LC oscillator - simple but temperature resistive

Hy,

I have to design LC oscilator with oscillating frequency around 1Mhz. I need oscillator for detecting small movements of object. This movements I must detect with changing position of ferrite core in coil. I wound coil by myself (15uH, 3ohm). Main advantages is that LC circuit must be simple as possible and temperature resistive (as much as possible). Temperature range must be up to 150 degrees celsius.

Because I have never work with oscillators before I will be gratefull also for any extra advice how to build such oscillator.

Thank you very much!

The major source of temperature effects in such a low frequency oscillator is the L and C elements that control the frequency changing with temperature. The usual way to minimize this is to use capacitors with negative temperature coefficients. They are designated by Nxxx where xxx is the ppm/degree change in value. The N means negative. An example would be N750. The NP0 type means no change either direction with temperature.

You may have to do some experimenting to find the value that works best.

Another method would be to have a second identical oscillator. It would track the temperature effect and you could subtract the frequencies of both oscillators to get the core movement effect.

Generally, this is done with two identical windings on the same coil former, plus one excitation coil: a transformer essentially, with one primary running the entire length of the bobbin and two secondaries, but placed each on one half of the bobbin.
The core moves inside the coil former and so both inductors change at the same time, but one increases in inductance, the other decreases. By measuring differentially the output voltage, the frequency of the generator is no longer so important, so you can have a less stable oscillator.

Among the greatest advantages are the large displacement range, sensibility, reproducibility and insensitivity to radial displacements of the core.

Take a look at this http://nees.buffalo.edu/pdfs/lvdt.pdf#search='lvdt%20principle'

www.rdpelectronics.com/displacement/lvdt/lvdt-principles.htm

Your L in the circuit will have a positive temp coefficient. You can match it with the proper N type capacitor that will result in very minimal drift with temperature. Another way to combat drift is to make up the C part with a few caps in parallel so the heating effect is spread over a larger plate area.

Thank you very much for your help.