEEPER PI DETECTION DEPTH

[Midas]

There is no room for misinterpretation. It is a PI transmitter.
Pulse Induction:
In terms of a pulse, yes, it is a current pulse (wave form need not be a rectangular form). This is given by the pulse timing: switching the coil on, switching the coil off for a specified duration of a time (single pulse width timing).
Wideband characteristics: Yes, it differs from the single frequency VLF system. It is a wide bandwidth current pulse like the conventional PI. The current pulse is exponential during the charging and recycling phase and cosine during the off period.

By just replacing the coil's tuning capacitor with the damping resistor, it's becoming a conventional PI transmitter. Well, it is a versatile PI transmitter with more benefits than disadvantages. This is the reason, why I like this TEM transmitter.

Aziz

OK then, let the time wasting begin.

There is a pulse within the circuit for sure, the LC tank IS being pulsed. But surely the pulse in 'Pulse Induction' refers to what is happening to the coil not to a circuit block that contains the coil.

Here's the definition of pulse:
3. Physics a. A brief sudden change in a normally constant quantity: a pulse of current; a pulse of radiation.
b. Any of a series of intermittent occurrences characterized by a brief sudden change in a quantity.

The current flow is never a constant quantity in this signal. You might be able to argue that you are pulsing the rate of change but I really don't think that's in the spirit of the definition. I believe it was called a PI because it pulsed the current to the coil intermittently, to distinguish it from other detectors that continuously varied the current.

Another way to look at it is that coil isn't being pulsed because it never stops being driven its simply swapping its drive source back and forth from the mosfet to the capacitor.

I understand what you saying in that its wide band characteristics make it more similar to a PI than a VLF but, IMHO, that's not enough to force it into that category.

[Aziz]

I realize arguments over terminology can be somewhat of a waste of time but surely this isn't PI by any sense of the word. Coil current is constantly flowing, there is no quiet time, so its really a waveform not a pulse. I still like the idea though. Very similar to how one might build an induction furnace and I suspect may be a key feature of the Nexus metal detectors.

Anyway my point is that Tinkerers idea of charging a capacitor (with, I assume, a circuit that doesn't require current to constantly flow through the search coil) and discharging it into low inductance coil is really quite different and may be worth looking at further. How about commandeering some automotive technology to speed development up a little, ie. using a capacitive discharge ignition modules?

There is no room for misinterpretation. It is a PI transmitter.
Pulse Induction:
In terms of a pulse, yes, it is a current pulse (wave form need not be a rectangular form). This is given by the pulse timing: switching the coil on, switching the coil off for a specified duration of a time (single pulse width timing).
Wideband characteristics: Yes, it differs from the single frequency VLF system. It is a wide bandwidth current pulse like the conventional PI. The current pulse is exponential during the charging and recycling phase and cosine during the off period.

By just replacing the coil's tuning capacitor with the damping resistor, it's becoming a conventional PI transmitter. Well, it is a versatile PI transmitter with more benefits than disadvantages. This is the reason, why I like this TEM transmitter.

Aziz

[simonbaker]

I realize arguments over terminology can be somewhat of a waste of time but surely this isn't PI by any sense of the word. Coil current is constantly flowing, there is no quiet time, so its really a waveform not a pulse. I still like the idea though. Very similar to how one might build an induction furnace and I suspect may be a key feature of the Nexus metal detectors.

Anyway my point is that Tinkerers idea of charging a capacitor (with, I assume, a circuit that doesn't require current to constantly flow through the search coil) and discharging it into low inductance coil is really quite different and may be worth looking at further. How about commandeering some automotive technology to speed development up a little, ie. using a capacitive discharge ignition modules?

Now we're talking... and it will double as a radar jamming system as well!

[Midas]

Hi Tinkerer,

the PI TEM transmitter we have discussed earlier does it already.

Well, it's a so-called "synchronous step-up converter" and charges the capacitor with high voltage, which is rapidly discharged in the coil then.
Oh yes, it recycles the energy in the coil then. Why we should waste battery power?
Efficiency: Can be made up to 95% (or even higher).

Aziz

I realize arguments over terminology can be somewhat of a waste of time but surely this isn't PI by any sense of the word. Coil current is constantly flowing, there is no quiet time, so its really a waveform not a pulse. I still like the idea though. Very similar to how one might build an induction furnace and I suspect may be a key feature of the Nexus metal detectors.

Anyway my point is that Tinkerers idea of charging a capacitor (with, I assume, a circuit that doesn't require current to constantly flow through the search coil) and discharging it into low inductance coil is really quite different and may be worth looking at further. How about commandeering some automotive technology to speed development up a little, ie. using a capacitive discharge ignition modules?

[simonbaker]

All ideas are welcome. By looking at them we can learn something.

What you propose is basically a Bucking coil. There are several ways to implement a Bucking coil, but the basic idea is to generate an identical waveform, 180 degrees different and then subtract.
An IB_PI

Tinkerer

Interesting perspective. Usually a bucking coil works by canceling the net magnetic field at the RX coil. My idea is purely electrical signal subtraction as you noted.

Cheers,

-SB

[Tinkerer]

A while back I had a kooky idea that go no response. But I'd like to see it tried just for fun. This is for mono-coil PI designs, but could be modified for balanced TX/RX designs probably.

The idea is to have a "dummy" TX coil circuit exactly like the real TX coil. The dummy TX coil is wound very small so it doesn't emanate much field.

Then fire both coils at once, and subtract the signal of the dummy TX coil from the real TX coil. You have to carefully tune the coils so they both have the same voltage profile.

By subtracting the signals, you don't need the diode/waiting period to start looking at the target signal, since the difference signal will be small enough to not overload your amplifier section for the whole time period.

I think I made a crude simulation of the idea once and it seemed to work as expected, but I'm not wise in the ways of PI circuits.

The downside is wasted energy in the dummy TX coil, but maybe with energy recovery designs that could be minimized.

Just my demented 2 cents!

-SB

All ideas are welcome. By looking at them we can learn something.

What you propose is basically a Bucking coil. There are several ways to implement a Bucking coil, but the basic idea is to generate an identical waveform, 180 degrees different and then subtract.
An IB_PI

Tinkerer

[simonbaker]

Increasing the TX voltage pushes more amps and energy into the coil in a shorter time. However, this is not where the limitation of the traditional PI lies.

The traditional PI is limited by the Flyback voltage. If we take the 1000V flyback voltage as mentioned in the thread, you will notice that this voltage is reached very quickly.

Another problem with traditional PI, is the Flyback decay. The more energy you got in the coil, the longer it takes for the Flyback decay to reach the level where you can sample.

This means that you can not detect very small targets anymore.

Attached is a LTSpice simulation to play with. It is quite accurate and also has 5 simulations of different size target.

You need to download the free LTSpice at: http://www.linear.com/designtools/software/#LTspice

Tinkerer

A while back I had a kooky idea that go no response. But I'd like to see it tried just for fun. This is for mono-coil PI designs, but could be modified for balanced TX/RX designs probably.

The idea is to have a "dummy" TX coil circuit exactly like the real TX coil. The dummy TX coil is wound very small so it doesn't emanate much field.

Then fire both coils at once, and subtract the signal of the dummy TX coil from the real TX coil. You have to carefully tune the coils so they both have the same voltage profile.

By subtracting the signals, you don't need the diode/waiting period to start looking at the target signal, since the difference signal will be small enough to not overload your amplifier section for the whole time period.

I think I made a crude simulation of the idea once and it seemed to work as expected, but I'm not wise in the ways of PI circuits.

The downside is wasted energy in the dummy TX coil, but maybe with energy recovery designs that could be minimized.

Just my demented 2 cents!

-SB

[Tinkerer]

Yes we can improve the PI design a lot for more deph with a easy method:
- switch between 12V to 24 V TX voltage....
- give more stability to the signal reception to "mix" 3 closed Frecuency (eJ: 515, 540, 593 PPS ) to cancel each other interferences and to forget the infamous potentiometer FCY adjusts of the majority of Classic PI
- for the battery ....still use the heavy but proved Gelcel battery7 amps 12V like the Lorenz...and PulstarII PI

Increasing the TX voltage pushes more amps and energy into the coil in a shorter time. However, this is not where the limitation of the traditional PI lies.

The traditional PI is limited by the Flyback voltage. If we take the 1000V flyback voltage as mentioned in the thread, you will notice that this voltage is reached very quickly.

Another problem with traditional PI, is the Flyback decay. The more energy you got in the coil, the longer it takes for the Flyback decay to reach the level where you can sample.

This means that you can not detect very small targets anymore.

Attached is a LTSpice simulation to play with. It is quite accurate and also has 5 simulations of different size target.

You need to download the free LTSpice at: http://www.linear.com/designtools/software/#LTspice

Tinkerer

[Aziz]

What are you guys smoking? I want to smoke the same $hit.

Did you take the parts with the parasitic features in your sims (resistance, capacitance, manufacturers spice model, ...)?

If you double the current in the TX circuit, the power loss
- will rise four fold on ohmic resistors (coil, coil leads, capacitors, mosfets, battery, inductors, ..)
- will rise two times on ideal diodes (constant voltage drop, but this is dependent on current flow -> power loss will be higher than two times usually)

The fact is, if you want to increase the eddy current in the target, you have to increase the TX circuit's current. Then you will have lost's of losses on the real parts.

It's always a compromise between a reasonable power consumption and target stimulation strength. If you start to saturate the TX coil current (which happens easily with lower inductivity coils), then forget the efficiency. You will contribute to the global earth warming.

Just my thoughts.
Aziz

PS: To have a controlable wide bandwidth response will give you more benefits.

[Tinkerer]

This is not the most efficient realisation of a target stimulation or power saving ... it is missing one component ( maybe deliberately ??? )

The circuit with the diode provides double the magnetic power and double the efficiency .. do the sims yourself or build real circuit. ( watch out ... circulating currents are so high in the coil that the capacitor gets hot )

The pulse repetition should be at the resonant frequency of the LC tank.

The duty cycle of the pulse will control peak current.

moodz

[ATTACH]17773[/ATTACH]

I simulated the circuit with the MUR460 and with a Schottky diode.
Amazing what a difference it makes.

The target response is indeed much higher. The response wave form is also much different.

I wonder which method gives more information for discrimination/ target ID, about the target?

Tinkerer

[Tinkerer]

This is not the most efficient realisation of a target stimulation or power saving ... it is missing one component ( maybe deliberately ??? )

The circuit with the diode provides double the magnetic power and double the efficiency .. do the sims yourself or build real circuit. ( watch out ... circulating currents are so high in the coil that the capacitor gets hot )

The pulse repetition should be at the resonant frequency of the LC tank.

The duty cycle of the pulse will control peak current.

moodz

[ATTACH]17773[/ATTACH]

Many variables to the basic circuit are possible.

For example, try an IGBT instead of the Mosfet and see how different the wave form turns out. (Caution)

As for the diode. It increases the resistance and therefore reduces the efficiency.

The diode also isolates the drain capacitance of the Mosfet. What is the influence of the drain capacitance? Is this capacitance in series? and therefore reduces the tank capacitance?

What is the capacitance of the diode itself? The resistance? the voltage drop? the speed?

A lot of things to check.

Tinkerer

[Aziz]

No! It's not better. A simple push-pull configuration on a LC tank achieves better results.

And I'm missing one important feature:
wide bandwidth response

Aziz

[Aziz]

This is not the most efficient realisation of a target stimulation or power saving ... it is missing one component ( maybe deliberately ??? )

The circuit with the diode provides double the magnetic power and double the efficiency .. do the sims yourself or build real circuit. ( watch out ... circulating currents are so high in the coil that the capacitor gets hot )

The pulse repetition should be at the resonant frequency of the LC tank.

The duty cycle of the pulse will control peak current.

moodz

[ATTACH]17773[/ATTACH]

Oh yes, this one should even be better.
Good one! I'll run the sims..

Aziz

[moodz]

I think it was one and a half year or so. Found the different phases of the transmitter now:

http://www.geotech1.com/forums/showpost.php?p=111998&postcount=819

Aziz

This is not the most efficient realisation of a target stimulation or power saving ... it is missing one component ( maybe deliberately ??? )

The circuit with the diode provides double the magnetic power and double the efficiency .. do the sims yourself or build real circuit. ( watch out ... circulating currents are so high in the coil that the capacitor gets hot )

The pulse repetition should be at the resonant frequency of the LC tank.

The duty cycle of the pulse will control peak current.

moodz

[Carl-NC]

How about charging a capacitor to 1000V and discharging it through a 100uH coil?

Now you're talking!