Pulse induction metal detectors work by pulsing the coil with a series of high current pulses and measuring the time it takes for the fly-back to decay to 0V
These pulses can last anywhere from 50 microseconds to several hundred, If there too long the ground gets saturated with the magnetic field and makes targets hard to locate, around 250 microseconds is adequate.
For the transmitter on this unit we'll use an Atmel Attiny25 (although anything in this series would work), we'll use the analog to digital converter on the chip to read an external potentiometer which will control frequency.
The Attiny25 can only deliver around 40ma, while enough to drive a small load it would be insufficient to switch the mosfet at any reasonable speed, to get around this we'll use a mosfet driver.
Switching a mosfet is a capacitive load which means we need to have a large amount of current to charge and discharge it fast enough.
For this were going to use the TC4420, It takes a logic level signal and outputs up to 6A at rail to rail voltage.
We also need to choose a power mosfet to switch the coil, It should have low capacitance, be able to deliver a good amount of current to the coil and have a fairly high breakdown voltage (At least 200V), more on this later.
The coil is built out of about 55' of wire wrapped in a tight bundle anywhere from 4-24", Providing an inductance of around 340 micro henry's (uh)
In order to prevent ringing in the coil (oscillations) we use a damping resistor, 680Ohms should be sufficient, This will go in parallel with the coil,
Here's what we currently have for the transmitter
Here's what happens:
The mosfet is switched on, the coil conducts and the magnetic field begins to build, after around 250 microseconds the coil is switched off and the current in the coil colapses and the voltage swings several hundred volts in the opposite polarity, this is called the flyback.
This is were the breakdown voltage of the mosfet is critical, the flyback voltage is limited by the breakdown voltage and the higher this flyback voltage can get the faster it can decay.
Faster decay of the flyback is critical for finding fine gold rings or gold nuggets buried deep in the earth.
This is also where the capacitance of the mosfet comes into play, the higher the capacitance, the more damping we need to stop ringing which puts more load on the eddy currents induced in the coil and makes targets harder to spot.
properly damped coil.
In order to measure the flyback voltage with normal electronics we need to filter out the high voltage, to do this we'll use two bi-directional diodes, they breakdown at around 0.7v so the final output voltage will be +/- 0.7v, a resistor will be used to limit current to the filter so we don't cap the whole coil at that voltage.
The eddy currents induced in the targets only become visible in the last few volts of the decay so this approach works fine.
Pulse induction Cycle:
The theory behind the receiver is that we use an analog comparator with a reference voltage of several millivolts, when the flyback starts the comparator's output goes high, when it decays below the comparator's reference voltage it goes low again.
We precisely measure this time.
For the implementation of the receiver we'll use another Atmel Attiny25.
The microcontroller's internal comparator continually compares the the coil voltage (protected by the two bi-directional diodes) with a reference voltage of several mv.
The microcontrollers timer/counter will be clocked at 64MHz, this provides adequate timing resolution to detect the small change in decay time as a target is put in range of the magnetic field.
When the microcontroller catches a rising edge on the comparator (starting of flyback) it clears the counter, when it catches a falling edge it gets the the time and adds it to a array of around 32 samples, which is averaged.
When that average rises above or below the value set by the variable resistor, read via the adc, it begins making sound, freq proportional to the target's signal
bi-directional diodes: 2X BAT45
multi turn trim-pot: 43P-502 or any other multi-turn trim-pot, just match the other resistor
capacitors: values don't really matter, there just used for filtering
DOWNLOAD GCC C SOURCE CODE FILES (WARNING COMPLETE PROJECT FOLDER DUMP... LOTS OF NONSENSE TO SORT THROUGH)
For my setup I used 50' of 24 guage single strand teflon wire wrapped at about 6"
more info coming soon!
To calibrate simple turn the sensitivity knob all the way in one directions and slowly turn the resistance down on the multiturn trim pot until you no loger get a clean sound but instead lots of noise.
The following results were obtained on a cruddy soldered bread board with componets socketed, you should be able to obtain better results by making a real pcb with proper ground planes etc..
PCB BOARDS WILL BE AVILABLE SOON!
1 Gram lead pellet: 2"
1oz gold: 10"
You might also want to check out are latest and most advanced Pulse Induction (PI) Metal Detector, XPI at http://www.miymd.com/index.php/projects/xpi/
To ensure continued developement please comment and consider donating! I need the encouragement guys!