Discharge curve PI with different materials

[andreapas79]

Hello I found this article ( https://www.lammertbies.nl/electroni...metal-detector ), it talks about the exhaust curve with ferrous and non-ferrous targets.
So I tried to do the same with my SurfPi 1.2, I read the preamp output, but I don't find correspondence with article, where am I wrong ?.
I used a piece of iron and a piece of pure copper (same size), placed them in the center of the coil and read the output. In the article the non-ferrous targets were found to decay before the no target curve, but during my experiment don't work in the same way. Why happen this??

[Tinkerer]

where am I wrong ?
You are not wrong. Yours is the better one.
What you see on your traces, is the faster initial decay of the magnetic part of the iron and the slower decay of the resistive part of the iron. This is why the iron decay curve crosses over the copper curve.
The decay curves we see on a PI, are compound exponential decays. That means there are several exponential decays superposed.

[Qiaozhi]

Hello I found this article ( https://www.lammertbies.nl/electroni...metal-detector ), it talks about the exhaust curve with ferrous and non-ferrous targets.
So I tried to do the same with my SurfPi 1.2, I read the preamp output, but I don't find correspondence with article, where am I wrong ?.
I used a piece of iron and a piece of pure copper (same size), placed them in the center of the coil and read the output. In the article the non-ferrous targets were found to decay before the no target curve, but during my experiment don't work in the same way. Why happen this??

What this clearly demonstrates is that this method cannot be trusted to provide accurate ferrous / non-ferrous discrimination, except in a carefully controlled environment with a small set of specific targets.

[andreapas79]

Thanks for your answer. What I didn't understand is why in the article the NOTARGET curve was in the middle of the two curves and in my experiment it was always above.

What this clearly demonstrates is that this method cannot be trusted to provide accurate ferrous / non-ferrous discrimination, except in a carefully controlled environment with a small set of specific targets.

I am studying discrimination in the Pi and I started following this path, because it would be the simplest and fastest, but I agree with you, Qiaozhi, that this method is not reliable, especially due to the many variables that come into play such as: the size of the object, shape, orientation, distance from the coil.
I did a test by simply varying one of these variables, the distance from the coil, and these are the results.
I will try the way of different frequencies as in the SD2000, I will start by studying the circuit, then trying to realize it in an analog way, and then in case manage it through a microcontroller.

[andreapas79]

For SD2000 i mean multipulse widths when I wrote different frequencies

[green]

Was never able to see effect with a mono coil. Can see with IB coil. Foil and no target peak at same time. Nickel and quarter go negative before going positive(non ferrous target signal is opposite during coil decay). Nail and ceramic magnet go positive(ferrous target signal is same polarity during coil decay). Hope I explained it right.

Looked at chart again. Think I should have kept recorded signal from over scaling. Tried recording without over scaling, trying to make sense of it.

[andreapas79]

I tried to see if anything changed with different pulse widths and different RDump values. The attached photo shows the result of my observation.

[green]

Was never able to see effect with a mono coil. Can see with IB coil. Foil and no target peak at same time. Nickel and quarter go negative before going positive(non ferrous target signal is opposite during coil decay). Nail and ceramic magnet go positive(ferrous target signal is same polarity during coil decay). Hope I explained it right.

Looked at chart again. Think I should have kept recorded signal from over scaling. Tried recording without over scaling, trying to make sense of it.

Traces without over scaling. US quarter appears to be only target that went negative before going positive. I'm confused again.

[andreapas79]

Traces without over scaling. US quarter appears to be only target that went negative before going positive. I'm confused again.

Sorry Green but I never build an IB Metal detector, so I don't know is behavior. For now I build only PI metal, because I love underwater research.
what kind of oscilloscope do you use?

I ask this because sometime with mine, looking in low Voltage and small time period give me some glich... another stuff I must improve it's quality of my electrical components, for now I'm using the cheaper

[green]

what kind of oscilloscope do you use?

Rigol DS1052E, I wanted all the target decays on the same chart so I recorded each one then copied them to Excel and charted.

[andreapas79]

what kind of oscilloscope do you use?

Rigol DS1052E, I wanted all the target decays on the same chart so I recorded each one then copied them to Excel and charted.

Nice, I'm looking for a new PC USB Oscilloscope, I wanna multiple channels to test multiple test points at same time, I found this one, Hantek 1008C. It have 8 channels, it's very limited Frequency range, 0-250kHz, but PI usually have from 400Hz to a couple kHz.. what do you think about it?

[waltr]

Nice, I'm looking for a new PC USB Oscilloscope, I wanna multiple channels to test multiple test points at same time, I found this one, Hantek 1008C. It have 8 channels, it's very limited Frequency range, 0-250kHz, but PI usually have from 400Hz to a couple kHz.. what do you think about it?

If you continue with various electronics projects then 250kHz BW will become a huge limitation.
Also, 8 channels is rarely needed.
Check the Hantek 2 or 4 channel scopes with BW of at least 10MHz but preferably 100MHz with 1Gsps.

[andreapas79]

If you continue with various electronics projects then 250kHz BW will become a huge limitation.
Also, 8 channels is rarely needed.
Check the Hantek 2 or 4 channel scopes with BW of at least 10MHz but preferably 100MHz with 1Gsps.

For now, I have a Hantek 6022BL ... 2 channels, 20Mhz (but I really think no more than 10) and logic analyzer (8ports) ... I thought about the 1008C because I would like to analyze all test points at the same time, study all variables and how they work ... is it stupid in your opinion?
PS: I'm not sarcastic I'm really asking for an opinion...=)

[waltr]

Ahh...if you already have a 2ch, 20MHz scope than maybe the 8ch 250kHz scope could be useful to you.

[green]

For now, I have a Hantek 6022BL ... 2 channels, 20Mhz (but I really think no more than 10) and logic analyzer (8ports) ... I thought about the 1008C because I would like to analyze all test points at the same time, study all variables and how they work ... is it stupid in your opinion?
PS: I'm not sarcastic I'm really asking for an opinion...=)

Two channel scope with external trigger is all I have so I can't compare with the capability of more channels. Mostly I look at amplifier out with channel 1, external trigger on Tx gate command and use channel 2 to look at coil signal(most of the time connected to wire insulation, effects decay less), delay or sample signals at 1 to 10us/division. Integrator out and other signals are lower frequency so use a different time base. Looking at more than 2 signals might be better but I'm thinking not necessary if scope is external triggered. Just my thought.