why not to sample the whole decay curve out of pre amp?

[Mr.Jaick]

hello
i'm wondering why we don't sample the whole decay curve out of the pre amp of a pi machine
because i know some targets with lower TC tends to die faster on early stages of the curve and some last to later stages
to detect them all at the same time why we don't feed the whole curve into the integrator to add up every target within the coil MF?
i know i'm missing something but can't realize...sorry

[waltr]

You can do this by adjusting the first sample duration.

Secondly, by having the second sample following one can detect low conductive verse high conductive targets. I do this on my HammerHead as per my build thread here:
https://www.geotech1.com/forums/show...ake-on-the-HH2

Third, I have experimented with sending the Pre-amp output into an ADC (PIC32) which samples the entire decay curve (0 to 128usec after TX OFF at 2usec rate). The code averages these over 4 to 16 cycles then does an exponential curve fit to obtain the target TC. Seems to be promising but never put this into a full build the field test.

If you use a micro-process for the Pulse timing then very easy to experiment with various sample timing.
Then setup some tests to learn if and how will this works.

[Carl-NC]

There are two ways to interpret your question:

1. Why don't we sample, say, 20 points on the decay instead of 1 or 2.

2. Why don't we widen the sample pulse to, say, 100us instead of 10 or 15 us.

I tend to think you are asking the latter. Yes, you can widen the sample pulse and definitely improve the response to high conductors which tend to be slow. However, I have found on some systems (like a security walk-through) that it makes it noisier. But not so on my pinpointer, it just equalizes the response to all types of coins.

This is only viable on simple single-sample PI designs. Subtractive ground balance designs require a close-in ground sample so the target sample has to remain fairly narrow.

[Mr.Jaick]

You can do this by adjusting the first sample duration.

Secondly, by having the second sample following one can detect low conductive verse high conductive targets. I do this on my HammerHead as per my build thread here:
https://www.geotech1.com/forums/show...ake-on-the-HH2

Third, I have experimented with sending the Pre-amp output into an ADC (PIC32) which samples the entire decay curve (0 to 128usec after TX OFF at 2usec rate). The code averages these over 4 to 16 cycles then does an exponential curve fit to obtain the target TC. Seems to be promising but never put this into a full build the field test.

If you use a micro-process for the Pulse timing then very easy to experiment with various sample timing.
Then setup some tests to learn if and how will this works.

I'm definitely talking about the first sample
i'll do microprocessor experiment as soon as i get some time and resources
thanks for the HH thread, i'll check that out

[Mr.Jaick]

There are two ways to interpret your question:

1. Why don't we sample, say, 20 points on the decay instead of 1 or 2.

2. Why don't we widen the sample pulse to, say, 100us instead of 10 or 15 us.

I tend to think you are asking the latter. Yes, you can widen the sample pulse and definitely improve the response to high conductors which tend to be slow. However, I have found on some systems (like a security walk-through) that it makes it noisier. But not so on my pinpointer, it just equalizes the response to all types of coins.

1. are you talking digital when you say 20 points because i think analog is continuous and has infinity points(within frequency limits) or you just mean to have 20 little sample windows
instead of like one big that covers most of the curve?

This is only viable on simple single-sample PI designs. Subtractive ground balance designs require a close-in ground sample so the target sample has to remain fairly narrow.

is it a plus to sample very wide on decay for single sample machines? maybe it'll be ground sensitive?
i personally just care about good conductive materials(EVERYBODY DOES!)

[Carl-NC]

1. are you talking digital when you say 20 points because i think analog is continuous and has infinity points(within frequency limits) or you just mean to have 20 little sample windows
instead of like one big that covers most of the curve?

Yes, 20 sample windows. You could do this with 20 demod circuits, or direct-sample with a fast ADC.

is it a plus to sample very wide on decay for single sample machines? maybe it'll be ground sensitive?
i personally just care about good conductive materials(EVERYBODY DOES!)

Might be noisy, might be sensitive to ground. You'll have to try it & see.

[Mr.Jaick]

yeah I have to try this with Felezjoo Pi cuz it has the capability for my tests for now; though it's an embarrassingly bad design in my opinion

[dfbowers]

I don?t know if this has been considered but if you are searching for small gold or micro jewelry, a narrow sample window is much better than sampling the whole curve. The fast decay from small gold will get averaged out. Cyril Smith sends me a lot of his designs and his goals are often small gold an micro jewelry. Just for that reason he will have a narrow & wide selection switch. I see your point on wanting to do both.