Musical Chairs and Magic Carpets - PowerPoint PPT Presentation

About This Presentation
Title:

Musical Chairs and Magic Carpets

Description:

MAS836 Sensor Technologies for Interactive Environments Second Nature Sensor Conditioning Electronics – PowerPoint PPT presentation

Number of Views:114
Avg rating:3.0/5.0
Slides: 40
Provided by: JoeP74
Category:

less

Transcript and Presenter's Notes

Title: Musical Chairs and Magic Carpets


1
MAS836 Sensor Technologies for Interactive
Environments
Second Nature Sensor Conditioning Electronics
2
Reference Sources
  • Jacob Fraden
  • AIP Handbook of Modern Sensors, gt2nd Edition
  • Ramon Pallas-Areny and John G. Webster
  • Sensors and Signal Conditioning, 2nd Edition
  • Thomas Petruzzellis (getting old)
  • The Alarm, Sensor, Security Cookbook

3
Auxilary References (signals)
  • Ramon Pallas-Areny John G. Webster
  • Analog Signal Processing
  • Paul Horowitz Winifield Hill
  • The Art of Electronics
  • Don Lancaster (online sources)
  • Active Filter Cookbook

4
Magazines
  • Sensors Magazine - Free!
  • Circuit Cellar - Best EE-hacker magazine out
  • NASA Tech Briefs - Free! (still there?)
  • Test and Measurement - Free!
  • IEEE Sensors Journal

5
Websites
  • http//www.sensorsportal.com/
  • References, hints, sources
  • http//www.sensorsmag.com/
  • Sensors Magazine site
  • Buyers guide, Archive articles
  • http//www.cs.cmu.edu/chuck/robotpg/robofaq/10.ht
    ml
  • Robotics sites often list sensor vendors, hints
  • http//www.billbuxton.com/InputSources.html
  • Bill Buxtons encyclopedia on input devices

6
Some Classic Sensor Module Sources
  • http//www.parallax.com/
  • http//www.sparkfun.com/
  • http//www.ramseyelectronics.com/
  • http//www.adafruit.com/

7
Basic Sources for Electronics
  • Digikey - www.digikey.com
  • Mouser - www.mouser.com
  • Newark
  • Allied
  • Hosfelt Electronics
  • JameCo
  • Mat Electronics
  • JDR
  • All Electronics
  • Radio Shack (mainly online now)

8
Trading Modality
  • Sensor modes are intrinsically synaesthetic
  • Use physics and constraints to couple a measured
    quantity into an unknown
  • Temperature can infer wind velocity (heat loss)
  • Displacement can infer
  • Pressure (with an elastomer or spring F kx)
  • Volume of fluid in a tank (V Ah)
  • Velocity (2 measurements at different times v
    dx/dt)
  • Temperature (thermometer level)
  • Angle from vertical (displacement of a bubble)
  • Measurements are used with a mathematical model
    to derive other parameters
  • Estimation and Kalman Filtering, etc.
  • Not covered here...

9
Signal Conditioning
Zo
i
Zi
Zo
Io
Zi
i
Wants Low Zi
Wants High Zi
Vo
  • Sensors produce different kinds of signals
  • Voltage output or current output
  • Cant necessarily take sensor output and put
    right into microprocessor ADC or logic input
  • Signal may need
  • High-to-low impedance buffer, current-to-voltage
    conversion, gain, detection, filtering,
    discrimination...

10
The Comparator
  • Makes an analog signal into a 1-bit digital
    signal
  • Directly drives logic pin on microprocessor
  • Detects when signal is above threshold

11
The Schmidt Trigger
Deadband
  • Suppresses jitter and spurious triggering from
    noisy signals
  • Deadband thresholds, V and V-, can be calculated
    via superposition
  • Ground VIN, and with Rf and Ri as a voltage
    divider on Vout , calculate the voltage at the
    OpAmps noninverting pin
  • Note that this assumes a low-impedance VIN
    (source impedance sums with Ri)

T
T
12
Negative Feedback
  • Transimpedance Amplifier
  • Voltage Follower
  • Non-Inverting Amplifier
  • Inverting Amplifier
  • Inverting Summer

13
The Voltage Follower
  • A unity-gain buffer to enable high-impedance
    sources to drive low-impedance loads

14
The Non-Inverting Amplifier
  • Like voltage follower, but gives voltage gain
  • Gain can be adjusted from unity upward via
    resistor ratio
  • High-Z input is good for conditioning High-Z
    sensors

15
The Transimpedance Amplifier
  • Converts a current into a voltage
  • Generates a proportional (w. Rf) voltage from an
    input current
  • Produces a low-impedance output that can drive a
    microcomputers A-D converter, for example

16
The Inverting Amplifier
  • Inverts signal, voltage gain varies from zero
    upward with the ratio of two resistors
  • Extension to summer is trivial with additional
    Ris
  • Input impedance is not infinite Zin Ri

17
The Summing Amplifier
  • No crosstalk between inputs because of virtual
    ground

18
Biasing
  • AC Coupling
  • Biasing noninverting input
  • Biasing at inverting input

Buffer the voltage dividers output and use it
everywhere...
19
Biasing an entire circuit with a Buffered Voltage
Bias Buffer
AC Coupling Capacitor
X10 inverting stage
X10 inverting stage
X11 noninverting stage
A 60 dB (x1100) high-impedance, AC-Coupled
amplifier with bias made from a quad OpAmp
AC Coupling Capacitor (decouple accumulated
offset errors)
20
The Simple Differential Amplifier
  • Subtracts two input signals
  • Input resistors must be equal, feedback and shunt
    resistors must be equal
  • Provides voltage gain
  • The input impedances arent equal, however
  • The amplifier is unbalanced!
  • A high-impedance sensor will produce common-mode
    errors (e.g., the system will be sensitive to the
    common voltage)
  • Differential sensors will be more sensitive to
    induced pickup signals (which tend to be high
    impedance)

21
The Basic Instrumentation Amplifier
  • Buffer each leg of the differential amplifier by
    a voltage follower
  • Impedance is now extremely high at both inputs
  • Impedance can be set by a shunt resistor across
    inputs
  • This is a balanced instrumentation amplifier

22
The Three-OpAmp Instrumentation Amplifier
  • Gain is varied by changing only one resistor, R1
  • No need to re-trim other components for a gain
    change
  • Gain at first stages is better for signal/noise
  • This is the instrumentation amplifier of choice

23
Commercial Instrumentation Amplifiers
INA2321 500 kHz, 94 dB CMRR, R-R, µA sleep
  • Analog Devices AD623
  • Analog Devices AD AMP01
  • BurrBrown (TI) INA series (INA2321)
  • TI TLC271

Can be fairly slow, but precise DC properties,
low drift, high gain, well matched
24
Passive RC Filters
  • Passive LP Filter RC network fc 1/(2pRC)
  • Passive HP filter RC network fc 1/(2pRC)

-3dB 0.707
25
Passive RC Filter Rolloff
Bode Plot Freq. Response as a log-log plot
Rolloff is 6 dB per Octave (2x)
20 dB per Decade (10x)
26
The First-Order Active High Pass Filter
  • Low impedance drive
  • Voltage gain via Rf/Ri

27
The First-Order Active Low Pass Filter
f
28
The Band-Select Filter
  • Cascaded high and low pass filters
  • Always follow high-pass with low-pass (noise)
  • Low-Pass cutoff needs to be below high-pass
    cutoff!
  • No Q, first-order rolloffs

29
Sallen-Key Filters Ref. Active Filter Cookbook
VCVS Filters
30
Multiple Feedback Bandpass
Single-OpAmp VCVS BP filter
31
Low Pass Filter Responses
Fr. Active Filter Cookbook
Response set by adjusting Rs and Cs
32
Or just run an applet
  • Analog Devices, etc.

http//www.analog.com/en/amplifiers-linear/product
s/dt-adisim-design-sim-tool/filter_wizard/resource
s/fca.html
33
Picking an OpAmp
High-Level Tree (AD) OLD
34
Picking a Particular OpAmp
Low-Level Tree (AD) OLD
35
Picking a Particular OpAmp
Interactive Parametric Search (AD) CURRENT
36
Sampling
  • Nyquist fin lt fs/2
  • Bandlimited (demodulation) sampling
  • Dfin lt fs/2
  • Loose absolute phase information
  • Dont know whether phase moves forward or
    backward
  • Quadrature sampling
  • Bandlimited sampling at t and a quarter-period
    later
  • Form the Analytic Signal
  • I.E., the Quadrature (complex) Amplitude
  • Can also do this with multipliers and quadrature
    demodulation
  • Synchronous undersampling for periodic signals

37
Peak Detector
Vs
t
Vo
t
Capacitor holds peaks! Need reset switch to
continue tracking
38
Pulse Stretcher
Vs
C
R
  • Resistor continually (and slowly) bleeds
    capacitor charge
  • Automatic reset
  • Tune time constant to match signal dynamics (so
    peaks are always followed)

Vo
e-t/RC
t
  • Enables lazy sampling to catch transients

39
Analog Multiplexers
Write a Comment
User Comments (0)
About PowerShow.com