Designing the perfect metal detector .......

While I haven't used the techniques I'm about the descrive to make a metal detector, here's what I have done many times in order to generate a low distortion sinusoidal drive for resolver coil excitation:

1) Generate a square wave at the desired frequency.  Pass it through a pre-filter to eliminate anything higher than 10F.  Now, use a high-Q (or many pole filter) to remove all but the first harmonic.  This technique works, but can be tricky.
2) Use a micro or an FPGA to generate PWM at a carrier frequency 20+ times the desired modulation frequency.  Again use a roughing fikter to remove anything above 10F.  Now Use a 4 or more pole LPF to recover the modulation.  This works EXTREMELY well and the technique has been used in numerous spaceflight and military designs I have done.

-130 db rejection of harmonics?  My feeling is that it's gilding the lily a bit but good work nonetheless  .  You might also consider using a sharp notch filter on the receive side to eliminate any IM.

In this day and age, I would use a DSP to process the receive signal.  With an FIR filter it's possible to get near brick wall filter response and linear phase response to boot.  For non-spaceflight designs, I use DSP heavily with success in motion control application.

john

Linkback:

You are not allowed to view links.
Please Register or Login

http://www.thunting.com/smf/index.php/topic,5368.msg34107.html#msg34107

Hi Moonshadow,
Wherabouts in the U.K. are you?
I'm north herts (Hitchin).....
 

Linkback:

You are not allowed to view links.
Please Register or Login

http://www.thunting.com/smf/index.php/topic,5368.msg34845.html#msg34845

hi . i was going to make my detecter just big enough to fit in my smaller back pack . but still quick to put back in to one .and go   

Linkback:

You are not allowed to view links.
Please Register or Login

http://www.thunting.com/smf/index.php/topic,5368.msg35122.html#msg35122

moonshadow;

I think there is an easier way, you are making something simple, really complex and throwing a PIC into the mix is definately complicating matters

First, I dont think a sine wave matters and you still need to square it (and use one-shots to equalize pulse length) for a phase detector, whether you employ an XOR or a D flip-flop to do the job. The PIC is not a sine wave user, either. Try using an LM567  to filter out noise, regenerate the signal and maintain your phase variations. (you can get a square wave from pin 5 or modified sine from pin 6 but dont load either pin with less than a 10k ohm load.) Using it this way, use 0.1 ufd on each of pins 1 and 2.

By the way, I just finished reading a PDF on CMOS crystal oscillators, they can be designed to produce a sine wave, even an LC CMOS is good. When you are using the Tx frequency for the reference frequency, super stability is not an issue.... let it drift! The source for reference and search (phase change) frequencies are the same.

Ive been considering phase change detection, for some time, (my get up and go got up and went) just never got to it but I have pursued  the idea from a different angle (no pun intended.)

I downloaded your site and will read it and make suggestions, if I see some .... points, otherwise I will just let you build your complex infernalmachine, as you wish. Ha.

I presume you will be adding to your site, as things develope?  I will look, now and then.
Brian AKA goldigger

Linkback:

You are not allowed to view links.
Please Register or Login

http://www.thunting.com/smf/index.php/topic,5368.msg38052.html#msg38052