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

[Aziz]

Ok guys,

I'm almost finished with the fine tune of the latest VLF osc. controller with active amplifier (the full-bridge/half-bridge osc. version). And I found, that some op-amps tend to oscillate and get unstable so we have to compensate the feedback loop and limit the upper frequency (LP-filter).

Oh yes, the active amplifier version makes truly sense in conjunction with anti-interference coils (Tophat(c)(r)(tm) coil). You get incredible high sensitivity and depth performance compared to the passive KISS version.

I'll post the final schematics soon. Just be patient.
I'm going to use a Tophat(c)(r)(tm) coil. I can use it with dual-frequency and PI/Hybrid VLF configuration as well. I have to build a wooden coil frame and some distance adjusting means... (I'm not a good craftsman.. it could take some time.. )

Cheers,
Aziz

[Aziz]

Hi all,

I'll change the latest VLF osc detector controller schematics. The power supply section and the transmitter are going to be changed slightly. I want to use the n-mosfets for the center-tapped TX coil configuration (half-bridge mode) so I have to make a positive charge pump converter instead of a negative. The half-bridge mode version is going to be supplied by a negative voltage (Bat+ at GND, Bat- and Vee, Vcc from dc/dc charge pump converter). The half-bridge mode with center-tapped TX coils gives more bang and the n-mosfets operate more efficiently (less resistance).

But I want to build and test it first before publishing the final schematics. It's really very trivial.
It could take some time. Sorry for being very lazy.

Cheers,
Aziz

[Aziz]

Ok guys,

I'm going to build a rectangular Tophat coil with 42 cm width and length.
Why 42 cm?
That's easy: "Answer to the Ultimate Question of Life, the Universe, the Metal Detecting, and Everything."

https://en.wikipedia.org/wiki/42_(number)

The inner coils? "Half of Everything"
21 cm width and length.

Co-axial height? Somewhere between 35 cm and 45 cm. The 13x13 mm wooden stripe should give enough stability I think. (The Tophat coil is going to be heavy weight. I probably have to use my infamous Al-strip wire. Or have to mount even two tyres to the coil frame. *LOL*)

Cheers,
Aziz

[sido]

Hi aziz, was the tophat antenna first discovered and developed for the Vhf/Uhf radio communications?

I recall something in the Amateur Radio field mentioning this antenna configuration and its extrodinary performance.

Sid

[Aziz]

Hi aziz, was the tophat antenna first discovered and developed for the Vhf/Uhf radio communications?

I recall something in the Amateur Radio field mentioning this antenna configuration and its extrodinary performance.

Sid

Sorry Sid, the tophat coil isn't an antenna. We are going to shield it to avoid it being an antenna. We don't want radiate EM waves. We want just emit local magnetic fields to the ground.
Aziz

[Davor]

Shield is not going to prevent it. Not long ago all the "frame" type antennas for amateur 73 kHz, 136 kHz and 160-190 kHz bands were shielded, and it did not prevent them from functioning. Centre tap and no shield is the not-so-new fashion there. That's one of my reasons for going FKK with my coils.

[sido]

Ok, misunderstood. I recall a article which referred to a tophat antenna in Amateur Radio which physically looks similar in appearance/set up and configuration to the way you have presented the Tophat coil for Metal Detectors.

I say antenna which in fact is a coil for Vlf frequencies as the diagram presented on the thread Advance PI coils. But in Vhf/Uhf frequencies it has 2 plates, one larger below and one smaller above due to the 2 different operating wavelengths (higher in frequency in this case) which in turn terminates a perfect 50ohm impedence match for RF frequencies in Vhf/Uhf Radio Comms with wide bandwidth operation.

Obviously for lower frequencies you require coils with wire to be constructed for those operating frequencies such as Metal Detectors.

I have to dig out the article so i can confirm.

Sid

[Aziz]

Guys,

we don't allow the electric fields coming outside of the coil. For a successful (and efficient) EM wave transmission, you have to allow both H (magnetic) and E (electric) field emission. Shielding blocks the electric field part.
We want to emit the H (magnetic) part into the ground.
Aziz

[sido]

Aziz, the tophat coil i am referring too has no shielding.

The difference is in the frequency of operation as your example of a tophat for LF/VLF.

Sid

[Aziz]

Aziz, the tophat coil i am referring too has no shielding.

The difference is in the frequency of operation as your example of a tophat for LF/VLF.

Sid

The tophat coil can even be used for a PI device (upto 1 MHz bandwidth, yep, don't forget PJ's harmonics... ) .
That's the reason, why it should be shielded. Even if it isn't an effective "antenna".
Shield everything to block electric fields.
Aziz

[sido]

I think i recall something about mutual capacitance, where capacitance is cancelled/eliminated in the physical design/construction of such coil.

Aziz i will follow this up and see if i can add further to its capabilities.

Sid

[Humanise]

Much of this thread seems (other than that recently on the coil) to have been about design of the amplifier to the transmitter coil. As I'm also trying to use a computer with sound card to drive a coil, I also have the requirement to make a similar amplifier. The problem is that I'm lazy and don't want to put together the electronics myself.

My current solution is to use one of the many already put together audio amplifiers based on the PAM8403 chip. The complete boards can be purchased from ebay for as little as US$2.84 delivered to your door from Hong Kong. I have gone fully upmarket paying US$9.99 to get a board that has the potentiometer with a nice knob plus a proper USB connector for getting the power from the USB plus the proper 3.5mm connector for getting the transmitter signal from the sound card etc. Everything to save me from having to do any soldering! You can see the board that I'm using at

http://www.ebay.com/itm/3W-3W-Amplif...item4ab7938e3c

Even getting a similar board with cheaper connectors and pot, you can get it delivered for US$4.09 from

http://www.ebay.com/itm/5V-mini-Digi...item2a24b00b16

Using the PAM8403 means that this amplifier is a class D amplifier which should make it very efficient as the output FETs are either fully on or fully off. The boards use a Bridge Tied Load (BTL) configuration on the drivers which means that even though they are only running from a very low voltage source being powered by the USB, they are able to supply up to 4 times as much power to the load than they would be able to without the BTL configuration.

The PAM8402 chips is giving me full short circuit protection, automatic shutdown and startup on overheating and protection against low voltage supply. In essence, a lot of sophistication all coming out of a very cheap device.

The specifications on this chip are available at

www.poweranalog.com/pdf/PAM8403.pdf

I still have some reservations about the use of this class D device. Class D amplifiers are described in a number of places such as

http://en.wikipedia.org/wiki/Class-D_amplifier

as requiring "A passive low-pass filter removes the unwanted high-frequency components, i.e., smooths the pulses out and recovers the desired low-frequency signal. To maintain high efficiency, the filter is made with purely reactive components (inductors and capacitors), which store the excess energy until it is needed instead of converting some of it into heat."

Now that seems quite sensible except that the PAM8403 describes the chip as a "filterless" class D amplifier and boards designed with it don't seem to have this filter. I don't understand how it can be "filterless" and still give an appropriate output. It seems to have an internal oscillator with a switching freq of 260 khz and yet the only suggestion for decreasing unwanted high-frequency components is to add ferrites to the output lines with high impedances in the MHz range. I don't fully understand this.

I would like to get the opinions of the people who follow this thread as to how appropriate this amplifier is for this PC based metal detector task

Thanks (and Merry Christmas)
Ken

[Aziz]

..
Now that seems quite sensible except that the PAM8403 describes the chip as a "filterless" class D amplifier and boards designed with it don't seem to have this filter. I don't understand how it can be "filterless" and still give an appropriate output. It seems to have an internal oscillator with a switching freq of 260 khz and yet the only suggestion for decreasing unwanted high-frequency components is to add ferrites to the output lines with high impedances in the MHz range. I don't fully understand this.

I would like to get the opinions of the people who follow this thread as to how appropriate this amplifier is for this PC based metal detector task

Thanks (and Merry Christmas)
Ken

Hi Ken,

that's a nice & small amplifier. It's cheap because the ferrite bead/output filter isn't on the board.
The ferrite bead at the outputs + small capacitor form a low-pass filter. The ferrite bead can be seen as a toroid ferrite coil core with 1 winding only.

Due to the input signal -> PWM conversion via the internal high switching frequency, you will get modulation of the output signal and thus at the TX coil (a large noise source, although not audible to us when using loudspeakers). It is important to use a good LC filter at the output of the amplifier.
Further more, you also need the TX coil reference signal to reduce futher noise. We somehow have to lock onto the amplifiers switching frequency.
Cheers & Merry Christmas
Aziz

[Humanise]

Hi Ken,

that's a nice & small amplifier. It's cheap because the ferrite bead/output filter isn't on the board.
The ferrite bead at the outputs + small capacitor form a low-pass filter. The ferrite bead can be seen as a toroid ferrite coil core with 1 winding only.

Due to the input signal -> PWM conversion via the internal high switching frequency, you will get modulation of the output signal and thus at the TX coil (a large noise source, although not audible to us when using loudspeakers). It is important to use a good LC filter at the output of the amplifier.
Aziz

Thank you for your reply Aziz

The parts that I didn't understand were

1) Why the data sheet for the chip itself says that this is "filterless"? If this chip still requires a filter, on the board or otherwise then it is not "filterless".

2) Why the only reference to a filter, near the end of the data sheet, says it "reduces EMI of around 1 MHz and higher" when the oscillator frequency is only 260khz?

I purchased some ferrite beads some time ago (on ebay) but they have yet to arrive. I was planning on just using the ferrite beads (one at each end, with several turns around each) without adding the capacitors. I would then be relying on the in-circuit resistance of the wire in the transmit coil to create a resistance for an RL low pass filter plus the stray capacitance between the wires in the winding of the coil and between the leads to create an LC filter. I see ferrites being used a lot like this where they are just in the leads and aren't used with any actual component resistors or capacitors.

After purchasing the ferrites, I found the actual specs for the ferrites that I had purchased. It turns out that the ones I had purchased are made optimal for frequencies over 5 Mhz. I'm now thinking of getting some lower frequency ones and putting one of each on both leads.

In the data sheet is shows a 220pf capacitor at each end being connected to earth. I presume by the earth it means Analog Ground, (pin 11 on the chip) not Power Ground. On this little board, it is just a pin on a surface mount device and it is not within my capabilities to make a connection to it. If I did put capacitors in, it would be between the two ends of the leads going to the transmit coil. (ie between the Out- and the Out+. Do you think that I need to add capacitors?

As I see it, strictly speaking, the main purpose of the filters is to stop EMI from affecting other people rather than changing anything relating to the operation of the metal detector. The receive coil will be symmetrically balanced (probably based on the figure 8 or similar) so it should see very little of it.

Further more, you also need the TX coil reference signal to reduce futher noise. We somehow have to lock onto the amplifiers switching frequency.
Aziz

I don't follow what you mean here. Are you saying that we need to read back to an input of the sound card the actual voltage waveform across the coil? I notice that you were doing this with one of your designs earlier in this thread. Why do you think that this is needed here?

Regards
Ken

[Aziz]

Hi Ken,

The parts that I didn't understand were

1) Why the data sheet for the chip itself says that this is "filterless"? If this chip still requires a filter, on the board or otherwise then it is not "filterless".

that's pure marketing. We can't use the amplifier without any filter in our application. I would even say, that such amplifiers will produce more noise. Better to use a simple class AB amplifier or one of the detector controllers here.

2) Why the only reference to a filter, near the end of the data sheet, says it "reduces EMI of around 1 MHz and higher" when the oscillator frequency is only 260khz?

The loudspeaker is an inductive coil. When it is switched in the amplifier (PWM modulation), the loudspeaker leads also carry the current, which radiate a high frequency EMF during the switching transitions. The fundamental frequency (260 kHz) and multiple of it called the harmonics (where is PJ? ) will be emitted and radiated.

I don't follow what you mean here. Are you saying that we need to read back to an input of the sound card the actual voltage waveform across the coil? I notice that you were doing this with one of your designs earlier in this thread. Why do you think that this is needed here?

If you want to make a good detector, the answer is yes. You have to process the signal across the TX coil as well.
Either for
- synchronising onto an external clock or
- for demodulation of a signal given by an external clock or
- recognition of external noise to eliminate it from the signal or
- external battery voltage tracking (low battery detection, on/off detection, etc.)
...

I for one wouldn't use this amplifier type. It will produce a lot of noise. A clean class AB amplifier would do it much much better.

I'll publish new and improved detector controllers next year.


Cheers,
Aziz