True All Metal Mode and decreased sensetivity in discriminate modes - GB 2 and others: Why?

[Reno Chris]

OK, so Its always been my experience that with a true all metal mode as in the Gold Bug 2, you have greater sensitivity in all metal than when you switch to discriminate mode. I've proved this over and over with with the GB2 and other machines that have a true all metal mode where the signal does not go through any processing for discrimination. My question is why.
I've always assumed that it was a matter of the additional signal processing - the more processing you do, the more you will wash out or lose the weakest signals. Would the folks who actually know about this type of circuit design agree?
So many makers are going to a phony all metal mode where the discrimination is still on, and the signal is being processed, but the rejection level is set to zero or acceptance notches are all set to accept.
By my thought above, as you still have the signal processing going on (even if you don't reject anything) you still have the loss of sensitivity created by the extra signal processing.
Standard use of the GB 2 by experienced users is to hunt in all metal, when you find a signal if it is strong, flip over to discriminate to see if its rejected. If the signal is weak, you need to dig on it until the signal comes in at least moderately strong before you flip to discriminate to see if its ferrous or not. This is gives the greatest sensitivity for finding targets while still exercising good discrimination. Unfortunately with the machines which do not have true all metal mode, there will be no real advantage to turning off the discrimination (as its always on anyway) - effectively the user is fooling himself (or the maker is fooling the buyer / user).
My main question is why - its the extra signal processing that causes the loss of sensitivity, right?

[Orbit]

You're absolutely right!

[Qiaozhi]

I assume your definition of "a true all metal mode" is the situation where the GEB channel is the only one in the signal path, and the DISC channel is physically disconnected. The other scenario is when the DISC channel is left connected, but the amount of discrimination is set to zero.

The crux of the problem is that the GEB and DISC channels sample the received signal in different ways. The GEB channel samples close to the zero-crossing point of the signal, and hence reacts to any phase-shift. Since the ground signal produces no phase-shift, the GEB channel eliminates that signal. For moderately mineralised homogeneous ground, the sample point can be adjusted so there is no response to the small phase-shift produced by the ground matrix. The DISC channel samples close to the peak of the received signal, and reacts to changes in amplitude. Inductively balanced coils are nulled in such a way that ferrous targets cause the amplitude of the signal to go one way (usually decrease) and non-ferrous signals go the other (an increase). By comparing the results from both channels, the detector can determine the type of target being detected.

In theory, assuming the two channels produce the same output signal for the target, it should be possible to simply adjust the DISC channel to zero discrimination to switch to all metal mode. However, this can only happen if the DISC control is able to move the sample point all the way back to the zero-crossing, which is not the case. There is a limited range of adjustment on the DISC control. Which means that the GEB channel is still reacting to phase-changes, and the DISC channel is still reacting to amplitude changes (but the DISC sampling now occurs a point where both ferrous and non-ferrous targets give the same response). Therefore, if you have a weak target that gives a decent phase-shift, but only a small amplitude change, the received signal will be weaker in scenario 2, and stronger in scenario 1.

[Koala]

Yes you are correct. Not sure if the GB2 discrimination goes to zero. If it does there is normally very little between zero disc and all metal mode. As you increase the disc you loose those weeker signals.

[Carl-NC]

Most VLF detectors also use high-pass filters to help distinguish fast responses (targets) from slow responses (ground). A high-pass filter mathematically produces a derivative of the input signal, so the output of the first filter is called a "first derivative" signal, and the output of a second filter is called a "second derivative" signal. The second derivative signal is what's used for disc mode as it has most of the ground removed. All-metal can use either first-derivative (motion-mode AM) or the unfiltered signal (zero-motion AM), but there are (progressively) more ground effects. Usually not a problem as all-metal is mostly used for pinpointing in a small area with more careful movement. Some detectors use first derivative for regular AM mode and unfiltered for pinpoint mode.

The depth difference arises because the filters also attenuate the signal. They attenuate slow signals a lot, and fast signals less, so they also tend to be somewhat sensitive to sweep speed. That's why disc mode (2 filters) gets less depth than AM mode (0 or 1 filter). Back when motion disc was invented (Payne, ~1977), the first machines had a 4-filter system, so the disc signal was even more attenuated, and was very sensitive to sweep speed. You had to whip the coil like a grass sling to get good depth.

There is also another depth issue alluded to by Koala. In the bad 'ol days of the analog linear discriminator, disc was accomplished by subtracting portions of one channel from another, so that the resulting disc signal got weaker as you turned up the disc setting. Modern detectors might still run the disc off the same kind of subtraction math, but they do so digitally and leave the original signals intact. That way they can still discriminate but with no depth loss, even between "zero-disc" and "full disc" settings.

So on a traditional design there still will be a depth loss between 0/1-filter AM and 2-filter disc mode, which is not a Bad Thing. Generally, you want a stronger signal when pinpointing; at the very least, you don't want the signal to disappear as soon as you switch to AM/pinpoint mode. If you run the "AM" signal off the 2-filter disc channel then, yeah, it's no stronger and is a motion signal to boot; I personally don't like motion-mode pinpointing, though motion-mode AM hunting is fine.

Finally, there is no reason disc mode can't be taken off the first derivative signal, or even the raw signal, if ground effects are minimal. Or an additional filter or 2 if ground effects are severe. I suspect that we're going to see more automation of filter selection in the future, where the signal processing is optimized for ground effects and sweep speed. Compass did a version of this in the 90's, called the Varifilter.

[Reno Chris]

The difference between a first derivative and second derivative signal can make a huge difference to an experienced operator. I will give an example with the GB2. A while back I was detecting an area and got a weak but repeatable target signal. I tested it with discrimination and got no response. Now the discrimination on the GB2 blanks out the response, so on weak signals you can't tell if you are getting a ferrous response or if you are just getting no signal because its too weak to respond (this is an advantage to multi tone systems which allow you to tell the difference). On this weak target other observations said I should dig - meaning I had already dug more than a dozen smaller nuggets within a 10 foot radius, plus the target seemed to be in a crevice in bedrock. So I dug down about 5 inches or so and the target was still in the hole but much stronger. I again tested the discrimination and got a very clear non-ferrous indication. Further digging showed it to be a nugget about 1/10th of an ounce or 2 dwt. In this case it seemed that the weak target was too weak for the discrimination to work and resolve a target. This sort of situation is common. For the gold prospector, having a true first derivative, all metal mode is very important. For a coin machine, maybe it is less so and a designer could get by with just setting the discrimination to the lowest setting and working with the second derivative signal, but for the prospector, that extra sensitivity is critical.

[Carl-NC]

My personal preference is mixed mode, with 1st derivative AM in one ear, and 2nd derivative disc in the other, with disc wide open and iron grunt turned on. I don't like iron blanking, I want to hear everything. A lot of people don't like this because it presents a very busy audio, which is true, especially in trashy areas. But then it forces you to slow down, which is what you should do anyway.

[ivconic]

My personal preference is mixed mode, with 1st derivative AM in one ear, and 2nd derivative disc in the other, with disc wide open and iron grunt turned on. I don't like iron blanking, I want to hear everything. A lot of people don't like this because it presents a very busy audio, which is true, especially in trashy areas. But then it forces you to slow down, which is what you should do anyway.

Nautilus...