I am trying to design an oscillator circuit which can produce a good sine wave with a frequency of at least 500 MHz and oscillating voltage of around 10 - 15 volts. The output of this oscillator will be just connected to a two conducting plates of one square inch (one to collector output and the other plate to ground, with 1 inch gap between the plates.) The resistive loading is therefore just the tuned circuit losses.

I tried learning to design my own oscillator but ended up getting confused reading the data sheet parameters, especially hfe (current gain) of various BJTs.

I have been told that in order to get 500MHz I need a transistor with a transition frequency of 5GHz. I thought that a gain above unity above circuit losses would be adequate for sustaining oscillations.

The things I do not understand are as follows:

  1. How do I set the operating point for the amplifier for a Colpitts oscillator to get the desired oscillating voltage?
  2. What is the voltage gain of the amplifier (voltage divider bias used for biasing of transistor,) thus I am stuck in calculating the feedback fraction for the amplifier to sustain the oscillations. I understand that the ratio of capacitors of the tank circuit is the feedback percentage, but without knowing how to calculate the voltage gain of amplifier I am unable to proceed further.
  • 1
    \$\begingroup\$ There are plenty of internet resources for these questions. What prior research have you done and what sites have you been to? \$\endgroup\$
    – Andy aka
    Jul 15, 2017 at 10:27

2 Answers 2


The transistor is used as a transconductor: voltage input, with current output. In which case, the GM (transconductance) computed at Ie / 0.026volts, give the value. At 1mA, GM is 1/39; at 10m1, GM is 1/3.9; at 26mA, GM is 1/1 (1 amp per volt).

Your bipolar's task is to work with AC input voltage on the emitter, and output a collector current into the tank.

Its also your task to fully recognize the entire circulating path of the RF energy; some Colpitts use the VDD as part of the circulating path, and forget about the bypass capacitor; this is not good; huge losses.


simulate this circuit – Schematic created using CircuitLab

What POWER are you asking this oscillator to provide?

Power = F * C * Vrms^2 = 500e6 * 10pF (minimum) * 15*0.707 * 15*0.707

Power = 0.5e+9 * 1e-11 * 225 * (0.707*0.707) = 0.5e-2 * 225 = 1.13/2 watts.

Power = 0.56 watts.

  • \$\begingroup\$ No way, that's not the power you asking this oscillator to provide that's the power which would be dissipated by a resitive driver (e.g. a MOS inverter) while feeding that ideal capacitance; all in all a totally different circuit . Power provided to that actual load depends only on its losses and radiation and cannot for worked out unless you account for them. \$\endgroup\$
    – carloc
    Dec 17, 2017 at 10:37
  • \$\begingroup\$ Should this works at 500 MHz? \$\endgroup\$
    – Antonio51
    Sep 21, 2022 at 18:07
  • \$\begingroup\$ @analogsystemsrf could you please comment on any impedance matching considerations, if any. Thank you.. \$\endgroup\$
    – ee_student
    Sep 22, 2022 at 0:30

I have been told that in order to get 500MHz I need a transistor with a transition frequency of 5GHz.

Not really.

To do ... First "evaluating" your measurement system (capacitance, inductance, and losses.
For some oscillators, you just "need" FT enough high (Common base). EE&O.

example of simulation :

enter image description here

Check specifications of 2N3904. FT is just 300 MHz.
Oscillation is ~ 295 MHz. Perhaps higher frequency can be checked ...
However, you have ~ 80 V peak. If lower voltage needed, just use an step-down "air" transformer.

Check with a 2N3866. FT=500 MHz.

enter image description here


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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