Design Methodology for Wireless Nodes with Printed Antennas

1
Design Methodology for Wireless Nodes with
Printed Antennas

• Jean-Samuel Chenard
• Chun Yiu Chu
• eljko ilic
• Milica Popovic
• McGill University
• Montréal
• DAC, June 15, 2005

2
Motivation

• New low-cost transceivers at microwave
  frequencies
• Very high level of Integration (PHY MAC)
• Few external components (crystal capacitors)
• Antenna cost becomes significant

3
Presentation Outline

• Introduction
• Antenna theory
• Wireless nodes constraints
• Printed circuit board (PCB) considerations
• Electromagnetic modeling 2.5D and 3D
• Design method
• Design flow
• Design of loaded dipole
• Conclusion

4
Problem Adding an Antenna to the PCB

• Follow manufacturer application note
• Gives a quick solution
• Will it work ?
• Yes but what is the performance, flexibility ?
• Antenna publications
• Amazing designs ! Complex
• Terminology differences PCB vs. Antenna
• Integration is left to the designer
• This Paper
• Method to integrate the antenna to PCB flows
• Explore 2.5D and 3D electromagnetic modeling (EM)
  in PCB design
• Design flow that reduces risk

5
Antenna fundamentals an antenna

• Sends and receives radio waves
• Provides a transition from a guided wave on a
  transmission line to a free-space wave

Free Space
Transceiver board
6
Antenna Theory Bandwidth Frequency

• Antenna bandwidth
• Resonant class (e.g. monopole, dipole)
• Narrow bandwidth (around 5)
• Help reduce coupled emissions
• Fit for current applications 2.4 GHz, 5 GHz
• Wideband class
• Future ultra-wideband transceivers
• Operating frequency
• Size of antenna ?/2 for dipole
• Frequency, dielectric constant
• Influence on the received power
• Double the frequency Half the received power

7
Antenna Theory Gain and Coverage

• Antenna Gain (dBi or dBd)
• Incorporates the efficiency (losses)
• More antenna gain More range, less angular
  coverage
• Need tools that predict the gain

Gain
8
Antenna Theory Polarization

• Polarization - Orientation of the E field
• Linear - Fixed orientation over time
• Smaller antenna
• Product orientation becomes important
• Horizontal / Vertical
• Circular or elliptic - Rotates over time
• Reduced range When used with linear antennas
• Larger antenna structure

Ex
Vertical Polarization (E Field Direction)
Hy
9
Antenna Integration Circuit level

• From the standpoint of circuit designers
• Characterize antennas using S-parameters
• Black box theory
• S11 (Input Reflection Coefficient) is the key
• Treat antennas as other RF components
• Like Filters, LNA, Mixers, etc.

Antenna
Return loss (dB)
10
Antenna Integration - Summary

• From the standpoint of antenna designers
• Achieve key antenna parameters
• Impedance matching to source
• Radiation patterns
• Optimized for the application
• Efficiency
• Aim for 100 - Cost tradeoffs
• Gain
• Follows previous two parameters
• Polarization
• Concentrate power in a single dimension

11
Wireless Node Constraints

• Low transmit power aim for
• Maximum efficiency
• Maximum power transfer
• Low standing wave ratio Low S11
• Antenna gain suitable for the application
• Better range or better coverage
• Lower node density Lower system cost
• Small board area
• Linear polarization
• Loaded antenna
• Suitable for extreme environments
• Thermal expansion coefficients match with PCB
• No fragile parts

12
Printed Circuit Board - Constraints

• Two main classes of PCB material
• Specialized dielectric materials
• Controlled dielectric Stable with temperature
• Controlled thickness
• FR4
• Inexpensive and widely available
• Design must tolerate variations in parameters
• Large increase in loss tangent for FR4
• Results in a 20-30 loss in antenna efficiency

13
Electromagnetic Simulation

• Current density explains antenna operation
• Good antenna- high density in radiating elements
• Assists in debugging problems
• Need to account for topology, metal, dielectric
• Comprehensive EM simulation

14
EM Simulation Tools 2.5 D

• 2.5D EM simulators (e.g. ADS Momentum, Ansoft
  Designer)
• Advantages
• Ideal for planar antenna designs
• Less demanding in computing time and resources
• Infinite dielectric substrate simplifies
  calculations
• Disadvantages
• Infinite ground plane limits classes of antenna
• Infinite substrate - Incorrect far-field
  modelling along dielectric plane
• Good for preliminary designs
• Designer should be aware of the limitations of
  2.5D Simulations

15
EM Simulation Tools 3D

• 3D EM Simulators (e.g Ansoft HFSS, IE3D )
• Advantages
• Finite substrate layer in calculations More
  accurate
• Better far-field radiation patterns No
  artifacts
• Include effects of near-field objects on antenna
  performance
• Validate device enclosure, batteries, etc.
• Disadvantages
• Computing resource requirements are very high
• Must simplify imported designs !
• Results are sensitive to mesh cell size and
  number of refinement cycles
• Trade off simulation time for accuracy
• Ideal for validation at later stages of the
  design

16
Design Flow
17
2.5D - Design Space Exploration (1)

• Monopole corner antenna
• Simple
• Small
• Extend PCB
• Reduces effect of connector
• Model frequency shift due to ground plane

18
2.5D - Design Space Exploration (2)

• Basic design from literature
• Better pattern
• Larger bandwidth
• Larger size
• Possible improvements
• Capacitive loading

19
Antenna Feed Network

• Losses may be significant
• Allow for test point
• 50O single ended
• Microstrip feed
• PCB thickness affect width
• 2.4 GHz 0.030 In ( 0.76mm ) is best compromise
• Prevent parallel-plate resonance
• Guard traces
• Stitch top/bottom planes
• Consider the transceiver DC biasing
• Shorted ?

20
Final Phase 3D Simulation

• Complex to set up
• Objects in X,Y,Z coordinates
• Boundary conditions
• Choose model simplifications carefully
• Trade-off simulation time for accuracy
• Imported circuit needs simplification
• Even simplified simulations are long
• Hours on high-end machines
• 3D simulation will confirm 2.5D
• More precise
• Radiation pattern without artifacts

21
2.5D VS 3D EM simulations results

• Quick results with 2.5D
• Very accurate pattern with 3D simulations
• Both tools can be combined in a good design flow

22
From the 3D Model to the Lab

• Input port launch near SMA port
• Note abstractions
• No digital lines
• Only planes remaining
• Keep the vias linking top/bottom
• Solder mask near antenna
• Optional
• SMA test port

23
Prototype Validation

• Accurate frequency match (1.8 mismatch)
• Measurement room limitations
• -15 dB min
• Room echoes
• Efficiency
• Costly to measure
• Can be inferred

24
Other Considerations EMI and Crosstalk

• Narrowband design helps reduce coupling

25
Other Considerations - Balun

• Balun losses
• Too high in first prototype
• Manufacturer application note
• Underestimated parasitic capacitance Bad
• Better EM model found the problem
• Not Simulated Problems
• Reduced range
• Range 60m in real world deployment (Charlotte,
  NC Convention Center)
• Can be doubled by improving balun circuit

26
Antenna Improvements

• Wider antenna bandwidth 400 MHz
• Loading element
• Tapered arms
• Bends
• Eliminate nulls
• New design targeted for FR4
• Use wider bandwidth design
• Tolerant to variations in dielectric constant
• Tradeoffs
• More losses
• More cross-polarization

27
Conclusions

• PCB replaces important components
• Higher frequencies better opportunity
• Matching network and balun usingmicrostrip lines
• Integrated EM modeling - becoming a necessity
• Printed Antennas up to 6 GHz No Problems
• Ultra Wide Band Antennas
• Microwave Design Methods Gaining Grounds
• In Embedded Systems
• In Very Low Cost Systems
• Applicable to EMI Analysis
• Techniques scale well with frequency

28
Questions ?
29
Antenna Integration I

• From the standpoint of circuit designers
• Characterize antennas using S-parameters
• Black box theory
• S11 (Input Reflection Coefficient) is the key
• Treat antennas as other RF components
• Like Filters, LNA, Mixers, etc.

Transceiver
S11
Antenna
Input VSWR
Return loss (dB)