In modern digital communication, Frequency synthesizers are widely used in wireless LAN, military radar and transceivers. In the military freaquency hopping system, the 16 Kbps data are modulated to 14.5 MHz, up-converted to 30∼88 MHz by frequency synthesizer, and transmitted through the channel. The received data are down-converted to IF band by frequency synthesizer and demoulated. Because transceivers rely heavily on frequency conversion using frequency synthesizer in hopping systems and ,therefore, the spectral purity of the internal oscillators in both the receiver and the transmitter is one of the factors limiting the maximum number of available channels and users. Also, theose are demanded for wide-band frequency scope and excellent frequency resolution. To satisfy this conditions, we must predict the oscillator's internal phase noise and carefully consider when it designed.
The phase noise model proposed in  is widely known as the Leeson model, and is by far the most well-known. It is based on a linear time-invariant (LTI) approach for tuned tank oscillator, but it really has the nonlinear time variant natures. LC-tuned oscillator is using band-pass charateristics to reduce phase noise and has good performance than ring oscillator which has switching effect in power supply. However, it has not been used widely, for not be integrated on synthesizer.
Thus, in this thesis, linear time-variant(LTV) CMOS inverter ring oscillator's model which can be integrated and has good perfomance in phase noise than relaxation oscillator is analyzed.
To predict the phase noise of oscillator very accurately, the oscillator is considered, which has the linearly time-varying nature when the input impulsive current into the oscillator is small. The performance which detect the corrupted signal by oscillator phase noise is compared with only affected by AWGN and analyze how much it degrade system performance for 64 QAM.
In accordance with phase noise level, QPSK system performance, using HP-ADS(Advanced Design System), has been analyzed and compared with the results which only affected by AWGN. Added -85 dBc phase noise at 10 kHz offset frequency into the system degraded the BER about 2 dB in QPSK and 4 dB in 64 QAM.