Manta Metal Detector

[Asgard]

whats the total current draw of your manta metal detector ???

For this test, battery Lithium Ion near full charge 16V, current 2.278A view : single pulse mode bipolar mode :

But I can increase the current with PIC32 Max 3.1A

And you ?

[6666]

17 inches on US nickel with slow sat speed (12 inches coil at 6us delay)
15.5 inches on US nickel with fast sat speed (12 inches coil at 6us delay)

13 inches on square foil aluminium 1" (medium sat speed, 12 inches coil at 6us delay)
16 inches on square foil aluminium 1.5" (medium sat speed, 12 inches coil at 6us delay)

Why is there such a large difference in detection distance with the nickel and foil when you alter just the sat speed ?

[6666]

Why is there such a large difference in detection distance with the nickel and foil when you alter just the sat speed ?

oops meant to just say nickel.

[Asgard]

[QUOTE = 6666; 217835]. Oops signifiait simplement dire nickel [/ QUOTE]

Since maximum Sat Speed caliber in 2 seconds and miniumum sat speed caliber in 0.3 seconds

A minimum sat speed you have a tendency to eliminate small signals because it think it's a parasite or ground effect, so it recalibrates

[6666]

Alexandre

In your TX schematic drawing in post number one you show a Pfet in series with an Nfet, what is the advantage of doing this configuration ?
Thanks.

[Sean_Goddard]

Surely one of the FETs would never turn on so you wouldn't get any current flow (except through the body diode).

I know the thread is old but I have recenetly been playing with this idea in a machine I am building for myself and I'm getting some good results. I'm using the Dave Emery "Worm Burner" ultra high tau / low noise front end with the diode switch and sampling at 4 to 7us. Depths of around 24" on a Victorian penny are what I am seeing so far and VERY HIGH sensitivity to gold. UNFORTUNATELY I am prohibited from posting any schematics as Dave has a patent on some of the methods I am using and I won't risk his wrath.

[techo_bob]

I believe there may be a couple of large caps missing from the original schematic going to the junction of the 2 fets creating a capacitive discharge setup
just my thoughts but may be Wrong.

Bob..

[moodz]

The PMOS blocks the negative going flyback ... the coils are closely coupled but in opposition. The drive circuit is over simplified.

[Asgard]

At the time I greatly simplified the diagram and I did not send the diagram which works well.
I shared the philosophy simply... ;-)

[kt315]

sorry, i understand just schematics' language. philosophy, ie Darwin's theory of evolution is GREAT LIE.

[Asgard]

I wanted to say it's for the principle.
My goal was not to give a detailed schematic.

[Elliot]

Surely one of the FETs would never turn on so you wouldn't get any current flow (except through the body diode).

I know the thread is old but I have recenetly been playing with this idea in a machine I am building for myself and I'm getting some good results. I'm using the Dave Emery "Worm Burner" ultra high tau / low noise front end with the diode switch and sampling at 4 to 7us. Depths of around 24" on a Victorian penny are what I am seeing so far and VERY HIGH sensitivity to gold. UNFORTUNATELY I am prohibited from posting any schematics as Dave has a patent on some of the methods I am using and I won't risk his wrath.

Hi Sean, without revealing any schematics, I was wondering if you could provide some more background information. For example, pulse on-time, frequency, coil Rx Tau?

[Sean_Goddard]

I was using 65us on time, a 3.6KHz frequency (but not critical). As for Rx coil, I went to a dual Rx channel system so basically TWO Rx channels with low johnson noise front end at the front end, diode switches as per Dave Emery's patent and different sample times between the two channels (5 and 12us) then each channel is normalised and the normalising factor used to calculate the probable conductivity of the target. There is a LOT of experimenting still to be done, but the results are VERY promising as regards discrimination. Certain items need characterising in a "live" database. Sadly it can NEVER be a commercial design as It uses information from various patents combined intyo one machine.

[Elliot]

I was using 65us on time, a 3.6KHz frequency (but not critical). As for Rx coil, I went to a dual Rx channel system so basically TWO Rx channels with low johnson noise front end at the front end, diode switches as per Dave Emery's patent and different sample times between the two channels (5 and 12us) then each channel is normalised and the normalising factor used to calculate the probable conductivity of the target. There is a LOT of experimenting still to be done, but the results are VERY promising as regards discrimination. Certain items need characterising in a "live" database. Sadly it can NEVER be a commercial design as It uses information from various patents combined intyo one machine.

Thanks Sean. It sounds very similar in principle to what Bernie is doing with the TineFPGA-Gold. His prototype discrimination system is working, but analyses the decay curve at many sample points much earlier than 5uS.