PI for hammered coins suggestions?

[RH-Lightning]

Hi,

So I have this site where I found about 25 hammered coins. Different years, but most are alike to: https://en.numista.com/catalogue/pieces78795.html
1-1.5cm in diameter, really thin and low content of silver.

Hunted the site initially with T2, then Equinox 800 (found some 4-5 more) and there are only shallow iron signals left. Equinox 800 found those coins on the limit of detection, barely there..
I conclude, that there should be more, just not reachable..

Soil is very easy IMO, yellow forest sand, could run equinox almost full throttle.

Is there a PI design that is DIYable that could be better than equinox 800 in these conditions?
Mirage, Goldscan maybe..?


I have built SurfPi, Pickini, Delta Pulse, QuasarARM and some other designs with success, so some experience is there. Just don't think that these detectors will perform better than equinox.

[Carl-NC]

I think the key with hammered silver coins is a very low pulse delay, maybe down to 6us. This means a lot of attention to detail in the design of the coil, TX, & preamp. You will also want a high-gain design so that will eliminate the direct-sampling projects.

[ionut_mtb]

Hi,

So I have this site where I found about 25 hammered coins. Different years, but most are alike to: https://en.numista.com/catalogue/pieces78795.html
1-1.5cm in diameter, really thin and low content of silver.

Hunted the site initially with T2, then Equinox 800 (found some 4-5 more) and there are only shallow iron signals left. Equinox 800 found those coins on the limit of detection, barely there..
I conclude, that there should be more, just not reachable..

Soil is very easy IMO, yellow forest sand, could run equinox almost full throttle.

Is there a PI design that is DIYable that could be better than equinox 800 in these conditions?
Mirage, Goldscan maybe..?


I have built SurfPi, Pickini, Delta Pulse, QuasarARM and some other designs with success, so some experience is there. Just don't think that these detectors will perform better than equinox.

Iron is masking the coins. Dig all targets and you will find more coins.

[RH-Lightning]

Thanks for the tip, that is what I am actually planning to do after the snow melts.

Place itself is very clean, above mentioned iron is very small nails mostly.

I'm more concerned about the depth. Equinox is a high performing detector, so I'm not sure DIY PI designs can beat that.
GPX series could, but it is not worth it with those coins..

[Elliot]

Thanks for the tip, that is what I am actually planning to do after the snow melts.

Place itself is very clean, above mentioned iron is very small nails mostly.

I'm more concerned about the depth. Equinox is a high performing detector, so I'm not sure DIY PI designs can beat that.
GPX series could, but it is not worth it with those coins..

Hi Mikhail, Carl is spot on. The Gold Max Pi is ideal for you here as it can sample at 6uS, is high gain, and is relatively straight forward to build using the supplied comprehensive instructions. I have sent you a separate email with much more detail to look over.

[RH-Lightning]

Thanks guys!

Maybe I'll try the Mirage with a 12'' coil. Anyway, it's about time for a new project.

[bernte_one]

Contact Sven1, maybe he can give you latest design files

[sinclairuser]

i dont know its specs, but someone i know does well on hammered coins with a garrett infinium land and sea, in fact he never uses anything else these days.
but its not a design you can build really, but finding out its specs could help you find another unit you can build.

[GeoMax]

I think the key with hammered silver coins is a very low pulse delay, maybe down to 6us. This means a lot of attention to detail in the design of the coil, TX, & preamp. You will also want a high-gain design so that will eliminate the direct-sampling projects.

Hi Carl,
can you explain it further, what you mean with ".... You will also want a high-gain design so that will eliminate the direct-sampling projects" I do not understand the meaning of the last part of this sentence. Please can you elaborate this further?

GeoMax

[Carl-NC]

In a direct-sampling design the preamp is the only gain stage so sensitivity is limited by the resolution of the ADC. Most direct-sampling designs I've seen have 10-12 bit ADCs. In a demodulator design you have the preamp gain, the demod gain, and usually a post-demod amp gain before you hit an ADC, often the same 10-12b ADC as before. This gives you a lot more gain before the ADC for more sensitivity. An all-analog design may have have even more gain stages.

[EL NINO]

Here is a comparison of my ..Test ATX on some targets of different conductivity- among others even on a small 14mm -0.5 gram silver hammered .. the farthest target in the test ../ 28cm / ..


Maximum range at the highest sensitivity 13..on small14mm silver hammered coin will be somewhere on the border12 ".. maybe a little more ....


Garrett ATX set in this to a sensitivity of 10 .. .. / 13 is the maximum /, Disc 1, -what is the minimum value of the disc / is the delay /, Threshold 9,volume 13-max.

The ATX test is without Ground balance ...

[GeoMax]

In a direct-sampling design the preamp is the only gain stage so sensitivity is limited by the resolution of the ADC. Most direct-sampling designs I've seen have 10-12 bit ADCs. In a demodulator design you have the preamp gain, the demod gain, and usually a post-demod amp gain before you hit an ADC, often the same 10-12b ADC as before. This gives you a lot more gain before the ADC for more sensitivity. An all-analog design may have have even more gain stages.

Hi Carl,

thanks for the explanation. Now it is much more clearer what you wanted to say in your previous post.

Geomax

[ivconic]

In a direct-sampling design the preamp is the only gain stage so sensitivity is limited by the resolution of the ADC. Most direct-sampling designs I've seen have 10-12 bit ADCs. In a demodulator design you have the preamp gain, the demod gain, and usually a post-demod amp gain before you hit an ADC, often the same 10-12b ADC as before. This gives you a lot more gain before the ADC for more sensitivity. An all-analog design may have have even more gain stages.

True.
But direct sampling at 24bit ADC is something what we would like to see.
For several years i've been using some audio modules with WM8738GED, to make audio links.
There is Atmega48 for simple control and WM8738GED ADC.
I am amazed how good it performs in live guitar playing, for example.
There is no noticable delay and quality is ultra high (44.1K sampling rate).
Though WM8738GED is already old chip.
For years i've been planning to reverse ee the module and see how is connected to Atmega.
Until recently i had not conditions to do that. Maybe in near future i will do that.
Problem with detector would probably be the Nyquist condition, such detector will have to operate not more than at 20kHz in theory and not more than 5-8kHz in practice.
Which is quite sufficient for PI detector, but maybe not for VLF I/B aimed for smallest and finest targets.
Will see.
I am just surprised that no one from the md world did not come up with similar design.
I am not sure about the Deus, i think it uses barely 18bit.
Would be interesting to know about Deus 2.
All other simillar attempts are like you said: 10-12 bit at most.
Probably there are traps and obstacles that i am still not aware of.
For example the later processor power. It may be the obstacle.
This may deserve separate topic to talk about.

[green]

What is the advantage or disadvantage of using a 24 bit ADC vs a 14 bit ADC with a 1000 gain amplifier? Maybe a 16 bit ADC and a 256 gain amplifier.

[Olly]

What is the advantage or disadvantage of using a 24 bit ADC vs a 14 bit ADC with a 1000 gain amplifier? Maybe a 16 bit ADC and a 256 gain amplifier.

I suspect that noise is very much the limiting factor.
For example, most 16-bit ADCs will have an equivalent number of bits of around 14 due to their own internal noise and non-linearities etc.
With a reference voltage of say 5V this means we can reliably discern voltage increments of around 300uV.

But even with very low noise opamps and 1000x gain we can expect the output noise voltage to be somewhere in the region of 6mV pk-pk.
So in this case 16-bits is pretty much overkill and we could actually get just as useful results with a 12-bit ADC which will have an ENOB of 10-bits.
Or we could reduce the gain to 50 and use the 16-bit ADC.

If we are trying to analyse the decay slope then I suspect that it's probably more useful to use a fast 12-bit ADC than a slow 16-bit.