Metal detector continuous sine wave (VLF) technology howto

[eneuro]

Hello,
Just looking for another explanation of this continuous sine wave (VLF) technology in metal detection, while read this Minelab technologies documentation  http://www.minelab.com/emea/consumer/knowledge-base/minelab-technologies and it is not clear  how this VLF method works and if It can be used with Monoloop coil or maybe I need Double-D coil or Concentric coil ? 
I have no idea so far what timeings in this VLF method could be and why signal send out and received not changeing? How to detect metal using this method?

VLF (Very Low Frequency) is a conventional single frequency sine wave transmission technology. This is the traditional type of technology used in most basic metal detectors. Single frequency sine wave detectors create an electro-magnetic field, which is transmitted into the ground in a continuous wave. While the basic way that the signal is sent out and received back has not changed, the receive signal processing has improved enormously. This has proven to be a reliable and easy to use technology for finding gold.

But I saw on youtube videos quite sucessfull metal detections with single coil (monoloop) and I'd like to design similar one maybe using VLF and implementing something Minelab names 3F-Selectable Three Frequency Transmission.

3F (Three Frequency) provides three different transmit frequencies in the one metal detector, selectable at the flick of a switch. Each transmit frequency optimises the detector for different size targets and conditions. The three transmit frequencies are:

    6.4 kHz - Best for large deep gold nuggets
    20 kHz - Best for general gold detecting
    60 kHz - Best for small gold nuggets

Having three selectable frequencies gives the versatility that is equivalent to three conventional single frequency detectors.explanation

I found a few patents used in art of metal detection Minelab E-Trac, but its a lot of  .

I'd like to try design circuit for VLF with 3F using monoloop coil if possible.
Of course I can put there even 3 microprocessors if needed and hardware PWM oscilator for wave generation.
Maybe instrumentation amplifier could help too, but what the timeings and requirements might be in VLF?
How to design metal detection circuit using those methods?
It could be nice have own made PCB, while you are able to fine tune circuit for what you need and what are you looking for and where

[KerryW]

There are basically 3 types of metal detectors, BFO, IB and PI.

BFO - a single coil that is part of an LC oscillator.  Frequency goes down in the presence of ferrous metals and up in the presence of nonferrous metals (or vice versa).  The cheapest and least effective type.  Usually has a 2nd tunable oscillator.  The user adjusts the 2nd osc. to match the 1st.  the 2 oscs are "beat together" and user listens for a change in frequency.  Might be better if a frequency counter was used to detect changes instead of a beat oscillator.

IB - Induction Balance uses 2 coils (Double D or concentric).  The coils are arranged so that a signal in the first coil  causes little or no signal in the 2nd when no metal is present.  Metal causes the magnetic lines to move, causing a signal in the 2nd.  This is the most popular type.  It can distinguish (to an extent) between different type of nonferrous metals.  If the coil frequency is low, it is called a VLF.

PI - Pulse Injection is a 1 coil design, although some use 2 coils.  It sends a large pulse of magnetic energy and measures the duration of the inductive spike generated in the coil.  The presence of metal causes the discharge of the spike to be extended slightly.  It cannot discriminate between ferrous and nonferrous metals.  It can detect at greater depths than IB, but is only useful where there is little crap metal.

[eneuro]

Thank you for metal detection overview 

IB - Induction Balance uses 2 coils (Double D or concentric).  The coils are arranged so that a signal in the first coil  causes little or no signal in the 2nd when no metal is present.  Metal causes the magnetic lines to move, causing a signal in the 2nd.

This 2nd coilin concentric setup cofuses me, why eddy currents do not induce in this 2nd coil (smaller receiver coil inside... bigger generator coil ) while 1st coil generates changeing magnetic field? 

Maybe, in presence of unknown metal we are looking for 2nd coil receives signal, but there is difference beetween generated one and thus they wrote something like this in their (Minelab) knowledge base:

Single frequency sine wave detectors create an electro-magnetic field, which is transmitted into the ground in a continuous wave. While the basic way that the signal is sent out and received back has not changed, the receive signal processing has improved enormously.

But this last sentence is strange: why "the receive signal processing has improved enormously" ?

Maybe it is time to dig those Minelab patents  http://www.minelab.com/patents  to see what do they mean, because of its knowledge  base and terms like VLF, 3F, BBS, FBS, etc. looks like marketing terms used to explain customer that he has to pay close to $2k for such "advanced" technology used in Minelab E-trac, while they simply named different common things 

The trick is to get the coil size just right

Just wondering why Minelab has no such coil in his products   

Hwo to build  XTR 11  elliptical search coil at home is here:


Maybe it could be another patent, to... make this elliptical search coil rotation ability, so we could track bigger field and then go down to the size of smallest ellipse dimension simply by rotation of this big elliptical coil and of course using different wave frequency like those specified in Minelab 3F and many more holding on our sholders big 12V VRLA battery used in... electric vehicle power supply

[NiHaoMike]

There are basically 3 types of metal detectors, BFO, IB and PI.

BFO - a single coil that is part of an LC oscillator.  Frequency goes down in the presence of ferrous metals and up in the presence of nonferrous metals (or vice versa).  The cheapest and least effective type.  Usually has a 2nd tunable oscillator.  The user adjusts the 2nd osc. to match the 1st.  the 2 oscs are "beat together" and user listens for a change in frequency.  Might be better if a frequency counter was used to detect changes instead of a beat oscillator.

IB - Induction Balance uses 2 coils (Double D or concentric).  The coils are arranged so that a signal in the first coil  causes little or no signal in the 2nd when no metal is present.  Metal causes the magnetic lines to move, causing a signal in the 2nd.  This is the most popular type.  It can distinguish (to an extent) between different type of nonferrous metals.  If the coil frequency is low, it is called a VLF.

PI - Pulse Injection is a 1 coil design, although some use 2 coils.  It sends a large pulse of magnetic energy and measures the duration of the inductive spike generated in the coil.  The presence of metal causes the discharge of the spike to be extended slightly.  It cannot discriminate between ferrous and nonferrous metals.  It can detect at greater depths than IB, but is only useful where there is little crap metal.

There's also the type that looks for a drop in Q from eddy currents. A cheap unit (that's also a stud finder, flashlight, and small tape measure) from Harbor Freight works on that principle. It's basically an oscillator and an adjustable comparator, all done with discrete transistors!

[eneuro]

There's also the type that looks for a drop in Q from eddy currents.

What Q there is? Some kind of energy or something else?

PI detection method to locate our dig field looks good.
However in this live hunting with Minelab E-TRAC it looks like 2 coils in Double D  setup are used and I guess, they use BI detection method maybe, while are able to find copper monet.

It is also used additional small metal detector which helps fine locate while digging at preselected place which is cool.
They also put some metal objects into the ground and explain how it affects this detector readings.
I've played with artifical neural networks at the university some time ago and it could be interesting to create training patterns, by putting different objects into the ground at known deep and teach neural network to recognize them before we go to real "battle" field 

I'm new to metal detection, so no expectations at all, but it will be fun to clean home outer space from unwanted steel nails, so any detection success will be acceptable at the begining 
It is interesting how many turns such Double D single coil might have and how much power they put into them to create this electromagnetic field?
Are there ampers in pulse or miliamps?

It is a long way to go, but it is challenge so it is always motivation to do so 

[G0HZU]

Many years ago I designed my own metal detector to look for something in the garden...

I made it using the double coil method. I think you can sub divide this type even further because some types use IQ 'motion' detection. i.e. you have to move the head across the metal to see a signal. The crude thing I designed just had manual balance controls.  I used many turns of Litz wire for the coils. It was extremely sensitive and could find a small metal pin or staple at several inches.

But it was pretty hopeless as a general metal detector as it had poor discrimination between target types and the balance controls were too fiddly. If I was to revisit this again I think I'd use a miniature netbook (in a shoulder pack?) with its soundcard chipset as the IQ back end of the detection system.

[NiHaoMike]

What Q there is? Some kind of energy or something else?

http://en.wikipedia.org/wiki/Q_factor
Basically, an air core inductor with a low series resistance connected to a low loss capacitor has an intrinsically high Q, but bring a conductive object near it and the eddy currents will make the circuit more lossy. Generally only metals are conductive enough to have much effect.

[KerryW]

Try this link:

http://www.geotech1.com/cgi-bin/pages/common/index.pl?page=metdet&file=projects.dat

[eneuro]

Try this link:

Metal Detector Projects 

Q_factor

I like it! It sounds like "Star Trek" future technology 

If I was to revisit this again I think I'd use a miniature netbook (in a shoulder pack?) with its soundcard chipset as the IQ back end of the detection system.

If we had Android tablet with internet connection, one could try write own application to connect our metal detection microcontrolers via optoisolated I2C interface and tablet could postprocess data and even send it via internet with GPS position to our home bigger "supercomputer" which could run neural network application and give us hints  it is under our coils  and if it is worth to dig there or forget about it, but we could have this place stored at home PC laptop and go back there later with better trained metal detector using Android GPS navigation software

So, it is time to dig good induction balance & pulse injection circuits 

[eneuro]

Just trying to remove two 555 in one IC (CMOS ICM7556IPD) from this induction balance "worlds simplest IB design" (matchless) http://www.geotech1.com/pages/metdet/projects/matchless/matchless300c.pdf and use AVR ATTiny85 (DIP8) 

Original assembly looks like this:

It uses 2 coils in Double D :   70 turns @ 120mm diameter before forming D shapes each.

My concept is to remove 555 and put ATTiny85, change  battery to 12V VRLA rechargable.

Left side (Tx coil) is quite easy part I think, while there is about 10mA current limited by this 680 ohm resistor @ about 700Hz 50% duty cycle.
But Rx coil detection is more tricky.
I found 555 funtional diagram truth table

so I think I could try to simulate in software 555 Thresold & Trigger influence on oryginal output send to speakers, based on ADC readings.

I have no idea for the moment why this capacitor 0.001uF in series with Rx coil has such value?
We have 70 turns in those 2 coils and they are made on 120mm diamater templates.
Is it some kind of radio receiver here and this capacitor is choosen for inductance of this Rx coil for resonance?
Just trying to look into different basic radio receiver circuits and guess what happends if i change Tx frequency lets say to 1kHz (duty cycle 50% no change)-do I need change this capacitor on Rx coil side or number of turns? 
Also I'm not sure how to scale those yellow resistors (on Rx coil side) If i'd rather power this Rx right side with 5V (not 9V) to fit into AVR adc range or maybe even use precision voltage reference 2.5V on AVREF, so need to scale it down to 2.5V range?

I'd like to start with something like "Hello to IB metal detection world" close to this oryginal circuit, but in AVR ATTiny85 version, while I'd like to experiment later with bigger Tx coil currents & voltages, different frequencies and another post processing.
Maybe I could be able skip tune and fine tune potentiometers and try hardcode in software this bias while I will have ADC moveing average and will try to send it to PC or scope to see what happends if I spot metal under the coils