Best circuit for search oscillator (quiz)

[mikebg]

2. How oscillators work?

If “AN AMPLIFIER CAN OSCILLATE, BUT AN OSCILLATOR CAN AMPLIFY”, then what signal it should amplify?
Without feedback, the oscillator amplifies microphone signal and thermal noise in input. We will test what happens with noise when change the gain (amplification).
You can use for experiments the souncard in your computer, microphone and loudspeaker. Since I have an audio amplifier with LM386, I connected to it 8 ohm loudspeaker and microphone of a headset. To the circuit diagram shown below must be added RC shunt for inductive loudspeaker because amplifiers with feedback hate inductive and capacitive load. Despite the headset has headphones, I use separate loudspeaker in order to change its distance to microphone and to realize regenerative "sound radar".
When your system oscillate, simply close microphone with cupped hands to stop it and hear the noise. To see noise in output with o-scope, disconnect the loudspeaker and measure voltage in output of amplifier.
Since sound radar must have maximal sensitivity to a remote obstacle, you should use directed microphone and loudspeaker?Increase the gain until output of amplifier starts to saturate with noise peaks. Visual the maximal gain in threshold of saturation seems as a band of noise having width about 1/8 to 1/6 of maximal possible swing as noted in right with Vp-p max.
Now you can reduce gain, connect the loudspeaker and test how changes frequency of oscillation with distance, with gain and when swap leads of loudspeaker or microphone.
Remains to see what happens when we connect TX coil instead loudspeaker and RX coil instead microphone. We need induction balanced coils, but for comparison we can build oscillating circuit with monocoil.

[mikebg]

NOISE INDUCTION METAL DETECTOR

3. Noise induction metal detector.

The principle shown in Fig. 2 above is specific case of NI (Noise Induction) metal detector. In this kind MDs, the TX excites target by wide band of frequencies as in PI metal detectors. The difference is that in the PI is used a pulse generator to drive TX coil, but in the NI is used noise generator . Our case is specific because there is no separate driving generator, TX coil is connected in a self-oscillating circuit.
However the self-oscillating principle is widely used in conventional metal detectors. The Colpitts circuit shown in posting #1 is mostly used as TX in single frequency metal detectors.

TX COIL CONNECTED IN SELF-OSCILLATING CIRCUIT IS BAD DESIGN SOLUTION.

You know that there are metal detectors, in which TX tank is connected as load to an efficient amplifier driven by rectangle pulses. This is the correct principle of operation.
I will try to explain why connecting TX coil in self-oscillating circuit is bad for sensitivity of metal detector. Let we see the spectrum of target signal. Below is shown an example when TX excites environment with frequency 6kHz and search head moves slowly. Without target, the RX receves AIR & GND signal as an unmodulated carrier frequency 6000Hz. When arises target signal, the movement of search head modulates the carrier frequency with two side bands USB and LSB. They contains modulating frequencies for example from 0.1Hz to 1Hz. In this region of extremely low modulating frequencies, every electronic amplifier has increased noise known as 1/f noise or flicker noise. The self-oscillating circuit contains amplifier which generates flicker noise. The circuit has regeneration (positive feedback), which increases these noise until amplitude of oscillation. That means our carrier frequency is modulated by flicker noise and demodulator will detect false target signals despite there is no target.

There is no generation of noise in TX with driven amplifier. TX amplifier operates as switch driven by rectangle pulses. There is no regeneration of noise. Despite known circuits are not designed properly because amplitude of TX current depends on distance coil - earth, this drawback can be avoided by amplitude stabilisation with P-I-D controller.

CONCLUSION: The TX circuit of every known sine induction metal detector can be improved.
Here is the block diagram of improved TX:
http://www.geotech1.com/forums/showt...eferrerid=2910

Remains to design a search oscillator for BFO metal detector according Fig. 2 above..

[mikebg]

4. Design of search oscillator

a) TX tank circuit
In the circuit diagram Fig 2 of posting #16 , the closed regenerative loop will have narrow pass band because there are two tuned circuits formed by TX and RX coils. That means we will have very steep phase characteristic and as result small frequency shift if both tanks are tuned to equal resonance frequency. To achieve large frequency shift, we need slant phase characteristic, ie both tanks should be tuned to different resonance frequencies and damped with resistors and increased coil resistances.
Let try with SPICE what happens when instead 16 ohm loudspeaker we connect a 16 ohm TX coil. To reduce GND signal, the resonance frequency is near to 8kHz almost as in VLF detectors.
The magnitude and phase characteristics shown below have dimension of transfer admittance because transforms output voltage of amplifier in current via TX coil.

[taliesin]

hi people,
very interesting discussion.i have learnt a lot from this forum.in the image i've added (thomas scarbourgh ) bbmd .i would like to ask which is tx and which is rx.L1 or L2 ?
to my way of thinking they are both rx and tx.i cant wait for your replies as i like all the semantic arguements you guys have.
73's

[maikl]

hi people,
very interesting discussion.i have learnt a lot from this forum.in the image i've added (thomas scarbourgh ) bbmd .i would like to ask which is tx and which is rx.L1 or L2 ?
to my way of thinking they are both rx and tx.i cant wait for your replies as i like all the semantic arguements you guys have.
73's

On your link Thomas says the following: """Notes
Various embodiments of the BB metal detector have been published, and it has been widely described in the press as a new genre. Instead of using a search and a reference oscillator as with BFO, or Tx and Rx coils as with IB, it uses two transmitters or search oscillators with IB-style coil overlap. The frequencies of the two oscillators are then mixed in similar fashion to BFO, to produce an audible heterodyne. On the surface of it, this design would seem to represent little more than a twinned BFO metal detector. However, what makes it different above all else, and significantly increases its range, is that each coil modifies the frequency of the adjacent oscillator through mutual coupling. This introduces the "balance" that is present in an IB metal detector, and boosts sensitivity well beyond that of BFO. Since the concept borrows from both BFO and IB, I have given a nod to each of these by naming it a Beat Balance Metal Detector, or BB for short. Happy hunting!""" I'd just added to L1 and L2 have different responses to metals (L1 detect for itself and for itself detects L2). But at the intersection of two coils (L1 and L2) is obtained by discrimination and greater depth of detection. Drawback is the stability of the ground (for those who do not want to dig in vain on hard ground it is very important)

[taliesin]

thanks for that maikl.i sort of get what he's saying he says "it uses two transmitters or search oscillators with IB-style coil overlap" then he says "I'd just added to L1 and L2 have different responses to metals (L1 detect for itself and for itself detects L2). But at the intersection of two coils (L1 and L2) is obtained by discrimination and greater depth of detection."now one's an rx ?.the two coils are exactly the same and configured the same so we could swap L2 detect for itself and for itself detects L1..dont get me wrong i built this and it's quite sensitive and works well for so few parts.he's very ingenius.
perhaps i'm splitting hairs here to me tho the two inductors just produce an emf

[ODM]

I've been hibernating for a while, but apparently there's some rather interesting discussion brewing in here. This was one thing that was bugging me for a while with VLF designs; the self-resonant designs are probably better off with battery life, but they suffer a similar shift with varied conditions as the receiving tank does, unlike reference-driven ones.

Analytic approach, got to love it. Thanks