Frequency Domain and Time Domain relationships

[vbeeeks]

To tame the smoke I've got to tame the pulse window to 3rd multiple of L/R so as not to avalance the FETs and detonate the Caps ... effectively to tame the flyback tidal energy which comes rushing in ... and then to break the waves by raising the barrier synchronously in a interfering manner such that wave's energy unpacks or demodulates itself in a smooth fashion ... oh no this is not a poem....

[Davor]

You just observe if there is a whiff of smoke escaping your design, if not - it is OK

You know the fact that it is the smoke that runs all the electronics. Once it escapes the electronics die. Quite literally.

[vbeeeks]

How to quantify all the timings as a function of voltages, currents, tao, frequency!!!!, amplitude, zero-crossings, phases, inductance, impedance, duty-factor ---so hell of a maths involved -- this all to derive the right/optimum sampling window, damping pulse, Tx pulse -PPR, SAT, power optimization & synchronization, Rx demodulation ---- exotic integrals and differential goes here --- then discrimination with FFT / DFT ... Moodz has already crossed over to image domain from the latter with fourier image, multichannel sampling all that with audio engineering, DSP and 8bit FPGA and 100 MHz clock. All in embedded technology. Praise the Lord.

[golfnut]

All of this reminds me of CDI unit on a car or primary pulse forming network on marine radar units. Storing high volts on a cap and switching it into an L is useful.


Charge up the cap, switch on a device - dump the high voltage stored cap energy into the L.

As the cap empties (within a cycle) its polarity flips - this resets the switch device to off - so cap can charge for next go.

S

[tony_av]

Close... Wait a little more, the energy starts to be returned to L. Then switch ON the main MOSFET for time needed to earn that energy.

[vbeeeks]

The circuit will be optimally efficient at 3 times L/R in terms of power recycling and may be maths!!!!

[tony_av]

You also need to discharge that capacitor. And, if it is a big, you need more time to get the energy = longer flyback. Instead, just wait a litle till all energy is in the self capacitance,
and at that very moment, discharge it. The moment is critical.

[moodz]

Geez guys don't be so harsh, if you design a metal detector that goes beep somewhere in the vicinity of metal then you've got competency. Its doesn't have to mean your any good, just that can perform the basic task. Non-optimum is probably a fairer term. Moodz.. still no idea.. give us the rest of the terms and operators and we'll arrange them for you.

Midas

I dont think orginal use of incompetant was true english meaning .. what is meant is "less than optimal" as you state .... it is a lesser used usage of the word incompetant.

The 0.5LI^^2 = energy in coil ... the rest is easy.

[Stefan]

Need to get rid of the energy in the inductor at turn off. ie the flyback voltage. A switch closed for a couple of u sec just after turn off and direct the current into a capacitor to store the energy and then turn the switch off.

Regards,

Stefan

[tony_av]

No, C is a self capacitance of the coil. We need to dicharge it. To do this we need a switch, closed for a microsecond or so after the flyback.

[Stefan]

0.5 * L * I * I ---- dump this energy into a capacitor----- power factor correction with a switch.

Regards,

Stefan

[tony_av]

0.5LI^2=0.5CU^2

[Midas]

Both designs are "INCOMPETENT" as they both rely on damping resistance ... damping resistors are for oscillators not wideband systems .... however typical IB design is not better than monocoil because it throws away a crucial piece of information that the monocoil has. Like antennas in radio the monocoil has wider "aperture" therefore better depth. The secret is in the damping .... not the coil topology ... what is the secret ??? .... As Aziz says "come on you guys .... put your thinking caps on" ... stop turning over the ploughed ground .... I will even give you a starter ( any coil topology you like ) ....

The LHS ( left hand side of the equation for non math ) is this >>>>> 0.5 * L * I * I <<<<<<

I am practically giving it away here LOL ... lets see who can figure it out .

Geez guys don't be so harsh, if you design a metal detector that goes beep somewhere in the vicinity of metal then you've got competency. Its doesn't have to mean your any good, just that can perform the basic task. Non-optimum is probably a fairer term. Moodz.. still no idea.. give us the rest of the terms and operators and we'll arrange them for you.

Midas

[vbeeeks]

My guess

Is it : L/R = (0.5*Inductance*Current^2)/(Resistance*Current^2)? "=" is proportional symbol. L/R time constant and somewhat sprinkling of D-class amplifier THD characteristics.

[moodz]

The circuit diagram in left shows an incompetent designed search head for PI or wideband metal detector. RX and TX windings have equal number of turns N1=N2. The windings are placed in close proximity, which makes coefficient of coupling almost K=1. This generates the maximal possible AIR signal in RX winding and RX receves strong GND signal. To avoid saturation of RX preamp, we should reduce its amplification (gain). This reduces sensitivity. Despite drawbacks, the incompetent designed search head has two advantages:
1. The form of received EMV is as derivative of TX current.
2. The phase of received signal is constant independent on frequency,
http://www.geotech1.com/forums/showt...552#post163552

The circuit diagram in right shows the worst case - both windings are combined to a self - inductance called MONOCOIL. Then voltage across coil resistance r is added to AIR signal. A Re component is added to received signal at all frequencies and the phase angle depends on frequency. The RX can not sample when flows current in TX coil and we have not correct reference voltage for synchronous demodulation:
http://www.geotech1.com/forums/showt...729#post163729

CONCLUSION: For best results, QED and all other (more primitive :-) metal detectors should use induction balanced search head.

Both designs are "INCOMPETENT" as they both rely on damping resistance ... damping resistors are for oscillators not wideband systems .... however typical IB design is not better than monocoil because it throws away a crucial piece of information that the monocoil has. Like antennas in radio the monocoil has wider "aperture" therefore better depth. The secret is in the damping .... not the coil topology ... what is the secret ??? .... As Aziz says "come on you guys .... put your thinking caps on" ... stop turning over the ploughed ground .... I will even give you a starter ( any coil topology you like ) ....

The LHS ( left hand side of the equation for non math ) is this >>>>> 0.5 * L * I * I <<<<<<

I am practically giving it away here LOL ... lets see who can figure it out .