Let's made a PC-base metal detector with usb interface !!!

[Aziz]

I've successfully used this Chinese company without any problems ->
http://www.pcbcart.com/

Thanks Q!


It makes sense to produce a small series (+100) so I can support the interested members here. It's darn cheap! Just right to the poor man's wallet.

Aziz

[Aziz]

Current PCB Layout

Hi all,

this is the latest PCB layout of the detector controller (80 mm x 50 mm). Notice, I have put all big caps back again (because, there was some space there). C1 and C14 can be soldered optionally (if you need more power buffer).
This PCB layout version is a hot candidate for manufacturing.

Cheers,
Aziz

[Aziz]

Latest Schematics & Eagle-Files of Version 1.2

Hi all,

this is it. Latest schematics & Eagle-files of the detector controller version 1.2.
Note: This reference hardware platform is an open-source hardware platform. This does not imply the software part however. So only the hardware part.

You are allowed:
- to share, copy & distribute it (at no cost, referring the documents)
- to modify & improve it (at your own risk)
- to manufacture & sell it (at your own risk, but don't be a greedy ba'$t'ard!)
- to develop & share your own developped detector software (at your own risk)
- to develop & sell your own developped detector software (at your own risk, but don't be a greedy ba'$t'ard!)

You know, I don't like bl00dy greedy ba'$t'ards. If you wanna make some money of it, don't forget the copyright owner. And don't be greedy too (Don't forget! Otherwise, I would make your business totally kaput!).

But I would like to see someone developping the detector software and sharing it at no cost.

The hardware platform offers many interesting configurations & detector modes: (all induction balance coil mode)
- VLF single frequency resonant mode
- VLF dual frequency mode
- VLF wide band mode
- VLF chirp modulation mode
- PI standard wide band mode (coil damping mode)
- PI TEM mode (wide band)
- True Hybrid VLF/PI mode (TEM+resonant mode)
- and more...

You can make a lot of interesting things with the hardware platform. Go and learn how to code a software and DSP (digital signal processing).


Cheers,
Aziz

[simonbaker]

Hi all,

this is it. Latest schematics & Eagle-files of the detector controller version 1.2.
Note: This reference hardware platform is an open-source hardware platform. This does not imply the software part however. So only the hardware part.

You are allowed:
- to share, copy & distribute it (at no cost, referring the documents)
- to modify & improve it (at your own risk)
- to manufacture & sell it (at your own risk, but don't be a greedy ba'$t'ard!)
- to develop & share your own developped detector software (at your own risk)
- to develop & sell your own developped detector software (at your own risk, but don't be a greedy ba'$t'ard!)

You know, I don't like bl00dy greedy ba'$t'ards. If you wanna make some money of it, don't forget the copyright owner. And don't be greedy too (Don't forget! Otherwise, I would make your business totally kaput!).

But I would like to see someone developping the detector software and sharing it at no cost.

The hardware platform offers many interesting configurations & detector modes: (all induction balance coil mode)
- VLF single frequency resonant mode
- VLF dual frequency mode
- VLF wide band mode
- VLF chirp modulation mode
- PI standard wide band mode (coil damping mode)
- PI TEM mode (wide band)
- True Hybrid VLF/PI mode (TEM+resonant mode)
- and more...

You can make a lot of interesting things with the hardware platform. Go and learn how to code a software and DSP (digital signal processing).


Cheers,
Aziz

Thanks Aziz for sharing your fine work!

I don't understand the purpose of it yet... is it to be used with a sound card and mobile computer primarily?

Regards,

-SB

[Aziz]

It's a KillerApp!

Thanks Aziz for sharing your fine work!

I don't understand the purpose of it yet... is it to be used with a sound card and mobile computer primarily?

Regards,

-SB

Yup.

A good external 24 bit/96 kHz USB sound card is for timing generation and signal data acquisition. A 16-bit/48 kHz sound card works also but with reduced capability of course (reduced ADC resolution and bandwidth). The sound card must support the full-duplex mode (while transmitting stereo sound data into line-output, you acquire stereo sound data from line-input). The sound card must share it's input/output channel ADC/DAC clock from same clock source (locked to it).

A second sound card (cheaper or internal one) is used for signal beeping only (beep, beeeep, beeeeeeeeeep! ).

The Netbook/Tablet/Laptop/PC is the CPU/DSP brain (the number cruncher). The detector software runs in the operating system environment (I'm using the Windows-XP).

The detector software generates real-time transmitter timing signal on one sound card output channel (CLK input). The other output channel is reserved for L-ch pre-amplifier compensation (CMP input, could be software controlled induction balance/phase compensation signal or other feature).

While generating real-time timing signal on output channel to drive the transmitter (TX coil), it also acquires the pre-amplifier output signals (target response on L/R channel) and processes them in real-time in the software. The RX coil(s) is/are connected to the pre-amplifier inputs. The demodulation and processing of the RX signals will be done purely in the software.

Provided that, you own a Netbook (>200 EUR) or a Tablet PC (expensive yet), a good sound card (50-80 EUR), an IB-coil + the detector controller (DIY project here) and the "rocket science" state of the art detector software (must be written yet, I'm inviting all the developers* to do this), you will have a good working metal detector.

Now fasten your seat belt:
It's a KillerApp*! One can implement the "World's Best Metal Detecting Technology" onto it.
It is a good metal detector development kit platform too (which I have been doing this for many years now).

Consider the simplicity of the detector controller and the simple available means (mass media products) to build a high-end DSP-based metal detector.


Cheers,
Aziz

PS:
Oh yes, it's the end of ..
it's the end of evil greed ..

*= Provided that, the software engineer knows the crucial point

[simonbaker]

Yup.

A good external 24 bit/96 kHz USB sound card is for timing generation and signal data acquisition. A 16-bit/48 kHz sound card works also but with reduced capability of course (reduced ADC resolution and bandwidth). The sound card must support the full-duplex mode (while transmitting stereo sound data into line-output, you acquire stereo sound data from line-input). The sound card must share it's input/output channel ADC/DAC clock from same clock source (locked to it).

A second sound card (cheaper or internal one) is used for signal beeping only (beep, beeeep, beeeeeeeeeep! ).

The Netbook/Tablet/Laptop/PC is the CPU/DSP brain (the number cruncher). The detector software runs in the operating system environment (I'm using the Windows-XP).

The detector software generates real-time transmitter timing signal on one sound card output channel (CLK input). The other output channel is reserved for L-ch pre-amplifier compensation (CMP input, could be software controlled induction balance/phase compensation signal or other feature).

While generating real-time timing signal on output channel to drive the transmitter (TX coil), it also acquires the pre-amplifier output signals (target response on L/R channel) and processes them in real-time in the software. The RX coil(s) is/are connected to the pre-amplifier inputs. The demodulation and processing of the RX signals will be done purely in the software.

Provided that, you own a Netbook (>200 EUR) or a Tablet PC (expensive yet), a good sound card (50-80 EUR), an IB-coil + the detector controller (DIY project here) and the "rocket science" state of the art detector software (must be written yet, I'm inviting all the developers* to do this), you will have a good working metal detector.

Now fasten your seat belt:
It's a KillerApp*! One can implement the "World's Best Metal Detecting Technology" onto it.
It is a good metal detector development kit platform too (which I have been doing this for many years now).

Consider the simplicity of the detector controller and the simple available means (mass media products) to build a high-end DSP-based metal detector.


Cheers,
Aziz

PS:
Oh yes, it's the end of ..
it's the end of evil greed ..

*= Provided that, the software engineer knows the crucial point

Thanks for explanation, that was very useful.

I see a MOSFET driver on the board -- is there a MOSFET somewhere driven by that, or are you using it to drive the coil directly?

Did I read correctly that this board supports both VLF (continuous wave) and PI designs, or is it primarily for one type?

Regards,

-SB

[Aziz]

Thanks for explanation, that was very useful.

I see a MOSFET driver on the board -- is there a MOSFET somewhere driven by that, or are you using it to drive the coil directly?

Did I read correctly that this board supports both VLF (continuous wave) and PI designs, or is it primarily for one type?

Regards,

-SB

The board is supporting one and two identical n-ch mosfets. I have soldered a connector (screwable pins) to the bread-board in the earlier posting, so there is only one mosfet connected. One can quickly change to other mosfets (that's a typical prototype work around for mode changes). I have choosen the big mosfet case (TO-247) as I use them quite often. If you have decided the operating mode, you can also solder the mosfet into the PCB of course. TO-220 mosfet pins must be slightly bent.

The single low-side n-ch mosfet mode:
The coil is driven between VCOIL and TXOUT. This is the high voltage PI/TEM mode. TXOUT is the high voltage connection (drain of the mosfet). The connection R0 is open to keep the high voltage out of the mosfet driver. Although it is supporting voltages upto 600V, the capacitive load is blocked with open R0. We don't need the high-side mosfet driver part of the IR2184 in this case. The bootstrapping of the high-side mosfet driver is disabled. This mode supports VLF/PI/TEM + Hybrid-VLF/PI mode.
BTW, it's cheaper. You only need one mosfet!

The dual (low/high-side) n-ch mosfet mode:
This is the H-bridge (Half-bridge) mode and you can use low voltage mosfets (30 V) as the TXOUT connection doesn't see a high voltage. The coil is driven between TXOUT and GND. Note, R0 must be shorted (zero Ohm resistor). We need the high-side driver of the mosfet driver in this case. This is destined for H-Bridge VLF mode operation.

I'll give you more infos later how to use all the different operating modes. Nearly a dozen of modes can be realised!!


Yes, you did read it correctly. It is supporting VLF (continious wave) and PI design (or both at the same time: Hybrid-VLF/PI).


Cheers,
Aziz

[Aziz]

Knock-off!

Hi all,

who dares and is willing now to blow the big players pants off?

Or give them an ?


(This isn't a joke guys! I'm serious about this.)

If you trust me and have the knowledge and confidence of the power of the DSP programming, you can easily blow their pants off. They will be way way behind you.

Perfect ground balancing? No problem.
Highly mineralized ground rejection? No problem.
Hotrock rejection? No problem.
Fully automatic ground balancing? No problem.
Better discrimination? No problem.
Perfect EMI rejection? No problem.
Automatic IB correction? No problem.
Automatic IB-coil mechanical shock rejection? No problem.
.. and much more features possible.


I will start the PCB manufacturing in China next month. He he he, a "small" guy can do it as well. I will do this to reduce the PCB costs for you (if interested). I very likely will have enough samples and will inform you, when they are available.

Let's fight against the:
- drip-feed technology release
- incredible huge greed (eat this )
- market monopolizing
- price dictators
- patent trolls


Cheers,
Aziz

[WM6]

If you trust me and have the knowledge and confidence of the power of the DSP programming, you can easily blow their pants off. They will be way way behind you.

Aziz, most of us "small guys" are too small for hi-tech DSP programming.

Apart of PCB we need pre-programmed (with your crypto protected ID) chip too.

If you intend to offer such "small guys" solutions, my pre-order is for 3PCS (kit?).

[Aziz]

Aziz, most of us "small guys" are too small for hi-tech DSP programming.

Apart of PCB we need pre-programmed (with your crypto protected ID) chip too.

If you intend to offer such "small guys" solutions, my pre-order is for 3PCS (kit?).

Be patient WM6. I have to code the final detector software application as well.
(After trying dozens of algorithms, I know what I need now.)

It's the beginning of the revolution. We will f$k the evil greed in this game!

Aziz

[simonbaker]

The board is supporting one and two identical n-ch mosfets. I have soldered a connector (screwable pins) to the bread-board in the earlier posting, so there is only one mosfet connected. One can quickly change to other mosfets (that's a typical prototype work around for mode changes). I have choosen the big mosfet case (TO-247) as I use them quite often. If you have decided the operating mode, you can also solder the mosfet into the PCB of course. TO-220 mosfet pins must be slightly bent.

The single low-side n-ch mosfet mode:
The coil is driven between VCOIL and TXOUT. This is the high voltage PI/TEM mode. TXOUT is the high voltage connection (drain of the mosfet). The connection R0 is open to keep the high voltage out of the mosfet driver. Although it is supporting voltages upto 600V, the capacitive load is blocked with open R0. We don't need the high-side mosfet driver part of the IR2184 in this case. The bootstrapping of the high-side mosfet driver is disabled. This mode supports VLF/PI/TEM + Hybrid-VLF/PI mode.
BTW, it's cheaper. You only need one mosfet!

The dual (low/high-side) n-ch mosfet mode:
This is the H-bridge (Half-bridge) mode and you can use low voltage mosfets (30 V) as the TXOUT connection doesn't see a high voltage. The coil is driven between TXOUT and GND. Note, R0 must be shorted (zero Ohm resistor). We need the high-side driver of the mosfet driver in this case. This is destined for H-Bridge VLF mode operation.

I'll give you more infos later how to use all the different operating modes. Nearly a dozen of modes can be realised!!


Yes, you did read it correctly. It is supporting VLF (continious wave) and PI design (or both at the same time: Hybrid-VLF/PI).


Cheers,
Aziz

Thanks, that really helps clarify.

It is really nice. I just have reservations about sound card as sampling tool, high-end and low-end roll-off. Probably not big deal for VLF (continuous wave) designs, but still...

In my dreams, I'd like a board with the 24 bit A/D onboard (broadband as possible) and USB interface. I realize that is much more complicated and expensive and would require a software driver as well. Just wishing... your board is a very practical way to get started and to develop cool software algorithms.

Any sound cards that are hackable to go down to 0 Hz?

-SB

[Aziz]

Working on Open Issues

Hi all,

I'm currently working on the gain of the pre-amps. The Gain of 100 in the current schematics is way too much of course. We don't need it for the 24-bit sound card.
A reasonable value should be between 10x .. 50x.
The coil is picking way more noise than the pre-amp is generating. So a high gain isn't really necessary. We need to rise the signal level (with RX coil) just above the noise floor of the sound card. That's enough.

I will make some noise measurements with different input impedances and gains and will let you know, how to set the optimal gain.
Remember: Gain = - RG / Rin

The AuxInput impedance depends on the RG too. A gain for AuxInput should be 1x .. 3x (enough).

The NE5532 isn't that bad. It's very cheap too.

We need some minimum input impedance (Rin >= 47 Ohm) to protect the pre-amps. Even the RX coil isn't seing the TX coil voltage (due to induction balance), it could pickup pulses from other detectors or EMI. The NE5532 has built-in protection diodes on its inputs (*LOL*). But double protection is better. (We could save 4 x 1N4148 but we could put a different dual op-amp, then we might need them again).

Well, you really get with NE5532 a lot for your money.

Aziz

[Aziz]

Thanks, that really helps clarify.

It is really nice. I just have reservations about sound card as sampling tool, high-end and low-end roll-off. Probably not big deal for VLF (continuous wave) designs, but still...

In my dreams, I'd like a board with the 24 bit A/D onboard (broadband as possible) and USB interface. I realize that is much more complicated and expensive and would require a software driver as well. Just wishing... your board is a very practical way to get started and to develop cool software algorithms.

Any sound cards that are hackable to go down to 0 Hz?

-SB

Hi SB,

I am wishing such an USB 24 bit A/D device too. Hey, we don't need them really. You could unscrew the USB sound card and could mount the PCB into the controller box too.


We don't really need the DC response capability (down to 0 Hz).
Our pulse measurements are modulated with the pulse frequency and we get therefore an AC response. But we need the upper frequency range (>20 kHz). Some dumb manufacturers limit the upper frequency and advertise it as a 96 kHz sampling rate sound card! Fine, but for what? You have to look at the input frequency response range.

Fortunately, there are sound cards available, which can be used upto 48 kHz signal frequency (96 kHz sampling rate, input frequency response corner frequency at approx. 46 kHz).

There will be sound cards, which will support even 96 kHz signal frequency in the future (192 kHz sampling rate). These sound codec chips are flooding the market at the moment.

We can use the 48 kHz sampling rate too. That's not the problem. We will miss the small targets only (or we will have less sensitivity to them).

And we should concentrate our TX frequency response energy to the supported input frequency response range. Otherwise, we would waste TX power. Better focus and concentrate it to the supported frequency range.

For frequency-domain metal detection, a frequency range of 100 Hz to 100 kHz is really enough. And I'm waiting to get the upper range soon with the 192 kHz sampling rate (SR) sound cards. But I can live with 96 kHz SR too. Or with 48 kHz SR.

Aziz

[Aziz]

Hi all,

pre-amplifier gain? Must delay a few days. It's very trivial.

But you must see the elegance and genius of the software coding. What I do in the software, you can't do it in the hardware (electronics). That's the reason, why the detector controller is so ridiculous simple and trivial.

I want to give the software engineer some hints:
- Visualising the signal data is very important and well worth. This is the number one tool for finding a convenient processing algorithm. Many different transformations and visualisations of the signal data is of benefit too.
- You need some (advanced) math:
DSP functionality (demodulation, filtering, etc.), statistics (mean, std deviation, variance, co-variance, regression analysis, etc.), vector math/complex number math (I/Q), matrix math just to mention some of the mathematical tools for the analysis.

Now hold your breath:
I have found a much more elegant way (algorithm) for ground balancing and discrimination recently. There exists many ways to solve this issue. And it makes the life of amateur coil builders easy too.


(Ups!, you may breathe further).

Cheers,
Aziz

[Alexismex]

Hello Aziz ,
you really " blast my old brain ", good job you make and i am sure that with deep software you can make a total analysis of the soil and target like our neuron brain in everyday life...here is the key of your smart detector...