Modeling Techniques for Rapid Antenna Evaluation and the Power of Field Visualization

Transcription

1 Modeling Techniques for Rapid Antenna Evaluation and the Power of Field Visualization John Locke Technical Specialist, RF and Antenna Subsystems Ford Motor Company

2 Introduction You are an Engineer Engineer it! Anonymous Antenna Team Sr. Manager when asked for budget to run an engineering antenna pattern test to validate a new test process (2001) There is no way we are going to buy any of those antenna simulation tools - Another Anonymous Antenna Team Manager (2001)

3 Problem Statement Engineers are requested to make rapid technical decisions, often on the fly Frequently the result is an engineering best guess and is sometimes incorrect Minor design changes often require days of testing out of scope with the level of change Antenna (Electromagnetic) Engineers often have trouble explaining the Black Magic to other engineering disciplines

4 Solution Electromagnetic Simulation Tools and Computing Resources have reached a point where they can replace many what we used to call engineering test Minor design change tests, proof of concept tests, investigative study tests The visualization capability of these tools is significantly better than it was many years ago Allows for the antenna/electromagnetic engineer to better understand and then explain the black magic The result is a antenna/electromagnetic engineer now can do quick first order verifications and visually demonstrate the concepts Reduce cost due to reductions in testing, reduce time in developing concepts, and easily research new ideas These Tools are Not just for the Antenna/RF designer or Electromagnetic Analyst All Antenna/RF engineers (Test, Release, Production) involved in the hardware production cycle to quickly validate engineering judgment calls

5 Outline Conceptualizing Waveguides Horn Antenna Test vs. Simulation Antenna Pattern Visualization Large Models and Field Strengths Time Domain Analysis

6 Conceptualizing Waveguides

7 Modeled WR975 TE10 Mode E-Field Surface Currents, Note Currents On Sidewall Are in Y Axis Direction

8 WR975 E-Field Movie

9 WR975 Evanescent Mode E-Field of Waveguide in Cut-Off Note that the Wave is Dissipating, Not Reflecting

10 WR975 Surface Currents Note Large Current With Y Component on Sidewall This is why waveguide filter designers try to avoid seams in this area Unfortunately, this is the easiest place to put the seam for machining X Y Z Do Not Do This

11 Problem: Request From Mechanical Mechanical Design Engineering has requested that they be allowed to put in small drill holes at the corners of the waveguide flange to allow machining Red areas are requested thru hole areas Should The RF Engineer Approve Of Design Change?

12 WR975 Will Drill Outs Simulation Engineering Analysis: Some disturbance to TE10 Mode, however, disturbance is minor field levels are low except near sharp edges, recommend rounding edges to reduce risk of corona E-Field Surface Currents

13 WR975 Will Drill Outs E-Field Movie

14 Engineering Recommendation The thru holes were only requested on test flange adapters Actual impact additional would be less than corrugated waveguide (Shown in Pictures to Right) Based on Engineering Judgment and Analysis OK Drill Thru Holes As Long As Edges Are Smoothed

15 Waveguide With Circular Polarization The sum of two orthogonal TE11 Modes Out of Phase by 90 deg

16 Horn Antenna Test vs. Simulation

17 S11 (db) 0.75 GHz SGH Why Doesn t The Measured Return Loss Agree With the Modeled Return Loss? Modeled S11 Measured S SGH Measured Inside EMC Lab Frquency (GHz) MI Technologies SGH

18 S11 (db) Inaccuracy In Measured Data Measured In Highbay Pointing At Celling, Cause Of Ripple In Data Data Shows Reflection At 22 ft. c/(2*delta(f)) SGH Measured Inside EMC Lab X: Y: X: Y: Frquency (GHz)

19 S11 (db) Differences In Model From Reality Modeled With Waveguide Port Below Cutoff Evanescent Wave In Model vs Mismatch At Coax To Waveguide Adapter Waveguide Adapter Only Designed To Excite TE10 Mode from.75 to 1.1 GHz Miss-Match at Adapter Occurs At Higher Frequencies Also Ideal TE10 Excited In Model, All Higher Order Modes Are Missing Ripple In Model Due To Time Domain Solution Some Error Due To Operator Inexperience In Setting Up the Model Modeled S11 Measured S11 No Reflections 0.75 SGH Measured In Free Space Frquency (GHz) Engineer Must Understand The Differences Between Simulation and Real World Test

20 Antenna Pattern Visualization

21 Antenna Pattern Visualization Simple ground plane impacts (ground plane edges, resonant structures) and arrays are easily evaluated Remember when combining antennas (Beamforming) reciprocity MUST be preserved Normalize signal with 1/sqrt(2) for every power split in hybrid couplers or power dividers This also applies to return loss, an open circuit will be -6 db at the input to the splitter When arrayed antennas are separated by large distances infinite nulls result Ideal Monopole Monopole on Finite Ground Plane Arrayed Monopoles

22 Sum Of Omni Directional Patterns + = Note Infinitely Deep Nulls, Represent Loss Of Signal The Sum of Two Omni Directional Antennas Does Not Provide Omni Directional Coverage* *with some exceptions (Multiple Receivers )

23 Large Models and Field Strengths

24 Reducing Model Elements Basic structure model of a High Gain Antenna used to estimate field strengths for EMC testing Model with Far Field Source For Feed Horn Model can be refined based on field strengths Unnecessary Structure Circled In Red

25 Providing E-Field For EMC EMI Susceptible Electronics Which Antenna Confutation Will Induce Larger E- Fields On The Electronics? Antenna With Direct Line Of Sight To Electronics Antenna Blocked From Electronics By Large Reflector

26 Answer As Expected But Line of sight is greater, but diffraction is closer than expected, within approx 10 dbm Antenna With Direct Line Of Sight To Electronics Antenna Blocked From Electronics By Large Reflector Note: When Doing This Analysis Do Not Use Physical Optics Diffraction Off Of Edge Can Have A Greater Impact Than You Expect!

27 Time Domain Analysis

28 Time Domain Analysis of Reflections CST is known for its time domain solver (Transient Solver) Question: Can we see the effects of surface treatments on dielectric sheets Could be a benefit to Radar Analysis and do things that Frequency Solvers cannot Simulation Setup Measures transmitted wave Measures incident and reflected waves Analysis performed by Patrick DeRoy (CST) and John Locke

29 Start With A Dielectric Sheet Model Setup Incident Time Domain Transmitted Reflected

30 Add Cones To Smooth Transition Model Setup Incident Time Domain Transmitted Reflected

31 Summary Results To This Point Incident Reflections Adding Cones Reduces The Initial Reflection, So What Will Adding Cones To Both Sides Do?

32 Added cones both sides (2 Cases) Cones Front To Back Lined Up Cones Front To Back Offset Expect This To Be The Best Case

33 Final Time Signals Comparison Incident Reflections Zoomed View Cones On Both Sides Have Least Amount Of Reflection Offset Cones Has An Additional Ripple Time Signals Can Be Used To Analyze Reflections

34 Conclusion Final Note: By 2008 Both Managers Quoted In The Beginning Were Expecting Their Engineers To Perform EM Analysis and Providing The Tools Thank you for your time Questions?