Metal Detecting Coil - Design Question. 1

[shansiden]

Hello, I am designing a metal detector coil for a project: it is long/narrow, perhaps up to 2 meters in length.
Here is the problem - How can the detection depth be kept uniform across the whole length, and detect at no more than 30-40 cm deep. Is the best approach to use lots of smaller DD coils or can a single coil 2 meters long be made to work?

is it possible to keep the same properties of a smaller coil by unwinding and elongating the coil windings over a larger area?

Many thanks in advance, S

 

[VEBill]

There are several slightly different concepts used in the design of metal detectors. In most cases, the design of the coil system and the design of the circuit need to be fully and tightly integrated. I doubt that the size of a coil can be dramatically changed without redesigning the entire system. And it's not cleat how you'd integrate many smaller coils.

Given that there are so many existing designs, including large versions to be dragged behind vehicles, can you find plans for one that meets your requirements?

Disclaimer: My expertise isn't in metal detector design. I'm not sure if there's a better forum on E-T.

 

[shansiden]

Thanks for your reply VE1BLL, I think it would be easier to used a group of smaller coils cycled/phaseed to avoid interference,

any ideas on a circuit that can sample multiple antennas at high speed? (the coil frequencies are generally 5khz-75khz)

Thanks Again

 

[VEBill]

Speculation: Rather than attempting to switch the coil circuits, perhaps it'd be easier to include with each coil its own oscillator and front end. Then you'd only have to deal with the individual enable signals and the baseband outputs.

 

[shansiden]

Do you think this may work? The coil design i mean.

 

[VEBill]

On the assumption that the Detector circuit is simple and cheap, then it might make sense to have multiple parallel Detector circuit (one for each Rx coil), and then combine their outputs in software further up the system. This would seem to be more likely; every metal detector that I've seen contained only cheap garden-variety analog components, sometimes plus a complicated digital processor further upstream. Modern digital processors could easily handle endless inputs.

On the other hand, if the Detector circuit was something highly complicated and very expensive, then it might make sense to switch the coils at the input as you've sketched. But that seems unlikely.

I'll admit that it's counter-intuitive, but it's still worth considering.

---

You might wish to search the USPTO (also available on Google Patents) patent database. As a first approximation, "everything" has already been invented. There seems to be about 20,000 patents related to metal detectors.

 

[MikeHalloran]

I get the impression that metal detectors almost universally use resonant circuits, i.e., oscillators, that for lack of a better word, misbehave, in a noticeable way, when metals are present in their magnetic field.

What may not be obvious is that oscillators can be cranky about starting to oscillate, so switching parts in and out to multiplex antennas is likely to be very much slower than you would like.

As VE1BLL wisely suggests, you would likely be better off to use a linear array of complete individual small detectors to accomplish what you want.

That strategy is still fraught with difficulty. Changing the output circuit of a detector in any substantial way is likely to affect the way the entire detector behaves, since the output is usually a change in the frequency at which the detector squeals into a speaker, and the circuits used are at least visually simple and use a small number of components. ... and parasitic components are likely to be as important as the real components. There is no convenient digital output that you can just wire into something else or place on a logical bus, and even just inserting a resistor or a photodiode may prevent the detector from working at all.

Okay, go ahead and try, but you may have to go to the trouble of adding a microphone or accelerometer to each individual speaker, and sampling those with a fairly fancy microcontroller, looking for changes in volume and/or frequency over time, maybe turning on a lamp to indicate which detector is providing an output, and maybe repeating that detector's output to some master speaker.

It sounds like an interesting project, that will ultimately cost you more than whatever is already commercially available, so I hope the education is worth something to you.



Mike Halloran
Pembroke Pines, FL, USA

 

[shansiden]

"you would likely be better off to use a linear array of complete individual small detectors to accomplish what you want."
Thanks for the advice there that is where ill go next. There definitely is a lot to be gleaned and time saved by browsing through patents. I think i will try and multiplex 2 or three coils and record the results with a voice recorder and analyse the tones on a spectrograph, then work towards lots of individual coils later.

Thanks so much, you have helped immensely, S