Practical Temperature Measurements

1
Practical Temperature
Measurements
Hewlett-Packard Classroom Series

• 001

2
Agenda
Hewlett-Packard Classroom Series

• Background, history
• Mechanical sensors
• Electrical sensors
• Optical Pyrometer
• RTD
• Thermistor, IC
• Thermocouple
• Summary Examples

• A1

3
What is Temperature?
Hewlett-Packard Classroom Series

• A scalar quantity that determines the direction
  of heat flow between two bodies
• A statistical measurement
• A difficult measurement
• A mostly empirical measurement

002
4
How is heat transferred?
Hewlett-Packard Classroom Series

• Conduction

• Metal coffee cup

• Convection

• Radiation

• 003

5
The Dewar
Hewlett-Packard Classroom Series

• Glass is a poor conductor
• Gap reduces conduction
• Metallization reflects radiation
• Vacuum reduces convection

• 004

6
Thermal Mass
Hewlett-Packard Classroom Series

• Don't let the measuring device change the
  temperature of what you're measuring.
• Response time
• fThermal mass
• fMeasuring device

• 005

7
Temperature errors
Hewlett-Packard Classroom Series

• What is YOUR normal temperature?
• Thermometer accuracy, resolution
• Contact time
• Thermal mass of thermometer, tongue
• Human error in reading

97.6 98.6 99.6
36.5 37 37.5

• 006

8
History of temperature sensors
Hewlett-Packard Classroom Series

• 1600 ad

• 1700 ad

• Fahrenheit
• Instrument Maker
• 12896 points
• Hg Repeatable
• One standard scale

• Galileo First temp. sensor
• pressure-sensitive
• not repeatable

• Early thermometers
• Not repeatable
• No good way to calibrate

• 007

9
The 1700's Standardization
Hewlett-Packard Classroom Series

• 1700 ad

• 1800 ad

• Thomson effect
• Absolute zero

• Celsius
• Common, repeatable calibration reference points

• "Centigrade" scale

• 008

10
1821 It was a very good year
Hewlett-Packard Classroom Series

• 1800 ad

• 1900 ad

• The Seebeck effect

• Davy The RTD

• Pt 100 _at_ O deg.C

• 009

11
The 1900's Electronic sensors
Hewlett-Packard Classroom Series

• 1900 ad

• 2000 ad

• 1 uA/K

• Thermistor

• IC sensor

• IPTS 1990

• IPTS 1968

• "Degree Kelvin"gtgt "kelvins"
• "Centigrade"gtgt " Celsius"

• 010

12
Temperature scales
Hewlett-Packard Classroom Series
Absolute zero

• Celsius

100
0
-273.15

• Kelvin

0
273.15
373.15

• Fahrenheit

32
212
-459.67

• Rankine

0
671.67
427.67

• "Standard" is "better"
• Reliable reference points
• Easy to understand

• 011

13
IPTS '90 More calibration points
Hewlett-Packard Classroom Series

• 273.16 TP H2O

• 1357.77 FP Cu

• 234.3156 TP Hg

• 1337.33 FP Au

• 1234.93 FP Ag

Large gap

• 933.473 FP Al

• 692.677 FP Zn

• 83.8058 TP Ar

• 505.078 FP Sn

• 54.3584 TP O2

• 429.7485 FP In

• 24.5561 TP Ne

• 20.3 BP H2

• 17 Liq/vapor H2

• 302.9146 MP Ga

• 13.81 TP H2

• 3 to 5 Vapor He

• 012

14
Agenda
Hewlett-Packard Classroom Series

• Background, history
• Mechanical sensors
• Electrical sensors
• Optical Pyrometer
• RTD
• Thermistor, IC
• Thermocouple
• Summary Examples

• A2

15
Bimetal thermometer
Hewlett-Packard Classroom Series

• Forces due to thermal expansion

• Two dissimilar metals, tightly bonded

• Result

• Bimetallic thermometer
• Poor accuracy
• Hysteresis
• Thermal expansion causes big problems in other
  designs
• IC bonds
• Mechanical interference

• 013

16
Liquid thermometer Paints
Hewlett-Packard Classroom Series

• Thermally-sensitive paints
• Irreversible change
• Low resolution
• Useful in hard-to-measure areas

• Liquid-filled thermometer
• Accurate over a small range
• Accuracy resolution f(length)
• Range limited by liquid
• Fragile
• Large thermal mass
• Slow

• 014

17
Agenda
Hewlett-Packard Classroom Series

• Background, history
• Mechanical sensors
• Electrical sensors
• Optical Pyrometer
• RTD
• Thermistor, IC
• Thermocouple
• Summary Examples

• A3

18
Optical Pyrometer
Hewlett-Packard Classroom Series

• Infrared Radiation-sensitive
• Photodiode or photoresistor
• Accuracy femissivity
• Useful _at_ very high temperatures
• Non-contacting
• Very expensive
• Not very accurate

• 015

19
Agenda
Hewlett-Packard Classroom Series

• Background, history
• Mechanical sensors
• Electrical sensors
• Optical Pyrometer
• RTD
• Thermistor, IC
• Thermocouple
• Summary Examples

• A4

20
Resistance Temperature Detector
Hewlett-Packard Classroom Series

• Most accurate stable
• Good to 800 degrees Celsius
• Resistance fAbsolute T
• Self-heating a problem
• Low resistance
• Nonlinear

• 016

21
RTD Equation
Hewlett-Packard Classroom Series

• R 100 Ohms _at_ O C
• Callendar-Van Deusen Equation

• RRo(1aT) - Ro(ad(.01T)(.01T-1))
• Ro100 _at_ O C
• a 0.00385 / - C
• d 1.49

For TgtOC
for Pt
R
300 200 100
Nonlinearity
T

• 0 200 400 600 800

017
22
Measuring an RTD 2-wire method
Hewlett-Packard Classroom Series
Rlead
Rx

100
V
I ref 5 mA
Rlead
Pt
-

• R Iref(Rx 2 Rlead)
• Error 2 /.385 more than 5 degrees C for 1
  ohm Rlead!

• d

• Self-heating
• For 0.5 V signal, I 5mA P.5.0052.5
  mwatts
• _at_ 1 mW/deg C, Error 2.5 deg C!

• Moral Minimize Iref Use 4-wire method
• If you must use 2-wire, NULL out the lead
  resistance

018
23
The 4-Wire technique
Hewlett-Packard Classroom Series
Rx

100
V
Rlead1
I ref 5 mA
-

• R Iref Rx
• Error not a function of R in source or sense
  leads
• No error due to changes in lead R
• Twice as much wire
• Twice as many scanner channels
• Usually slower than 2-wire

• 019

24
Offset compensation
Hewlett-Packard Classroom Series
Voffset

100
V
I ref (switched)
-

• Eliminates thermal voltages
• Measure V without I applied
• Measure V I applied

With
V
R
I

• 020

25
Bridge method
Hewlett-Packard Classroom Series
100

• High resolution (DMM stays on most sensitive
  range)
• Nonlinear output
• Bridge resistors too close to heat source

• 021

26
3-Wire bridge
Hewlett-Packard Classroom Series
100
1000
Rlead 1
3-Wire PRTD
Sense wire
1000
Rlead 2
100

• Keeps bridge away from heat source
• Break DMM lead (dashed line) connect to RTD
  through 3rd "sense" wire
• If Rlead 1 Rlead 2, sense wire makes error small
• Series resistance of sense wire causes no error

• 022

27
Agenda
Hewlett-Packard Classroom Series

• Background, history
• Mechanical sensors
• Electrical sensors
• Optical Pyrometer
• RTD
• Thermistor, IC
• Thermocouple
• Summary Examples

• A5

28
Electrical sensors Thermistor
Hewlett-Packard Classroom Series
Rlead1

V
5k
I 0.1 mA
Rlead1
-

• Hi-Z Sensitive 5 k _at_ 25C R 4/deg C

• Limited range

• 2-Wire method R I (Rthmr 2Rlead)
• Lead R Error 2 /400 0.005 degrees C
• Low thermal mass High self-heating
• Very nonlinear

023
29
I.C. Sensor
Hewlett-Packard Classroom Series
AD590
I 1 uA/K

• High output
• Very linear
• Accurate _at_ room ambient
• Limited range
• Cheap

100
-
5V
1mV/K
960

• d

024
30
Summary Absolute T devices
Hewlett-Packard Classroom Series

• 025

31
Agenda
Hewlett-Packard Classroom Series

• Background, history
• Mechanical sensors
• Electrical sensors
• Optical Pyrometer
• RTD
• Thermistor, IC
• Thermocouple
• Summary Examples

• A6

32
Thermocouples The Gradient Theory
Hewlett-Packard Classroom Series

• The WIRE is the sensor, not the junction
• The Seebeck coefficient (e) is a function of
  temperature

• 026

33
Making a thermocouple
Hewlett-Packard Classroom Series

• Two wires make a thermocouple
• Voltage output is nonzero if metals are not the
  same

Ta
e dT
B
A
Tx

• 027

34
Gradient theory also says...
Hewlett-Packard Classroom Series

• If wires are the same type, or if there is one
  wire, and both ends are at the same temperature,
  output Zero.

Ta
e dT 0
A
A
Tx

• 028

35
Now try to measure it
Hewlett-Packard Classroom Series

• Theoretically, Vab fTx-Tab

• But, try to measure it with a DMM

• Result 3 unequal junctions, all at unknown
  temperatures

029
36
Solution Reference Thermocouple
Hewlett-Packard Classroom Series

• Problems a) 3 different thermocouples,
  b) 3 unknown temperatures

• Solutions a) Add an opposing thermocouple
  b) Use a known reference temp.

Isothermal block

• 030

37
The Classical Method
Hewlett-Packard Classroom Series
Cu
Fe

• If both Cu junctions are at same T, the two
  "batteries" cancel
• Tref is an ice bath (sometimes an electronic ice
  bath)
• All T/C tables are referenced to an ice bath
• V fTx-Tref

Tx
Con
V
Tref 0 C
o
Cu
Fe

• Question How can we eliminate the ice bath?

• 031

38
Eliminating the ice bath
Hewlett-Packard Classroom Series

• Don't force Tref to icepoint, just measure it
• Compensate for Tref mathematicallyVf Tx
  - Tref
• If we know Tref , we can compute Tx.

Tice
Tref
Tice
Tice

• 032

39
Eliminating the second T/C
Hewlett-Packard Classroom Series

• Extend the isothermal block
• If isothermal, V1-V20

Fe
Cu
Tref
Tx
V
Con
Tref
Cu

• 033

40
The Algorithm for one T/C
Hewlett-Packard Classroom Series

• Measure Tref RTD, IC or thermistor
• Tref gt Vref _at_ O C for Type J(Fe-C)
• Know V, Know Vref Compute Vx
• Solve for using Vx

o
Tx
Compute VxVVref
Vx
V
Vref
Tx
0
Tref
o
034
41
Linearization
Hewlett-Packard Classroom Series
V
Small sectors
Tx
T
0
Tref
o
2
9
3

• Polynomial Ta a V a V a V .... a V
• Nested (faster) Ta V(a V(a V(a
  .......)))))))))
• Small sectors (faster) TT bVcV
• Lookup table Fastest, most memory

0
1
2
3
9
0
1
2
3
2
0

• 035

42
Common Thermocouples
Hewlett-Packard Classroom Series
Platinum T/Cs
Base Metal T/Cs

• All have Seebeck coefficients in MICROvolts/deg.C

• 036

43
Common Thermocouples
Hewlett-Packard Classroom Series
Seebeck Coeff uV/C
Type
Metals
Fe-Con Ni-Cr Cu-Con Pt/Rh-Pt Ni/Cr-Con Ni/Cr/Si-Ni
/Si

• Microvolt output is a tough measurement
• Type "N" is fairly new.. more rugged and higher
  temp. than type K, but still cheap

J K T S E N
50 40 38 10 59 39

• 037

44
Extension Wires
Hewlett-Packard Classroom Series

• Possible problemhere

Large extension wires
Small diametermeasurementwires

• Extension wires are cheaper, more rugged, but not
  exactly the same characteristic curve as the T/C.
• Keep extension/TC junction near room temperature
• Where is most of the signal generated in this
  circuit?

• 038

45
Noise DMM Glossary
Hewlett-Packard Classroom Series

• Normal Mode In series with input
• Common Mode Both HI and LOterminals driven
  equally

• 039

46
Generating noise
Hewlett-Packard Classroom Series
Normal Mode

• Large surface area, high Rlead Max. static
  coupling
• Large loop area Max. magnetic coupling

• Large R lead, small R leak Max.common mode noise

• 040

47
Eliminating noise
Hewlett-Packard Classroom Series
Normal Mode
dc SIGNAL

• Filter, shielding, small loop area(Caution
  filter slows down the measurement)

• Make R leak close to

• 041

48
Magnetic Noise
Hewlett-Packard Classroom Series

• Magnetic coupling

DMM InputResistance
Induced I

• Minimize area
• Twist leads
• Move away from strong fields

• 042

49
Reducing Magnetic Noise
Hewlett-Packard Classroom Series

• Equal and opposite induced currents

DMM InputResistance

• Even with twisted pair
• Minimize area
• Move away from strong fields

• 043

50
Electrostatic noise
Hewlett-Packard Classroom Series
AC Noise source
Stray capacitances
DMM InputResistance
Inoise
Stray resistances

• Stray capacitance causes I noise
• DMM resistance to ground is important

• 044

51
Reducing Electrostatic Coupling
Hewlett-Packard Classroom Series

• 045

52
A scanning system for T/Cs
Hewlett-Packard Classroom Series

• One thermistor, multiple T/C channels
• Noise reduction
• CPU linearizes T/C
• DMM must be very high quality

• OHMs
• Conv.

Isolators
HI
uP
uP
ToComputer
ROM Lookup
LO
Integrating A/D
Floating Circuitry
Grounded Circuitry
046
53
Errors in the system
Hewlett-Packard Classroom Series
Ref. Block Thermal gradient
T/C Calibration Wire errors
Thermal emf
Ref. Thermistor cal, linearity
Reference Thermistor Ohms measurement
Extension wire junction error
Linearization algorithm

• OHMs
• Conv.

Isolators
HI
uP
uP
ROM Lookup
LO
Integrating A/D
Floating Circuitry
Grounded Circuitry
047
DMM offset, linearity, thermal emf, noise
54
Physical errors
Hewlett-Packard Classroom Series

• Shorts, shunt impedance
• Galvanic action
• Decalibration

• Sensor accuracy
• Thermal contact
• Thermal shunting

• 048

55
Physical Errors
Hewlett-Packard Classroom Series

• Hot spot causes shunt Z, meter shows the WRONG
  temperature

• Water droplets cause galvanic action huge
  offsets

• Exceeding the T/C's range can cause permanent
  offset
• Real T/C's have absolute accuracy of 1 deg C _at_
  25C Calibrate often and take care

• 049

56
Physical error Thermal contact
Hewlett-Packard Classroom Series
Surface probe

• Make sure thermal mass is much smallerthan that
  of object being measured

• 050

57
Physical errors Decalibration
Hewlett-Packard Classroom Series
350 C
300 C
975 C
200 C
1000 C
100 C
This section produces theENTIRE signal

• Don't exceed Tmax of T/C
• Temp. cycling causes work-hardening,decalibration
  
• Replace the GRADIENT section

• 051

58
Agenda
Hewlett-Packard Classroom Series

• Background, history
• Mechanical sensors
• Electrical sensors
• Optical Pyrometer
• RTD
• Thermistor, IC
• Thermocouple
• Summary Examples

• A7

59
The basic 4 temperature sensors
Hewlett-Packard Classroom Series
Thermocouple

• Wide variety
• Cheap
• Wide T. range
• No self-heating

• Hard to measure
• Relative T. only
• Nonlinear
• Special connectors

• 052

60
Summary
Hewlett-Packard Classroom Series

• Innovation by itself is not enough...you must
  develop standards
• Temperature is a very difficult, mostly
  empirical measurement
• Careful attention to detail is required

• 053

61
Examples
Hewlett-Packard Classroom Series

• 054