PI discrimination

[green]

I've looked at amplifier out during coil on time with IB coils with a scope. Thought I could see how to discriminate ferrous-nonferrous. Added steps to my last controller. Turn coil on,10usec delay, 10usec sample, 50usec delay, 10usec sample, turn coil off. First sample(integrate amplifier out normal) second sample(integrate amplifier out inverted). Adjust gain of second sample to GB. I get a target signal out of the integrator and can GB but haven't seen a way to tell ferrous-nonferrous. Flipping nails or coins flat or on edge changes things. One thing I see, using an o0o coil a single target nulls in the center of the coil on both (coil on and coil off) integrators, with more than one target they don't. My question, is there a way to discriminate with a PI and should I look for another way or try more with the coil on integrator?

[württemberger]

I think, it´s not possible to eliminate ferrous Metal with simple PI Detectors.
Shift the 1. Sample Impuls, and you can eliminate all non-ferrous, but not the
ferrous.
I make the "Ferrous-Test" very simple. I push an hold my Test-Switch > If the Signal
lost > then is the target = non-ferrous. If the Signal the same, then = ferrous.

My Test-switch shift the 1. Sample Puls.

[Davor]

Ferrous component reacts immediately to a Tx activation, so you need to find a way to sample at a very flyback, or alternatively during a charging period.

[green]

Been playing some more. It's possible to discriminate on the bench sampling during charging period. Getting it to work on the ground might could be more difficult.

[green]

Been playing some more. It's possible to discriminate on the bench sampling during charging period. Getting it to work on the ground might could be more difficult.

I scoped some targets this morning. I'm thinking, integrate amplifier out for about 10usec just before coil turn off. Sensitive to ground, I'm hoping with a o0o IB coil to help cancel ground and operating non motion, slow poking the target looking for a negative change might work. Got some replies on the diff amplifier I'm using, hoping to get some ideas to try for discrimination

[Chet]

Hi Green
Your results are consistent with your previous tests done during coil on time. I think some kind of ferrous / nonferrous probability meter or readout would work well with the kind of measurements that you are getting. Sound could also be used such as a low grunt or high chirp based on the measurements. You could also use a track and hold circuit to capture the peak measurement as the coil is swung across the target.
Thank you and have a good day,
Chet

[sinclairuser]

hi green, nice work, being able to identify/discrim non ferrous metals with a pulse induction, but not being able to do the same with ferrous metal types has always bothered me.
its counter-intuitive, what is possible in an analogue circuit is usually stated as a limitation here, but i do wonder if it is possible but as yet unknown.

[green]

I tried adding the target sample and a gained sample just before turn off with another integrator. I could adjust the gain to GB the California clay. Looked good, still went opposite direction for ferrous non ferrous. Problem is, the Alabama cay didn't GB at the same setting. Detecting the target while not detecting a change in ground signal is a problem.

[Tinkerer]

Mr. Green, I really like your test graphics. I would like to donate one pre-amp board to you for all the excellent work you have been doing to the benefit of all of us.

Pre-amp board with electronic Induction Balance correction.

Today I received my first made in China. It usually takes quite a long time toreceive the goods from China. This was the great exception.
I sent the Gerber files, output of the DESIGNSPARK software off on Tuesday and received the boards less than a week later, on Monday.
The price was unbeatable and the quality looks very good. I can really recommend PCBWAY for the excellent service. Pictures annexed.

This is the pre amp board, with 1 output of the target signal to the ADC and a second output to an ADC for the MCU, IB field adjustment input.
There are 2 analog inputs to the board from coil A and coil B. These coils can be very different in their characteristics. An on-board impedance network makes it possible to achieve induction balance.

The RX signal is LP filtered at 1Mhz. Sampling is possible extremely early.

[Chet]

I tried adding the target sample and a gained sample just before turn off with another integrator. I could adjust the gain to GB the California clay. Looked good, still went opposite direction for ferrous non ferrous. Problem is, the Alabama cay didn't GB at the same setting. Detecting the target while not detecting a change in ground signal is a problem.

Hi Green

Two possible answers; re-ground balance frequently or constantly track the ground and let the circuity automatically adjust the ground balance.

Have a good day,
Chet

[green]

Hi Green

Two possible answers; re-ground balance frequently or constantly track the ground and let the circuity automatically adjust the ground balance.

Have a good day,
Chet

Hi Chet, Think I'll wait until I try it on the ground. I like the second integrator if ground isn't a problem. Has a higher output for low TC targets and doesn't have a hole.