Pulse Induction Metal Detector With GreenPAK




Introduction: Pulse Induction Metal Detector With GreenPAK

About: Silego provides a development platform based on an easy-to-use hardware and software GUI that allows users to easily create custom ICs. We provide design files and application notes that allow readers & e…

A pulse induction (PI) metal detector can be very easily implemented with a single GreenPAK configurable mixed-signal IC (CMIC). Pulse induction designs have the advantage of tolerance to ground effects, and work even when submerged in water. It also performs well on non-ferrous metals. This instructable uses an analog comparator, digital timing vernier, and digital averaging to provide an LED bar display and audio indication.

You can go through all steps to understand how the GreenPAK chip has been programmed to control the PI metal detector. However, if you just want to easily create the PI metal detector without understanding all the inner circuitry, download GreenPAK software to view the already completed Pulse Induction Metal Detector GreenPAK Design File. Plug your computer to the GreenPAK Development Kit and hit program to create the custom IC to control your PI metal detector. Once the IC is created, you can skip to Step 3.

Steps 1 and 2 will discuss the logic that is inside the Pulse Induction Metal Detector GreenPAK design file for those that are interested in understanding how the circuit works.

Step 1: Pulse Induction (PI) Basics

An electromagnetic field is produced by building up a current in a single search coil, and then switching off that current to allow that field to collapse. As the field collapses, it induces a voltage back into the coil (appearing as high voltage undershoot or flyback) and also into objects near the coil. Ferrous or non-ferrous objects will have eddy currents induced in them producing a small magnetic field which then opposes the decay of the original field. So when the search coil is near metal, the magnetic field around the search coil decays differently, as does the voltage that was induced back in the search coil. The tail end of this decay voltage is what is analyzed by this circuit. Fig. 1a shows the change in coil voltage when metal is brought near .

Step 2: GreenPAK New Circuit Implementation

Typical PI designs incorporate a high gain op amp to amplify the last mV of decay signal, then try to deal with offset, saturation, noise, bandwidth, and filtering issues. The design shown in this instructable is a new circuit that does not use a high gain op amp. The circuit uses an analog comparator to compare the decay voltage against a fixed threshold of 50-100mV. The resulting timing edge (ACMP- OUT) is compared against a fixed delay time (DLY-LAT) from when the coil was de- energized. Since the change in delay is very small, a timing vernier is used with a 4nS resolution.

The timing vernier is constructed with a delay chain that propagates the decayed coil voltage logic “0” past the inputs of a parallel input shift register. The schematic is shown in Fig. 2a.

The tuning potentiometer-ADC-DLY2 path generates the (dly lat) signal which latches the register inputs. So the longer the decay(closer to metal object), the more 1’s will still remain to get latched. The contents of the shift register then get shifted out(by SR CLK) to a counter(CNT4) in Fig. 2b, which accumulates how many 1’s. This cycle continues to repeat. Counter CNT4 binary output bits 0,1,2 are ignored, while bits 3,4,5 are sent to a 3 to 8(actually 6 in this case) line decoder. This results in digital averaging to reduce the effects of voltage and timing noise. These averaged/decoded bits drive an LED bar graph display, and also go to a muxed input divider chain to generate audible beeps. The closer a metal object, the faster the beeps will sound. The beeps are generated by a small piezo transducer driven directly from the GreenPAK output pin. 2kHz was chosen because that frequency was sufficiently loud with only 5V drive signal.

You can recreate this schematic in the GreenPAK designer software, or just use the GreenPAK Pulse Induction Metal Detector design file. Insert a GreenPAK IC into the socket, hit program and your chip can now control the PI metal detector. We will keep the IC in the development kit to allow easy and fast access to the chip pins. However, in a real world design, you can create a tiny PC to make the connections for an ultra-small metal detection device.

Step 3: Other Circuitry

The coil drive requires a high voltage MOSFET connected to a 12V power supply. In this case I have chosen to leave the analog comparator input ground based, so the coil will actually be driven to the -12V connection. This way the induced coil voltage at shutoff will go positive, and the decay voltage will fall from there down to GND. The vertical scale units in Fig 1. are correct... the coil voltage does swing up to 400V, and is the reason for using a high voltage NMOS MOSFET for Q1. The diodes clamp the input to the ACMP within the region of interest. The remaining circuitry is simply level shifting from +5V logic levels down to -12V gate drive level as shown in Fig 2c.

The potentiometer at pin 8 could be any voltage divider between 0-5V supply so that it can provide a tuning voltage to the AD converter which subsequently tunes the metal detector.

This circuit can be easily implemented on a small breadboard or via dead bug style circuit wiring.

Step 4: Sensing Coil

The coil used was wound from speaker wire, a total of 34 turns and measured 320uH inductance. The insulation is kept on to reduce the self capacitance of the coil. Of course smaller coils and different shapes can be used, such as oblong for a security wand. When different coil inductance is designed in, it may be necessary to adjust the DLY5 delay setting to reasonably center dly lat signal. The coil windings should be held securely in place or embedded in epoxy to eliminate any vibration, which causes loss of efficiency. The coaxial cable is RG62, also having low capacitance. Another benefit of using an NMOS for Q1 is that its parasitic capacitance is lower than an equivalent PMOS. All this consideration to minimizing parasitic capacitance increases the amount of remaining signal we are trying to detect.

Resistor R1 in Fig 2c is a damping resistor and determined empirically so that the coil ACMP input won’t have undershoot.

Step 5: Assembly of the Metal Detector

Connect the the deadbug circuit board to the GreenPAK development kit and to the sensor coil. The sensor coil can be taped to the top of a cardboard box lid. Once the lid is closed over the box, you have a nicely concealed sensor coil, and your super cool looking metal detector is ready for use. Apply power and start looking for loose change around the house. Happy treasure hunting!

For more complete instructions on making this staff with additional design considerations, download the Pulse Induction Metal Detector App Note

Be the First to Share


    • Laser Challenge

      Laser Challenge
    • Paint Challenge

      Paint Challenge
    • Edible Art Challenge

      Edible Art Challenge



    Question 4 months ago on Introduction

    Hello, my problem lies with pic 2.c. Sorry, I showed the photo to an electric engineer and we were both wondering why that board looks so ''burned''. I don't know how to build this and would appreciate your help.


    4 years ago

    Now someone just needs to make a diy 3d metal detector. I'm terrible with electronics but know physics and magnetic field sciences well. If you use a pair of cone coils running a pulse or pump z wave(better) or short among 2 grounds in a 3 ground (3 dynamo 3 phase)circuit, and aim those cones at the same point from different perspectives, you create a grid zone and can detect to fantastic depths using no power at all thanks to recent discoveries in zero point energy. The most over efficient designs use no physical connection between dynamos.

    A 3 phase modulation among them will naturally occur. To get more power out you simply extract energy off one of dynamo 3 using a pickup xformer, while gnd1 is given a timed short to gnd2 just as each 3rd neg cycle peaks. Gnd1 should have some way and reason to absorb from a larger radius with ferric material, so charge can properly implode. One way is using a pipe with 3 coils close inside it that is larger than a pipe with 3 coils inside it for Gnd3, so you've formed a fractal of amps/volts among them 2/1.

    A fellow on youtube demonstrates how to gain 300% and more using a simple modification to a stovetop induction heater to provide the pulse/pump z wave. Any pump z carrier will cause carried waves to follow suit if those carried waves implode through a nonlinear medium and focus out again to become a sum of wave peaks. This is what drives the conversion of lf into hf in any zpe circuit.

    ahmet ozdemir
    ahmet ozdemir

    Reply 1 year ago

    please could you explain your idea? ı want to learn about. Thanks


    1 year ago

    Has anyone got a reply, seems like this site is un monitored


    Question 2 years ago on Step 3

    Hello, can you give details of the "Dead Bug" circuit?,
    I have currently just coded a multi frequency, variable duty cycle and amplitude pulse, I would love to incorporate your concept into a commercial product.....are you interested?


    Question 2 years ago on Step 4



    3 years ago

    Hi Silego, my name is Dimitris Charir. I have reade a guide for making a PI metall detector "Pulse induction metal detector with DSP". I am interested in making one , but i have some quesion about the coil and the ground penetration. In the guide said that the detection depth of a coil is five times the diameter and that stronger magnetic fields can penetrate deeper in the soil . If i have a small coil 50cm but we puss a large amount of current throw the coil, the magnetic field of the coil will go deep in to in to ground because it is very strong. You will over saqurato the soul and maybe loss the ability to find small obche you may over-saturate the ground which makes small objects invisible due to background noise but that's ok. In my mind the depth is dependent of the magnetic field and you can have big filds with small coils. Why the depth is dependent of the diameter and if i use a small coil and puss throw high currents, will i have problems? Also am planing to use a simple arduino and a ADC for the discharge graph analysis. I will charge the coil with defernet currents while the coil is stationery so i will have the info to make a 3D image. Can you give any suggestions on this to?


    Question 3 years ago on Step 5

    I like the idea to use such tiny devices, I have a question though, How much it costs an entry level board in order to try out the greepack system? and is it possible to implement it without the board in a real device ?

    Thank you