I'm trying to make a Collpits oscillator. I've made a design that fits my resources and works in a simulation but does not work when assembled. I've checked every element and have no idea what to do but give up. What can I do to possibly make any progress?

the circuit in question

The voltage source is 9 volts, the transistor is BD135-16, the expected frequency is 2.75 MHz (that's what the simulation showed), but I'm not aiming at a specific frequency at this point, just trying to achieve any oscillation. I measured every point of the circuit with an oscilloscope and everything was in a steady state. Since I don't currently have a way to do any more advanced assembly I just soldered the elements directly.

  • \$\begingroup\$ Add a circuit that shows your real oscillator circuit and, name the transistor and power supply voltage. Also what did you use to measure the real circuit? What operating frequency did you expect? \$\endgroup\$
    – Andy aka
    Dec 19, 2023 at 22:24
  • \$\begingroup\$ The base resistors are too low values. So you are putting bjt into deep saturation so bjt is latching( not oscillating). \$\endgroup\$ Dec 19, 2023 at 22:48
  • \$\begingroup\$ What are you building this on, all bets are off if it's a breadboard. The difference is in parasitics. \ \$\endgroup\$
    – Voltage Spike
    Dec 19, 2023 at 22:57
  • \$\begingroup\$ Please clarify your specific problem or provide additional details to highlight exactly what you need. As it's currently written, it's hard to tell exactly what you're asking. \$\endgroup\$
    – Community Bot
    Dec 19, 2023 at 22:57
  • \$\begingroup\$ Sorry, I put in the missing information into the question. \$\endgroup\$ Dec 19, 2023 at 23:05

1 Answer 1


This oscillator might benefit from a few additions, and some component value changes. Also, you might consider adding a load somewhere, since an oscillator usually drives a bit of power out.

  • add a bypass capacitor from top of inductor to GND, with short lead lengths. This ensures that long wire paths back to the battery isn't part of the resonant circuit.
  • add a bypass capacitor from transistor base to GND. This is a common-base circuit where oscillations are fed into emitter, and out collector...base should be at a steady DC voltage.

Bypass capacitor value should be roughly a fraction of an ohm reactance at the oscillation frequency. Too much capacitance can cause "squegging" where oscillation amplitude comes in short spurts. A capacitor of roughly \$ {1}\over{2 \pi fX_c}\$. For 2.75 MHz and \$X_c=0.1\$, works out to about 0.5uf.
0.1uf or 1uf might be acceptable.

Another issue is choosing capacitor-inductor ratio for the resonator. Sometimes you want a power oscillator, other times you want stable frequency that is noise-free. It isn't clear which is required here - the transistor chosen is a power transistor, yet it is DC biased at low power.
For a stable low-power oscillator, you might try to choose an inductor whose reactance is in the 50 - 100 ohm ballpark. This would be about 5 uH for 2.75 MHz, somewhat larger than OP's choice of 0.5uH. For 5uH, resonating capacitance drops from 6.6 nf to 660 pf, so that 2.75 MHz is maintained.

A fraction of available power is taken for feedback, while a larger fraction might be delivered to a load. Feedback is determined by capacitor ratio. Since common-base transistors have a low input impedance, the capacitor from emitter-to-GND is the larger one, while capacitor from emitter-to-collector is smaller. OP might simply try swapping those 20nf, 10nf capacitors.
For the 5uH inductor L1, resonating with 670pf, one might choose capacitor ratios so that collector impedance is higher...so that a load gets a larger share than feedback. I'd guess that a 1:5 ratio might work here. Using standard values, C1 of 820pf with C2 of 3900pf might be OK.
To the 5uH inductor, a dissipative resistor (R1) of 1.8 ohms is added, since no inductor is perfect. This inductor-resistor has an unloaded Q of about 50. A load resistor (Rload) is added - in LTspice this oscillator was able to drive Rload=150 ohms, and oscillated more robustly as Rload moved toward infinity...but with distortion.
Colpitts schematic

DC bias has been modified so that the transistor collector can swing over a large voltage range without swinging lower than the base voltage. Note that bypass capacitor C3 of 0.5uf has no impact on LTspice simulation, since the DC supply V1 is connected to GND and to L1 with perfect inductance-less wires. In real life, such wire connection are often long, and have enough inductance to influence circuit behaviour. A short-leaded C3 mitigates this effect.
For this oscillator, it is especially important to establish a good ground plane at a single point, where C2, C3, C4 all connect together.


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.