Super BFO

[Tec]

I'm working on a different design.

Much simpler,
(the SPI MX is complex,;complex software and complex circuit)


Hopefully as I progress I will make a new project title 'Super BFO'.

But here is a summary of progress so far, with thoughts.


The detector is like a BFO, it has a free-running oscillator based on the coil (see attached schematic).
It is an 'Armstrong oscillator'.

However the frequency shift is being measured by a processor running an 80mHz, nice and accurate (rather than by beating frequencies together)

I am using a PIC32 processor on a low cost off the shelf development board ' PIC32-PINGUINO-OTG'.

At the moment the frequency is 16.6 kHz.

Coil:
Not much different to a PI coil, all the same rules apply: low capacitance, shielded, inductance (~ 300uH).


Caps * use temperature stable MKT capacitors for the oscillator.

Results:
Air testing only, oscillator is very stable.
Just as sensitive as my PI detector, discriminates between iron bottle cap and coins.
Iron = decrease in frequency, Non Fe (eddy currents) = increase in frequency.


General Thoughts.

A pulse induction metal detector sends a burst of frequencies through the target.
This sends a single frequency.

The coil is designed to be best for the target and ground conditions, based on work in pulse induction detector.
So, really shouldn't we end up with a coil that is almost the same as that used in a mono pulse induction detector ?


Targets ranging between a small 1cm square piece of al-foil to large copper coin, roughly equates to a response of '10uSec to 70uSec' in a pulse induction detector.
This equates to a frequency range of 100kHz to 14kHz.
So, depending on target size, there will be a specific frequency that works best for that target.


Tec

[6666]

discriminates between iron bottle cap and coins.

Sounds good, keep up the good work.

[joop]

Never seen a BFO controlled with a microprocessor. Should be interesting!

[Davor]

What you describe is called a "pulse counting FM" detection, and is easily achieved by less sophisticated methods. Fun part is that pulse counting approach is far more linear and has a potential of better S/N than more common FM detection methods, so you'll find it in high-end FM receivers. Even funnier, it dates back to valve era. The only downside to this method in use as FM detectors is that they require strong enough carrier signal, but it is not a problem with metal detectors - they are hooked directly to the oscillator.

At the moment I'm not at liberty to visit my "lab" often enough to explore this principle, yet IMHO this is a worthy principle, especially if combined with some off-resonance approach.

Just a little suggestion on the schematic... You may wish to include some 10-22ohm emitter degeneration resistor, as it is recognised as a simplest method of reducing phase noise produced by the transistor 1/f noise.

[Tec]

Better Oscillator and Bottle Cap Trivia.

Improved oscillator attached (just built and tested), This is more stable and sensitive.
'Hartley' Oscillator.


Good Air tests.

Thanks Davor, yes I will try a resistor on the emitter for more stability.


I am counting the pulses, and measuring the period with the micros' xtal based 40MHz clock.
i,e Measure the time for 1000 pulses.

So the resolution is 1 part in 40 million !


Bottle Cap Trivia:

At 28kHz, the bottle cap responds like a non ferrous target (coin) = increase frequency
At 14kHz, it responds like a Ferrous target = decrease frequency.


Also,

28kHz much smaller targets (foil cap) are detectable, where as they were invisible at 14kHz.
As expected.

[Tec]

Wet sand, test at the beach, always learning.

Here's an update on my progress.

Test at the beach was good on Dry sand, found some coins, can hear iron and bottle caps easily (different from non ferrous).
Wet sand is no good


Lowering the frequency from 20kHz to 14kHz, made it quieter on wet sand.


Pulse Induction vs Single frequency:

I believe single frequencies can be compared to 'sample delay' on a Pulse Induction:

A sample delay of 20uSec corresponds to a frequency of 12.5kHz..

i.e roughly the pulse delay x 4...
Think of the target discharging as one quarter of a sine wave...

This is the rule of thumb I am using to compare the performance of the BFO to experiences with the PI:

eg, Say Wet-sand becomes a problem at 15uS sample delay and smaller,..this would corresponds to 16.6 kHz and above..



TWO FREQUENCIES AT ONCE

I was hoping to switch between two frequencies as a way of cancelling out mineralization...

And accidentally got the oscillator to oscillate at two frequencies at once:
Frequency1 and Frequency2 (3rd harmonic)... never thought of that.
Could be useful to someone..


Tec

[Tec]

Printed Circuit boards

I have designed some circuit boards and assembled one.

There were a few minor changes, but the boards work.

I will do some more testing and then post the final schematic.
I intend to sell the assembled boards ~$35.

Summary of the 'Super BFO':

It's a resonant oscillator using a single mono coil.
Similar to a traditional BFO detector.

It measure tiny frequency changes.
It has a USB port it can be plugged into a PC and the frequency can be displayed (to one part in 80 million)

It is useful for creating a BFO-Style ('Super BFO') detector, with different coil shapes and sizes.
Also can make an excellent pinpointer (small coil).


The audio is varying frequency and volume like high-end detectors.

Fe targets decreases the frequency, Non-Fe increases the frequency.

Firmware can be updated via USB.
It uses a PIC32



Tec

[6666]

Hi Tec
sounds good please keep us informed, and I would like a board thanks.
What value of coil inductance are you useing ? my intention would be to use this at the beach.

[WM6]

Good work from you Tec again. Pro board, very nice. Congratulations.
Why you avoid to count on third harmonic? It should be more sensitive.

[joop]

What kind of depths are you getting in the field?

[6666]

Thanks to Tec, my pre made pcb arrived today.

Looks a good job, now just have to figure which is the correct coil to make.

6666

[6666]

I have done a couple of quick tests only on the bench with a 400uH unshielded 25cm mono coil connected with a 1 meter length of coax.

It works, loud audio, coil is a bit sensitive to my hand, but needs more precise testing.

[Shawkins]

Hi, my mx3b has somthing going wrong - all targets sound exactly the same, 24c gold ring, iron file, lead sinker, copper,
pliers, magnet etc. All metals create a decreasing tone I'm using a 398uH coil and a damping resistor of 260ohm. Any help
would be appreciated.

[6666]

What damping resistor ?

I am just useing a straight coil 400uH
and I can hear a difference in some targets, I have not documented them yet
just had a few blockers like xmas in the way.

[Shawkins]

The unlabeled resistor parallel with L2 (coil) from pin 1 to 2 in "sheet 2 of 3" circuit diagram. Above the coil whare it is marked connector (on diagram) and adjacent to the coil pins on the board. I assume it's a damping resistor? (or serves a similar purpouse). Thanks for the timely reply, merry xmas!