Why is my LC oscillator PCB design not working?

Question

I'm building a simple FM transmitter to learn more about LC oscillators. I've successfully designed and built the following transmitter on a bread-board and perf-board using standard through hole components. For the 100 nH, I used this through-hole component.

It works adequately, not great quality, but sound comes through none the less. As an experiment, I designed a PCB using the same schematic.

Crucially, the conductor that I've used for my PCB is the TDK NLV25T-047J-EF. It has a Q value of 20 at 100 MHz and self resonant frequency at 1.2 GHz. I also used this 68 pF capacitor.

I should expect an oscillation around 89 MHz, however, I'm getting nothing near that on my oscilloscope.

Any tips on what I could change? I'm wondering if the traces are causing a lot of extra capacitance. Or perhaps, this is just the wrong type of inductor.

Answer 1

A couple of things:

First, if you haven't already done so verify your board layout, you may have missed something, so go over it carefully and make sure all the connections are correct.

Second, you prototyped it with one value inductor and then used one with half that value in your finished design, meaning that you had to double the capacitance. The difference in the L/C ratio could be part of the problem.

Third, not absolutely necessary, but you have no means of tuning it. Usually there's a variable capacitor so you can adjust it to an unused frequency in the FM band. A good quality ceramic trimmer cap is recommended for this, cheap plastic ones will mostly just frustrate you.

I built some transmitters like this many years ago, I tried surface mount inductors similar to what you're using and for some reason had no luck at all with them. What I ended up using was a hand wound toroidal coil, it was something like 10 turns on a very small core and you have to get the right material mix for the frequency you're using. The advantages were that I could tweak the value by adding or removing a turn, and a toroid is self shielding so the frequency was barely affected by nearby fingers or metal objects the way a solenoid coil is. I think I have an image of it, will post it if I can find it.

Answer 2

I ran an LTspice simulation of your circuit, although I added a 10k resistor R3 to provide a better bias point and a higher output voltage. Frequency is about 85 MHz as expected. The 10 kHz input signal seems to modulate the RF a bit.

Here is the modulation with a 100 kHz input signal, which shows a more visible effect because of the higher voltage through C1: