[SOLVED] Colpitts oscillator problem

Any idea why the attached Colpitts oscillator won't start ? I built the oscillator circuit and I get a flat signal with an offset close to the input voltage (5V).

thanks

You are looking for 750 kHz?

Your LC ratio is a bit extreme. Q of a 4 uH coil at 750 kHz is going to be fairly low.

I would increase inductance by factor 5 to 18 to 22 uH. Drop C2 & C3 value by corresponding amount. C1 you can turn into a DC block.

You are running less then 1 mA of collector current. You may want to drop emitter resistor by half to increase current.

thanks RcinFLA,

the coil size I is fixed and can only vary between 4 and 6uH . Now Fc on the hand I can be flexible
let say i change Fc to 200Khz and I calculate new C2 and C3 ( I will make them equal) and remove C1
will the oscillator start ?
What if I want a higher frequency such as 1.5MHz keeping in mind the coil stays in that range Can I recalculate C2 and C3 and it will work ?
thanks

I guess, you are talking about a real circuit rather than a simulation? Unfortunately, the point is often unclear in edaboard posts. Most likely, the coil's Q is too low. Increasing the transistor current (rather by a factor of 10 than 2, I think) allows oscillation also at lower Q values.

Generally, operating an oscillator at "extreme" characteristic impedance is possible, but involves more design action than just dropping some arbitrary component vlaues.

elisemark said:

thanks RcinFLA,

the coil size I is fixed and can only vary between 4 and 6uH . Now Fc on the hand I can be flexible
let say i change Fc to 200Khz and I calculate new C2 and C3 ( I will make them equal) and remove C1
will the oscillator start ?
What if I want a higher frequency such as 1.5MHz keeping in mind the coil stays in that range Can I recalculate C2 and C3 and it will work ?
thanks

Click to expand...

Going down in frequency will make matters worse. The reactance of the coil and associated capacitors are so low at 750 kHz it will have low Q. Go for higher frequency, 1.5 MHz, adjust the coil up to 5 uH and change C2 & C3 to 4.7 nF. C1 as DC block 100 nF.

If you want to drive a CMOS buffer directly, drop emitter resistor to 330 ohms and feed CMOS buffer from emitter. At emitter, you should see about a 3 volt pk-pk sinewave centered at about 2.5 volts from ground. You may want to put two buffers in cascade to square up output.
.

---------- Post added at 08:52 ---------- Previous post was at 08:39 ----------

FvM said:

I guess, you are talking about a real circuit rather than a simulation? .

Click to expand...

Difference between real circuit and simulation is the result of bad simulation.

Difference between real circuit and simulation is the result of bad simulation.

Click to expand...

Taking this statement strictly, you say that any simulation is bad, because it involves some differences to a real circuit, so I fear, it won't get you anywhere.

In a practical view, if someone doesn't manage to operate a real oscillator circuit, it's unlikely that he's able to simulate it correctly, e.g. introduce the necessary parasitic component properties. But simulation also has some particular problems caused by the simulator behavior. To be able to understand the reported problems, I simply prefer to exclude, that it's possibly a pure simulation issue.

FvM said:

Taking this statement strictly, you say that any simulation is bad, because it involves some differences to a real circuit, so I fear, it won't get you anywhere.

In a practical view, if someone doesn't manage to operate a real oscillator circuit, it's unlikely that he's able to simulate it correctly, e.g. introduce the necessary parasitic component properties. But simulation also has some particular problems caused by the simulator behavior. To be able to understand the reported problems, I simply prefer to exclude, that it's possibly a pure simulation issue.

Click to expand...

Do you have anything of value to add to this discussion?

Thanks all for your great feedback .