Best way to divide frequency

[Paul Andrulis]

First, I know that it is next to impossible to provide a solution, without first knowing what is being attempted. Therefore, I will give a heads up of my project.

I do not know how many here are familiar with metal detection circuits, but a common problem with these circuits is discrimination of "trash" signals produced by undesirable objects, such as common iron nails, pop tabs, etc. On many vlf machines, this is easily accomplished. However, I am trying to rig up a unit that can discriminate for a pulsed induction metal detector, not an easy task. A pulsed induction unit sends DC pulses through the coil, waits a period of time for the back emf to decay, and then "listens" for inductive replies from metal buried within the ground. Not much information is sent back by the buried object, and the entire section which distinguishes one metal from another tends to be in the short decay section of the received signal. Some metals have a very short decay time, and others, like iron, have a long decay. However, the decay is an oscillation, and according to what I have tested on my B&K O-scope, it oscillates slightly differently from one metal to the next.

What I need to do is isolate this oscillation from the decay section, spread it out to give a PIC controller time to analyze the signal, and from there it is a matter of programming the PIC.

Could I do this using frequency division, and still save the information from the signal? Would there be a better way? The decay signal is a very short oscillating ramp signal, which has changes in both voltage and frequency as it dies out.

Any ideas, thoughts, or comments on the matter would be welcome.

Paul

 

[Paul Andrulis]

3VO pointed out something to me in chat. The decay lasts for a few milliseconds on the short decay signals. (Unless I misread my scope). For a short decay, there are usually 20-30 oscillations. Therefore, I would need to take a sample at least 30 times per millisecond. (Again, thank you 3VO. I was thinking real time, and not decay time.)

 

[Ian Rogers]

3V0.... Not 3D0!

 

[Paul Andrulis]

3V0.... Not 3D0!

Thanks for pointing that out. No insult was intended, and I changed it.

 

[ronsimpson]

Can you post a picture of your o-scope traces.

 

[Paul Andrulis]

Can you post a picture of your o-scope traces.

I wish. I have an old non-digital 100mhz quad channel B&K, and it doesn't have storage features. I used to use an old Emerson digital camera to take screen shots, but it has gone the way of the dodo. My new one just catches screen glare.

I keep thinking I have seen some shots online though, showing the decay section. I will try to find one, and provide a link for reference if I can.

 

[Paul Andrulis]

This guy has a lot more "snow" in his decay signal than mine has, but it is suitable to demonstrate the principle. Looking over his site, it looks like he is using a similar approach as I am thinking of taking, as far as using a MPU to analyze the signal for discrimination principles.

**broken link removed**

And I was thinking I might have come up with something new... I should have known better. However, he doesn't post any details of how he was going about the situation, as far as I can tell, so I may be taking a different approach at the same problem.

EDIT: ADDED

What is a pain is that he doesn't list his scope settings, so it doesn't really tell much.

EDIT: ADDED MORE

Been reading his site more, and he has a schematic. I think I know where the extra fuzz is coming from. He is using three ATTINY MPUs, as the detector. His filter caps are located by his 7805s, and not close to the voltage inputs on the MPUs. It may be causing hiccups. His claims towards depth in free space truly suck for a PI unit. (No insult, just a statement of fact. PI units are known for deep field penetration, and seven to ten inches into the ground for a silver dime is not uncommon. He should be getting 15 to 17 inches or more in free air using an ounce of gold.)

However, I will give him huge kudos on the concept of using JUST MPUs.... Simple circuit, and I never would have thought of that.