Transcript Mica Sensor Board Review - University of California, Berkeley

Mica Sensor Board Review
Alec Woo
November 21, 2001
NEST Meeting
What can we do with it?






Light sensor
Temperature sensor
Microphone sensor
2.6kHz Sounder
2 Axis Accelerometer
2 Axis Magnetometer
Board Layout

Size

2.25 x 1.25 square inches


2 Layer Board (Top and Bottom)


Same size as the Mica Board
No ground plane
One sided 51 pin connector

top in a stack of daughter cards
Sensor Physical Placements
Magnetometer (bottom)
Accelerometer (top)
Light
(top)
Temp
(top)
1.25 in
Sounder (top)
2.25 in
Microphone
(top)
Signal and Power Interface
Sensors or
Actuators
Light
Temperature
Sounder
Microphone
Mic. Bandpass
2 axis
Accelerometer
2 axis
Magnetometer
ADC
Power
Channel Control
ADC1
PW0
ADC2
PW1
PW2
ADC5,
PW3
ADC5,
PW3
ADC3,4 PW4
ADC6,7
PW5
MUX
Setting
INT3 = 1
INT3 = 0
Potentiomenter
Control Interface
• Same Interface as adjusting the Radio Signal Strength.
Gain
Adjustment
Microphone
Gain
Magnetometer
1st Axis
Magnetometer
2nd Axis
Resistor
INC
DEC
0 – 100k LED1 LED2
0-50k
LED1 LED2
0-50k
LED1 LED2
Select
PW6,
PW7
PW6,
PW7
PW6,
PW7
Alternative Control Interfaces
(ie. Bus Interface)

I2C

2 pins instead of 4 pins


I2C_BUS_1_CLK, I2C_BUS_1_DATA vs. LED1,LED2,PW6, PW7
Higher overhead



1 Wire


Only need 1 pin instead of 4 pins
Higher overhead




I2C commands vs. flipping pins
Chip addressing is hardwired
Basically implement a 1 wire protocol Bus Master
Every 1 wire chip has unique 64 bit addressing
Use a 1 wire Bus Master Chip with serial interface
In any case, at least one extra pin can be freed to
avoid using INT3 pin
Light Sensor



Clairex CL9P4L 10kohm
Voltage Divider Design
Same as before
Temperature Sensor

2 options

The usual one




Alternative






YY14406
10kohm, 0.2C accuracy, 0-75C
$6/unit
ERT-J1VR103J
Negative temperature thermistor
10kohm, 1C accuracy, -40 to 125C
$0.43/unit @1000 or $0.96/unit
Voltage Divider Design
Populate the one you want
Sounder








Piezoelectric
Resonant at 2.6kHz +/- 500Hz
85dB sound pressure
Weight 4grams
Diameter 29mm
$ pending from Taiwan
Circuit + sounder draws 1.5mA from
measurement
Sine Wave as Output waveform
Microphone



Panasonic WM-62A
500uA max
<5kHz is good as observed




20Hz to 16kHz from spec
omni directional
6mm in diameter
$2.12/unit @1k
Amplification and Filtering
Pre Amp

Pre Amp


159Hz – 6.4kHz
Amp
~4.7mA for circuit prototype
+ microphone
Adjustable Amplification = -1 to -101
Active Bandpass filter

2.6kHz +/- 500Hz
Active
Bandpass
Filter
mic_out
mic_bandpass_out
Amp


Amplification = -100
Passive RC


Passive RC
High pass and
Low pass filters
MIC_BANDPASS_OUT Signal

Biquad Active Filter


P. 278, The Art of Electronics
Tunable





center frequency(fo) and bandpass bandwidth (BW)
Bandpass bandwidth determines quality of the filter
Center frequency can shift while bandpass bandwidth
remains the same
fo = 1/2RFC
BW = 1/2RBC
Example of
Time of Flight Estimation
Sender

End of
RF Signal
t1
e.g. t1 = 865us for 1 foot
Peak Detection wrt periodicity identifies sounder
signal


Receiver
t2
Period of 2.7kHz = 370us first peak (1/4 period) = 92.5us
t1 = t2 – 92.5us + 



Sampling rate determines granularity of t2 and affects 
Uncertainties in frequency also affects 
Distance = (speed of sound) * t1
Accelerometer



2 Axis
ADXL202E
2mg resolution at 60Hz



600uA current consumption
Uses the raw analog output channel for
both X and Y


Filter capacitors set to be 50Hz bandwidth
Duty cycle output are not used but fanned out
$13.38/unit @1k
Magnetometer






2 Axis
HMC1002, $20 @100
+/-6 gauss (earth’s field +/-0.5 gauss)
27ug at 10Hz
~5mA
2 stage amplification





29 * 41 = 1189
Digital Pot to adjust 2nd stage amplification to avoid
railing on both axis
Amplifiers are not Rail-to-Rail (0.66V to 2.33V)
Added a virtual ground chip to give better
voltage reference/(roll back to voltage divider)
Fan out reset pins for demagnetizing the chip