Ground Balance and ID

[Daves]

We see the ground balance feature in a normal detector, it usually comes at 90 as standard. When we hold a target and pass it through the air test, we see ID=52 as an example. When you change the ground balance and pass the target again, the id does not change.

Could you briefly share what ground balancing actually does?
Note: We talk about vlf metal detectors

[Carl-NC]

There are different ways to design a GB system. The simplest is to rotate the R channel until ground is nulled. If the X channel is also rotated along with the R channel then target IDs will change some. If the X channel is not rotated then IDs will change but not as much. The better way to do GB is to make a dedicated G channel that is rotated to null ground, while X & R are left fixed. Then target IDs stay the same no matter the ground setting.

[Daves]

There are different ways to design a GB system. The simplest is to rotate the R channel until ground is nulled. If the X channel is also rotated along with the R channel then target IDs will change some. If the X channel is not rotated then IDs will change but not as much. The better way to do GB is to make a dedicated G channel that is rotated to null ground, while X & R are left fixed. Then target IDs stay the same no matter the ground setting.

Ok Carl, i understand and applied this step, after than?

[Carl-NC]

I don't know what you are asking.

[Daves]

I don't know what you are asking.

Carl, after receiving the X and R signals, we cannot process them statically and we pass them through a filter. That's where the problems start. Double response etc. What do you suggest here?

[Carl-NC]

After the second filter, apply a rectifier so you only get the center lobe.

[Daves]

After the second filter, apply a rectifier to you only get the center lobe.

Thanks Carl,
I will research and try to implement it, I will let you know.

[Daves]

After the second filter, apply a rectifier so you only get the center lobe.

Carl, I couldn't find anything as "rectifier". Can you explain a little more?

[Carl-NC]

If you look at, say, a Tesoro circuit a comparator is often used to trigger on the (positive) main lobe while ignoring the (negative) side lobes. If you are getting a double beep then your signal may be upside-down, and the comparator polarity must be inverted. If this is a software design then an absolute value function can be used to truncate negative signals. There are lots of ways to do this, it's useful to study other circuits.

[Daves]

If you look at, say, a Tesoro circuit a comparator is often used to trigger on the (positive) main lobe while ignoring the (negative) side lobes. If you are getting a double beep then your signal may be upside-down, and the comparator polarity must be inverted. If this is a software design then an absolute value function can be used to truncate negative signals. There are lots of ways to do this, it's useful to study other circuits.

Carl, I'm asking about direct sampling. Naturally, there will be positive and negative signals in ferrous and non-ferrous metals.
If we cancel any response, we won't detect the sample ferrous or non-ferrous metal, right?

Can the coil response be adjusted to be unidirectional on every metal?

[Carl-NC]

The things you need to do in direct sampling are basically the same as what's done in analog designs. Study the analog designs to see how they do things. At the demod outputs, the R-channel is normally the all-metal signal and is positive for both ferrous & non-ferrous. The X-channel is negative for ferrous, positive for non-ferrous (polarities depend on the design). But then you pass those signals through motion filters and you get bidirectional responses no matter what. You can get rid of the motion filters but then performance in mineralized ground will suck.

[Daves]

The things you need to do in direct sampling are basically the same as what's done in analog designs. Study the analog designs to see how they do things. At the demod outputs, the R-channel is normally the all-metal signal and is positive for both ferrous & non-ferrous. The X-channel is negative for ferrous, positive for non-ferrous (polarities depend on the design). But then you pass those signals through motion filters and you get bidirectional responses no matter what. You can get rid of the motion filters but then performance in mineralized ground will suck.

yes, the problem is bidirectional signal I made fft and produced X and R signals. I tried and created the 1st derivative and 2nd derivative. But the problem is that it is a bidirectional signal. What kind of digital procedure do you recommend for the Center lobe in the 2nd derivative?

[Altra]

Here is an old patent that illustrates what Carl suggesting. See fig7. feeding the peak 2nd derivatives into a AND gate. The negative going signals are ignored.

https://www.geotech1.com/pages/metde.../US4700139.pdf

[Daves]

Additionally, how can I create the audio signal?

[pito]

bidirectional signal = for that we using DC bias, or you can use abs() in program

Additionally, how can I create the audio signal? = magnitude of X and R + VCO