Walk Through Metal detector Circuit digaram

[jamil_tech]

Finally I made hammerhead

Hi Geo,

I made Hammerhead and it is working. But what I found it is absulately not suitable for work through coz its unstability and detecting depth is not meet with walk through.

[Qiaozhi]

Hi Geo,

I made Hammerhead and it is working. But what I found it is absulately not suitable for work through coz its unstability and detecting depth is not meet with walk through.

Please tell us more about how you have constructed the coil(s).

[jamil_tech]

I made two coil , 25cm and 60cm coil. For 25 cm i made ir 26 turn of 25 swg. and for 60 cm coil 13 turn of 25 swg on a pvc pipe. I found for small object 25 cm coil is sensitive and with 60 cm coil no sensitivity in center. With 25cm coil i found 14cm detection range for a 5 inch pencil cutter. and with 60 cm coil it is not detected at center of the coil. My detector circuit is well calibrated and all output is almost same as carls article.

So I am still in problem. Does anybody have any good solution for this plz ?

[jamil_tech]

Here i attached my circuit.

[Qiaozhi]

I made two coil , 25cm and 60cm coil. For 25 cm i made ir 26 turn of 25 swg. and for 60 cm coil 13 turn of 25 swg on a pvc pipe. I found for small object 25 cm coil is sensitive and with 60 cm coil no sensitivity in center. With 25cm coil i found 14cm detection range for a 5 inch pencil cutter. and with 60 cm coil it is not detected at center of the coil. My detector circuit is well calibrated and all output is almost same as carls article.

So I am still in problem. Does anybody have any good solution for this plz ?

Your detection depth seems very poor. For a 25cm (9.8") coil you are only achieving 14cm (5.5") for a 5" metal target! This is worse than a BFO.
I would have expected about 4x that distance.
Any Hammerhead constructors want to comment?

[bbsailor]

Check IC3 and Q3

jamil_tech

On the photo, it looks like the MOSFET (Q3) coil driver is mounted where IC3 (voltage regulator) should be. IC3 should be a small negative 5V regulator and look like IC4.

bbsailor

[Leto]

My detector circuit is well calibrated and all output is almost same as carls article.

If you are absolutely sure in no parts misplaced first thing I would do is to clean with alcohol all flux residue throughout the PCB..

[jamil_tech]

Hi bbsailor,

My MOSFET IRF740 is just under the PCB as I change it continuously like IRF840, IRF 640 etc.

Yes I use Big size regulator for 7905 coz in my box there is no smaller version. But all my voltage is ok.

[bbsailor]

jamil_tech

This is the best I can do by just looking at your circuit board photo.

Here is a little trick I learned to change MOSFETS easily. Obtain a flat-pin type IC 8-pin DIP socket. The pin spacing is the same as MOSFET lead spacing. Try to insert the MOSFET pins into the IC socket to make sure it fits. This will not work with round pin sockets. Once you find a DIP socket that fits your MOSGET leads use a fine-tooth saw and cut the DIP socket length-wise and then cut off one pin. Now you have a MOSFET socket that you can make quick MOSFET changes without destroying your circuit board.

If you move the large capacitor (C10) to under the board or to the side a little, you can mount the MOSFET socket on the top of the circuit board; bend the MOSFET leads 90 degrees, drill a hole in an empty space, place a small teflon or fiber washer about .250" high to mount the MOSFET on a small "U" shaped heat sink.

If you want to make a sensitive coil for small objects, use stranded hook-up wire to make your coil (19 to 21 turns) such as AWG 24 to AWG 28. Teflon insulated wire makes a slightly better coil as it has a little less capacitance than PVC insulated wire. Also, mount your coil far enough away from your circuit board to not get any interference on the coil. Use about 3 ft/1 meter of coax to connect your coil to the circuit board and suspend your coil in mid-air and orient the coil away from metal objects.

You can use about 50 ft of wire to make a small 4 inch diameter coil (45 to 47 turns) that is better to use in confined spaces for testing. Make sure you shield the coil to minimize noise pickup.

bbsailor

[Unregistered]

WoW, lots of information for walokthrough in this forum but all are uncomplete

[Qiaozhi]

WoW, lots of information for walokthrough in this forum but all are uncomplete

Dear Unregistered - why are you mad? :confused:

The information is quite likely incomplete because no-one here has ever built one, but Jamil is trying hard and we are trying to assist to the best of our ability.

[Sean_Goddard]

OK Try this......

You need.. ONE BIG Tx coil this is the coil on the outside of the picture you have, then you have FIVE or SIX SMALLER Rx coils on the INSIDE of that. Each one is INDIVIDUALLY BALANCED for zero output under no metal conditions.

Next, you need as many RECEIVE circuits as you have coils (what you are doing is making 5 or 6 INDIVIDUAL metal detectors with one common TX system. Simply poll or multiplex the outputs from these RX's into a PIC and VOILA, you can drive a flashing LED to show which area the metallic object has been detected in.

You can use digital potentiomenters to set the threshold and sensitivity levels automatically by looking at the outputs and adjusting via SPI inteface on PIC until all is stable.

You could (AHEM** cough**..copy a Silver sabre type diagram to get the ideas and speed up design, but you will have to make the "search head" yourself, HEY WAIT UP.. What about doing this for a portable system, simply lay the coil down. Hows that for WIDESCAN!!

OR.....

One PI system (dual coil type- one for Tx and another for Rx) then repeat the mulitplexed Rx modules. That's why the diagram you show had many boards. LOTSA RX channels.

If you make one side operate at say (3KHz) and the other at say 22KHz you can have a VERY good system (you are duplicating th above systems one on each side of the walkthu' area).

Hope this gives you some ideas.

Oh BTW, if you use the PI type system you COULD just make an 6 channel Tx/Rx PI and fire them in sequence, top t obottom. The PI approach (one Tx, multi Rx is best as no hard to balance IB coils).

[jamil_tech]

Is there anybody who have its schematic ?

Hi Sean_Goddard

Thank you very much for sharing your Idea with us. Yes you may right with your idea but it is very much hard doing anything without proper schematic and guideline.

I got a PCB with component lay out one of forum member but unfortunately he lost it schematic diagram. Is there anybody who have its schematic or anybody can recover its schematic will help me.

[Sean_Goddard]

Try this, it might help.

It should not be too difficult to work out whichc bits you need to "multiple up". If it is, let me know and I'll try to do the job for you .

[Carl-NC]

Jamil,

If this is intended to be a design project for engineering credit, then you really ought to focus on coming up with something that is not simply a direct copy of an existing circuit. My suggestion is to work with PI circuits, for several reasons. First, it is what commercial walk-throughs use. Second, you don't have to deal with induction-balanced coils. Third, you can set up zones that can be run independently without interference. Fourth, the PI circuit & coil is relatively simple.

I suspect that a walk-through design is a pretty difficult challenge. You are probably expecting too much if you think you will end up with something that can detect tooth fillings, or even coins. Shoot for something easier, like a machete. And don't assume you will end up with 100% zonal coverage... you will probably have gaping holes that an Uzi can get through.

Start with designing & building a very basic PI circuit. Any of the PI projects on this site are sufficient as a starting point, and most can be simplified somewhat. Since you will likely be running several zones, go ahead and assume a PIC-controlled clocking circuit. You could use std gates, but you'll quickly wish you hadn't.

The point of these projects is usually to demonstrate an ability to actually understand a problem, and design a solution. It is not necessary to end up with a world-class solution, unless it's a PhD thesis.

- Carl

P.S.--Try contacting Garrett, tell them exactly what you are doing (provide the name of a directing professor if you can), and see if they will provide any insight on these things. I've visited their factory, and their walk-through detectors are impressive.