Winland Electronics, Inc'

1
Winland Electronics, Inc.

• Wireless Sensor Networks
• A Survey of Design Options
• November 16, 2009

2
Winland Electronics Overview

• Winland is a full service EMS (Electronic
  Manufacturing Services) provider
• Design and manufacture for customers in consumer,
  industrial and medical markets

• 35M in sales
• 140 employees
• Mankato, MN
• 58,000 sq ft.
• Public AMEX
• Symbol (WEX)

Slide 2 of 29
3
Winland at a glance

• Wide range of customers
• Consumer products
• Medical electronics
• Computer hardware
• Industrial / Utility Equipment
• Wireless devices
• Sensors and security products

4
Why Deploy Wireless Networks

• Installation labor costs reduced
• Attractive appearance
• Increased number of sensors
• Some applications are impossible to hard wire

Slide 4 of 32
5
Network Solutions Available

• Drop-in Solutions
• All electronics complete
• Module Solutions
• RF section complete
• Discrete RF IC Solutions
• Full custom

Slide 5 of 32
6
Wireless Frequency Bands

• Common sensor network license free bands
• 418 MHz (Canada and U.S.)
• 433 MHz (Europe)
• 868 MHz (Europe)
• 902 928 MHz (Canada and U.S.)
• 2400 2483 MHz
• 2.4 GHz advantage
• World wide

Slide 6 of 32
7
Drop-in Solutions

• Example manufacturers include Digi and
  SensiCast
• Gateway devices
• used to convert wireless signals to a PC
  or local display
• Size 7.7 x 4.1
• Wireless sensors
• used to convert sensor data
• (temp, water detection, etc)
• to a wireless signal
• Size 3.9 x 1.8

Slide 7 of 32
8
Module Solutions

• Example manufacturers include MeshNetics and
  Digi
• Meshnetics ZigBit Module
• FCC cerrtified, high and low power
  versions available
• Silicon Atmel
• Size 0.9 x 0.5
• Digi Xbee Series 2 Module
• FCC certified, high and low
• power versions available
• Silicon Freescale Ember
• Size 1.0 x 1.1

Slide 8 of 32
9
Discrete RF IC Solutions

• Example manufacturers include Atmel and
  Freescale
• Atmel AT86RF230 Radio transceiver
  targeted for IEEE 802.15.4 Zigbee
  apps Size 0.2 x 0.2
• Freescale - MC13193 Radio transceiver
  targeted for IEEE 802.15.4 Zigbee
  apps Size 0.2 x 0.2

Slide 9 of 32
10
Available Products Overview

• Estimated line of sight range based on
  specified link budget using Friis range equation
  with transmitter and receiver gain set to -4dBm.
• Battery life based on 0.1 duty cycle with
  2 AA alkaline batteries fed into a 85 efficient
  regulator during active mode. Iq of regulator
  set to 10uA. Life listed is approximate only.
  Other external circuitry not included in
  calculations.

Slide 10 of 32
11
Drop In, Module or Discrete?

• Application and EAU will drive your decision
• Drop-in solutions
• Quick time to market
• Lowest product development costs
• Certifications complete (country specific)
• Limited application
• Module solutions
• Middle of the road
• Certifications complete (country specific)
• Discrete RF IC solutions
• Lowest per unit cost (13 - 28 cheaper then
  module)
• Diverse application ready
• Form factor
• Design control

Slide 11 of 32
12
Wireless Challenges To Overcome

• Embed application with wireless protocol
• Network deployment ease
• RF Tuning / Layout Considerations
• Form Factor
• Power Consumption
• Bit Error Rate Testing
• Certification Testing
• And many more!

Slide 12 of 32
13
Wireless Challenge 1 of 7

• Embed application with wireless protocol
• Drop-in
• N/A
• Module
• Generally your application will send serial
  commands to the module (module dependent)
• Discrete
• Interface directly with protocol
• Protocol learning curve commands to call
  functions
• Availability of protocol support (3rd party vs.
  manufacturer)

Slide 13 of 32
14
Wireless Protocols

• Zigbee
• Mesh network
• Self healing
• Interoperable
• IEEE 802.15.4
• Star network
• Proprietary
• Custom network topologies
• faster data rates (not limited to 250K bits/s)

Slide 14 of 32
15
Wireless Challenge 2 of 7

• Network deployment ease (depends on application)
• Drop-in, Module and Discrete
• Design user interface to add and remove sensors
• Ensures foreign sensors dont attach to your
  gateway
• Ensures your sensors dont attach to foreign
  gateways
• Design user interface to accommodate sensor
  diagnostics
• Examples Link quality, Received power, Bit Error
  Rate, etc.
• Quick start guide

Slide 15 of 32
16
Wireless Challenge 3 of 7

• RF Tuning / Layout Considerations
• Drop-in
• N/A
• Module
• Ideally place module near corner of PCB
• Follow guidelines per the module datasheet
• Discrete
• Impedance matching
• Transmission lines
• Antenna design (multipath considerations)
• Current return paths

(Discrete RF IC Layout)
Slide 16 of 32
17
Wireless Range

• Range Variations from
• RF Design / Layout
• Environment
• Protocol (star network vs mesh network)
• Repeaters can increase coverage
• Range (without repeaters)
• Low power (105dB link budget maximum)
• 35 to 100 feet indoors
• 500 to 1000 feet outdoors (line of sight)
• High power (110 dB link budget minimum)
• 100 to 200 feet indoors
• 1 to 2.5 miles outdoors (line of sight)

Slide 17 of 32
18
Estimating Wireless Range

• Friis range equation (dB form)
• Pr Pt Gt Gr 10log(4p/?)2 10log(R)n
• Where
• Pr Receive power sensitivity (dBm)
• Pt Transmission power (dBm)
• Gr Receive gain amplification, antennas,
  component loss (dB)
• Gt Transmission gain amplification, antennas,
  component loss (dB)
• ? Wavelength speed of light / frequency
  (meters)
• R Range (meters)
• n Propagation constant accounts for
  obstructions, reflections, etc
• n 2 (outdoors - line of sight at 2.4GHz)
• n 5 (average indoors at 2.4 GHz)
• n 7 (Indoors with 3 walls at 2.4GHz)

Slide 18 of 32
19
Estimating Wireless Range

• Example
• Calculate the average indoor range of the Atmel
  AT86RF230. Assume the
• antennas have no gain and are 50 efficient. 2dB
  of loss exists between
• the transceiver and the antenna due to component
  loss.
• Answer
• From the Atmel AT86RF230 data sheet
• Pr -101dBm
• Pt 3dBm
• Frequency 2.4 GHz
• ? speed of light / frequency 3x108 /
  2.4x109 0.125 meters
• n 5 (average indoor propagation constant at
  2.4GHz)
• Gt -3dB (50 efficient antenna) -2dB
  (component loss) -5dB
• Gr Receive gain is the same as above
  transmission gain -5dB

Slide 19 of 32
20
Wireless Challenge 4 of 7

• Form Factor
• Drop-in
• Limited choices
• Module
• Module size (length x width) may cause issues
• Many modules require headers that increase height
• Discrete
• Full control

Slide 20 of 32
21
Wireless Challenge 5 of 7

• Power Consumption
• Drop-in
• Limited choices
• Module and Discrete
• Generally very good
• Total power consumption will increase from
• External MCU
• Watch dog reset circuits
• Brown out detection circuits
• Quiescent current from all ICs on the board
• Etc

Slide 21 of 32
22
Power Consumption / Battery Life

• Battery life varies from
• Sleep mode power consumption - most critical
• regulator current consumption
• Watchdog reset brownout detection circuitry
• Active mode power consumption
• RF TX/RX consumption antenna switches
• MCU application specific circuitry
• Duty cycle active time / (active sleep time)
• Typical active time 3 100ms
• Typical sleep time 1s 1h
• Battery capacity
• Remember to supply a given power requirement
  Battery current ? as battery voltage ?

Slide 22 of 32
23
Estimating Battery Life

• Where
• D duty cycle () L battery life (years)
• Ia active current (mA) Is sleep current
  (uA)
• Ta active time (ms) Ts sleep
  time (s)
• C battery current capacity (mAh) 8760 hrs
  per year
• Note
• To predict Ia Is (currents sourced from
  batteries) you will need to know
• Battery voltage vs. service life at specified
  active current (battery data sheet)
• Regulator output voltage (regulator data sheet)
• Regulator efficiency vs. input voltage (regulator
  data sheet)

Slide 23 of 32
24
Estimating Battery Life

• Example
• Using 2 AA alkaline batteries in series specified
  at 2530mAh, how long will the
• batteries last if a wireless sensor consumes 35mA
  at 3.3V while active and 75uA
• at 3.3V while asleep. The sensor sleeps for 30
  seconds between active durations
• of 25ms. (2 batteries in series double voltage
  while capacity remains constant)
• Answer
• Batteries must source more then 35mA (voltage
  conversion efficiency)
• Assume 90 efficiency from regulator average
  battery voltage of 2.7V
• 3.3V 35mA 2.7V i 90 -gt i 48mA
• During very low current draw regulators become
  less efficient
• Assume 68 efficiency from regulator using
  similar math to above -gt i 125uA
• Plug the values into the equations

Slide 24 of 32
25
Wireless Challenge 6 of 7

• Bit Error Rate (BER) Testing
• Drop-in
• No control
• Module and Discrete
• Testing required as external circuitry will add
  noise to the board causing reduced sensitivity
• This noise will likely be channel specific and is
  caused by resonant frequencies from crystals, MCU
  core speed, bus speed, etc.

Slide 25 of 32
26
Bit Error Rate Testing

• Why perform Bit Error Rate Testing
• Noise on your PCB from other components will
  cause the sensitivity to degrade from what is
  published on the data sheet
• Sensitivity may vary across the different
  channels
• How to perform Bit Error Rate Testing
• Measure the output of a transmitter and attenuate
  the signal to the sensitivity value that you hope
  to obtain (include all cable losses).
• Connect your device under test
• (DUT) as shown below
• If signals reach the DUT with less then 1
  (typical) BER, increase the attenuation if it is
  more then 1 BER, decrease the attenuation.

Slide 26 of 32
27
Duration of BER Testing

• BER accuracy is increased with duration of the
  test
• Confidence Level - Std Deviation
  Confidence Level - Std Deviation
• 80.00 1.2816 68.27 1
• 90.00 1.6449 95.45 2
• 95.00 1.9600 99.73 3
• 99.00 2.5758 99.994 4
• 99.90 3.2905 99.9999 5
• Example
• How many bit errors would you need to obtain
  before you can stop the test
• assuming your BER is accurate to within 5 with a
  confidence level of 99? How
• much time would this take at 5 samples / second
  and an expected BER of 1?
• Answer

Slide 27 of 32
28
Wireless Challenge 7 of 7

• Certification Testing (U.S., Canada Europe)
• Drop-in
• Complete (country specific)
• Module
• Intentional radiator testing complete (country
  specific)
• The following testing may still be required on
  your product
• Product Safety
• Unintentional emissions
• Immunity
• Discrete
• Certification testing required

Slide 28 of 32
29
Typical Certification Costs

• United States only FCC certification costs
• Intentional Radiator test costs - 2,500
• FCC submittal costs - 2,600
• Europe, U.S. Canada Total certification costs
• Intentional Radiator test costs - 4,500
• Total submittal costs - 7,150
• Assuming 2 unique models, NRTL cert mark, rates
  subject to change

Slide 29 of 32
30
Drop-in vs. Module vs. Discrete

• ? good ? ? better ? ? ? best
• All costs listed include BOM, development labor,
  certification and production costs. All costs
  are approximate only and do not reflect any
  specific Drop-in, Module or Discrete solution.
• System defined as 1 gateway and 4 wireless
  temperature sensors.

Slide 30 of 32
31
Winland Wireless Design Example

• EA800 Design Project

(EA800) (EA800
close up)

• Network solution Discrete RF IC
• Protocol 802.15.4
• Temperature Sensor BOM cost 16.45
• Includes plastics, PCB, antenna, electronics,
  etc
• EAU 1K

(Temperature Sensor)
Slide 31 of 32
32
Wireless Sensor Networks

• Questions?

Contact Information
David Liverseed Electrical Engineer E
dgliverseed_at_winland.com www.winland.com
Greg Burneske VP of Engineering E
gwburneske_at_winland.com www.winland.com
All registered and unregistered trademarks used
herein are the property of their respective
owners. Winland Electronics is not affiliated
with, or endorsed by, any of the owners of these
marks.
Slide 32 of 32