1
\$\begingroup\$

enter image description here

Here is a schematic of a simple FM transmitter. I understand that this work in the following way: the sound picked up by the mic is translated into a voltage at the base of the Q1 transistor. As the internal base-collector capacitance of a BJT depends on the base-collector voltage, the B-C capacitance changes with the sound signal. This internal capacitance forms a resonant circuit with C1, C3 and L1. The resonant frequency of this circuit changes as B-C capacitance changes, and so the signal is FM modulated with the sound signal.

But how do I work out the frequency of the FM signal by reading this schematic? Let's say there is no sound signal. I set VC1 to a known capacitance. I know that C1 is large enough so it offers no impedance for the FM signal and therefore the base is effectively grounded for the FM signal. C3 is even bigger. So therefore we have, in effect, L1 directly parallel with VC1 and B-C capacitance. To get the total capacitance in parallel with L1, I need to add VC1 and B-C capacitance.

But how do I know the capacitance of the B-C junction? Let's say I want to know the base frequency (no sound signal), how much capacitance is there? I don't see C-B capacitance as a function of C-B voltage on datasheets.

\$\endgroup\$
  • 1
    \$\begingroup\$ As far as I know, those things are never calculated. You get "close enough" by ignoring the modulation. Those things drift all over the place. You must always tune them manually (both VC1 and L1) - and you have to retune them during operation. So, really, there's no point in trying to precisely calculate the operating frequency. \$\endgroup\$ – JRE Aug 15 at 10:46
  • \$\begingroup\$ Which things are drifting all over the place and why? \$\endgroup\$ – S. Rotos Aug 15 at 10:58
  • \$\begingroup\$ The transmit frequency drifts. The transistor oscillates differently with temperature. The capacitance varies with temperature. Just moving your hand near the circuit changes capacitances and inductances. Metallic objects near by will change the inductance. The coils are usually hand wound, with whatever wire is at hand so the inductor has the wrong value. \$\endgroup\$ – JRE Aug 15 at 11:06
  • \$\begingroup\$ "Those things" in this case refers to FM transmitter circuits like the one you found. There are many variations on the "single transistor FM transmitter" out there. They mostly work, but are very finicky in operation. \$\endgroup\$ – JRE Aug 15 at 11:09
  • \$\begingroup\$ Alright, thanks for the info! I was wondering, is there any way to diminish the effect the effect of surroundings on the circuit, such as your hand and metallic objects nearby? Perhaps some kind of shielding? I would like to be at least send short messages (so that the effect of components heating won't cause too much drift) on a consistent frequency that I have once found using my receiver. \$\endgroup\$ – S. Rotos Aug 16 at 10:21
2
\$\begingroup\$

In this circuit, collector-to-base capacitance would be close to \$C_{ob} \$. The base is almost grounded by C1 as far as radio frequency is concerned.
\$ C_{ob} \$ is typically 2.5pf, maximum of 4pf when the DC collector-to-base voltage is 5V.(From a National Semiconductor data sheet). This measurement was done at 1MHz, not at 100MHz where this circuit is usually meant to operate. From the 2N3904 data sheet by ON semiconductor, \$C_{obo} \$ corresponds to collector-to-base capacitance:
ONsemi 2N3904 capacitances
This graph will not be useful if you wish to include \$ C_{ob} \$ in calculation of oscillation frequency, since \$C_{ob} \$ has a relatively minor affect on frequency, compared to other capacitances. Not shown on the schematic are stray capacitances and inductances, and the reactive component added by the antenna. Besides strays, the inductor's self-resonant frequency is likely unknown. Tolerance of small-value inductor and capacitors is often poor, and their values are modified by inductance of connecting wires.
At 100MHz, the 2N3904 hasn't much gain, and is likely running close to its gain*bandwidth limit: a carefully-done internal model would be required.
As others have said: you build it and verify that it oscillates by searching for its frequency with measurement tools. Then you vary component values in an attempt to move frequency up or down. (VC1 variable capacitor is meant to modify frequency, but 100pf seems somewhat too large). This is by no means a stable and predictable oscillator at 100 MHz.

\$\endgroup\$
  • \$\begingroup\$ "verify that it oscillates by searching for its frequency with measurement tools." AKA: Tune your FM radio receiver up and down the band in hopes of finding your transmitter. \$\endgroup\$ – JRE Aug 15 at 15:27

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.