Metal detector with depth measurement

[Jackie]

Hi,

Is there any metal detector project that can able to measure the depth of the object underground? Beside that, it also can able to detect different types of metal such as gold, aluminium etc. Thanks...

[Carl-NC]

Most of the projects can detect all types of metals. A few will discriminate. Don't think any of them have depth meters. A depth meter only works for certain coins in certain conditions.

- Carl

[Max]

Hi,

Is there any metal detector project that can able to measure the depth of the object underground? Beside that, it also can able to detect different types of metal such as gold, aluminium etc. Thanks...

Hi,
there are some that could give you that indication (I'm thinking at a garret model with 3 coil levels). But then there are programs that runs inside a processor that gives you an estimation of depth.
Many times it is very few accurate...and target id also lacks of accurancy if certain conditions aren't meet.

Best regards,
max

[Qiaozhi]

Metal detectors that have depth indication are making an "intelligent guess". This is usually based on a known type of target, such as a coin.

[Leto]

Intelligent Guess

I doubt of any intelligence nor any guessing in Garret GTI series.

And I even firmly belive (..as non EE) that it could be possible with "pure" analogue design and with two Rx coils of different sizes achieve same results, given that the target is well centered below both coils.
Probably this could be even easyer to achieve with PI type detector (less delicate coil tunings).

To affirm this: one should make a detector with double (twin) Rx part and measure ratios of signal strength at the end of recievers for targets with different dimensions.

I imagine results like this:
small target: reciever with small coil detecting.
large target close: both recievers detecting.
large target far away: reciever with big coil detecting

True problem with this is that there are two main variables: dimension and depth --> so Qiaozhi probably you are right! there is some guessing involved but surely no intelligence

[Jackie]

Is there any equation that can calculate the depth of the object underground?

[Carl-NC]

The way detectors calculate depth is to first determine the probable target type. This is done by reading the RX phase shift. Different coins give slightly different phase shifts, so once you know that the target is a quarter and not a dime, you can then look at the RX signal strength and determine how far it is from the coil.

To make this work, you will need some sort of look-up table or algorithm for depth vs. signal strength for each type of target you want to consider. The depth vs signal strength for a silver dollar is quite a bit different than for a quarter. Some are nearly identical, like the cent and dime, and could share a look-up table.

This system works OK when the number of target types is limited, such as with US coinage, which is what US detectors are calibrated to read. But it assumes the coins are buried flat; if they are tilted, it throws the depth reading off.

Wildly varying targets, such as you get in nugget hunting, don't produce reliable depth readings, or even phase readings for that matter.

- Carl

[Max]

I doubt of any intelligence nor any guessing in Garret GTI series.

And I even firmly belive (..as non EE) that it could be possible with "pure" analogue design and with two Rx coils of different sizes achieve same results, given that the target is well centered below both coils.
Probably this could be even easyer to achieve with PI type detector (less delicate coil tunings).

To affirm this: one should make a detector with double (twin) Rx part and measure ratios of signal strength at the end of recievers for targets with different dimensions.

I imagine results like this:
small target: reciever with small coil detecting.
large target close: both recievers detecting.
large target far away: reciever with big coil detecting

True problem with this is that there are two main variables: dimension and depth --> so Qiaozhi probably you are right! there is some guessing involved but surely no intelligence

Hi,
yes probably is true that not real intelligence is involved at now. Just some MCUs programs and lookup-tables, all right.
But maybe in future this kind of stuff could be done by "intelligent" systems e.g. a small neural network (implemented in some newer and powerful MCU) could be trained to detect different signals from a known test garden , also in different soils etc and then replicated in software for production in newer MDs series. This way could be an evolution too in neural network algorithm that don't need user programming (like in actual detectors) to identify new signals/target/soil conditions etc and then users could share (if not locked by manifactures) this evoluted software too.
Many used neural networks do now much more complex tasks than this and I think that some manifacturer is thinking to add one to newer design but at now no one (as I know) has done, cause maybe there's no competition out there pushing out this approach.
Just all are traditional designs that use phase to id target and amplitude at coils to guess depth by lookup or formula(s), and that's all.
Neural network would be more flexible and fast updating on that , much more like what really happens e.g. in animals (hammerhead shark is a good example) to locate remotely by organic senses.

Best regards,
Max

[okantex]

I am not agree with you all sorry.
I will repeat my old claim again.
WE had a detector could discriminate and give dept of metal.
and we found agold coin at 4 meters with identificatin.
and coil was small as mallboro cigaretet box.I mean the one with 20cigars.

how did it work I do ot know.I was a child at that time.and did not have interest at that moment.then it was stolen during earthquake chaous.

we have to search ,search ,search
nothing to say much

[JC1]

Garrett

Max is correct on the GTI

Believe it is 1500 or 2500,

three rx coils which you must have

The patent explains the math.

It is real depth.

the others are guesses at depth

based on coins.

Find the patent and you will find

the depth technique and math

it is real, give garrett that one.

[JC1]

this one

US5786696 (656K) Metal Detector for Identifying Target Electrical Characteristics, Depth and Size Garrett's target ID patent,

http://geotech.thunting.com/cgi-bin/...le=patents.dat

[Qiaozhi]

Intelligent Guess

I doubt of any intelligence nor any guessing in Garret GTI series.

And I even firmly belive (..as non EE) that it could be possible with "pure" analogue design and with two Rx coils of different sizes achieve same results, given that the target is well centered below both coils.
Probably this could be even easyer to achieve with PI type detector (less delicate coil tunings).

To affirm this: one should make a detector with double (twin) Rx part and measure ratios of signal strength at the end of recievers for targets with different dimensions.

I imagine results like this:
small target: reciever with small coil detecting.
large target close: both recievers detecting.
large target far away: reciever with big coil detecting

True problem with this is that there are two main variables: dimension and depth --> so Qiaozhi probably you are right! there is some guessing involved but surely no intelligence

OK - you've taken my words in a literary manner.
By "intelligence", I simply meant the "intelligence" of the designer, not the machine.
The high-end Garrett detectors use two receive coils. By comparing the target phase-shift from these two coils it is possible to "estimate" the depth. Lookup tables are then used in conjunction with the signal amplitude (as described previously by Carl) to "estimate" the target size. Other mid-range Garrett detectors (like the GTAx1000) only have one receive coil. In this case, it's even more of an "intelligent guess". These measurements are nowhere near being precise. Target composition and orientation also have a major effect. I'm not saying that the results obtained are useless. Just that they are not 100% reliable.

[Geo]

This system works OK when the number of target types is limited, such as with US coinage, which is what US detectors are calibrated to read. But it assumes the coins are buried flat; if they are tilted, it throws the depth reading off.


- Carl

Hi. I think that it says All