Study on Bluetooth Antenna Integration into Metal Environment Julnar Musmar, Dr. Ing Denis Sievers, Ralf Kakerow Continental, Paderborn University

Transcription

1 Bitte decken Sie die schraffierte Fläche mit einem Bild ab. Please cover the shaded area with a picture. (24,4 x 7,6 cm) Study on Bluetooth Antenna Integration into Metal Environment Julnar Musmar, Dr. Ing Denis Sievers, Ralf Kakerow Continental, Paderborn University IIC SP RD EE

2 Introduction/ Motivation Bluetooth attracted the Automotive device developers due to its outcome behavior & cost efficiency. The study aims to find a suitable simulation topology for a Bluetooth antenna enclosed in a metal environment by: Studying the antenna behavior. Model simplification. Complexity reduction. Developing a good antenna is not the main purpose here. All simulations were conducted using CST MWS 2

3 Aims & Challenges The study was performed with the following aims in mind: Res. Freq. S11 db Tot. Effici. db rlzd. Gain db Sim. Time Memory 2.45 GHz < -10 max. max. min. min. S11: Return loss or Reflection coefficient, is an indication how much power is reflected and how much power reaches the Antenna. R.L = 20log 2 (IΓI), Γ=(Z input -Z 0 )/(Z input +Z 0 ). Total Efficiency: Relationships between the radiated power to the net power accepted by the antenna from the connected transmitter. Realized Gain: The ratio of the radiated power intensity in certain direction to the radiated intensity obtained by an isotropic antenna (the later is given by Power accepted. /4π) and it include the impedance mismatch loss. G(θ,Φ)=η Radiation D(θ,Φ), G match = η match G(θ,Φ) = η match η Radiation D(θ,Φ), η Radiation = Gain/Directivity, η mismatch =1-I Γ I 3

4 Antennas used Two Antennas were within our main prospect due to their Low profile & Robustness Low cost Efficiency Easy Manufacturability PIFA (Planar Inverted F Antenna) Meander Antenna 4

5 PIFA in the presence of PCB layers PIFA Aim: study the effect of design parameters. Taking the res. frequency & S11 Decreasing the length of the free end shifts the frequency to higher values. S11 amplitudes don t change significantly. 5

6 PIFA in the presence of PCB layers: cont. Manipulating the distance D between the ground point and feeding point: D Both resonance frequency and S11 parameters are affected. Increasing the feeding point increases the resonance frequency and improves S11 amplitude. 6

7 Meander Manipulating the free end: The resonance frequency changes S11 amplitude change is low S11 and radiation of meander antenna are sensitive to geometry changes PIFA antenna design is more straight-forward 7

8 Surface current in the presence of PCB and antenna 8

9 Introducing the Model Used material PCB FR-4 Antenna & PCB ground Copper Metal housing Steel Boundary conditions : Open (add space) Frequency range: 0-3GHz. Solver used : Time domain transient Housing & front Mechanical department / CATIA. PCB Electrical development department / E-CAD ZUKEN Computer architecture: 2x CPU E GHz, 64GB RAM, NVidia acceleration card Tesla K40 9

10 Meshing Hexahedral meshing Meshing controls cells numbers time step accuracy & simulation run time. Initial mesh generation results in 1.5 billion mesh cells Meshing approach High mesh resolution around antenna (high field gradients) High mesh resolution of PCB layer stack Low mesh resolution on GND plane topology (vias, etc.) Three meshgroups are defined 10

11 The reference case A reference model has been defined including Meander antenna PCB ground layers (4 layers in this testcase) Metal housing Front assembly/pcb Res. Freq. S11 db Tot. Effici. db rlzd. Gain db Sim. Time Memory 2.7 GHz GHz h, 30m 53G Byte 11

12 Reference case Results All simulation were performed without antenna matching unless indicated otherwise, in order to study the effect of the antenna itself. 12

13 Reference case Results with matched antenna 13

14 Surface current in the presence of metal housing frontside view 14

15 Metal housing effects on the surface current backside view 15

16 Testcases for simulation time reduction Suggested steps for simulation runtime reduction: Material change (copper PEC) for PCB ground layers and antenna. Material change (steel PEC) for metal housing. Reduction of the PCB ground layers. Filling the structure between the upper & lower ground layers with copper. Filling the surrounded space with a low conductive material. 16

17 PCB ground layers and antenna material changes into PEC PCB ground layers and antenna: copper PEC Res. Freq. S11 db Tot. Effici. db rlzd. Gain db Sim. Time Memory 2.7 GHz h 48G Byte reference 2.7 GHz h, 30m 53G Byte Performance factors do not change significantly Time saved : 8%, Memory saving: 12% Time saving because PEC modeling is less expensive 17

18 Metal housing material changing into PEC. Steel of the metal housing: steel PEC PCB ground layers and antenna: copper PEC Res. Freq. S11 db Tot. Effici. db rlzd. Gain db Sim. Time Memory 2.7 GHz h 58G Byte reference 2.7 GHz h, 30m 53G Byte Explanation: Housing is highly resonant No attenuation in the PEC material leads to even stronger housing resonance Energy needs more time to decay 18

19 Energy reference PEC housing 19

20 PCB layer stack testcases Reference design Only L2 L2 & L7 Copper block between L2 & L7 20

21 Layer stack testcase 1 All ground layers except upper layer are removed from the model Ground layer material: copper Housing material: steel Copper Res. Freq. S11 db Tot. Effici. db rlzd. Gain db Sim. Time Memory 2.47 GHz h 51G Byte reference 2.7 GHz h, 30m 53G Byte This model simplification leads to wrong results Frequency shift 21

22 Layer stack testcase 2 Keeping upper and lower layers Ground layer material: copper Housing material: steel Copper Res. Freq. S11 db Tot. Effici. db rlzd. Gain db Sim. Time Memory 2.68 GHz h 33m 51.8G Byte reference 2.7 GHz h, 30m 53G Byte This model simplification leads to wrong results Frequency is accurate, but antenna impedance changes significantly 22

23 Layer stack testcase 3 One thick ground brick instead of individual ground layers Relaxed mesh requirements (smallest mesh cell increased larger time step) Res. Freq. S11 db Tot. Effici. db rlzd. Gain db Sim. Time Memory 2.85GHz h, 40m 45G Byte reference 2.7 GHz h, 30m 53G Byte Ongoing investigations 23

24 Next steps Optimize the Antenna performance in the presence of PCB layers. Shift the resonance frequency in housing to 2.45GHz. Filling the surrounding space with a low conductive material. Replacing the GND planes with layout topology by thin solid blocks. Effect of the antenna type on the structure. Comparing different solvers results. Create a simulation guideline based on the results 24

25 Conclusion First priority: simulation runtime optimization Allows fast simulation of antenna integration into products Material replacement effects on simulation time PCB ground - PEC instead of copper: reduced simulation time, good accuracy Housing PEC instead of steel: increased simulation time Model simplification Elimination of selected PCB ground layers affects the simulation accuracy Replacing ground layer structure by copper brick investigation is ongoing 26