Metals have what is called the frequency of the metal, signature frequency, etc, but boils down to a metal can be simulated with by an inductive/resistive circuit. The inductance is charged through the resistance of the metal. The self inductance charge and discharge through the resistance is frequency sensitivity. (As you know the other side of the coin where resistance is concerned is conductivity.) The charge and discharge is supports eddy currents which is how a target is identified. Gold will respond better to a higher frequency and small gold, say the size of a BB, will be detected better with a higher frequencies and a smaller coil. This all woks out fine but if we take a small link chain a foot long and coil it up we have a lot of gold but it still will be rejected many times. If we are in all metal then the gold chain looks like soil minerals in that it blends in with the soil.
The Explorer does not have front end filters as such. The receiver coil is direct coupled to 6 demodulators. They work in pairs of two to output three signals. The outputs of the demodulators are then filtered and averaged for three time domain signals that are process by the microprocessor. A frequency domain detector is much more concerned with filtering so we see as many as 6 filters in some of them.
The Explorer looks at the outputs of the demodulators, averages the outputs, filters out any ground noise, and then uses sensitivity, manual or semi-auto, to control the demodulators so the Explorer is not going to have soil mineral problems. What is often posted as soil mineral problems is actually discrimination of refined iron, non-linear electronics, Q of coil changes as it is moved over the ground, and ageing of electronics. This can be corrected simply by using semi-auto or manual which adjusting one leg of the demodulators. The end results are three ground balanced signals that consist of long, short, and medium time constants that are used to id the metal.
The links and self inductance of gold chain will oppose the electromagnetic field of the coil so the end results are a long time constant. The soil has a TC of about 1us so the gold chain has a long TC which causes it to look like the soil. The bucking of the transmitter generated magnetic field by the eddy current in the links is based on Lenz