Effect of Coupling Capacitors/Inductors

[dalarev]

More specifically, I have an NMOS oscillator, where the output is taken from the source terminal.

I have to couple the output (50 Ohm load) with a coupling capacitor, but is this going to affect my calculations of the resonant frequency of the circuit?

In my case, I have a Clapp oscillator (Capacitive feedback divider). When the load is connected, however, the Coupling cap + R_load appear, in effect, in parallel with my feedback capacitor. How should I implement this in my calculations?

 

[berkeman]

More specifically, I have an NMOS oscillator, where the output is taken from the source terminal.

I have to couple the output (50 Ohm load) with a coupling capacitor, but is this going to affect my calculations of the resonant frequency of the circuit?

In my case, I have a Clapp oscillator (Capacitive feedback divider). When the load is connected, however, the Coupling cap + R_load appear, in effect, in parallel with my feedback capacitor. How should I implement this in my calculations?

I wouldn't think you'd want a 50 Ohm load associated with the oscillator section at all. I think more traditionally you will use a buffer stage between the oscillator output and your load.

 

[waht]

What is the operating frequency?

I have to couple the output (50 Ohm load) with a coupling capacitor, but is this going to affect my calculations of the resonant frequency of the circuit?

generally yes, but in any case, you can always drive an emitter follower buffer stage which has low output impedance.

Although, there are some microwave oscillators which are specifically designed around 50 ohm transmission lines, and load.

 

[dalarev]

I wouldn't think you'd want a 50 Ohm load associated with the oscillator section at all. I think more traditionally you will use a buffer stage between the oscillator output and your load.

50 Ohm load is the equivalent circuit of a "port" in the simulator I'm using; also, this port must be used for power measurements.

I have seen several schematics with a secondary buffer stage for that reason, perhaps I can implement that later.

What is the operating frequency?

Although, there are some microwave oscillators which are specifically designed around 50 ohm transmission lines, and load.

2.4 GHz. No transmission lines in this oscillator, at least not in the designing stage. I've got it to oscillate, but the sine wave I get in output looks like a perfect single-tone sinusoidal. Not only does that immediately raise suspicion, but the spectrum doesn't back that up at all.

I've attached an image of my schematic, my output waveform, and the spectrum. Thanks for the help.

 

[waht]

2.4 GHz. No transmission lines in this oscillator, at least not in the designing stage. I've got it to oscillate, but the sine wave I get in output looks like a perfect single-tone sinusoidal. Not only does that immediately raise suspicion, but the spectrum doesn't back that up at all.

The second harmonic is about 30 db below the fundamental. You wouldn't see much harmonic distortion on the scope. That's why it looks like perfect sine wave.

If you need to match the output to a 50 ohm load, consider designing an L, Pi, or a T match as a last resort. This type oscillator design definitely needs a buffer as any load impedance will affect the frequency, and power output.

 

[dalarev]

The second harmonic is about 30 db below the fundamental. You wouldn't see much harmonic distortion on the scope. That's why it looks like perfect sine wave.

If you need to match the output to a 50 ohm load, consider designing an L, Pi, or a T match as a last resort. This type oscillator design definitely needs a buffer as any load impedance will affect the frequency, and power output.

Good point; that would rid my need for the RF choke at the source at the very least.

I've been trying to break this oscillator up into functional parts. I'd like to isolate the negative resistance generator portion of the circuit, which I'm assuming is everything except the series L+C3, and plot port parameters to confirm negative resistance behavior.

Any ideas how I could analyze, on paper, the negative resistance portion?

 

[waht]

I've been trying to break this oscillator up into functional parts. I'd like to isolate the negative resistance generator portion of the circuit, which I'm assuming is everything except the series L+C3, and plot port parameters to confirm negative resistance behavior.

Any ideas how I could analyze, on paper, the negative resistance portion?

Once you break up the circuit into its small-signal equivalent, it's very easier to identify the feedback path, verify Barkhausen's criterion for oscillation, calculate resonant frequency, calculate how load will affect frequency, or write it in negative resistance form.

Negative resistance can be modeled with an RLC circuit hooked to a non-linear dependent voltage or current source (nmos in this case). Write out the differential equation and identity what part needs to have negative resistance in order to oscillate. The simplest case that works is a coefficient of linear term of the non-linear dependent voltage or current source (nmos in this case). There is a whole derivation of this in Ludwig/Bretchko.

Once you worked out the negative resistance model, it becomes much easier to compare it with the circuit that you have. Although negative resistance model is rarely used in oscillator design. Barkhausen's criterion is used in practice. But still it's a great exercise.