TineFPGA-GOLD Metal detector

[SaltyDog]

@walter @fb_bf

Good luck with your respective projects .. hope they work out for you.

... but my approach is still the same, and won't change cause it works and the performance per $ is much better and much lower cost in energy, my 2c

[SaltyDog]

The only part of your block diagram that needs to be FAST is the ADC.

Disagree: If the "Shape" of the coil discharge is modified by a slow op-amp, then you won't get discrimination ... you NEED a fast op-amp as well to preserve the shape ..

You need to understand ... discrimination info is in the EARLY part of the discharge, not in the tail end where all existing designs work (and therefore don't get discrimination) ....

[waltr]

@walter @fb_bf

Good luck with your respective projects .. hope they work out for you.

... but my approach is still the same, and won't change cause it works and the performance per $ is much better and much lower cost in energy, my 2c

No, I do not mean for you to change. FPGA's are great way to do this sort of thing. Only pointing out that with the right uController/processor this can be done without an FPGA.

Disagree: If the "Shape" of the coil discharge is modified by a slow op-amp, then you won't get discrimination ... you NEED a fast op-amp as well to preserve the shape ..

You need to understand ... discrimination info is in the EARLY part of the discharge, not in the tail end where all existing designs work (and therefore don't get discrimination) ....

Did not mean the 'analog' stages didn't need to be fast. Was only discussing the digital sections as per FPGA needed verse can be done with a uC.

I think you have hit on why most PI detector projects never get to do discrimination.
The 'X' component is only "the EARLY part of the discharge" and most PI detectors do not have a fast enough coil & pre-amp.
I think we know how to build fast coils but the pre-amp is tough.
Most op-amps take some time to recover from overload and this spec is very rarely in the data sheet.

You have gotten my attention and will follow this thread.

[SaltyDog]

Thanks mate ... appreciate the support

[6666]

If you have it working and discriminating why do you need another preamp, I am puzzled.

[SaltyDog]

If you have it working and discriminating why do you need another preamp, I am puzzled.

Don't need another preamp ... just want to speed up/optimize the existing preamp to get more sensitivity/depth .... I am currently getting 200mm on my small gold ring, after a few tweaks ...

I want to improve even on that ..

I am a perfectionist ..... and I am after small gold nuggets WITH discrimination ... This old guy wants to dig only the good stuff ....
( ... and yes, I know there is quite often lead and iron with the gold when prospecting, but in those locations where that happens, will dig it all .)

... also, when hunting on the beach, don't want all the bottle caps/pull-tabs, I am also aiming this to be IP68 waterproof for use in the surf ..

[fb_bf]

One quick question. You said that the adc you were using was sampling >200khz. Then you said that to accurately measure the slope change you needed ns resolution. How fast is the adc sampling with your FPGA? I'm just trying to understand your system. Like I said, I was thinking that what you are ALREADY doing was a great idea. Congratulations on getting to where you are today. It is impressive.

Frank

[SaltyDog]

One quick question. You said that the adc you were using was sampling >200khz. Then you said that to accurately measure the slope change you needed ns resolution. How fast is the adc sampling with your FPGA? I'm just trying to understand your system. Like I said, I was thinking that what you are ALREADY doing was a great idea. Congratulations on getting to where you are today. It is impressive.

Frank

Hey Frank,

The 200khz is a "thumb suck" estimate of what I need to pass through a 2us slope with a reasonable degree of resolution, I am actually sampling the ADC at a much higher rate. (Mhz)
The "ns" resolution I am referring to, is the rate the slope detection module samples ..

The ADC specifics as well as the Slope detection module details I will be keeping close to my chest as they are unique idea's (IP) I want to keep ..
My intention for this project is to provide a module kit for others to use ... I won't be releasing the source ... sorry

[fb_bf]

Thanks for the reply. I understand and again congratulations on getting that running. Very cool idea.

Frank

[green]

Fascinating experiment.

I did investigate this idea in the past (but not with an FPGA) and concluded that attempting to discriminate different targets by measuring the slope was equivalent to discrimination by conductivity. I'm interested in finding out what happens when you get the design off the bench and into the real world.
Lack of sensitivity is most likely due to only using 10-bit ADC sampling directly at the preamp output.
Good luck. Watching with interest.

Also, here's your block diagram the correct way round.

Trying to understand conductivity. A chart of some targets made with copper wire. If conductivity of copper is a constant why does the decay slope change. Does a target have conductivity or does the material have conductivity?

[dbanner]

Oh oh. Hear comes Mr. green.

I forgot about those pesky charts.

[F117]

We also do not need updates to LEDs or audio at the TX pulse rate so I have summed 8 to 32 TX pulse/RX curve cycles before processing and output to user

One thing I also learned with the Pickini project: try to make n samples x pulse interval = equal to or a multiple of the mains period.
Eg here in Europe 50 Hz = 20 ms. Suppose your pulse period is 2 ms. Accumulate 10 pulse samples, instead of the easier to calulate 8 or 16.

[waltr]

One thing I also learned with the Pickini project: try to make n samples x pulse interval = equal to or a multiple of the mains period.
Eg here in Europe 50 Hz = 20 ms. Suppose your pulse period is 2 ms. Accumulate 10 pulse samples, instead of the easier to calulate 8 or 16.

That is a good idea.
Here in USA 1/60Hz = 16.666...ms so accumulate and update user every 33.333..msec would work well.
ADC sampling, TX pulse rate, etc can all be easily adjusted to match the AC Mains frequency.
In the PIC32MX code I do TX rate at 666.6us so 25 cycles = 1/60Hz

[godigit1]

Hi Salty dog
Nice work .
Check out the map lamp is vastly improbable thread
There are a couple ideas there.
Just switching to a 4562 helps.
I have not tried the amp designed for the barracuda but intended to.

[SaltyDog]

Hi Salty dog
Nice work .
Check out the map lamp is vastly improbable thread
There are a couple ideas there.
Just switching to a 4562 helps.
I have not tried the amp designed for the barracuda but intended to.

Thanks, I have already a LM4562 ordered ... taking awhile due to the covid-19 issue ...