AWR eBooks

TECHNICAL MEMORANDUM Page 8 of 12 Figure 12: Dynamic Load Line The dynamic load line of the Push-Push oscillator and the single common collector oscillator discloses profound non- linear operation of the Push-Push oscillator. Operation at high collector current and low collector voltage has been correlated with poor phase noise. Figure 13: Base Voltage Waveforms The voltage waveform at the base of each of the push-push stages discloses the out-of-phase signals – as expected. Discussion of Simulation Results "… It is typical of modern physicists that they will erect skyscrapers of theory upon the slender foundations of outrageously simplified models …" – J. M. Ziman Like the modern physicist, the engineer should be cautious in the use of overly simplified models. Prediction of Oscillator Center Frequency The simulation utilized a published Spice model of the AT420 bipolar transistor; attempts to validate the model from the manufacturer were not successful. In fact, S-parameter data did not correlate well between the manufacturer's published data and S-parameters simulated with the Spice model. It is this lack of correlation that is believed to be responsible for the poor correlation between the measured (10 GHz) and simulated center frequency (9.2 GHz). Prediction of Oscillator Phase Noise The measured oscillator phase noise was -45, -73 and -100 dBc/Hz at the respective offset frequencies of 1, 10 and 100 KHz. The simulated phase noise data is: - 61, -84 and -104 dBc/Hz at the respective offset frequencies. Phase noise of a free-running oscillator near to carrier is closely related to the Spice model noise parameters. Unfortunately, Spice noise parameters are generally not published and when published, disclaimers sometimes follow with respect to production lot variations and measurement methods. The Spice noise parameters – AF, KF and FFE – for the AT420 transistor are not available from the manufacturer and therefore, the author relied on measured data from Gris [3.]. The deviation from measured and simulated phase noise was not unexpected. Measurements of a second push-push oscillator disclosed better correlation as the graphic of Figure 14 illustrates. Figure 14: Phase noise of prototype Push-Push oscillator Prediction of Oscillator Output Power The measured oscillator output power included a saturated buffer amplifier and therefore, correlation with simulated output power could not be immediately determined. Prediction of Oscillator Voltage Tuning The oscillator frequency versus voltage tuning was 250 MHz (simulated) compared to 575 MHz (measured) for the push-push prototype – clearly, 0 1 2 3 4 5 6 7 8 9 10 11 12 Voltage (V) AT42000 IV Curve -40 0 40 80 120 p8 p7 p6 p5 p4 p3 p2 p1 IVCurve() (mA) IV Curve IVDLL(S1\S1\GBJT.QINT@ce,S1\S1\GBJT.QINT@ce)[1,4] (mA) CC PP Oscillator A IVDLL(S1\S1\GBJT.QINT@ce,S1\S1\GBJT.QINT@ce)[1] (mA) CC Oscillator A p1: FREQ = 1 GHz p2: FREQ = 1 GHz Vdc = 4 V p3: Istep = 0.5 mA p4: Istep = 0.4 mA p5: Istep = 0.3 mA p6: Istep = 0.2 mA p7: Istep = 0.1 mA p8: Istep = 0 mA 0 0.1 0.2 0.3 0.4 0.434 Time (ns) Voltage Waveforms -10 -5 0 5 10 15 p2 p1 Vtime(V_PROBE.VP1)[1,4] (V) CC PP Oscillator A Vtime(V_PROBE.VP2)[1,4] (V) CC PP Oscillator A p1: FREQ = 1 GHz Vdc = 4 V p2: FREQ = 1 GHz Vdc = 4 V X-Band Push-Push Oscillator Simulation and Measurement 8 www.cadence.com/go/awr

• Cover