VCO Oscillation Frequency

[umberabbas]

How to make the output swing of an LC VCO a smooth sinusoidal waveform?
Any reference circuit?
Also the VCO is not achieving it's intended frequency.

 

[BradtheRad]

LC and VCO in one oscillator? Is the purpose so that your vco creates a sweeping frequency, derived from the momentary voltage from an LC pair (inductor & capacitor)?

The LC pair generally has just one resonant frequency, and its natural waveform is a sinewave. There are several oscillator circuits which can be built around LC combinations. Usually careful adjustment of values is necessary, to get a clean smooth sinewave.

 

[umberabbas]

LC and VCO in one oscillator? Is the purpose so that your vco creates a sweeping frequency, derived from the momentary voltage from an LC pair (inductor & capacitor)?

Yes I'm designing a cross-coupled VCO at 60 GHz and the purpose is to achieve this frequency through voltage provided to LC. The thing is I'm achieving around 54 GHz and the output swing is skewed at some point. It is not a smooth sine-wave.

The LC pair generally has just one resonant frequency, and its natural waveform is a sinewave. There are several oscillator circuits which can be built around LC combinations. Usually careful adjustment of values is necessary, to get a clean smooth sinewave.

The circuit is a cross-coupled VCO topology.

 

[BradtheRad]

Is a skewed sinewave the same thing as a sine that 'leans' to the left or right?
The L:C ratio has something to do with the shape of a sine. Also the resistance and Ampere level in the network.

When you use a low L value, it causes the waveform to rise quickly. The L/R time constant is important here.
When you have a high L value, it slows the rise of the waveform.

Then to obtain your desired frequency you can calculate a capacitor value that is compatible with your given L value. However the capacitor plays its part in shaping the waveform too. The RC time constant is a guiding factor.

To create symmetry of the sine shape, you want the rise to be the same rate as the drop. So the inductive time constant ought to match the capacitive time constant. (Or at least it stands to reason.) These are not necessarily easy to calculate, since the resistance may be contributed by semiconductors, and their conductivity can change at different times in the cycle. Or a change in supply voltage can change the shape of the sine.

 

[umberabbas]

This the output I'm getting at 54 GHz. The value for L is 50pH and for that of the capacitor I'm not using any external ones so far but the capacitance should be 141fF as calculated from the equation.

 

[BradtheRad]

Cross-coupled implies 2 inductors. Without a capacitor the inductors turn on and off via the cross-coupled transistors. The output is something other than a sinewave.

By including a capacitor you form an LCL tank loop which usually creates sine waves. It's easy to change the frequency by changing the capacitor. Furthermore it is easier to get oscillations going.

Your waveforms look like normal sines. Probably your devices contain sufficient capacitance to create the LCL tank behavior by some means.

With value 50pH my simulation is never faster than 50 GHz. Even if I make the capacitor a fraction of a pF, the frequency does not reach 60 GHz. Therefore if you wish to make oscillations faster, try reducing inductor value. I don't know how feasible it is to achieve that fast a frequency in real hardware.