[Simetrix] Colpitts Oscillator

[guitarnoob123]

Hello,
This particular circuit provides a good periodic waveform when I implemented it in the lab

However, I can't seem to make a transient analysis in simetrix. I wanted to add a few components and simulate before I test it again.

Any ideas? Is it because I didn't use any resistors for biasing? Take note that the BJT was biased properly in the lab (it works!).

 

[FvM]

This particular circuit provides a good periodic waveform when I implemented it in the lab

Really? The transistor is just driven into perfect saturation without any chance to oscillate.

Connecting R2 to the collector instead of positive supply makes the circuit oscillate, but heavily overdriven. Some corrections towards a regular colpitts oscillator seem badly necessary.

 

[WimRFP]

I agree with FvM.

In your circuit there is no intentional feedback. Connecting R2 to the collector avoids saturation and provides at least positive feedback. I would expect larger C1, smaller C2 and larger R2 to get more output voltage swing. Is this just to experiment, or will you drive some other circuit?

 

[guitarnoob123]

Oh sorry I sent a wrong snapshot. Resistor R2 was indeed connected to the collector of the BJT when I implemented it.

I would expect larger C1, smaller C2 and larger R2 to get more output voltage swing. Is this just to experiment, or will you drive some other circuit?

I will try this next time I work. No particular circuit to drive. We are asked to filter the oscillator output to approximate a sine wave then do the measurements through a 50 ohms load resistor.

 

[WimRFP]

Where do you put the filter and the load?

The oscillator may experience problems when driving a filter and or a 50 Ohms load. When it remains oscillating, you may expect some 200mV output voltage.

Weird things may/will happen when you directly connect an LC filter to the collector (don't forget DC decoupling capacitor). Virtually all filters show wild impedance behavior outside the pass band. This will affect the phase shift and gain of the BJT amplifier. The oscillator may oscillate at another frequency and/or with unstabie amplitude.

I would first try to understand how the circuit operates (so you can make good educated guesses) and then play with the design so that you have a reasonable sinusoidal wave shape. It will not be easy as the loaded Q factor for the LC circuit is relatively low. When done, use a second stage (buffer) that provides some isolation between the oscillator and the load+filter.

As a buffer stage, you may use a common collector stage (emitter follower) where you put a 50 Ohm resistor between the emitter and the actual output. Both the input and the output of the filter (that you have to design) "sees" about 50 Ohms. You may know that if you have a single transistor emitter follower running at 10 mA in class A, you can only get about 400mVp into 50 Ohms.

If you need several volts of output into 50 Ohms. you may use a complementary class B output stage (with PNP and NPN transistor).

 

[guitarnoob123]

That's very helpful.
Now I want to vary the oscillation frequency a little bit by varying the capacitance of C2 using a varactor. This will look like a voltage controlled oscillator since I can change the capacitance of the varactor using an external voltage source.

This was my circuit so far: **broken link removed**

I succeeded a while ago (I obtained plus-minus 300kHz frequency deviation from a 12 Volts plus-minus 1 varactor voltage) and now the circuit doesn't work anymore. I am wondering if I just got lucky. Any ideas with what is wrong in my setup?

 

[godfreyl]

V2 is short circuiting D1 at AC. They should be connected in series, not parallel. Preferably with one end of V2 connected to ground, so you don't need a floating voltage source.

 

[guitarnoob123]

Is this what you mean?
**broken link removed**

 

[godfreyl]

Almost, I was thinking more like this: