0
\$\begingroup\$

I'm having trouble getting my colpitt's oscillator to give me nice sinusoids. It seems to be clipping the output for for reasons I don't understand. The circuits needs to oscillate at 7MHz (Body capacitance does zilch at 1.5MHz). So here's my process. I choose \$I_c = 150 \mathrm{mA}\$ since that's what they tested the transistor at according to the datasheet. Also i want to bias \$V_{c}\$ in the middle of the supply voltage. With this in mind here is my process:

  1. \$R_c = \frac{12-6}{150\mathrm{mA}} = 40 \Omega\$
  2. Equally split the 6V across the transistor and emitter resistor
    • \$R_e = \frac{3V}{150\mathrm{mA}}\ = 20\Omega\$
  3. If \$V_e = 3\mathrm{V}\$, \$Vb = 3.7\mathrm{V}\$
  4. Assuming \$h_{fe} = 100, I_b\$ has to be \$1.5\mathrm{mA}\$
  5. If 10 times \$I_b\$ flows through the voltage divider, \$R_1\$ and \$R_2\$ can be calculated to be \$533.4\Omega\$ and \$246.7\Omega\$ respectively.
  6. Oscillation frequency is \$f = \frac{1}{2\times \pi\times \sqrt{800\times 10^{-9}\times 0.6\times 10^{-9}}} = 7.2\mathrm{MHz}\$

LTSpice says the frequency is \$5.47MHz\$ and it seems to be quite distorted and clipping at the bottom. Any help in fixing this or understanding this would be awesome.

Thanks!!

schematic

simulate this circuit – Schematic created using CircuitLab

LTSpice Simulation

EDIT : As you guys suggested, I biased the transistor at 10mA instead and added a schottky diode between the forward and feedback path. Someone pointed out to me that the higher the current, the lower the BW so that might have been messing it up too. Everything works now. Thanks!!

\$\endgroup\$
  • \$\begingroup\$ 150mA is a huge bias current for an oscillator. Typically, you want the total loop gain to be just greater than 1 at the frequency of oscillator for a good oscillator. Otherwise, you are going to get the nonlinear output that you are seeing. Remember that the total gain is proportional to the transconductance of the BJT, which in turn is proportional to the bias current. Also, a large bias current means the transistor (and resistors) will be dissipating a lot of power. Specifically, the BJT will be dissipating 450mW -- it's gonna get pretty hot. \$\endgroup\$ – LetterSized Mar 8 at 7:35
  • \$\begingroup\$ As an example, here is a Colpitts oscillator I made a while ago to oscillator at 27MHz: i.imgur.com/tlS90YO.png \$\endgroup\$ – LetterSized Mar 8 at 7:45
  • 1
    \$\begingroup\$ Multiply all your resistor values by ten and report back with hopefully better news. \$\endgroup\$ – Andy aka Mar 8 at 9:43
  • \$\begingroup\$ You are missing a bunch of design specs like sensitivity pF*MHz product or some max deviation or what you mean by THEREMIN. You are wrong about body capacitance vs freq. It works all the way up in the GHz range due to polar moisture content . Theramins were Unijunction audio oscillators of the 70's \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 Mar 8 at 15:22
  • \$\begingroup\$ @TonyStewartSunnyskyguyEE75 My bad on the wrong phrasing. I know it'll work on any higher frequency. I just choose to do it at 7MHz. That's the frequency that the NorCal 40A transmit/receives at and I've noticed frequency change while i was tuning the air capacitor in it. So i know it works well at 7MHz. I don't need to be super precise either. As long as there is an audible change with hand waving, i'll be happy \$\endgroup\$ – Pratik Kunkolienkar Mar 9 at 17:01
0
\$\begingroup\$

For lower distortion, you need the PI network to provide some filtering.

I think that requires the PI network be an effective low-pass network.

I think that requires the amplifier Rout be much higher than the PI capacitive reactance.

At 7MHz, the 1.2nanoFarad is

(1uF at 1MHz is -j0.16 ohms; 1nanoFarad is 160 ohms; at 7MHz, Xc is j20 ohms)

I agree with Andy ----make the resistor values at least 10X higher.

| improve this answer | |
\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.