Mod TX of Coinmaster, TM808, Eagle

[mikebg]

Mod TX of Coinmaster, TM808, Eagle

The purpose of the modification is to suppress parasitic AM of TX when the height of sensing head above conductive ground changes.
The new element is a Schottky diode SD. Resistor R3 is only required during setup.
For setup, TX coil is placed above salt water. The resistance of R2 is reduced untill SD start to conduct short current pulses. This is energy that returns to the supply rails, but not to overdo because it appeared harmonic frequencies in the TX spectrum.

[WM6]

Mod TX of Coinmaster, TM808, Eagle

The purpose of the modification is to suppress parasitic AM of TX when the height of sensing head above conductive ground changes.
The new element is a Schottky diode SD. Resistor R3 is only required during setup.
For setup, TX coil is placed above salt water. The resistance of R2 is reduced untill SD start to conduct short current pulses. This is energy that returns to the supply rails, but not to overdo because it appeared harmonic frequencies in the TX spectrum.

Hi mikebg

What mean here parasitic AM (parasitic amplitude modulation)?

[pebe]

Mod TX of Coinmaster, TM808, Eagle

The purpose of the modification is to suppress parasitic AM of TX when the height of sensing head above conductive ground changes.
The new element is a Schottky diode SD. Resistor R3 is only required during setup.
For setup, TX coil is placed above salt water. The resistance of R2 is reduced untill SD start to conduct short current pulses. This is energy that returns to the supply rails, but not to overdo because it appeared harmonic frequencies in the TX spectrum.

Component values and waveform amplitudes would have helped to understand the circuit. The circuit on its own is quite meaningless, as is the unqualified term 'parasitic AM of TX'.

[mikebg]

Parasitic amplitude modulation of TX

If you connect an AC voltmeter across TX coil, you'll notice that when TX coil is close to the conductive ground, the amplitude of oscillation decreased. This amplitude modulation is parasitic, because ground absorbs energy from LC tank circuit.
The TX transistor operates as current pump, charging LC tank with energy. White's uses the worst circuit diagram for TX, because each current pulse contains stabile portion energy. When the height of sensing head above conductive ground changes, the amplitude of oscillation in TX tank modulates.
To avoid this are used amplitude stabilizers for TX (see attached TX circuit of Fisher 440). Garrett uses P-I controller for amplitude stabilisation. I posted several TX circuit diagrams of Garrett in this section.
If TX signal is modulated by ground, RX coil receives two ground signals:
1. AIR signal, modulated by change of height. The AIR signal is due to imperfect induction balance.
2. GND signal, modulated by change of height. The GND signal is due to soil properties.
Damage from parasitic amplitude modulation for detection of deep targets, will be described in detail in other postings.
In the modification, current pump carges tank circuit with a large portion of energy. Thus the LC tank overflowing. Indeed, excess energy is returned to the supply rails. When the TX coil is close to the conductive ground, the excess energy becomes less and supply rails receive less energy returned.
Return of energy to the supply rails is not the best solution to the problem of "amplitude dismodulation". I tried to find the best solution, and since experimented circuit works very well prepared for it publication of the Forum.
The REMI group warned me that if I publish this imperfect circuit diagram, they will nominate me for the prize "carbon paper", which gives designers incompetents reckless copying of foreign structures.
The bad thing is that they refuse to improve my circuit, I've posted in the forum. Improvement was my task for the implementation of which should have read the literature on P-I-D controllers and their methods of adjustment. Moreover, should make mathematical analysis for antenna matching and analysis for efficiency of TX coil with limited weight. I have no desire to do so, because my TX circuit is working well. I hope that someday they will do and this job because their hobby is to create superior designs.

[WM6]

If you connect an AC voltmeter across TX coil, you'll notice that when TX coil is close to the conductive ground, the amplitude of oscillation decreased. This amplitude modulation is parasitic,

.

Hi mikebg,

I mean here we can talk about dumping not about modulating signal. Maybe we have problem with terminology.

[pebe]

I don’t see any likeness to your first circuit and I don’t agree with many of your terms

If you connect an AC voltmeter across TX coil, you'll notice that when TX coil is close to the conductive ground, the amplitude of oscillation decreased. This amplitude modulation is parasitic, because ground absorbs energy from LC tank circuit.

The amplitude simply ‘varies’ – it is not ‘modulation’.
It is not caused by anything ‘parasitic’ it is due to the loading on the tuned circuit – the same way any supply is affected by loading.

If TX signal is modulated by ground, RX coil receives two ground signals:
1. AIR signal, modulated by change of height. The AIR signal is due to imperfect induction balance.
2. GND signal, modulated by change of height. The GND signal is due to soil properties.

The Rx coil will receive a single sine wave whose amplitude and phase are the vector sums of the two signals you mention.

In the modification, current pump carges tank circuit with a large portion of energy. Thus the LC tank overflowing. Indeed, excess energy is returned to the supply rails. When the TX coil is close to the conductive ground, the excess energy becomes less and supply rails receive less energy returned.
Return of energy to the supply rails is not the best solution to the problem of "amplitude dismodulation".

This is not so. The circuit is that of an oscillator whose amplitude is controlled by negative feedback – as in normal practice. When the amplitude reaches a predetermined level, feedback reduces the forward bias of the cross-coupled pair (and hence the gain of the oscillator) to prevent any further increase. Excess energy is not returned to the supply rails because there was never any excess to start with.

You have introduced a new word ‘dismodulation’. I have never heard of it!

[mikebg]

To WM6

WM6, look into your English dictionary to see that "modulate" means change. If we use the term "dumping" of TX tank circuit, then the real process is CHANGE OF DUMPING due motion of TX coil relative to a halfspace core (ground) and change electromagnetic properties of the core (because the soils are different). A thorough visual explanation will be made with pieces that we have now begun to plot.

[WM6]

WM6, look into your English dictionary to see that "modulate" means change. If we use the term "dumping" of TX tank circuit, then the real process is CHANGE OF DUMPING due motion of TX coil relative to a halfspace core (ground) and change electromagnetic properties of the core (because the soils are different). A thorough visual explanation will be made with pieces that we have now begun to plot.

Probably you use general dictionary. Here we use AM or Amplitude Modulation as electronic term. So you have to use dictionary of electronic term not general dictionary. On web you can find free online electronic dictionaries. Attenuation of signal is not the same as modulation, although in both cases the changes are going on.

[mikebg]

To pebe

Pebe, before modification and after modification, the transistor operates as current pump. This is not oscillator circuit, but an amplifier class C. It transforms the input voltage (square wave) in current pulses with duty cycle 50%. The current of pulse is determined by expression Ic=(Vz-0.6)/R2. The LC tank circuit transforms current pulses in oscillation.
Energy return in original circuit is possible via collector junction of transistor and ZD, however this requires voltage excess 1.2V. The SD reduces this excess to 0.4V only.
What you say for oscillator, refers to TX circuit diagram of Fisher 440 in posting # 4. The task of "feedback" in this circuit is to dismodulate the amplitude of oscillation. For this purpose a transistor is connected as diode and operates as amplitude detector. It generates control signal for amplitude dismodulation. The dismodulator is P (proportional) controller. Garrett uses P-I controller, but for best results it should be P-I-D controller.

[pebe]

Mikebg,

Pebe, before modification and after modification, the transistor operates as current pump. This is not oscillator circuit, but an amplifier class C. It transforms the input voltage (square wave) in current pulses with duty cycle 50%. The current of pulse is determined by expression Ic=(Vz-0.6)/R2. The LC tank circuit transforms current pulses in oscillation.

Energy return in original circuit is possible via collector junction of transistor and ZD, however this requires voltage excess 1.2V. The SD reduces this excess to 0.4V only.

I suggested you put component values on your first circuit. Instead, you ignored my suggestion and came up with a second circuit. So I assumed you comments referred to that and I replied accordingly. But now it seems you are commenting on the first one. And now your last post shows yet another, different, circuit!

Why don’t you number your circuits and refer to their numbers when making comments? That way, we would both know we are talking about the same thing.

What you say for oscillator, refers to TX circuit diagram of Fisher 440 in posting # 4. The task of "feedback" in this circuit is to dismodulate the amplitude of oscillation. For this purpose a transistor is connected as diode and operates as amplitude detector. It generates control signal for amplitude dismodulation. The dismodulator is P (proportional) controller. Garrett uses P-I controller, but for best results it should be P-I-D controller.

You are talking about ‘dismodulation’ again. As you obviously mean ‘varying’ when you talk about ‘modulation’, do you mean non-varying or ‘amplitude level stabilisation’ when you talk of dismodulation.

As a newcomer to metal detection, I am not familiar with P-I and P-I-D controllers. Could you briefly explain what they are?

[mikebg]

"A proportional–integral–derivative controller (PID controller) is a generic control loop feedback mechanism (controller) widely used in industrial control systems. A PID controller calculates an "error" value as the difference between a measured process variable and a desired setpoint. The controller attempts to minimize the error by adjusting the process control inputs. The PID parameters used in the calculation must be tuned according to the nature of the system." - Wikipedia
This is not the best definition, but works. An excellent explanation for beginners was made by Tietze and Schenk:
"Semiconductor Circuit Technology" Tietze and Schenk, Springer Verlag Publishers.
The TX circuit diagram of Garrett in posting #9 contains P-I controller. The proportional part is an amplifier with U1B. The integral part is a capacitor C2 connected as feedback of amplifier. It makes gain extremely high at zero frequency. This reduces error at setpoint to minimum.

[pebe]

"A proportional–integral–derivative controller (PID controller) is a generic control loop feedback mechanism (controller) widely used in industrial control systems. A PID controller calculates an "error" value as the difference between a measured process variable and a desired setpoint. The controller attempts to minimize the error by adjusting the process control inputs. The PID parameters used in the calculation must be tuned according to the nature of the system." - Wikipedia
This is not the best definition, but works. An excellent explanation for beginners was made by Tietze and Schenk:
"Semiconductor Circuit Technology" Tietze and Schenk, Springer Verlag Publishers.
The TX circuit diagram of Garrett in posting #9 contains P-I controller. The proportional part is an amplifier with U1B. The integral part is a capacitor C2 connected as feedback of amplifier. It makes gain extremely high at zero frequency. This reduces error at setpoint to minimum.

The control system described in Wikipedia relates to servo-operated systems commonly found in digital proportional radio control systems for models and robots. It ‘looks ahead’ and computes present and targeted servo position and speed (taking into account servo overshoot) and comes up with the most suitable correcting signal at any one time.

Whether the term PID should be applied to #9 circuit is debatable. While initially correcting for an error it behaves as a simple AGC system, as used in radio receivers for over 70 years. Once in lock, the full 1 million gain of the op-amp comes into action. I think the addition of that feature is ‘gilding the lily’ and I cannot see how the circuit benefits from such a high gain. It would forever be correcting for an insignificant variation in amplitude.

[mikebg]

It would forever be correcting for an insignificant variation in amplitude.

Pebe, Eureka!
You already rediscovered the "dismodulation".
In modern metal detectors there are two or more demodulators. Often demodulators are called "detectors" because their job is to detect changes of the received signal. Whether a signal is modulated or not, decide demodulators. The signal from deep target appears as very low modulation index because it is against a background of two strong signals AIR and GND. Synchronous demodulators in metal detector are adjusted so as to eliminate signal GND, which generally appears modulated. AIR signal differs in phase from the GND signal, therefore demodulators are sensitive to its modulation. The AIR signal should be "dismodulated" so that we can detect deeper targets, because its modulating spectrum coincidents with modulating spectrum of target signal.
Theory of PID control is fundamental and used everywhere in electronics: AGC, AFC, frequency correction of operational amplifiers.
In the aforementioned book by Tietze and Schenk, is described the tuning of PID controller to achieve minimal settling time.

[pebe]

You insist on continually referring to 'dismodulation'. Not only is it not a technical term - it is not even in the English dictionary! So you are the only one who knows what you mean. Why do you use this nebulous waffle?

If you want to communicate, please use standard electronic terms, otherwise I cannot take you seriously.

[mikebg]

DISMODULATION

Posted by Pebe
"Separate Oscillator. Can anyone point me to a circuit that uses a driven TX output stage - rather than a self-oscillating one?"
Pebe, below is the worst working TX circuit with driven output stage used by a great company. Output stage operates as class C amplifier with current limit by equivalent emitter resistance Re. It pumps tank circuit with current pulses, which are independent from energy absorbtion in ground. Pumping TX tank with calibrated current pulses causes parasitic modulation of TX field when sensig head moves. Spectrum of modulating signal are frequencies from 0,1 to 6Hz. The same spectrum occupates modulating signal of the target, so you need to remove the parasitic modulation if you want to find deeper targets.
Pebe, I use one term expressing the removal of parasitic modulation. If you do not like "dismodulation", then offer some other.
Now you know a simple method to suppress parasitic modulation at mode Hi Power, when Q11 is on. Put a Shottky diode and adjust tank to overflow to supply rails when sensing head is over salty water.
However in mode Low Power, when Q11 is off, we can not use the supply rails as amplitude limiter because the amplitude should be lower than rails voltage. But you know a complicated metod for removal of parasitic modulation - the PID controller.
The (R)EMI group has a detailed description of the PID controller by Tietze and Shenk as a PDF file 996KB, but it is in German. If anyone wants this information in German, let me know.