Improved resonator phase shift in passive and active coupled LC Quadrature Oscillator

A new passive and active coupled CMOS LC Quadrature Oscillator (LC-QO) in the complementary topology is proposed by this paper. The noiseless passive elements are employed to couple two identical LC oscillators. The cores of LC oscillators are coupled via four identical capacitors which are placed in the gates of NMOS cross-coupled switching pairs. Using this technique leads to lower noise and lower dissipation power since the passive coupling techniques are known as noiseless methods for generating quadrature signals. However, the direct injection into the output nodes which leads to the oscillator's cores be synchronized and prevents output signals to be generated, should be avoided and therefore an extra RLC filter placed between the gates and drains of switching pairs. The capacitor provides AC path and the inductor provides DC path for the passive coupling network, they act as a Bandpass filter for the side-band's noise and result in the phase noise improvement. However, when mismatches occurred between LC tanks, phase accuracy is low. Therefore, another coupling technique is employed in order to reduce the circuit's sensitivity and removes the phase error due to mismatches. So, with the additional active coupling network, phase accuracy of the proposed LC-QO improved further compared with other passive coupled QOs. The coupling performance also performed not at zero-crossing points (the most vulnerable instants), so doesn't degrade phase noise performance. To evaluate the validity of the improved-phase noise proposed QO with the reduced phase sensitivity, it is simulated at 3.038 GHz central frequency in the practical 0.18 TSMC CMOS technology. The LC-QO consumes 5.988 mW from a 1.8V power supply, phase noise is -123.264 (dBc/Hz) @1 MHz and -132.9 (dBc/Hz) @3 MHz offset frequencies with the quality factor (Q=12) and is desirable. The Figure of Merit (FOM) of - 185.14 (dBc/Hz) achieved. The obtained simulation results confirm the reliability of the proposed LC-QO.