# How can I drive an LC tank circuit to get a nice and clean sine wave?

As advised by Andy in this post, I constructed single loop transmit and receive coils:

I use this circuit to transmit a powerful field:

The Tx signal has a large spike that is reflected in the Rx signal, too:

How can I drive the Tx tank circuit to get a nice and clean sine wave?

• By using some of resonant switching techniques. You would need a current transformer, zero cross detector and FPGA circuit. Example: highvoltageforum.net/index.php?topic=188.80 Dec 22 '20 at 13:09
• Thanks a lot @JRE. Dec 27 '20 at 8:35

The way I drove the MOSFET (I used the very same MOSFET, the IRFZ44) was in a linear fashion. The MOSFET was also involved with other transistors and the whole parallel LC (plus transistors) was a highly controlled oscillator. I didn't use a signal generator; whatever tuning capacitance was added to the inductive loop, caused it to operate at a different frequency. This is how thousands per year were made and shipped (and still are).

However, I can't give you the full circuit details but I can advise you on how to drive that MOSFET in a more linear manner: -

• Set the DC bias current through the coil/MOSFET to be about 100 mA.
• This requires choosing resistors or using a pot
• AC couple your input waveform to the MOSFET gate and control its amplitude to get about 22 volts p-p.
• Use C0G/NP0 tuning capacitors - yes they are big and expensive but there is no option if you want really decent sinewave purity (and you do need that purity when looking for really small pieces of metal.
• The operating frequency for your sized coil should be around 500 kHz to 1 MHz.

Now it should be a reasonable sinewave.

If the driving Tx signal mismatches the self resonant frequency of the tank, then you get your result. You could try with a signal generator and find the resonant frequency, but I advise you that the circuit would most probably burn due to energy accumulation, if there is not enough resistance in the tank to dump the energy.

Now, even if you find the resonant frequency it will drift when you insert some magnetic material. So, basically you would need an air coil and additionally a ferrite core inductor to act as tank.

simulate this circuit – Schematic created using CircuitLab

I have drawn a series resonant circuit with Lr (ferrite core) >> Lw (work coil) and resonance capacitor.

@Andy provided few points to ponder, study and experiment with. I took few days to study driving mosfet in linear fashion and then applied suggestions of @Andy to my circuit. Added 4.7ohm resister to source and added a 47ohm drain resister before LC tank.

I provided 330kHz sine wave from my signal generator but 22vpp could not be achieved at input due to limitations of signal generator. This generated following source to drain switching wave form. Although its not a perfect sine wave (yellow) yet its very much clean from ringing noise. Blue signal is received at Rx. Following is the signal across tank circuit. The vpp at yellow signal is 1.64V only.

To increase to signal strength i removed 47ohm resistor between supply and tank. This caused a rise in Vpp of transmit signal.

The circuit worked at 47ohm and lower values but ceased to switch at higher resistance values.

The reception of this transmitter was very clean as appears in blue signals. It sensed stainless steel forcep with very small rise in amplitude. This was not happening in my previous attempt.

I used a push pull driving circuit mentioned here. Designing a *linear* MOSFET driver stage But it did not serve my purpose.

I could not find C0G class I ceramic capacitors in local suppliers. I shall manage to source them soon.

The question is If both rx coils cancel each other why am I getting a signal of few millivolts and how to increase sensitivity of system to sense 2mm object of any metal?

I don't know if its appropriate to ask another question here or should I start a new thread.