Active integrated oscillator antenna

We had a school project to build an active integrated oscillator antenna at 3GHz. We used the AT42085 as the active device and patch antenna as load.

The result from the spectrum analyzer showed it oscillated close to the design frequency about 7% off, which is fine due to inaccurate transistor model.

The question that I've been wondering is that when looking at the spectrum at a large frequency scale, it looks clean. When I zoom in close to the first harmonic, then I see a lot of small spikes forming the envelop of the first harmonic (kind of like a modulated signal).

I wonder whether this is because the design is not very stable due to poor antenna Q (patch antenna has Q in the range of hundreds). What can be done to make it more stable ?

Any feedback/comments are welcome. Thanks in advance !

If I understand what you are saying, you are using an antenna element as part of the oscillator circuit. Since the antenna is radiating, it is guaranteed to be a pretty low Q circuit. If the antenna mostly forms the frequency determining part of the oscillator, then that is your problem--a low Q resonator. The oscillator frequency is jumping around slightly.

Other things could be going on, though. Oscillators are VERY susceptable to noise pickup induced frequency modulation. Try replacing your power supply, and its long leads, with a battery with short leads. Did the "modulation" go away? If so, it was just AC hum pickup.

There are other practical considerations. Do you have adequate voltage bus capacitive filtering? a 4.7 uF tantalum capacitor in parallel with a 0.1 uF ceramic is a good idea.

Finally, sometimes you have faulty components. I have seen microcracks in chip resistors that were causing small frequency jumps. Sometimes loose mechanical connections can cause vibration induced frequency modulation. Sometimes cooling fans can cause some modulation.

Good luck

Thank you for your reply.

Yes, the antenna element is using as the frequency selective element as well, it is a patch antenna.

Reducing the power supply helps, but I still observe small "frequency jumping". I did have a 0.1uF ceramic cap for the power supply, but I don't have the 4.7uF cap.

Btw, I was able to get it oscillate very stably at a much lower frequency (i.e. 2GHz) by replacing/cutting some microstrip line at the base of the transistor.

It is a common emitter with inductive load at base and load (antenna) at collector topology.

Well, try all that stuff. I have built VCO's where I had 100 uF of capacitance on the power supply line. But DO use more than one size of capacitor, so you get a broad frequency range of noise reduction.

The oscillator fundamentally oscillates where its resonator reactance cancels the active device reactance. If you have a high reactance slope in the resonator (high Q), and a high reverse slope in the device, you get a very stable oscillation. If there is little slope, it is just going to jump around! Blame it on thermdynamics!