AMX RX

[detectormods]

https://www.analog.com/media/en/tech...s/623637fb.pdf

this is in the gold monster as the adc driver, diff to single line.

I think this is the adc that that they use in the coil itself, https://www.analog.com/en/products/l...oduct-overview

[ivconic]

Do you know of any ADC with +/- 5V input that we could use?
Would it help if I redraw and simplify the RX ?

I am impressed with an obsolete one, WM8738, I have dozen modules with it and made audio links with it. Impressed with what I saw in practice. Live instrument playing over it, no latency that may be easy noticed especially in live playing.
You know; dealing with guitar preamps and effects for guitars is pretty similar as with dealing with some parts of metal detectors. And the speed and resolution of such ADC is playing crucial role in that story.
Fast guitar strings picking and no latency at all; that's what I call "The ADC"!
It is single supply, I tested it with 3.7-4.2V battery. It has I2S and that's limiting factor to use some dev boards. But many better and more powerful dev boards do have I2S.
As for the RX frontend; I planned (and obtained in local) LME49720, look its datasheet, looks also impressive and quite up to this task.
But... !!!
Than I randomly stumbled on this article and was completely shocked:

https://www.audiosciencereview.com/f...-family.10687/

[Tinkerer]

detectormods and ivconic, thank you for the valuable contributions. We are making progress.

Maybe we should start a new thread for the digital part which is really going to be the main part of the design.
"Does AMX Digital" sound OK?
There are many decisions to be taken for the digital part. My knowledge in this field is about zero, so please somebody with knowledge step in.

We need to decide between direct sampling and analog integrated amplifier and then ADC.

With The analog integrated signal, a slow 12 bit ADC might be good enough.

With direct sampling, like a preamp with a gain of 1000 input to the ADC, we should use a high Msps 24bit ADC.

In any case, the choice of MCU and ADC should be for relatively new products because it is no easy task to change the software once countless hours have been spent to get everything just right and tested and the right algorithms designed and tested and debugged.

[ivconic]

detectormods and ivconic, thank you for the valuable contributions. We are making progress.

Maybe we should start a new thread for the digital part which is really going to be the main part of the design.
"Does AMX Digital" sound OK?
There are many decisions to be taken for the digital part. My knowledge in this field is about zero, so please somebody with knowledge step in.

We need to decide between direct sampling and analog integrated amplifier and then ADC.

With The analog integrated signal, a slow 12 bit ADC might be good enough.

With direct sampling, like a preamp with a gain of 1000 input to the ADC, we should use a high Msps 24bit ADC.

In any case, the choice of MCU and ADC should be for relatively new products because it is no easy task to change the software once countless hours have been spent to get everything just right and tested and the right algorithms designed and tested and debugged.

"AMX Digital" is alright as far as I am concerned.
Yes that's most important question, which is about to decide the whole project destiny; direct sampling or analogue integrator.
In the "ideal world", (if I would be able to easily and cheaply get everything I need); I would always choose direct sampling.
But almost none of the conventional development systems have a processor with a good ADC (24bit is ideal).
So for good direct sampling, it is necessary to choose a good external ADC.
And it is... a can with many worms!
In the current conditions (this "current" has lasted decades for me);
it might be smarter to choose an option with an analog integrator that will "chew" the signal to the extent that even
a 12bit ADC is more than enough for further processing.
I say "12bit" on purpose, because the vast majority of various development systems have a processor on them, mostly with such an ADC.
And that choice will make this project a "people's" project. For every pocket, for every taste, for the vast majority of participants.
But! As in the previous case with your RX design; one should try to avoid "exotic" component values.
One Dave Johnson is enough for this life of mine!

[Willy Bayot]

"
One Dave Johnson is enough for this life of mine!

I fully agree

[ivconic]

Having said that what I said in previous post; there is very next question to answer instantly, before we start any work.
Power supply!
Another large can with very fat worms!
Analogue integrator shall provide such "chewed" signal to 12bit ADC, to be in common range 0-5v.
Alternatively in range 0-3v3.
But most of the "easy" development systems are taking 0-5v to ADC.
Very IMPORTANT detail to consider!
This would further lead to a delicate selection of opamps that should ideally be single supply.
In range, again 0-5v, alternatively +/-2.5v (LME49720 seemed to me ideal choice because of that, along with other nice specs).
To avoid additional conversions. But it is not mandatory. Additional conversions can be done with a few compromises.
Having this in mind; it is obvious that TX part would need separate power supply. Not a big deal though.
As you can see; it matters much from which "end" you start planning the project.
Because one choice directly determines the next in order.
I am in favor of the described approach.
Because I always keep in mind that it is preferable device to be powered by a LiPo 18650 battery. One or more.
Because today we have really many good solutions available when it comes to such power supplies.
I would follow that logic, while folding the whole project, stage by stage.

[Tinkerer]

"AMX Digital" is alright as far as I am concerned.
Yes that's most important question, which is about to decide the whole project destiny; direct sampling or analogue integrator.
In the "ideal world", (if I would be able to easily and cheaply get everything I need); I would always choose direct sampling.
But almost none of the conventional development systems have a processor with a good ADC (24bit is ideal).
So for good direct sampling, it is necessary to choose a good external ADC.
And it is... a can with many worms!
In the current conditions (this "current" has lasted decades for me);
it might be smarter to choose an option with an analog integrator that will "chew" the signal to the extent that even
a 12bit ADC is more than enough for further processing.
I say "12bit" on purpose, because the vast majority of various development systems have a processor on them, mostly with such an ADC.
And that choice will make this project a "people's" project. For every pocket, for every taste, for the vast majority of participants.
But! As in the previous case with your RX design; one should try to avoid "exotic" component values.
One Dave Johnson is enough for this life of mine!

Have you looked into possibilities of having the PCB's assembled in China?
What components do they have available? Electronic Components Distributor - LCSC Electronics
How can the components they have not available be sourced?

Potentially this might solve a lot of problems for you?

Have you tried JLCPCB? How does it work for you?

[ivconic]

I am using JLPCB and PCBWay services for few years.
Only once I oredered pcb assembly and there was small annoyance with them, about few components which they didn't manage to obtain.
So we exchanged few emails back and forth until agree on some substitutions.
Assembly is solution for small SMD components. But looking at the average prices of some opamps (for example some from your schematic), some ADC chips etc; I guess, if they are to obtain them: prices will not be so affordable.
So, before I decide to use their pcb assembly service; first I must be 100% sure it is fully functional, bug free, design.
Long way to that. But yes, that's the option for sure.
Usually goes like this; you provide them with BOM and pick 'n place files, they put up Excell like sorted BOM with precise component list and send you back to aprove.
If there is (usualy is some) component they have hard time to obtain; they note it in the file. And it is up to you to point them on source, where they can obtain it.
You add direct link to that component in file and send it back to them.
Once the whole list is negotiated (usually one to two steps, back and forth emailing); they start producing. It last short, and in few days all is finished.
Good service, I am very satisfied with them. So far no problems.
But first things first; to complete the design, make prototype and proof of concept. Test it and check for mistakes, bugs etc.
That part of process is much easier to do at home and if possible in through hole arrangement. Or at least using 1206 SMD size components.

[Tinkerer]

Having said that what I said in previous post; there is very next question to answer instantly, before we start any work.
Power supply!
Another large can with very fat worms!
Analogue integrator shall provide such "chewed" signal to 12bit ADC, to be in common range 0-5v.
Alternatively in range 0-3v3.
But most of the "easy" development systems are taking 0-5v to ADC.
Very IMPORTANT detail to consider!
This would further lead to a delicate selection of opamps that should ideally be single supply.
In range, again 0-5v, alternatively +/-2.5v (LME49720 seemed to me ideal choice because of that, along with other nice specs).
To avoid additional conversions. But it is not mandatory. Additional conversions can be done with a few compromises.
Having this in mind; it is obvious that TX part would need separate power supply. Not a big deal though.
As you can see; it matters much from which "end" you start planning the project.
Because one choice directly determines the next in order.
I am in favor of the described approach.
Because I always keep in mind that it is preferable device to be powered by a LiPo 18650 battery. One or more.
Because today we have really many good solutions available when it comes to such power supplies.
I would follow that logic, while folding the whole project, stage by stage.

Thank you for the very useful feedback.
Yes, the PSU thread is next to be started

[Tinkerer]

I am using JLPCB and PCBWay services for few years.
Only once I oredered pcb assembly and there was small annoyance with them, about few components which they didn't manage to obtain.
So we exchanged few emails back and forth until agree on some substitutions.
Assembly is solution for small SMD components. But looking at the average prices of some opamps (for example some from your schematic), some ADC chips etc; I guess, if they are to obtain them: prices will not be so affordable.
So, before I decide to use their pcb assembly service; first I must be 100% sure it is fully functional, bug free, design.
Long way to that. But yes, that's the option for sure.
Usually goes like this; you provide them with BOM and pick 'n place files, they put up Excell like sorted BOM with precise component list and send you back to aprove.
If there is (usualy is some) component they have hard time to obtain; they note it in the file. And it is up to you to point them on source, where they can obtain it.
You add direct link to that component in file and send it back to them.
Once the whole list is negotiated (usually one to two steps, back and forth emailing); they start producing. It last short, and in few days all is finished.
Good service, I am very satisfied with them. So far no problems.
But first things first; to complete the design, make prototype and proof of concept. Test it and check for mistakes, bugs etc.
That part of process is much easier to do at home and if possible in through hole arrangement. Or at least using 1206 SMD size components.

How does their shipping work for you?

[Tinkerer]

I think this is the adc that that they use in the coil itself, https://www.analog.com/en/products/l...oduct-overview

detectormods, thank you for your valuable input.

You have seen Carl's wish list for the AMX.
What is your take on such a detector? Would it be useful and desirable in Australia?
We would appreciate your comments and opinion. Thank you.

[David_1]

I am using JLPCB and PCBWay services for few years.
Only once I oredered pcb assembly and there was small annoyance with them, about few components which they didn't manage to obtain.
So we exchanged few emails back and forth until agree on some substitutions.

The way Jlcpcb works you pick the component from their list, well there is an option for them to buy in parts aswell..
once you're have got all the correct parts chosen, they build what you ask, having 5 pcb built seems expensive so check 25 pcb price
it surprised me how little the overall cost ended up, but then you realize most of the cost is in the setup buying a few extra PCB cost very little..

building ten pcbs by hand takes some time, getting them done for you is often less expensive in time and faster in the end !

[Willy Bayot]

The problem we found before was with the specific active components for which there was no alternative to be found anywhere in stock.
Even if the PCB factory would find them locally, we were not sure that they were not fakes or bad copies.
The solution we found was to buy them ourselves from our own suppliers and send them to the PCB service like PCBWay.
That procedure extended the lead time but it was giving exactly what we needed with high quality components.

[detectormods]

detectormods, thank you for your valuable input.

You have seen Carl's wish list for the AMX.
What is your take on such a detector? Would it be useful and desirable in Australia?
We would appreciate your comments and opinion. Thank you.

The detector business in Australia is 40% marketing and 60% performance, there is always room for another detector, if it can be shown to perform better than brand ML then it will be a winner. There are additional technologies that are yet to be discussed and it may not be wise to put these ideas into the public domain as they can easily be taken by commercial interests. But each to there own, i have seen discussion over the years where design ideas on this forum are mysteriously converted into commercial patents not long after being published.

[Tinkerer]

The detector business in Australia is 40% marketing and 60% performance, there is always room for another detector, if it can be shown to perform better than brand ML then it will be a winner. There are additional technologies that are yet to be discussed and it may not be wise to put these ideas into the public domain as they can easily be taken by commercial interests. But each to there own, i have seen discussion over the years where design ideas on this forum are mysteriously converted into commercial patents not long after being published.

The world is full of stories of stolen ideas.
Is it true that in Australia a patent for the wheel was issued?

We all know that everything that is posted on a forum is in the public domain.
We all know that the path from idea to actual functioning product is long and arduous. Without investing all that hard work and time into the idea, an idea itself is worthless.