X vs R

[Sergey_P]

The process of formation X R of the target in post #31
For proper detection of the target is very important to eliminate any external factors: M(tx-rx)=0, const; M(tx-T)=0; M(T-rx)=0. Detector should just accept a secondary field from the target ... and all. Tx, Rx should not change anything in the signal. Low Q, without resonant system - do it
I do not view technology: monocoil, IB,… and other..

[Davor]

I understand your point, however, there is no detector around that uses exact resonance at both Tx and Rx, except maybe Nexus. Either way, I prefer low Q at Rx. In fact, I prefer series resonance supplied to hi-z input (hence Q is well below 0.5) of a real low noise preamp - there is no flux return from such Rx back to Tx and no way to spoil IB or anything else. IMHO a parallel tank at Rx input is a bad idea, and it does promote Tx-Rx linking (current circulates within a tank), and is prone to funky phase shifts due to poor component tolerances etc.

Lowering Tx Q can not help you there But loading a Rx coil with high impedance in low Q configuration can.

[Sergey_P]

Lowering Tx Q can not help you there

compromise

IB Rx do with low Q at the resonance frequency - it is impossible ... only the shift of the resonance frequency ...

[Davor]

Well, you can't say resonance for a series contraption with Q much lower than 0.5 ... It's -3dB points are very wide apart.

Frequency shifting a parallel tank is a compromise that fixes only a severe phase shift, and ideally improves your S/N a little bit. Current flows within a tank at a much greater rate than to the outside. The worst thing yet is that such resonance does not reduce low frequency pickup.

[Sergey_P]

you can't say

I can, for ‘Verator4’ :

Q~Ri/(wL) <<1 , Ri - resistance switches Tx

[mikebg]

AIR signal in TX coil

AIR signal in TX coil

Each TX coil is self-inductance, ie it operates as MONOCOIL and receives maximal possible AIR signal. The simple circuit in post # 42 shows TX coil as two-port network.

The attached image illustrates how NETWORK ANALYZER displays transfer function Z21 of MONOCOIL (the worst sensing head for metal detectors ;-). It has Re component at all frequencies equal to resistance of TX coil. To calculate Re and Im component of AIR signal, you should multiply Z21 (shown in upper left angle) to Zo (scale factor).

[Tepco]

I can, for ‘Verator4’ :

Q~Ri/(wL) <<1 , Ri - resistance switches Tx

Verator4 is built using square wave driven TX, coil only, no LC resonant tank, so speaking about Q factor is inappropriate.

[Tepco]

It is not technically correct to define monocoil parameters using network analyzer and transfer function. With good enough model, RF transmission line response can be derived this way, but that is useless. In normal PI operation, coil is “time divided” between TX period and receiving interval, only signal remaining is coils own LRC decay and tail due to eddy current decay in coil material. Nothing remotely similar to AIR signal, rather COIL signal, so to say.

[mikebg]

AIR signal as PHASE REFERENCE

It is not technically correct to define monocoil parameters using network analyzer and transfer function. With good enough model, RF transmission line response can be derived this way, but that is useless. In normal PI operation, coil is “time divided” between TX period and receiving interval, only signal remaining is coils own LRC decay and tail due to eddy current decay in coil material. Nothing remotely similar to AIR signal, rather COIL signal, so to say.

Tepco, the AIR signal in a conventional PI machine is the flyback voltage at step-down response. Its value is very high at high frequencies (short sample delay) and becomes low at low frequencies (large sample delay). This is shown by NETWORK ANALYZER in frequency domain (see post #27):
http://www.geotech1.com/forums/showt...552#post163552
For a sensing network (coil & target) there is something common between
- step-up response,
- step-down response,
- impulse response,
- ramp response and
- frequency response.
The common is that if we know one of them, we can calculate all other responses.
Frequency domain is the most powerful tool for calculating, analyzis and design of signal systems. Frequency domain shows that the conventional PI metal detector is simply an incompetent designed wide band metal detector. Please read this:
http://www.geotech1.com/forums/showt...374#post163374

AIR signal as PHASE REFERENCE

Most designers of metal detectors prefer to use voltage across TX coil as phase reference for synchronous demodulation. For this purpose, they use a wire between TX coil and reference SIGNAL processor in RX. This makes the RX "nonwireless". However the NETWORK ANALYZER shows that voltage across TX coil is not the genuine AIR signal.
We need a genuine carrier wave to make proper demodulation and discrimination.

Dave J. uses in FRL 1265 a small audio transformer as current sensor. It generates carrier wave for two anti-phase reference voltages.

To make RX operation "wireless", Bulgarian designers use a DD search head with Ref winding as shown below. It receives the genuine carrier wave and this voltage is used for reference. However this design solution requires an additional wire in connecting cable and additional pin in cable connector.

[Davor]

Dave J. uses in FRL 1265 a small audio transformer as current sensor. It generates carrier wave for two anti-phase reference voltages.

Please note that a transformer is not needed because a common mode of LM339 inputs includes a negative supply rail. I didn't test it though.

[mikebg]

Please note that a transformer is not needed because a common mode of LM339 inputs includes a negative supply rail. I didn't test it though.

I think Dave J. is participant of our forum and can explain this. However I prefer the Bulgarian solution because it involves GND signal as carrier wave for TGT signal.

[mikebg]

Conductive GND signal

Conductive GND signal

When a TX coil is above conductive half-space, the eddy currents produce secondary magnetic field. The configuration of secondary field is as if a mirror current loop is placed (buried) symmetrical to conductive surface.
Virtual current loop has the same diameter as the TX coil. This is illustrated below.

The inductance and resistance of current loop form a time constant or cutoff frequency that can be calculated with formula given in other thread:

http://www.geotech1.com/forums/showt...462#post104462

The measurements confirm a more accurate theory that there are more virtual current loops with 9 times and 25 times smaller diameter. Since their influence is weak, in order not to complicate the analysis, we will ignore them.
The received signal that secondary magnetic field induces in a coil, we will call GND signal. To get a clean GND signal, the coil must be plced in induction balance with the TX coil. In practice we receive a combination called AIR&GND signal. Most powerful AIR&GND signal receives the TX coil, therefore the MONOCOIL proved most inopportune search head for metal detectors.

We can use NETWORK ANALYZER to explore the conductive GND signal. The method of study of conductive nonferrous object with a time constant was explained in another thread:
http://www.geotech1.com/forums/showt...773#post100773

[Dave J.]

The transformer seemed like a good idea at the time: and, it did work well.

--Dave J.

[Davor]

Hi Dave, I'm so glad to see all the key people in metal detecting design are here on this forum.
BTW, I do not doubt that transformer did well - my idea is to avoid parts that are difficult to come by, and in this case it is a transformer. Detecting a coil current by comparators to set a 90° phase is a brilliant idea.

[Tepco]

Probably there is not much difference between current transformer and extra coil approach for reference generation. In order to get any valid reference at all, extra coil must be relatively closely coupled to TX coil, then ground influence to resonant tank parameters will be about the same, except usual 90deg phase shift generated by current sensing. (Also, in Fisher, TX is quartz derived, Bulgarian design is self-oscillating)