New Idea

Ok I'm back and thanks again for all of your replies to my post. 

You will notice I use the term CEMF Pulse here because that is what it is. It is not a reflected pulse because it does not bounce off of anything and it is not a signal from the target metal. The target signals only delay the pulse's return to ground or quicken the time that the pulse does  return to ground. This is why I only work with the CEMF pulse width.

If your are happy with what you have for your PI detector, I am happy for you. Don't change anything. If you want something different then maybe what I have to say will help you. I did not mention delay times, because what I was talking about is processing the CEMF (counter electromotive force) pulse directly. The time after the CEMF pulse (or "spike") is great if you only want to find ferrous items. Good for relic hunting though or finding buried pipes. I have attached a block diagram of a workable non-MCU PI circuit that is very stable and has good sensitivity. This was my first attempt at making my own design when I got fed up with other circuits where I was constantly twiddling knobs, poor sensitivity, and other problems. It worked very well, and then I thought Hmmmm "Good Circuit....Can I simplify it or make it better?" So I designed an MCU version of it which eliminated a number of parts, but used the same principles. My latest version also has an 16 x 2 Parallax LCD panel, 2 MCU's, dual tones, a super-bright led that lights when any type of target is found, and wireless headphones. A switch allows me search for ferrous or non-ferrous metals. Also I designed a different multi coil head that has nearly the same sensitivity for small items as large items. Going to seek a patent on that part. The rest I might make a kit.

In the block diagram I used a 4 MHZ clock oscillator module and divided it down with a few 4017 divide by 10 ICs (division ratio was 25000:1) I "and'ed" two of the outputs together to get a 160 Hz pulse repetition rate and a 4% 250 µS pulse. I used this to drive a Power Mosfet into a 7-3/4" diameter coil. A smaller secondary coil provided a replica of the pulse. The lower end of the coil was grounded and the other end was fed to 2 series connected 1N4148 diodes and a 2k resistor to ground. This eliminated the negative going drive pulse and left only the positive going CEMF pulse. From there it goes to a resistive voltage divider. the output of that is a 10 volt 10µS positive pulse that I feed into a comparator circuit. The comparator squares up the sides of this rectangular pulse as well as letting one adjust the width of the pulse. I found that a 12 µS pulse was most sensitive to non-ferrous metals and an 18-20 µS was most sensitive to ferrous metals. A non-ferrous metal reduces the output pulse width of 12 µS to about 5-7 µS while a ferrous metals increases the 18-20µS pulse by quite a bit more (up to 145 µS)
So what I am doing here is to determine whether the pulse width has changed width up or down. A simple way to determine that is with a very simple circuit. A low pass filter. It doesn't need any amplification, because it is at 10 volts already. I made a simple temperature stable low pass filter by connecting a 5.1 volt zener in series forward conducting direction to a resistor and then a capacitor to ground. The output is a ramp voltage at 160 Hz  that goes up or down in peak voltage according to what type of metal is under the coil. I then feed this to 2 more comparators. One is connected in inverting mode and the other is connected in non-inverting mode. The non-inverting comparator drives an electronic switch that passes a 400 Hz tone to the headphones to indicate a ferrous target, while the inverting comparator drives another electronic switch that sends a 800 Hz tone to the headphones when a non-ferrous target is detected. Also included is an optional meter circuit for those that would like a meter instead.

I also attached the circuit diagram if someone wants to build it. I will leave building it up to you. I only made one prototype on a punchboard so there are no kits. Good luck to all that build it. I will try to give technical help when I can, but don't expect me to build it for you. You guys are pretty smart from what I can see.

Golddustcarl

Linkback:

You are not allowed to view links.
Please Register or Login

http://www.thunting.com/smf/index.php/topic,56782.msg275499.html#msg275499

A bad day at gold prospecting or metal detecting is better than a good day at work.

Thanks for the reply. No the coil is not IB. The closest definition is concentric wound. The outer coil is the larger and the inner coil is 5-3/4". 1 inch spacing between inner and outer coil. Hand wound is ok. No shield needed. Be sure to bind coils tightly and secure firmly in your coil form. I am using the 2 coils in a similar fashion to a transformer to achieve level transformation. This makes the detection process easier.
I search for everything of value. I am not choosey. Lately though I have been searching mostly for gold.
It makes me happy to see that there are still people interested in home-brew electronics in this day and age when almost all of the electronics houses have closed down.
What part of the world do you hail from if I may ask? I am located in Phoenix, Arizona, USA. 50 years of electronic experience.


Linkback:

You are not allowed to view links.
Please Register or Login

http://www.thunting.com/smf/index.php/topic,56782.msg275520.html#msg275520

A bad day at gold prospecting or metal detecting is better than a good day at work.

Hi 6666
 Glad to make your acquaintance. No I have not tried it on the beach as of yet. I live in a desert area. Plenty of high desert mountains with a gold band that runs NW by SE through the state. I would love to be a the beach even just for a vacation. lol

That is correct about what I am doing at the coil. 10 volts is what I need to drive the comparator. This is a real comparator, and not an op-amp. The secret lies in the shape of this pulse. it curves upward at the base on the leading edge and downward at the falling edge. Rail-to rail operation is needed to get a clean rectangular pulse at the output. A metallic object at the coil not only changes the width of the spike, but also changes the peak amplitude of the spike. When you set a the inverting input to a voltage that is near the base of the pulse (about 1 to 3 v) the comparator will change the width of the output pulse according to the type of metal that is being detected. Narrower for non-ferrous and wider for ferrous objects. If you set this pulse to 12 µS width, it is more sensitive to non-ferrous metals. If you adjust the control to 18-20 µS it will be more sensitive to iron objects.  This is a way to discriminate the two types of metals. If you detect only for a narrowing pulse, no ground balance circuitry is needed, as most ground balance problems are ferrous in nature. Pi is kind of immune to that anyways.
I have included a scope photo of the input and output pulses of the comparator so that you can see what I am saying. F trace is the divider output and the H trace is the comparator output. As you can see 3 volts gave a 12 µS pulse. A lower voltage will give a wider pulse.

Hope this helps
Golddustcarl

Posted on: May 17, 2014, 12:06:24 am

---

Sorry meant to send this one.

Linkback:

You are not allowed to view links.
Please Register or Login

http://www.thunting.com/smf/index.php/topic,56782.msg275527.html#msg275527

A bad day at gold prospecting or metal detecting is better than a good day at work.

Alex1

Sorry, No PCB was designed for it. I used a perfboard and point to point wiring. I later graduated to a microcontroller version of it and built it also on a perfboard. I only design a pc board when I am relatively sure there are no more bugs, and that I do not plan to make any more features. In this case I just moved forward.

Let me know if you build it.

Thanks,
golddustcarl

Posted on: May 18, 2014, 08:47:17 pm

---

I am currently designing a mini PI detector for detecting non-ferrous metals that has only 1 control, is controlled by a micro, uses a buzzer or LED for signaling, and you can hold it in one hand like a probe. This will not be like the cheap Centech circuit, which buzzes on all metals and is drifty as all get out.  Nor does it use that fragile ferrite rod. I plan to use it for sniffing out nuggets in cracks in bedrock and deposits of black sand. Iron or ground effects will not set it off. I have drawn a schematic and written some software ( need to burn it yet, build the circuit, and do testing). If it pans out, I will sell a partial kit of a pre-programmed PIC, schematic, parts list and plans). I will share the schematic on the site, but not the software. I expect the semi-kit will probably be  around $10 with free shipping in the US. Overall parts count is low.
Wish me luck.
golddustcarl


Linkback:

You are not allowed to view links.
Please Register or Login

http://www.thunting.com/smf/index.php/topic,56782.msg275567.html#msg275567

A bad day at gold prospecting or metal detecting is better than a good day at work.

You have got the right idea johnny23! I for one hate to have to spend 1/2 an hour on warm up time, twisting knobs to get it balanced just right and then bump it on a twig and have to do it all over again. Jeesh! One knob for me is one knob too many, so maybe I will try to eliminate that when I get things working good. I have an idea for a auto threshold adjust for it already without adding any other parts. (well maybe a resistor or two).
What you are doing is a great idea. Wished I had thought of taking an old detector and modernizing it myself. Maybe I will try that next. Keep me posted on how it is going.
golddustcarl


Linkback:

You are not allowed to view links.
Please Register or Login

http://www.thunting.com/smf/index.php/topic,56782.msg275569.html#msg275569

A bad day at gold prospecting or metal detecting is better than a good day at work.