Competition!

[Tinkerer]

I am always in mood to see new schematics, kinda type of "insanity", already have tens of GB of such documents.
But those are never enough!
So yes, post it.
At this moment I have here in my "collection" Barracuda, Pulse Star 2, probably few other simpler PI detectors.
Barracuda is nice little diy. Pulse Star 2 is "heavyweight" category, yet without GEB.
I was thinkig to figure out GEB circuit for both, also auto-zeroing circuit too.
Long story short; yes, do post what's on your mind.
Next question; describe in steps, using oscilloscope: hot to determine/measure accuratelly TC at targets?
I don't have digital one with textual values on the screen. I do have analog one, good though, brand new, bought it back in 2019. 2X20MHz.
With such oscilloscope you can see the signal fine even which those cheaper digital ones can't.
But it will need math to add to what you see on the display.
It would be good for all other layman members too; somebody to explain step by step such measurements and math.
You see flyback voltage, you see what's on RX frontend output, all in form of oscillograms.
But exact numbers you don't see on such simpler analog oscilloscopes.
You need to use simple math to figure out accurate TC.
So, one good explanation and step by step instruction is welcome here!

IVICA, I did the work for you. Here are the screenshots of some targets. I just slapped a 2 stage preamp together on a breadboard, quick and dirty.
The TX is a bipolar CCSQW, Continuous Current Square Wave
The coil is a DD coil, about 38cm diameter
The cycle is about 100us, that is about 10,000 Flybacks per second, 5000 positive and 5000 negative
The cursor shows the beginning of sampling at about 4.5us
Yellow trace is the TX Flyback
These are static or non motion tests

You can see that the Eddy currents decay faster or slower than the blue trace which is a US 5 cent, also called a NICKEL. The Nickel has a TC of about 10us. We use it often as reference.
The white trace is the NO Target reference.
Targets on picture 1
Red= 1 Gram gold bar
Green= very small gold ring, 0.35Grams
Blue= US Nickel

Picture 2
White trace=No target
Blue=Nickel
Red=a coin about the same size as the Nickel, made out of lead
Green, tiny high conductive coin from Netherlands Antillen
Turquoise= 26mm diameter silver coin

The pictures are not very good, first try, if there is enough interest, I will try to make better ones.
Since this is a DD coil, many iron targets give a positive response, opposed to the negative response you see with colored metals, at the same place under the coil.

[ivconic]

I have Barracuda.
I can do that. These days though, next few days will be busy!
I will shot the pics and post here.

[Tinkerer]

Analog scope is fine.
Do you have a PI where you can connect your scope probes at the following places:
Output to TX coil, so you can see the TX ON, OFF and Flyback
TX Mosfet gate and timing signals
Input to preamp
Output of preamp
Input and output of integrator

Then we can look at the signals without target and with different targets and I will explain what to look for.

I could also set up an LTSpice simulation to look see how close the simulation compares with the real signals.

[ivconic]

It might be useful to build a simple, basic PI and run some tests with it, looking at the decay curve and time constants of your desired targets.
Then you have a much better idea of what you really want.
All the targets you show, have relatively long time constants.
Silver and high carat gold usually has a long TC. Low carat gold is much more variable and often has a short TC.
A 3mm diameter lead bird shot has a very short TC. A gold nugget of similar size has a much longer TC.

If you desire a PI with the capability of detecting a TC of less that 1us, I can help you with a tested circuit, but it does not have ground balance.

I am always in mood to see new schematics, kinda type of "insanity", already have tens of GB of such documents.
But those are never enough!
So yes, post it.
At this moment I have here in my "collection" Barracuda, Pulse Star 2, probably few other simpler PI detectors.
Barracuda is nice little diy. Pulse Star 2 is "heavyweight" category, yet without GEB.
I was thinkig to figure out GEB circuit for both, also auto-zeroing circuit too.
Long story short; yes, do post what's on your mind.
Next question; describe in steps, using oscilloscope: hot to determine/measure accuratelly TC at targets?
I don't have digital one with textual values on the screen. I do have analog one, good though, brand new, bought it back in 2019. 2X20MHz.
With such oscilloscope you can see the signal fine even which those cheaper digital ones can't.
But it will need math to add to what you see on the display.
It would be good for all other layman members too; somebody to explain step by step such measurements and math.
You see flyback voltage, you see what's on RX frontend output, all in form of oscillograms.
But exact numbers you don't see on such simpler analog oscilloscopes.
You need to use simple math to figure out accurate TC.
So, one good explanation and step by step instruction is welcome here!

[Tinkerer]

3 time constants?
I don't know how relevant this is; with such detector and coil I will hunt only for described targets.
I am not interested in any different and larger targets then what I described already.

It might be useful to build a simple, basic PI and run some tests with it, looking at the decay curve and time constants of your desired targets.
Then you have a much better idea of what you really want.
All the targets you show, have relatively long time constants.
Silver and high carat gold usually has a long TC. Low carat gold is much more variable and often has a short TC.
A 3mm diameter lead bird shot has a very short TC. A gold nugget of similar size has a much longer TC.

If you desire a PI with the capability of detecting a TC of less that 1us, I can help you with a tested circuit, but it does not have ground balance.

[ivconic]

[bbsailor]

Joseph, I thought about it, what I could suggest as a common target that could be found easily anywhere in the world.
It seems that we have no choice but to cut squares of several dimensions with sheet metal scissors.
Squares or circles... or both. 2mm,5mm, 10mm etc

ivconic,

Aluminum foil and masking tape is commonly available. Just cover the aluminum foil with masking tape and mark the tape with lines that match a specified size and cut the foil to be specified targets. You can even use thicker aluminum targets such a soda can that is cut to the same size targets. The main point is that everyone who submits a design has access to specified targets and provides information about their coil such as coil diameter, inductance, resistance, pulse length, current, coil discharge time constant, damping resistor value, target size and time constant and target distance to coil.

Just pick the things that most people can agree to submit with their circuit and coil design.

Joseph J. Rogowski

[pito]

[ivconic]

Joseph, I thought about it, what I could suggest as a common target that could be found easily anywhere in the world.
It seems that we have no choice but to cut squares of several dimensions with sheet metal scissors.
Squares or circles... or both. 2mm,5mm, 10mm etc

[bbsailor]

[ATTACH]n417773[/ATTACH] The MiscEl app is a good tool to calculate almost everything on a PI detector. It is old and I can not open the homepage, but the app in the Zip file should work. It does here. It starts giving some error messages, but just ignore these.

The UXO research that the military did was about optimizing the detection of small metal parts found in recovered UXOs. Their main point was that they tried to stimulate many different size and time constant metal parts and found that if they kept increasing the coil discharge slope, there was no better full stimulus of a particular time constant target beyond five times faster than the target time constant. This sets a reasonable goal for specified targets.

The PI coil performance is a real challenge for small low time constant targets. When a fully stimulated target signal decays to 2 target time constants it has lost about 86 percent of its induced charge. At 3 target time constants the target has lost about 95 percent of induced charge. This is why making a pulse induction coil to operate a very low delays requires a higher damping resistor value to make fully stimulating small targets possible. Then reducing the delay so the RX circuit can turn on as quickly as possible to capture the small target signal before it dies out and cannot be detected.

The optimization of detecting low time constant targets then becomes a balancing act between these things.
1. Coil diameter and distance from target
2. Coil inductance governed by the number of coil turns.
3. Coil current governed by coil wire size and voltage.
4. Coil seen capacitance which affects the damping resistor value.
4.1 MOSFET and other circuit capacitance.
4.2 Coil turn to turn capacitance governed by coil wire insulation dielectric constant value and thickness.
4.3 Coil to shield capacitance governed by the thickness and dielectric constant of the spacer between the coil and the shield plus the area of the shield either being solid or a thin wire mesh.
5. Damping resistor value which governs the coil discharge slope and ability to fully stimulate a low time constant target as well as reducing the delay between the coil TX mode and RX mode.

This is why we need some commonly available targets to specify so the final Pulse Induction performance can be more easily and accurately compared too each other.

Joseph J. Rogowski

[GeoMax]

The small coil had about 300uH and 20cm diameter. 20AWG wire. Best to use 300V or 600V plastic insulated, tinned, 1x19 or even finer stranded wire, but for a cheap quick and dirty trial, magnet wire will work.
This detector was for underwater use, so it did not need shielding on the coil. For use on land, shielding will reduce EMI noise.
This circuit ran at about 3000 cycles per second. I think about 30us TX on - 30us delay to first sample- 30us sample time - second sample same 30us just before TX on again.
If you change the clock capacitor to a smaller value, the clock will run faster and all timings will be shorter, but remain relative.

Thanks for the coil info....

[Tinkerer]

[ATTACH]n417773[/ATTACH] The MiscEl app is a good tool to calculate almost everything on a PI detector. It is old and I can not open the homepage, but the app in the Zip file should work. It does here. It starts giving some error messages, but just ignore these.

[Tinkerer]

Hi Thinkerer,

I have a question in regards to the coil of the KW-1. From the circuitry, I would guess the search coil has about 200-220uH. What coil inductance/specification has the actual search coil of the KW-1?
Can you still remember?

The small coil had about 300uH and 20cm diameter. 20AWG wire. Best to use 300V or 600V plastic insulated, tinned, 1x19 or even finer stranded wire, but for a cheap quick and dirty trial, magnet wire will work.
This detector was for underwater use, so it did not need shielding on the coil. For use on land, shielding will reduce EMI noise.
This circuit ran at about 3000 cycles per second. I think about 30us TX on - 30us delay to first sample- 30us sample time - second sample same 30us just before TX on again.
If you change the clock capacitor to a smaller value, the clock will run faster and all timings will be shorter, but remain relative.

[GeoMax]

Hi Thinkerer,

I have a question in regards to the coil of the KW-1. From the circuitry, I would guess the search coil has about 200-220uH. What coil inductance/specification has the actual search coil of the KW-1?
Can you still remember?

[GeoMax]

Here is the KW_1 detector on a shipwreck. If I remember right, it was about 1996 on the wreck of a French galeon that wrecked in 1678 on the reefs of Aves Islands, off the coast of Venezuela.
This is the small coil, about 20 cm diameter. I could also change to a 32cm diameter coil for deeper targets.

Nice to see the actual device in use!!! I like it. Thanks again for posting the picture(s)