ARO MURI Kband Spatial Power Combiner Using Active

ARO MURI K-band Spatial Power Combiner Using Active Array Modules LY. Vicki Chen, Peng. Cheng Jia, Robert A. York PA Workshop, San Diego 2002

Presentation Outline • Passive System Antenna array, Higher order mode problem, Performance • Amplifier Design Two-stage Amplifier, Flip-Chip IC, CPW-line • Power Combiner Design considerations, Performance • Conclusions ARO MURI

Spatial Power Combining ARO MURI • Normal incident/outgoing waves • Limited bandwidth in general • Easier for monolithic design • Challenge in thermal management Tile Approach • Parallel incident/outgoing waves • Broadband characteristics • Good heat-sinking property • Consuming more substrate area Tray Approach

System Overview • • ARO MURI Extended work from the X-Band Spatial Combiner Design Oversized Waveguide Environment – TE 10, TE 20 from 18 to 22 GHz Fin-line to CPW line Transition Monolithic Circuit Design – Flip-Chip IC Oversized waveguide (TE 10, TE 20) Gradual transition from WR 42 to the oversized waveguide Active Devices Tapered-Finline Antennas

Antenna Design ARO MURI Finline to CPW line transitions • Design is based on the optimal taper design of the X-band system. • Finline to CPW line transitions – Eliminate the bond-wires. Klopfenstein Taper Ground Signal • Air-bridges are needed to provide good grounding in the middle ground plane. Ground • Use HFSS for simulation. Signal Ground Al. N substrate CPW line Reference paper: Design of Waveguide Finline Arrays for Spatial Combining. Submitted to IEEE transaction on MTTs

HFSS Simulation PEC metal air Et metal Et ARO MURI By forcing the PMC boundary condition, the even mode does not exist in the system. PEC Fin-line Metal PMC Dielectric Et Waveguide Wall PMC • Simulate for 2 x 4 system. • Simplify the problem by applying the boundary conditions. • Impedance & Gamma vs. Gap-size • Finline – CPW Transition • Reflection coefficient for the taper design.

Effects of Mounting Grooves Single mode unilateral finline d b ARO MURI The mounting grooves affect the optimal values of many parameters: • • • Operating frequency Effective dielectric constant Substrate thickness Small slots are affected more severely than broader slots. short circuit grooves d The depth of the short circuit grooves has huge effect on the return loss Reflection Coefficient (d. B) d d=λ/5 d=2λ/15

Combining Efficiency ARO MURI • Symmetrical loading is necessary to avoid TE 20 mode and achieve efficient combining. • ~ 76% combining efficiency is achieved. Measurement for one card (asymmetrical) and two cards (symmetrical) system Frequency (GHz) 6 cards (4 x 6 system) 50 ohm Termination & Through-line measurement Frequency (GHz)

Two-Stage Amplifier Design Flip-Chip Technology using CPW-line • Thermal Management – FCIC • Gain Enhancement – pre-amplifier • Optimal load matching • ADS/Momentum Simulations Design Challenges – • Substrate modes are excited easily. • Biasing circuitry is complicated. • Good grounding should be maintained. 13. 5 mm Ground Signal 8. 56 mm Ground Signal Ground Pre-amp ARO MURI Power device

Two-Stage Amplifier Design CPW-line Substrate Mode Substrate mode is excited easily! eff Al. N > eff > Air Quasi-TEM ARO MURI Increasing the substrate thickness or reducing the width of the CPW-line could reduce the effect of the substrate mode. Al. N S 21(d. B) of the Two-stage Amplifier

Two-Stage Amplifier ARO MURI Performance • Return loss <-10 d. B for the operating frequencies. • 27. 3 d. Bm output power with 20% PAE and 9. 5 d. B power gain was obtained. • Thin film (on-chip) and chip capacitors were both needed for biasing circuitry. Pout PAE Gain

Combiner Performance Two Cards Measurement ARO MURI Small Signal Performance & Power Measurement @ 18 GHz • Two cards system (8 amplifiers) with 34 d. Bm output power. • 12. 5% PAE – 62% Combining Efficiency (optimal: 76%) • Phase difference between cards degrades the output performance the most. Pout (d. Bm) PAE Gain (d. B)

Measurement ARO MURI

Statistical Errors in Arrays Output voltage: G 2 Combiner Splitter G 1 A ARO MURI Bout ri = 0 or 1 Probability of device survival GN Output Power: Change in power due to errors: Ensemble average: Phase errors and device failures are most important in large combiners Ref: R. York, “Some considerations for Optimal Efficiency and Low Noise in Large Power Combiners”, IEEE Trans. Microwave Theory Tech.

Combiner Performance 6 Cards Measurement ARO MURI • Six Cards System (6 x 4) – 2 devices failed. • 10 d. B small signal gain. • Absorbing material were added to prevent the in-band oscillations. • Each cards was biased individually. • Improper grounding could result in oscillations. • Some chip resistors were added to prevent bias-line oscillations. Signal cross-talk between bond-wires could induce in-band oscillations Absorbing Material

Phase Noise Reduction G Combiner Splitter G G Noise contributed by the ensemble is reduced by 1/N (-13. 4 d. B) compared with a single amplifier ARO MURI Amplifiers degrade phase noise due to internal nonlineariities which up convert low-frequency amplitude and phase noise to the carrier

Power Measurement • 37. 3 d. Bm output power @18 GHz was obtained. • Non-uniform excitation. • Chip resistors were added for bias-line stabilization. • Combining Efficiency: 53. 7%. ARO MURI Non-uniform excitation profile

Conclusion ARO MURI • Passive combiner achieved 76% combining efficiency. • 34 d. Bm output power has been obtained for 2 -tray system. • 62% combining efficiency was obtained. • 37. 3 d. Bm output power has been obtained for 6 -tray system. • 53. 7% combining efficiency was obtained. • Phase noise reduction compared with single amplifier. • Stabilization should be maintained for all frequencies.