IB-Metal Detector (IB-nulling problem)

[Aziz]

Has anyone some ideas about the nulling (zero induction) of the IB-Coil. I found that die Induction-Balance is frequency dependent. If a metal detector works with multiple frequencies, only one frequency can be nulled optimally.
How does the other detectors with multiple frequency null the coil?
(RC-Network?, mechanically?, additional wire loop?)

And what is the minimum signal level (peak to peak mV) at the balanced coil (receive-coil)?

Aziz

[Carl-NC]

Re: IB-Metal Detector (IB-nulling problem)

The frequency dependency is probably due to capacitive parasitics, rather than inductive coupling. So I'd look for ways to minimize capacitive coupling. Or ways to cancel it.

As I understand it, detector manufacturers don't null the coil for an absolute minimum RX signal. They set it just off of the null, I guess to establish a repeatable reference point. Or so I'm told.

- Carl

[Aziz]

Re: IB-Metal Detector (IB-nulling problem)

Hello Carl,

I think so too, to have a minimum signal on the RX coil. Especially to have a defined phase offset. On the optimally balanced coil system (RX signal = 0), it would be very difficult to have a stable phase offset.

Nevertheless, the more the IB-system is balanced, the more sensitive would be to very small changes in change of magnetic field disturbation.

I have many difficulties getting an optimal and noise free amplifier. Do you know a good operational amplifier? I have tried with OP07 and OP27. But I am searching for a better one.
I also tried to supply the amplifier with a very stable reference voltage regulator. It brought better result.

Indeed, on my search coil, there must be a high capacitive load (PCB-IB-Coil, 6 coils, 3 on each side, most of the area is copper-wire). I will try with fewer windings and more gain of the amplifier.

An another idea to minimize IB-signal is, to adjust the received signal with an additional decoupled single loop, which is electronically adjusted and real-time feeded correction signal. This seams to be solved with more computational power (DSP or Micro-Controller).

Aziz

[Siliquae_Sid]

Hi Aziz, try some of the LT series Op-Amp from Linear Technologies. Some of them are so ultra low noise, thet it's silly!

Also, what about your PCB layout, and input filtering? I have reached noise figues of signal amplification on soem of my designs of -123dBm (Earth noise figures) This gives a sensitivity of 1.25nV and thus can troretically pick up a target up to 28" from the coil.

There is a trade-off. this level of sesitivity means your ground balance must be 100% accurate, and this means delaying the signal returning from the head. Why? Because as the user sweeps, the auto ground balance lags the actual ground conditions, therefor the main Rx signal is delayed and times to reach the demodulators at EXACTLY the same time as the compensated signal, this means a MUCH more accurate GEB than is currently available on any machine.

Even Minelab would be proud of the performance I can achieve, and I'm going commercial with my design later in the year ;-).

I'm not interested in patents, certain people who work for large companies and design machines for a living, seem to think that detector design is some form of secret black art. Well they are just sad retentives. Most of them know very little about electronics other than analogue, and so are not very good Engineers.