Let's improve this oscillator.

[Teleno]

As it is, it delivers high voltage and current to the coil, plus the amplitude has excellent temperature stability.

This circuit contains no frequency-dependent components other than the resonant LC tank. The driving transistors are switched at the exact zero-crossing of the tank's current in perfect quadrature with the tank's voltage regardless of the resonant frequency. Just replace L and C and you get a perfect sine wave signal at any frequency. The ratio L/C must be kept between 15.000 and 120.000 (H/F).

Q9 senses the current through lower driving transistor Q6, R9 converts it to an amplified voltage and this is used to turn the upper driving transistor off. The same is applicable to Q7, Q8 and R12. The Schottky diodes prevent saturation that would delay the switching relative to the zero-crossing. If they're omited the quadrature is less than perfect but still very good and you get a 5% more power.

I share this circuit idea so that you guys can make it better, for example: making less dependent on resistor values and transistor tolerances by using op-amps, comparators etc. where appropriate.

[Skippy]

Resistors R10, R11 are different values? Intentionally, because of the differences between the npn and the pnp parts of the circuit?

Perhaps put Schottky diode clamps on Q1,Q2, like on the main switching transistors? You may need some resistor in series with the base of Q1,Q2

[Teleno]

Resistors R10, R11 are different values? Intentionally, because of the differences between the npn and the pnp parts of the circuit?

Yes, the difference is to compensate for the slightly higher base voltages in PNPs. The idea is to get the same voltage drop across R12 and R9. Also, R9 should be about 5%-10% higher than R11 for the same reason in regard to Q1 and Q2.

[green]

I share this circuit idea so that you guys can make it better, for example: making less dependent on resistor values and transistor tolerances by using op-amps, comparators etc. where appropriate. reply#1

.

Won't say it's any better but after a lot of tries I finally got something to oscillate with an op amp. Change R1 to adjust current.

[Teleno]

I share this circuit idea so that you guys can make it better, for example: making less dependent on resistor values and transistor tolerances by using op-amps, comparators etc. where appropriate. reply#1

.

Won't say it's any better but after a lot of tries I finally got something to oscillate with an op amp. Change R1 to adjust current.

That's great stuff, green. I love it. It's not very temperature stable but crossover is just perfect.

After adding the lump resistive elements (RL and Resr n my circuit) the output is still more powerful than my transistor circuit.

I've also removed one of the two voltage dividers because a slight mismatch between the two will prevent oscillation. Here are my modifications:



I'll checking the day price for an LTC6244 chip, cross my fingers.

[Davor]

Now it looks astonishingly like a mikebg's regenerative detector.

[Teleno]

The point is to provide the phase shift necessary for oscillation without the need of any passive components (which are dependent on frequency) and without a resistor in series with the tank (which increases losses).