Multidisciplinary Senior Design Conference Kate Gleason College of Engineering Rochester Institute of Technology Rochester, New York 14623 SINGLE-BALANCED MIXER PROJECT FINAL PRESENTATION RIT Senior Project Jared Burdick May 17, 2012 INTRODUCTION What is a mixer? Where are they used? Communication systems. Radar applications. How does a mixer work? A device used to convert frequencies. Mixer is a term generally associated with converting higher frequencies to lower frequencies. RF IF LO They take advantage of the non-linear properties of diodes. The signal (RF) is “mixed” with another fixed (or tunable) frequency (LO) and a “difference” frequency (IF) is produced along with a number of predictable inter-modulation products. There are several different configurations for mixers. A single-balanced configuration was selected for this project. PROJECT GOALS Research Mixers Understand theory, applications, configurations, and design trade-offs. Design, Simulate, Prototype Mixer Choose an appropriate configuration. Develop design and simulation skills. Mitigate risks and follow project plan. Test mixer and compare simulated to actual performance. Analyze results and offer possible future improvements / implementations. CUSTOMER NEEDS SPECIFICATIONS Several specifications were modified during the development (with customer approval) SYSTEM BLOCK DIAGRAM COMPONENTS USED Anaren 90º Hybrid Coupler (XC0900A-3S) Avago Schottky-Diode (HSMS-2822) Coilcraft Chip Inductors (0805HT-12NTJB) DLI Chip Capacitors (C06UL120G and C04UL2R7) Gigalane SMA Connector (PAF-S05-007) Murata Chip Inductor (LQW18AN39NG00D) Rogers Substrate Material (RO4003C) Lumped Element Filter AWR MODELS Lumped LPF 5th Order LPF Maximally Flat Degree= 5 Fp= 400 MHz PORT P=1 Z=50 Ohm INDQ ID=L1 L=12 nH Q=40 FQ=0.1 GHz ALPH=0 INDQ ID=L2 L=39 nH Q=40 FQ=0.1 GHz ALPH=0 CAPQ ID=C1 C=12 pF Q=50 FQ=0.1 GHz ALPH=0 Single-Balanced Mixer INDQ ID=L3 L=12 nH Q=40 FQ=0.1 GHz ALPH=0 CAPQ ID=C2 C=12 pF Q=50 FQ=0.1 GHz ALPH=0 PORT P=2 Z=50 Ohm DESIGN TRADE-OFFS & DECISIONS MADE Configuration Use commercially available components wherever possible. Removed BPF’s from the RF and LO paths due to not readily available. Went to lumped-element LPF in the IF path for the same reason. LO Leakage (LO to IF Isolation) Increased to 5th order of LPF at IF port Better rejection (approx. 20dB more) at 1GHz, which improved LO/IF isolation (SBM configuration offers no natural reduction of the LO). Conversion Loss Flatness Added micro-strip quarter-wave transformer to help match the impedance coming out of the diodes and going into LPF Varied width of micro-strip line to see which gave the best conversion loss result Changed the radial RF micro-strip choke into a shorted quarter-wave micros-trip stub Tried Various angles for the radial choke and line width and found there was little improvement Finally went to a true shorted quarter-wave stub Gave the best result in simulation Easy to provide ground to stub for physical layout Added RF bypass capacitor shorted to ground after the diode provide additional filtering prior to the impedance transformation. SIMULATION RESULTS Mixer Conversion Loss Mixer Isolation -4 -20 -30 -5 Output Level (dBm) Conversion Loss (dBm) [email protected] (dBm) -6 -7 [email protected] (dBm) -40 -50 -60 -8 -70 0.8 0.9 1 Frequency (GHz) 1.1 1.2 0.8 0.9 Mixer Return Losses 1 Frequency (GHz) 1.2 IF Output Power Spectrum 0 0 RF = 0.8 GHZ LO = 1.0 GHz 0.2 GHz -16.86 dBm -20 Power (dBm) -10 Return Loss (dB) 1.1 -20 -30 0.4 GHz -61.92 dBm 1 GHz -40.24 dBm -40 0.8 GHz -51.21 dBm -60 0.6 GHz -63.94 dBm LO Return Loss, dB -40 RF Return Loss, dB -80 0.8 1.2 Frequency (GHz) 0 1 2 3 Frequency (GHz) 4 4.8 SIMULATION RESULTS Power Compression Approximate 1dB Compression Points LPF IL RL 0 DB(|S(2,1)|) (L) LPF -2.8 -10 -60 -80 0.5 1 1.5 Frequency (GHz) 2 2.5 Coupling (dB) -20 DB(|S(3,1)|) Coupler -3 Return Loss -40 0 1.1994 GHz -2.8599 dB 1.0006 GHz -3.0466 dB DB(|S(1,1)|) (R) LPF -20 Insertion Loss XC0900A3 Coupling 0 DB(|S(4,1)|) Coupler -3.2 1.0003 GHz -3.1613 dB -30 -3.4 -40 -3.6 1.1995 GHz -3.5242 dB 0.8 0.9 1 1.1 Frequency (GHz) 1.2 1.3 CIRCUIT LAYOUT & ASSEMBLED UNIT Assembled Unit Circuit Layout SMA Conn Launch (RF In) λ/4 shorted stub Diode Pair λ/4 transformer Coupler SMA Conn Launch (LO In) RF Bypass Cap LPF λ/4 shorted stub SMA Conn Launch (IF Out) TEST RESULTS – SUMMARY COMPARISON Spec Specification Unit of Spec Simulated # (metric) Measure S1 RF Frequency GHz 0.8 - 1.2 OK S2 LO Frequency GHz 1 OK S3 IF Frequency MHz DC - 200 OK Maximum S4 dB 10 7.2 Conversion Loss S5 RF/IF Isolation dB 30 42 S6 LO/IF Isolation dB 35 50 S7 Minimum LO Power dBm 10 OK S8 Maximum RF Power dBm -10 OK Minimum 1dB S9 dBm 5 8 Compression S10 Maximum RF VSWR Ratio 2.0:1 1.40:1 S11 Maximum LO VSWR Ratio 2.5:1 1.09:1 S12 Maximum IF VSWR Ratio 1.5:1 - Measured Comments Unit 1 Unit 2 OK OK All parameters tested over these ranges. OK OK Several tests looked at performance OK OK beyond specified. 7.6 8.0 37.0 41.0 OK OK 38.5 43.0 OK OK 8.6 8.6 1.29:1 1.05:1 1.46:1 LPF rejection less than simulated. All parameters tested at these levels. Approximate Value 1.25:1 LO & RF Port VSWR's response similar - LO 1.08:1 port defined at 1.0 GHz only (null) 1.43:1 Not simulated All specifications were met by both units built. Both units had very similar performance. TEST RESULTS Conversion Loss RF to IF Isolation -4.0 60 Conversion Gain (dB) Unit #1 50 RF - IF Isolation - Unit #1 -5.0 Isolation (dB) Conversion Gain (dB) -4.5 -5.5 -6.0 -6.5 40 30 20 -7.0 Measured Conversion Gain (dB) -7.5 Simulated 10 Simulation (dB) -8.0 0 0.800 0.900 1.005 1.100 1.200 RF Frequency (GHz) RF Frequency (GHz) RF Input Power Level (dBm) -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 0 6.0 -10 0.8 GHz 8.0 0.9 GHz 1.005 GHz 10.0 1.1 GHz 12.0 1.2 GHz Spur Level (dBc) Conversion Loss (dB) 4.0 -20 Frequency (GHz) Highest Spur Output Level Unit #1 -30 -40 14.0 16.0 Conversion Loss vs. RF Input Power Unit #1 -50 Measured -60 18.0 1-dB Compression Spurious Output Simulated TEST RESULTS IF Output Power Spectrum 0 RF = 0.85 GHZ LO = 1.0 GHz 0.15 GHz -16.27 dBm -20 Power (dBm) Unit #1 IF Output Spectrum LO = 1000 MHz RF = 850 MHz Horiz. Scale: 200 MHz/div 1 GHz -40.24 dBm -40 0.3 GHz -72.65 dBm 0.85 GHz -54.26 dBm -60 0.7 GHz -77.27 dBm -80 0 Spurious Output 1 2 3 Frequency (GHz) 4 4.85 CONCLUSIONS The prototype mixer met the target specifications. There were differences between the simulated performance and the actual measured performance. In general, the actual measured performance was consistent with the model. LO to IF Isolation about 7-9 dB less. Suspect that the LPF roll-off (rejection at higher frequencies) was less than modeled – this will need further evaluation to confirm. RF to IF Isolation 3-5 dB less – LPF roll-off would contribute here as well. Conversion Loss was slightly higher – connectors not modeled could be a contributor. Future Iterations / Investigations. Add BPF to the LO and RF input paths. Investigate LPF performance. Refine the AWR model (connectors, HFSS sub-models, etc.). Work on final mechanical packaging concept. LESSONS LEARNED Do your homework before starting to design Time is a scare resource There are many trade-offs that need to be considered and decisions that need to be made in order to best match the expected performance to the application and requirements. Ability to model the circuits accurately was key and greatly increased the probability of success. Valuable lessons can be learned even in non-ideal circumstances. Figuring out project limitations early on in the process helped reduce risk and deliver the final product on time. Look at contingency plans (alternate parts, fabrication alternatives etc.) Identifying concrete action items helped to focus efforts and reduce wasted time. Make use of all available resources Eliciting feedback from other knowledgeable people proved invaluable. There was a significant amount of information available on-line (technical papers, forums, etc.).