THE PERFECT GROUND BALANCE

[moodz]

Which is not unlike step voltage after all. No recovery time, and no delay, hence less juice required for detection.

I am toying with a matchless detector by rev Scarborough. OK, it is a toy, but it has a potential. All metal readings are related to mutual inductance between search coil and a target, and in case of "beat balance" it is merely acting as a free running PWM. Perhaps the PWM principle will yield a better PI after all.

BTW, there is an arrangement that uses a capacitor and analog switches that yields an opamp with over 120dB CMMR. Having PI sample as a windowed pulse, this seem as a perfect candidate, and it is not complicated at all. Any interest?

Bring it on

[Aziz]

Which is not unlike step voltage after all. No recovery time, and no delay, hence less juice required for detection.

I am toying with a matchless detector by rev Scarborough. OK, it is a toy, but it has a potential. All metal readings are related to mutual inductance between search coil and a target, and in case of "beat balance" it is merely acting as a free running PWM. Perhaps the PWM principle will yield a better PI after all.

BTW, there is an arrangement that uses a capacitor and analog switches that yields an opamp with over 120dB CMMR. Having PI sample as a windowed pulse, this seem as a perfect candidate, and it is not complicated at all. Any interest?

120 dB CMMR at high frequencies?
Bring it on!!!

Aziz

[Aziz]

Triangular Bipolar Current Pulse

Hi all,

I have prepared an LTspice simulation file for your convenience, which is showing the triangular (+quasi triangular) bipolar current pulse.
The source is a forced step voltage coming from a simple H-bridge driver (rectangular step source voltage).

Note, that the blocking capacitor reduces the TX impedance further. It is essential if you use an unipolar voltage source to block the DC voltage.
The blocking capacitor isn't required if you use a symmetrical bipolar source. But you will have a higher TX impedance. So using the blocking capacitor doesn't hurt you.

To overcome the high TX impedance, just increase your source voltage. Note, that the voltage rating of the capacitor is important, when you come into the resonance frequency region (high voltage!!!).

You can also use standard bipolar electrolytic capacitors but you have to select the capacitor value appropriate to the frequency and coils inductivity not to create a high negative and positive voltage. Set the capacitor value just to produce a voltage over the capacitor between 0 - source voltage level. Take care to the polarity.

Have fun.
Cheers,
Aziz

PS: Patent Trolls! Keep Out!

[Davor]

That's more like it. If you push it just a tiny bit, it is all the same as if you produce high voltage across a high L in PI style, except that the classic PI transient is suffering from high Z.

The high CMMR thing is based upon the LTC1043 switching capacitor controller. I prepared a somewhat more common solution using 4066 for your convenience. Enjoy.

LTC1043 has very interesting specs, and its clock can be completely controlled from outside source, which makes it a very interesting sampler for PI or otherwise.

Point is that you don't have to use a chopper as a mere chopper, and instead you can use it as a PI sampler/demodulator. Doing so you get instrumentation amplifier benefits as a bonus.

[Davor]

I'm thinking - what goes around, comes around.
There are only two options in PI world: monocoil or otherwise.
In case of monocoil, and a monopolar pulse it is more or less straightforward thing.
In case of IB Rx coil we may encounter everything, so bipolar pulse or not - there is a need for a demodulator that can accept both polarities. In case of bipolar pulses there is a potential benefit of Earth field cancellation, but only in case Rx is capable of demodulating a bipolar signal.

There are also some conflicting details that just might be in favour of VLF IB which Aziz nicely pointed out in his spice solution: the coil + system resistance. Targets respond much better with low resistance in Tx, and a LC tank in resonance does it spectacularly well.

There are step voltage low resistance options for supplying a coil even without a capacitor in resonance. This still requires a separate Rx coil and a demodulator that works well with bipolar signal. Argh! it works so well in simulation.

[moodz]

I'm thinking - what goes around, comes around.
There are only two options in PI world: monocoil or otherwise.
In case of monocoil, and a monopolar pulse it is more or less straightforward thing.
In case of IB Rx coil we may encounter everything, so bipolar pulse or not - there is a need for a demodulator that can accept both polarities. In case of bipolar pulses there is a potential benefit of Earth field cancellation, but only in case Rx is capable of demodulating a bipolar signal.

There are also some conflicting details that just might be in favour of VLF IB which Aziz nicely pointed out in his spice solution: the coil + system resistance. Targets respond much better with low resistance in Tx, and a LC tank in resonance does it spectacularly well.

There are step voltage low resistance options for supplying a coil even without a capacitor in resonance. This still requires a separate Rx coil and a demodulator that works well with bipolar signal. Argh! it works so well in simulation.

Good analysis Davor ... however slight correction .... the differential coil ( or monocoil with centretap ) does bipolar beautifully because it is symmetrical already. Just add switching mosfet/diode on both sides of coil and pulse on alternative times. Rx and demod is off same coil ( RX of both sides ) Thats why the UNIPI chip I am working on has two PWM Tx outputs for bipolar tx pulsing.

moodz.

[hdphilip]

moodz

i was wondering that, when i was experimenting with the differential front end, now, if their was a slight "overlap" of the timing pulse for the mosfets, would this help the damping of the fly back pulse without he need of a resistor? or in the differential front end, the mosfet?

philip

[Midas]

Good analysis Davor ... however slight correction .... the differential coil ( or monocoil with centretap ) does bipolar beautifully because it is symmetrical already. Just add switching mosfet/diode on both sides of coil and pulse on alternative times. Rx and demod is off same coil ( RX of both sides ) Thats why the UNIPI chip I am working on has two PWM Tx outputs for bipolar tx pulsing.

moodz.

Tough to maintain your mosfet blocking scheme though. You may have to go back to series resistor\diode blocking.

Midas

[moodz]

moodz

i was wondering that, when i was experimenting with the differential front end, now, if their was a slight "overlap" of the timing pulse for the mosfets, would this help the damping of the fly back pulse without he need of a resistor? or in the differential front end, the mosfet?

philip

Have not tried it but worth investigation ... somehow due to energy conservation laws unless you dissipate it or store it it won't work.

[moodz]

Tough to maintain your mosfet blocking scheme though. You may have to go back to series resistor\diode blocking.

Midas

It's hard at the bleeding edge ....the resistor diode schema is pretty much foolproof however the blocking scheme does work ..I will shortly return to it ..its probably easier than everest.
Moodz

[Davor]

But somehow your MOSFET blocking scheme reminded me of diode limiters. They are very low impedance within a conducting window, maybe an order of magnitude higher than MOSFETS, but still several orders of magnitude lower than the rest of the frontend.

[Midas]

But somehow your MOSFET blocking scheme reminded me of diode limiters. They are very low impedance within a conducting window, maybe an order of magnitude higher than MOSFETS, but still several orders of magnitude lower than the rest of the frontend.

Hi Davor

As I understand it diode limiters are already the more usual way to do it with perhaps a series resistance of 500-1000 ohms. MOSFET can be well under an ohm which makes it more like 2-3 orders of magnitude. Also I'm not sure where the rest of the input impedance is that your referring to, typically the signal goes straight into the preamp next. Although switching first seems to be coming into fashion. In theory the MOSFET blocking should be beneficial and Minelab use them to good effect, but in reality no one here that I know has really proved them to be any better. My guess is a MD needs to be VERY good before resistor noise start to become at all significant.

Midas

[Davor]

As usual I jump just a bit ahead of myself. I thought of series configuration diodes, not parallel, the way they were used in olden headphone limiters. The obvious solution would be a plain vanilla series resistor, and a pair of diodes, but in this case diodes are normally biased, with low series impedance and all, and they become switches only in case of too large signals. In this configuration they do not cut off abruptly, hence they do not suffer from recovery problems either.

I arranged a limiting voltage as floating, but it works even better when split and ground referenced in between. Works any way you want it.

Have fun

[Tec]

Hi This is interesting,

Ground Balance, mmmm.
I think the 'noise from the ground' consists of the following different effects (as already mentioned by others)

Earth Magnetic field (as coil moves in it)
Capacitance to the ground
Eddy currents in conductive ground
Distortion of the field due to magnetic field permeability (how easy it permeates the ground)

Have I missed something ?..

Tinkerer do you work for a metal detector company ?, I don't want to give away any secrets on here.

My metal detector cancels the Earths field by using Bi-pulse (pulses in one direction followed by in the other direction).

It prevents the capaciitive effects by using shielding around the coil.

It isn't affected by distortion of the field affecting balance (because it is a mono coil)..

Which just leaves eddy currents in the the soil to worry about I guess..

[Tinkerer]

Hi This is interesting,

Ground Balance, mmmm.
I think the 'noise from the ground' consists of the following different effects (as already mentioned by others)

Earth Magnetic field (as coil moves in it)
Capacitance to the ground
Eddy currents in conductive ground
Distortion of the field due to magnetic field permeability (how easy it permeates the ground)

Have I missed something ?..

Tinkerer do you work for a metal detector company ?, I don't want to give away any secrets on here.

My metal detector cancels the Earths field by using Bi-pulse (pulses in one direction followed by in the other direction).

It prevents the capaciitive effects by using shielding around the coil.

It isn't affected by distortion of the field affecting balance (because it is a mono coil)..

Which just leaves eddy currents in the the soil to worry about I guess..

Hi Tec,

thanks for the feedback.

Am I working for a metal detector company? No, I am just an old man who is curious about how things work. Maybe there are other people out there who like to know, so I make my knowledge "Open Source", for anybody to use. So If you have a trade secret that you think is worth a lot of money, don't tell me about it.

Bi-polar pulsing is a good way of taking care of the EF effect.

Does the shielding eliminate coil to ground capacitive effects? Could you explain how that works?

Tinkerer