Parent Category: 2014 HFE
By Jiang Liu, Hugo Morales, and Larry Dunleavy
Overview
This application note documents results from the characterization and modeling project of a broad-band 0.5 to 18 GHz tuner/receiver system intended for defense applications. Measurements such as S-parameter and power/frequency swept conversion gain were performed during the project. Equipment used in acquiring measured data include RF source, power meter, spectrum analyzer and DC sources. All measured data is properly calibrated at each frequency and de-embedded to the coaxial RF test port of the receiver system. The receiver system model was developed using the measurement-based behavioral modeling approach for usage in Agilent Technologies’ Advanced Design System (ADS) software. This note highlights the good agreements achieved between the measured and simulated conversion gain as a function of both input RF power and frequency. This leads to good TOI performance as validated in given result.
Figure 1 • Candidate ADS amplifier and mxer system models that can be used for construction of behavioral models for receivers and transmitters².
Basic Description of the Receiver to be Modeled
The receiver to be modeled is a two-stage superheterodyne down converter with an RF input frequency range of 0.5 to 18 GHz, and an IF output frequency of either 160 MHz or 1 GHz, with corresponding IF output band of either 100 MHz or 500 MHz, respectively. The noise figure is rated at 20 dB, and gain 20 or 40 dB. Maximum input CW power is +20 dBm, and the unit has two LO inputs required, that are supplied by an adjacent (modular) unit. The system uses coaxial SMA connectors on all inputs and outputs. Maximum output power is approximately +10 dBm.
Technical Approach
Modelithics performed a series of measurements to characterize the receiver in terms of conversion loss vs. frequency and power, noise figure vs. frequency, and non-linear intermodulation distortion levels. The developed model uses built-in ADS system level amplifier and mixer model with parametric inputs derived and conforming to measurement observations. For the purposes of this project, the LO input sources are considered to be integral to the receiver.
ADS provides several built-in behavioral models that can be utilized to characterize the receiver of interest. Figure 1 shows some of the candidate models. The core of the developed model utilizes the built-in Mixer2 ADS model which provides key features to characterize typical mixers’ nonlinearities, such as conversion gain. An additional filter module is added at the output side to provide better representation of the IF spectrum contents.
Figure 2 • Subject receiver system includes an IF module, an LO module, a power supply and an interface controller (not shown), mounted together on a large metal carrier for convenient handling.
Figure 2 shows the ADS symbol representation of the receiver model. The receiver model requires the RF input frequency to be entered as a parameter, with a valid range of 0.5 to 18 GHz in 0.5 GHz step. The total RF power into the receiver is also passed along as an input parameter. The LO frequency is automatically set in the receiver model as (RFfrequency + IFfrequency), with the IF frequency at 1 GHz. Harmonic balance simulations must be evaluated at both the RF frequency and LO frequency.
Figure 3 • Schematic symbol of the 0.5 to 18 GHz receiver model. The input RF frequency and power are passed in as parameters to the model.
Simulation Results
The presented results shown in Figures 4-8 compare measured and simulated conversion gain and power compression characteristics for various frequency setups. Based on S-parameter results, not specifically shown in this note it was found that the RF port reflection, RF to IF leakage, and IF to RF leakage are dependent upon the frequency band used (i.e. 0.5-6 GHz band or 6-18 GHz band). The receiver model automatically detects from the input parameter “RFfreq” what reflections and leakage levels to use, and properly predicts the simulated port reflections and leakage. The model can predict third-order two-tone intermodulation behavior, input and output reflection coefficients and noise figure as exemplified in Figures 9 and 10 and Table 1.
For spectral analysis, spur measurements were made under narrow bandwidth settings and low noise floor. The noise floor on the spectrum analyzer was lowered to -95 dBm, and the receiver was driven up to 2 dB compression. No significant spurs were detected on the spectrum analyzer. Therefore, the model was configured to suppress all spurs (with the help of a filter module).
Figure 4 • Measured (symbols) and simulated (solid line) conversion gain of receiver. Input signal is at RF frequency of 10 GHz, output signal is at IF frequency of 1 GHz.
Figure 5• Measured (symbols) and simulated (solid lines) conversion gain of receiver. Input signal is at RF frequency of 0.5 GHz, output signal is at IF frequency of 1 GHz.
Figure 6 • Measured (symbols) and simulated (solid lines) conversion gain of receiver. Input signal is at RF frequency of 18 GHz, output signal is at IF frequency of 1 GHz.
Figure 7 • Measured (symbols) and simulated (solid lines) conversion gain vs RF input power for frequencies 0.5 GHz to 18 GHz in 0.5 GHz steps (excluding RF frequency at 1GHz).
Figure 8 • Measured (symbols) and simulated (solid line) conversion gain at RF input power = -30 dBm. Frequency is from 0.5 to 18 GHz with 0.5 GHz steps.
Figure 9 • Measured (symbols) and simulated (solid line) input (RF) port and output (IF) port reflection coefficients for the case of RF frequency of 10 GHz and IF frequency of 1 GHz and an input RF power of -30 dBm.
Figure 10 • Measured (symbols) and two-tone simulated model predictions (solid lines) for the case of RF frequency of 15 GHz and IF frequency of 1 GHz and an RF input power swept from -22 to -10 dBm. Shown are results for third-order intercept (TOI), third order power level (IM3) and carrier power level at the receiver output (IF Carrier).
Table 1 • Measured and simulated noise figure.
Conclusions
This note has presented results of a custom system level modeling project that demonstrates how a fairly complex receiver system can be reduced to a simple and accurate behavioral model within Agilent ADS. The resulting system can reproduce many relevant system linear and non-linear performance characteristics. Although not part of the scope of the described modeling effort, further improvement may be possible with the use of the X-parameters2 approach. This could enable, for example, better accuracy in combining the described system with other non-linear system components in terms of the phase representation of various non-linear frequency components.
About this note:
Jiang Liu, Hugo Morales, and Larry Dunleavy are with Modelithics, Inc., Tampa, Florida. The authors would like to thank Marvin Marbell, now with Infineon Technologies, Rick Connick, now with TriQuint Semiconductor, previously with Modelithics for their contributions. We would also like to thank Paul Watson, of the Air Force Research Laboratory for his helpful collaboration.
Footnotes:
¹ Based on a project funded by Air Force Research Laboratory, Sensors Directorate, WPAFB, OH.
² Included in this list of model icons (as the “X2P” element) is the Agilent Technologies X-parameters file-based modeling approach, which can also be considered for modeling amplifiers, mixers , receivers and transmitters. X-parameters is a trademark of Agilent Technologies