Stability of the VDI scale

[JoyJo]

Greetings! Gentlemen, I have a question. A question about the stability of the VDI scale.
In homemade metal detectors, a piece of ferrite (a piece of ferrite magnetic circuit of a transformer or antenna) is used to set the beginning of the VDI scale. At the same time, there is no such thing in branded metal detectors: they are configured from the factory and the user does not use additional items for customization.
How can this be achieved? The laws of physics have the same effect on homemade metal detectors and branded ones.
After calibrating the VDI scale with ferrite (which has its own VDI minus 90 degrees) we set the beginning (reference point) of the VDI scale for the remaining metals.
But depending on the ambient temperature, the electronic components of the metal detector change their parameters. because of this, the reference point of the VDI scale is shifted and as a result, all VDI values for other targets are shifted.
How can I stabilize the VDI scale setting? I assume that the trend of change can be traced by a change in the phase of the current in the TX circuit and by a change in the phase of the voltage in the RX circuit.
But. I understand how to implement the measurement of the signal phase in the TX circuit, it is enough to install a current measuring shunt. But how do I directly measure the voltage phase in the RX circuit (of course, if the receiving circuit is resonant and the RX capacitor is located directly in the coil)?
How do all branded metal detectors ensure the stability of the VDI scale to the user?
Maybe there is some kind of patent in this area or a scientific article?

[Carl-NC]

In commercial detectors a ferrite is normally used to set a nominal ground balance phase, but the VDI scale is normally set using certain eddy targets. At White's, we used a target emulator with equivalents of multiple US coins. Probably just a US nickel (low VDI) and a US quarter (high VDI) are sufficient to get the whole scale right.

VDI is calculated by a ratio by a ratio of the R signal to the X signal. If you design the 2 channels identically, then any tempco will cancel. For this it is important to the match resistors and capacitors in the motion filters to a tight tolerance, say 1%. Then you are left with the tempcos of the preamp, TX, and demod clocks. We are concerned with phase tempco, and usually the preamp is pretty good. If you see some phase movement, try adding a temperature-dependent resistor as needed (Spice can help figure this out). The demod clocks are similar, pay close attention to getting the phases stable over temperature (Spice again). The TX has a coil with wire resistance that shifts over temperature and this alters the TX voltage phase reference used for both ground balance and VDI. Usually the shift is enough to mess up ground balance (which needs 1/10th degree accuracy) but not VDI (1 degree or more resolution). One way to minimize the effect is to design the demod clock circuits so they cancel the tempco of the TX coil. Another is to use a current sensing resistor on the TX coil.

Since it seems you are seeing a significant shift in the VDI the first step is to identify where the problem is. This will require an oscope, a heat gun, and a can of cold spray.

[JoyJo]

It turns out, if I understood correctly, the vdi scale in commercial metal detectors is calibrated at the manufacturer's factory at two points on the scale and is taken into account in the software. Right?

Regarding the phase departure... The phase of the TX current will change as from a change in the resistance of the TX coil (due to temperature) this is how the inductance of the TX coil will change depending on its proximity and distance from the ground (and the properties of the soil, of course). At the moment, I'm using a sequential oscillating circuit. Do you propose to turn on the thermocompensated resistor afterward with the TX coil? And what about the discharge phase in the RX circuitcircu?
Or do you want to say that commercial devices use capacitors with different temperature coefficients of capacitance to combat the influence of temperature?

[Carl-NC]

In older analog commercial detectors the VDI (or tone regions, if there was no display) might have several calibration points, potentially one for each region. In digital commercial detectors the VDI might not be calibrated at all, it is correct "by design." Or perhaps there is a single calibration point after the GB cal is set.

In your case, the first thing to do is figure out what is causing the problem. Capacitors can be a problem which is why detectors usually have high quality film caps with low TC where it's critical. For you, I have no idea where it's critical because I don't know what circuit your talking about.

[JoyJo]

Yes, it is possible. Please excuse me: a discussion without a diagram is not correct.
Digital processing. There are three channels: X, Y and the soil channel.
ferrite sets the beginning of the vdi scale. and the soil balance is performed by a separate procedure before starting the search.

[JoyJo]

This is a fully digital device. I Q channels are already allocated in digital form. There is also a third channel. Detection channel. It turns to the corner of the ground. Based on the signal level of this channel, the presence of a target is detected. VDI is calculated from the I and Q channel data. In this setup, only one VDI scale calibration is needed. Ferrite calibration. VDI of ferrite minus 90. How can you avoid ferrite calibration altogether? In commercial instruments, users do not calibrate the VDI scale at all and it is stable. Even after 10 years it remains in place.

[Artyom]

Hi. Carl, if I'm not mistaken, it's about this device.

[Carl-NC]

OK, direct sampling eliminates the demods and motion filters. Now you are left with the preamp and TX circuit. I would look closely at the series resonance cap on the TX, that could cause a decent amount of phase shift over temp. I assume demodulation in the micro is done with a fixed reference to the TX clock, you may need to add a sense resistor to the TX and use a capture channel to make a new demod reference.

[JoyJo]

It turns out that all branded detectors use parts of the circuit to evaluate the phase of the TX current?

[Carl-NC]

Until recently, most detectors did not. The TX coil voltage was used directly to drive the demod clock generators. This meant the coil current phase could shift over temperature. To deal with this, the TX coil would be designed so the phase shift was minimized, or an NTC resistor was added in series. Some newer detectors have a TX coil sense resistor to get the TX current phase directly.

[JoyJo]

And if you use a full H-bridge? There will be no resonant capacitor, there will be a small inductance that the soil will not affect.

[Carl-NC]

Sorry, I don't understand the question.

[JoyJo]

I meant, will the conversion of the TX cascade instead of the resonant circuit to a full-bridge (H-bridge) circuit change the stability situation?

[pito]

This is what I think.
We are shifting Tx phase to have ground balance and if we use that shifted signal as reference for VDI = 0 deg then we have stable VDI and there is no need for ferrite .


TX coil sense resistor to get the = I will try that

[JoyJo]

Can I tell you more about the "sensitive element of the TX coil"? I didn't understand.