Metal Detector inside trashbin?

[Wbdsgnr]

Hi there

We are doing a project for school and we decided to make a automated trashbin. Very funny, but we thought of including a metal detector so if we would throw a can inside it that thing will start to make noise. But I have no clue wether I can just wind a coil around that thing (it is square shaped). The only schemes I find on the internet include 2 coils, but since it is only a trashbin I though a single coil would do well. Or am I being foolish If I would use this scheme (2 coils ) http://www.geocities.com/tomzi.geo/metal_detector/metal.jpg I guess I would need to make some minor adjustments? Could someone help us out

 

[Johnson777717]

Heya!
I truly love your autosignature! †

The link you posted cannot be viewed for some reason. Possibly the page is being developed.

 

[Johnson777717]

From my limited knowledge about metal detectors, the two rings are required in order to create a "conversational" signal.

A BFO (Beat-frequency oscillator) metal detector uses two oscillating frequencies, one through each ring. Each oscillating frequency is set to match the other. Through circuitry, a beat is produced, which is constant. When a metal object is introduced into the field, the frequencies are distorted from their normal oscillation, which, through circuitry, generates a quicker beat and thus lets the user know that a metal object is within the field.

Honestly, I think it would be pretty difficult to scan a trash can for metallic elements, for a couple of reasons:
1. Trash can contain small fragments of metal that will give a false reading that something of importance, in this case a can, has been thrown away. For instance, foil gum wrappers, pop can tabs, bottle caps etc.

2. I'm not sure if the other items within the trash will impede in the metal detector signal. For instance, a piece of cardboard, or plastic may keep the metal detector signal from reaching the can which is behind the plastic. etc.

This doesn't mean the project isn't possible, I just think developing a metal detector that will work efficiently in a trash can may be beyond what you may be willing to develop.

 

[Russlk]

Geocities does not allow links so you have to copy and paste the URL.
The theory behind the metal detector is that there are two oscillators of slightly different frequencies. The difference "beat" frequency is audable so the operator will know when metal near one coil has changed its frequency. One oscillator has to be isolated and stable, the other oscillator coil will be the one wound around the trash bin. Paper and plastic will not be detected, so your project should work well, but I think the design is beyond your capabilities ( or maybe not, give it a shot). You can get help on this forum.

 

[Nigel Goodwin]

This is all quite correct about how BFO metal detectors work, but it's probably not required for this very simple application. He's not wanting to find a small coin under six inches of soil - the metal object is actually passing through the coil - which makes a huge difference.

How about the coil in the bin been used as the coil in a simple oscillator, as the metal actually drops through the coil it should give a fairly large change in frequency - this should be easily detectable, without using high sensitivity BFO techniques. You could use a 555 as monostable, as you would in an analogue frequency meter (or rev counter) - this gives a DC voltage out, and you could use a comparator to give a signal when it goes above (or beow) a certain range.

Or, being a 'PIC person', use a PIC to measure the frequency, and give a warning when it's outside a certain range.

Experimentation with the coil design and frequency used would help to find the optimum design.

 

[samcheetah]

being a 'PIC person'

lolzzzz :lol: :lol: :lol: :lol:

thats a nice name

ive learnt alot about PICs from this site and most of the info comes from u. man how do u do that

 

[Nigel Goodwin]

ive learnt alot about PICs from this site and most of the info comes from u. man how do u do that

Perhaps I ought to 'get a life' :lol:

Actually I already have!, I teach martial arts 3 nights a week as well as having an interest in Electronics. I've been using PIC's a long time now, my first version of PicProg ran under DOS, and (as far as I'm aware) my original Win3.1 version was the first Windows PIC Programmer software for David Tait style programmers.

 

[Wbdsgnr]

This is all quite correct about how BFO metal detectors work, but it's probably not required for this very simple application. He's not wanting to find a small coin under six inches of soil - the metal object is actually passing through the coil - which makes a huge difference.

How about the coil in the bin been used as the coil in a simple oscillator, as the metal actually drops through the coil it should give a fairly large change in frequency - this should be easily detectable, without using high sensitivity BFO techniques. You could use a 555 as monostable, as you would in an analogue frequency meter (or rev counter) - this gives a DC voltage out, and you could use a comparator to give a signal when it goes above (or beow) a certain range.

Or, being a 'PIC person', use a PIC to measure the frequency, and give a warning when it's outside a certain range.

Experimentation with the coil design and frequency used would help to find the optimum design.

Ahh that is what I mean! a very basic and simple detector for large objects And I'm a beginning pic person But the clue is you give that coil a frequency and it should change if metal comes trough it? But I have no clue how to do that with a pic Is just making output high and low many times? But how than would you detect a change? I'm a kinda noob I ordered a few sample pics, including the 16f628 and one with a/d convertor, maybe I need that? Thank for your help!

 

[Nigel Goodwin]

Ahh that is what I mean! a very basic and simple detector for large objects And I'm a beginning pic person But the clue is you give that coil a frequency and it should change if metal comes trough it? But I have no clue how to do that with a pic Is just making output high and low many times? But how than would you detect a change? I'm a kinda noob I ordered a few sample pics, including the 16f628 and one with a/d convertor, maybe I need that? Thank for your help!

There's an old MicroChip application note for building a 50MHz frequency counter using a PIC. Basically you can do it in two ways:

1) Count the number of pulses over a certain time.

2) Measure the time between two pulses.

Option 2 is faster, and for your purposes probably easier as well, you don't need to actually know the frequency - just to detect a change in it.

For a simple way, feed the oscillator signal into a PIC pin (you may need to square it first, or if you choose an I/O pin with a schmitt it will probably be OK without). Wait for the pin to go high and then start counting, stop counting when it goes low - the value in the counter will be a representation of the frequency. This is only counting half a cycle, to count a full cycle you would need to carry on counting until it went high again. However, it should be plenty good enough for your needs.

 

[Exo]

being a 'PIC person'

lolzzzz :lol: :lol: :lol: :lol:

thats a nice name

ive learnt alot about PICs from this site and most of the info comes from u. man how do u do that

In dutch, PIC sounds much like a word for Dic* :lol:

 

[Wbdsgnr]

Ahh that is what I mean! a very basic and simple detector for large objects And I'm a beginning pic person But the clue is you give that coil a frequency and it should change if metal comes trough it? But I have no clue how to do that with a pic Is just making output high and low many times? But how than would you detect a change? I'm a kinda noob I ordered a few sample pics, including the 16f628 and one with a/d convertor, maybe I need that? Thank for your help!

There's an old MicroChip application note for building a 50MHz frequency counter using a PIC. Basically you can do it in two ways:

1) Count the number of pulses over a certain time.

2) Measure the time between two pulses.

Option 2 is faster, and for your purposes probably easier as well, you don't need to actually know the frequency - just to detect a change in it.

For a simple way, feed the oscillator signal into a PIC pin (you may need to square it first, or if you choose an I/O pin with a schmitt it will probably be OK without). Wait for the pin to go high and then start counting, stop counting when it goes low - the value in the counter will be a representation of the frequency. This is only counting half a cycle, to count a full cycle you would need to carry on counting until it went high again. However, it should be plenty good enough for your needs.

Ok but how do I attach every thing? I still don't quiet get it. A coil has two ends right? One end gets pulses from a output pin and the other end goes onto a input pin to measure the time? So basicly all you need is a PIC and a coil, and a LED or something to show the output? Or am I still don't getting the idea

 

[Nigel Goodwin]

Ok but how do I attach every thing? I still don't quiet get it. A coil has two ends right? One end gets pulses from a output pin and the other end goes onto a input pin to measure the time? So basicly all you need is a PIC and a coil, and a LED or something to show the output? Or am I still don't getting the idea

You need to build an oscillator using the coil as the inductor, then feed the output from the oscillator into the PIC to measure the frequency. You can construct an oscillator in loads of different ways, using a transistor, CMOS gates - all sorts of things. There's a PIC based LCF meter in the current EPE magazine, that uses a CMOS 4011 oscillator to measure inductors.

 

[john1]

When an L.C. oscillator runs the frequency depends
on the coil(s) and the capacitor(s).

The coil's characteristics can be modified by various
materials in its immediate vicinity.

In particular, metals in the field of the coil will
have an effect, thereby altering the frequency of
the oscillations produced by an L.C. oscillator.

Such alterations can be used as an indication to show
the presence of metals.

Hope this clarifies it a little ...

John

 

[plot]

just make sure it's not a metal trashcan

 

[Wbdsgnr]

When an L.C. oscillator runs the frequency depends
on the coil(s) and the capacitor(s).

The coil's characteristics can be modified by various
materials in its immediate vicinity.

In particular, metals in the field of the coil will
have an effect, thereby altering the frequency of
the oscillations produced by an L.C. oscillator.

Such alterations can be used as an indication to show
the presence of metals.

Hope this clarifies it a little ...

John

Well, a little I guess the 555 isn't a LC oscillator? But i have trouble finding the right sheme for it I made this one according to another sheme (don't notice the values) **broken link removed** but this wouldn't be sufficient? And it wouldn't differ if it is a metal trashbin right? the frequency would still change if another metal would come trough.. (i hope )

 

[Nigel Goodwin]

I don't think that would work, a 555 works by charging a capacitor up, it's not 'really' an oscillator at all.

Here's the oscillator circuit from the EPE LCF meter, with the values shown it runs about 500KHz, it's nice and simple!. The preset is adjusted to ensure reliable oscillation.

As you are at school, presumably you have access to scopes and frequency counters - you could use either (or both) to check the amount of frequency shift you get, and find the optimum frequency and size and shape of coil. Your teacher should love you for it :lol:

 

[Wbdsgnr]

Ok, maybe a very stupid question. But how would you attach that thing? I mean, if you charge a capicitor it won't oscillate forever... don't you need some amplifier or something? I'm sorry , but i'm just a first grade student I found this sheme **broken link removed**, can't I use that? thx in advance

 

[Nigel Goodwin]

Ok, maybe a very stupid question. But how would you attach that thing? I mean, if you charge a capicitor it won't oscillate forever... don't you need some amplifier or something? I'm sorry , but i'm just a first grade student I found this sheme **broken link removed**, can't I use that? thx in advance

Sorry, do you mean 'where are the supply rails?', it's normal not to bother showing them on logic circuits - you do (of course) need power for it. The CMOS NAND gate is the amplifier, it's a standard circuit.

I can't comment on the web page you gave, as I can't get it to work.

 

[Wbdsgnr]

Yes that I feel really stupid you know :? but erhm... bottom is plus ? It dc right? And ehm.. where would be the output then, after the nand gate?? pff, i should really try this with a oscilloscope, much easier than guessing Where would I be without you

 

[Nigel Goodwin]

Yes that I feel really stupid you know :? but erhm... bottom is plus ? It dc right? And ehm.. where would be the output then, after the nand gate?? pff, i should really try this with a oscilloscope, much easier than guessing Where would I be without you

Perhaps I should have labelled the diagram :lol:

Here's a slightly better diagram, you will need to sort out the pin connections for the chip yourself.