Application Note 28

February 1988

Thermocouple Measurement
Jim Williams

Introduction Thermocouples in Perspective

In 1822, Thomas Seebeck, an Estonian physician, acci- Temperature is easily the most commonly measured
dentally joined semicircular pieces of bismuth and copper physical parameter. A number of transducers serve tem-
(Figure 1) while studying thermal effects on galvanic ar- perature measuring needs and each has advantages and
rangements. A nearby compass indicated a magnetic dis- considerations. Before discussing thermocouple-based
turbance. Seebeck experimented repeatedly with different measurement it is worthwhile putting these sensors in
metal combinations at various temperatures, noting relative perspective. Figure 2’s chart shows some common contact
magnetic field strengths. Curiously, he did not believe that temperature sensors and lists characteristics. Study reveals
electric current was flowing, and preferred to describe the thermocouple strengths and weaknesses compared to
effect as “thermo-magnetism.” He published his results in other sensors. In general, thermocouples are inexpensive,
a paper, “Magnetische Polarisation der Metalle und Erze wide range sensors. Their small size makes them fast and
durch Temperatur-Differenz” (see references). their low output impedance is a benefit. The inherent volt-
Subsequent investigation has shown the “Seebeck Effect” age output eliminates the need for excitation.
to be fundamentally electrical in nature, repeatable, and
quite useful. Thermocouples, by far the most common
transducer, are Seebeck’s descendants.

JUNCTION
•

COPPER BISMUTH

•
N JUNCTION

W E

S
COMPASS AN28 F01

Figure 1. The Arrangement for Dr. Seebeck’s Accidental

Discovery of “Thermo-Magnetism”

an28f

AN28-1
 RANGE OF SENSITIVITY AT SPEED IN

AN28-2
TYPE OPERATION 25°C ACCURACY LINEARITY STIRRED OIL SIZE PACKAGE COST COMMENTS
Thermocouples –270°C to Typically Less ±0.5°C with Poor Over Wide Typically 1 Sec. 0.02 In. Bead Metallic $1 to $50 Requires Reference. Low
(All Types) 1800°C Than 50μV/°C Reference Range, Better Some Types are Typical. 0.0005 Bead, Variety Depending Level Output Requires
Over ≈100°C Faster In. Units are of Probes On Type, Stable Signal Conditioning
Available Available Specifications Components
and Package
Thermistors –100°C to ≈5%/°C for ±0.1°C Standard ±0.2°C for 1 to 10 Sec. is Beads Can be as Glass, $2 to $10 Highest Temperature
and Thermistor 450°C Thermistors. from –40°C to Linearized Standard; 3ms Small as 0.005 Epoxy, Teflon for Standard Sensitivity of Any Common
Composites ≈0.5%/°C for 100°C; ±0.01°C Composite Units to 100ms Types In., But 0.04 to Encapsulated, Units. $10 Sensor. Special Units
Linearized Units from 0°C to 60°C Over 100°C are Available 0.1 In. is Typical. Metal to $350 for Required for Long-Term
Available Ranges “Flake” Types are Housing, Etc. High Precision Stability Above 100°C
Application Note 28

Only 0.001 In. Types and

Thick Specials
Platinum –250°C to Approximately ±0.1°C Readily Nearly Linear Typically 1/8 to 1/4 In. Glass, Epoxy, $25 to $1000 Sets Standard for Stability
Resistance Wire 900°C 0.5%/°C Available. ±0.01°C Over Large Several Typical. Smaller Ceramic, Depending On Over Long Term. Has Wider
in Precision Spans; Typically Seconds Sizes Available Teflon, Metal, Specs; Most Temperature Range Than
Standards—Lab Within 1° Over Etc. Industrial Thermistor, but Lower
Units 200°C Ranges Types Below Sensitivity
$100
Diodes and –270°C to –2.2mV/°C ±2°C to ±5°C Over Within 2° Over 1 to 10 Sec. is Standard Diode Glass, Metal Below 50¢. Require Individual
Transistors 175°C (Approx. –55°C to 125°C Operating Range Standard. Small and Transistor Cryogenic Calibration. Must be Driven
0.33%/°C) Diode Packages Case Sizes. Glass Units More from Current Source for
Permit Speeds Passivated Chips Expensive Optimum Performance.
in ms Range Permit Extremely Extremely Inexpensive.
Small Sizes Calibrated Cryogenic Types
Available
Integrated Circuit –85°C to 0.4%/°C Typical Over –55°C to Within 1° (0.2° Several TO-18 Transistor Metal, Plastic $1 to $10 Current and Voltage Outputs
125°C 125°C from 0°C to Seconds Package Size. Available
Typical 70°C) Typical Also MiniDIP

Figure 2. Characteristics of Some Contact Temperature Sensors (Chart Adapted from Reference 2)

an28f
 Application Note 28
APPROXIMATE APPROXIMATE
SENSITIVITY IN VOLTAGE SWING OVER
JUNCTION MATERIALS μV/°C AT 25°C USEFUL TEMPERATURE RANGE (°C) RANGE LETTER DESIGNATION
Copper—Constantan 40.6 –270 to 600 25.0mV T
Iron—Constantan 51.70 –270 to 1000 60.0mV J
Chromel—Alumel 40.6 –270 to 1300 55.0mV K
Chromel—Constantan 60.9 –270 to 1000 75.0mV E
Platinum 10%—Rhodium/Platinum 6.0 0 to 1550 16.0mV S
Platinum 13%—Rhodium/Platinum 6.0 0 to 1600 19.0mV R

Figure 3. Temperature vs Output for Some Thermocouple Types

Signal Conditioning Issues 0°C in an ice bath. Ice baths, while inherently accurate, are
impractical in most applications. Another approach servo
Potential problems with thermocouples include low level
controls a Peltier cooler, usually at 0°C, to electronically
outputs, poor sensitivity and nonlinearity (see Figures 3
and 4). The low level output requires stable signal condi- simulate the ice bath (Figure 6). This approach* eliminates
ice bath maintenance, but is too complex and bulky for
tioning components and makes system accuracy difficult
most applications.
to achieve. Connections (see Appendix A) in thermocouple
systems must be made with great care to get good accuracy. *A practical example of this technique appears in LTC Application Note
AN-25, “Switching Regulators for Poets.”
Unintended thermocouple effects (e.g., solder and copper
create a 3μV/°C thermocouple) in system connections
+ VOUTPUT =
make “end-to-end” system accuracies better than 0.5°C MEASUREMENT VMEASUREMENT –
THERMOCOUPLE VCOLDJUNCTION
difficult to achieve. –

0 KmSCALE 0
– +
ERROR FOR TYPE E AND T (°C)

ERROR FOR TYPE J AND K (°C)

2.5 1
“COLD
ICE BATH
5.0 2 JUNCTION”
(0°C)
JmSCALE THERMOCOUPLE
7.5 3
AN28 F05

10.0 4
12.5 SCALEkE 5 Figure 5. Ice Bath Based Cold Junction Compensator
15.0 6
17.5 7
+V
20.0 SCALEkT 8

0 50 100 150 200 250 300 350 400 +

TEMPERATURE (°C) SERVO
AN28 F04
AMPLIFIER
–
Figure 4. Thermocouple Nonlinearity for Types J, K, E and T Over TEMPERATURE
0°C to 400°C. Error Increases Over Wider Temperature Ranges SENSOR MATED TO +V
PELTIER COOLER POWER
STAGE
– +
Cold Junction Compensation AN28 F06

PELTIER
COOLER
The unintended, unwanted and unavoidable parasitic ther-
mocouples require some form of temperature reference
–
for absolute accuracy. (See Appendix A for a discussion MEASUREMENT VOUTPUT =
on minimizing these effects). In a typical system, a “cold THERMOCOUPLE + VMEASUREMENT –
VCOLDJUNCTION
junction” is used to provide a temperature reference
(Figure 5). The term “cold junction” derives from the Figure 6. A 0°C Reference Based on Feedback Control of a
historical practice of maintaining the reference junction at Peltier Cooler (Sensor is Typically a Platinum RTD)
an28f

AN28-3
Application Note 28
AMBIENT Figure 8 shows a monolithic cold junction compensator
TEMPERATURE
COLD SENSOR IC, the LT®1025. This device measures ambient (e.g., cold
THERMOCOUPLE
JUNCTION junction) temperature and puts out a voltage scaled for
use with the desired thermocouple. The low supply cur-
TEMPERATURE COLD
JUNCTION COMPENSATED
rent minimizes self-heating, ensuring isothermal operation
TO BE
MEASURED COMPENSATION OUTPUT with the cold junction. It also permits battery or low power
CIRCUITRY
operation. The 0.5°C accuracy is compatible with overall
THERMOCOUPLE COPPER AN28 F07

WIRES WIRES achievable thermocouple system performance. Various

(e.g., IRON-CONSTANTAN, ETC.) compensated outputs allow one part to be used with many
Figure 7. Typical Cold Junction Compensation Arrangement. thermocouple types. Figure 9 uses an LT1025 and an
Cold Junction and Compensation Circuitry Must be Isothermal amplifier to provide a scaled, cold junction compensated
output. The amplifier provides gain for the difference
Figure 7 conveniently deals with the cold junction require- between the LT1025 output and the type J thermocouple.
ment. Here, the cold junction compensator circuitry does C1 and C2 provide filtering, and R5 trims gain. R6 is a
not maintain a stable temperature but tracks the cold typical value, and may require selection to accommodate
junction. This temperature tracking, subtractive term R5’s trim range. Alternately, R6 may be re-scaled, and R5
has the same effect as maintaining the cold junction at enlarged, at some penalty in trim resolution. Figure 10 is
constant temperature, but is simpler to implement. It is similar, except that the type K thermocouple subtracts from
designed to produce 0V output at 0°C and have a slope the LT1025 in series-opposed fashion, with the residue
equal to the thermocouple output (Seebeck coefficient) fed to the amplifier. The optional pull-down resistor allows
over the expected range of cold junction temperatures. readings below 0°C.
For proper operation, the compensator must be at the
same temperature as the cold junction. R1 R3 R4
10k 1M 10k
1% 1% R5 FULL-SCALE
E
60.9μV/°C + – 2k ADJUST
C1
VIN R6
J 0.01μF
51.7μV/°C 8.4k
+ V+ TYPE J
K, T
BUFFER 40.6μV/°C V+
10mV/°C – VIN –
10mV/°C R, S
TEMPERATURE
OUTPUT 6μV/°C R7 VOUT
SENSOR LT1025 LT1001
6.8k 10mV/°C
+ AN28 F09
J
C2
GND R–
0.01
V–
0.5°C ACCURACY GND VO RESISTOR
4V TO 36V OPERATION 10mV/°C COMMON AN28 F08

80μA SUPPLY CURRENT

COMPATIBLE WITH TYPE E, J, K, R, S AND T THERMOCOUPLES
AUXILIARY 10mV/°C OUTPUT Figure 9. LT1025 Cold Junction Compensates a Type J
Thermocouple. The Op Amp Provides the Amplified Difference
Figure 8. LT1025 Thermocouple Cold Junction Compensator Between the Thermocouple and the LT1025 Cold Junction Output

an28f

AN28-4
 Application Note 28
R2 current exceeds 500μA. These leads can generate both DC
100Ω R3
FULL-SCALE 255k and AC offset terms in the presence of thermal gradients
TRIM 1%
in the package and/or external air motion.
R1 C2
1k 0.1μF In many situations, thermocouples are used in high noise
1%
environments, and some sort of input filter is required.
V+ To reject 60Hz pick-up with reasonable capacitor values,
V+ – input resistors in the 10k to 100k range are needed. Under
LTKA0x
VOUT these conditions, bias current for the amplifier needs to
VIN 10mV/°C
K
+ AN28 F10 be less than 1nA to avoid offset and drift effects.
– + C1
LT1025 0.1μF
V– To avoid gain error, high open-loop gain is necessary
VO
GND R– V–
for single-stage thermocouple amplifiers with 10mV/°C
TYPE K *R4 ≤ . R4 IS NOT REQUIRED (OPEN)
R4*
30μA or higher outputs. A type K amplifier, for instance, with
FOR LT1025 TEMPERATURES ≥ 0°C
100mV/°C output, needs a closed-loop gain of 2,500. An
V–
ordinary op amp with a minimum loop of 50,000 would
Figure 10. LT1025 Compensates a Type K Thermocouple. The have an initial gain error of (2,500)/(50,000) = 5%! Although
Amplifier Provides Gain for the LT1025-Thermocouple Difference
closed-loop gain is commonly trimmed, temperature drift
Amplifier Selection of open-loop gain will have a deleterious effect on output
The operation of these circuits is fairly straightforward, accuracy. Minimum suggested loop gain for type E, J, K
although amplifier selection requires care. and T thermocouples is 250,000. This gain is adequate for
type R and S if output scaling is 10mV/°C or less.
Thermocouple amplifiers need very low offset voltage and
drift, and fairly low bias current if an input filter is used. Additional Circuit Considerations
The best precision bipolar amplifiers should be used for
Other circuit considerations involve protection and com-
type J, K, E and T thermocouples which have Seebeck
mon mode voltage and noise. Thermocouple lines are
coefficients to 40μV/°C to 60μV/°C. In particularly critical
often exposed to static and accidental high voltages,
applications, or for R and S thermocouples (6μV/°C to
necessitating circuit protection. Figure 11 shows two
15μV/°C), a chopper-stabilized amplifier is required. Linear
suggested approaches. These examples are designed
Technology offers two amplifiers specifically tailored for
to prevent excessive overloads from damaging circuitry.
thermocouple applications. The LTKA0x is a bipolar design
The added series resistance can serve as part of a filter.
with extremely low offset (30μV), low drift (1.5μV/°C), very
Effects of the added components on overall accuracy
low bias current (1nA), and almost negligible warm-up
should be evaluated. Diode clamping to supply lines is
drift (supply current is 400μA).
effective, but leakage should be noted, particularly when
For the most demanding applications, the LTC ®1052 large current limiting resistors are used. Similarly, IC bias
CMOS chopper-stabilized amplifier offers 5μV offset and currents combined with high value protection resistors
0.05μV/°C drift. Input bias current is 30pA, and gain is typi- can generate apparent measurement errors. Usually, a
cally 30 million. This amplifier should be used for R and S favorable compromise is possible, but sometimes the
thermocouples, especially if no offset adjustments can be circuit configuration will be dictated by protection or noise
tolerated, or where a large ambient temperature swing is rejection requirements.
expected. Alternatively, the LTC1050, which has similar drift
and slightly higher noise can be used. If board space is at Differential Thermocouple Amplifiers
a premium, the LTC1050 has the capacitors internally. Figure 12a shows a way to combine filtering and full dif-
Regardless of amplifier type, for best possible performance ferential sensing. This circuit features 120dB DC common
dual-in-line (DIP) packages should be used to avoid mode rejection if all signals remain within the LTC1043
thermocouple effects in the kovar leads of TO-5 metal supply voltage range. The LTC1043, a switched-capaci-
can packages. This is particularly true if amplifier supply tor building block, transfers charge between the input
an28f

AN28-5
Application Note 28
+VS

RLIMIT

CIRCUITRY CIRCUITRY

OPTIONAL AN28 F11

ACCEPTABLE WHERE
FILTER
GROUND INTEGRITY IS OPTIONAL OPTIONAL
ASSURED OR FOR –VS
FILTER FILTER
BATTERY OPERATION
USEFUL WHERE GROUND INTEGRITY IS UNCERTAIN.
INCLUDING OPEN THERMOCOUPLE LINE

Figure 11. Input Protection Schemes

100Ω 255k*
+V
1k*
9.1k 0.1μF VIN
OUTPUT
VO
10mV/°C
LT1025
5V 470k
– GND R–
–15V
1/2 LTC1043 LTKA0x†
+
6 5

–5V
2

TYPE K 1μF 1μF * = METAL FILM

† = LTC1050 CAN BE USED

18 15
AN28 F12a
1M

16
0.01μF

Figure 12a. Full Differential Input Thermocouple Amplifiers

“flying” capacitor and the output capacitor. The LTC1043’s Isolated Thermocouple Amplifiers
commutating frequency, which is settable, controls rate In many cases, protection networks and differential
of charge transfer, and hence overall bandwidth. The dif- operation are inadequate. Some applications require
ferential inputs reject noise and common mode voltages continuous operation at high common mode voltages
inside the LTC1043’s supply rails. Excursions outside with severe noise problems. This is particularly true in
these limits require protection networks, as previously industrial environments, where ground potential differ-
discussed. As in Figure 9, an optional resistor pull-down ences of 100V are common. Under these conditions the
permits negative readings. The 1M resistor provides a thermocouple and signal conditioning circuitry must
bias path for the LTC1043’s floating inputs. Figure 12b, be completely galvanically isolated from ground. This
for use with grounded thermocouples, subtracts sensor requires a fully isolated power source and an isolated
output from the LT1025.

an28f

AN28-6
 Application Note 28
R2
100Ω R3
FULL-SCALE TRIM 255k

R1
1k 0.1μF

5V
–
VOUT
1/2 LTC1043 LTKA0x*
10mV/°C
+
6 5
TYPE J
–5V
2
*LTC1050 CAN BE USED
V+ 1μF 1μF

VIN 3

LT1025
J 18 15
GND R– AN28 F12b

16
0.01μF

Figure 12b

signal transmission path to the ground referred output. settles, the clamp value is stable. This stable clamp value
Thermocouple work allows bandwidth to be traded for DC represents A5’s thermocouple related information. Inverter
accuracy. With careful design, a single path can transfer I4 generates a clock delayed pulse (Trace C) which is fed
floating power and isolated signals. The output may be to A3, a sample-hold amplifier. A3 samples L1’s primary
either analog or digital, depending on requirements. winding clamp value. A4 provides gain scaling and the
Figure 13 shows an isolated thermocouple signal condi- LT1004 and associated components adjust offset. When
tioner which provides 0.25% accuracy at 175V common the clock pulse (Trace A) goes low, sampling ceases. When
mode. A single transformer transmits isolated power and Trace B’s stretched clock pulse goes low, the I5-I6 inverter
chain output (Trace D) is forced low by the 470k-75pF
data. 74C14 inverter I1 forms a clock (Trace A, Figure 14).
I2, I3 and associated components deliver a stretched pulse differentiator’s action. This turns on Q1, forcing substantial
to the 2.2k resistor (Trace B). The amplitude of this pulse energy into L1’s primary (Trace E). L1’s secondary (Trace
is stabilized because A1’s fixed output supplies 74C14 F) sees large magnetic flux. A2’s output (Trace G) moves
power. The resultant current through the 2.2k resistor drives as it attempts to maintain its loop. The energy is far too
L1’s primary (Trace E). A pulse appears at L1’s secondary great, however, and A2 rails. The excess energy is dumped
(Trace F, Q2’s emitter). A2 compares this amplitude with into the Pin 1-Pin 4 winding, placing a large current pulse
A5’s signal conditioned thermocouple voltage. To close (Trace H) into the 22μF capacitor. This current pulse occurs
its loop, A2’s output (Trace G) drives Q2’s base to force with each clock pulse, and the capacitor charges to a DC
L1’s secondary (Pins 3 to 6) to clamp at A5’s output value. voltage, furnishing the circuit’s isolated supply. When the
Q2 operates in inverted mode, permitting clamping action 470k-75pF differentiator times out, the I5-I6 output goes
high, shutting off Q1. At the next clock pulse the entire
even for very low A5 outputs. When L1’s secondary (Trace
cycle repeats.
F) clamps, its primary (Trace E) also clamps. After A2

an28f

AN28-7
AN28-8
15k

390k

+VREG
I1

0.01 I2 I3
330k 2.2k*

75pF +VREG 1.2M*

470k 4.7k +VISOL 4.7k (TYPICAL
I5 I6 6.2k SEE TEXT) 1k*
Q1 39pF 150pF
150pF
Application Note 28

2N3906
–
2k A2 0.1
150pF 470k Q2
2N3904 1/2 LT1013
+VREG +
430k* 2 • 3
–15V
• +VISOL
100k*
15V 6 –
10k 430k* +VISOL
100k* 50mV • 1 (≈10V) A5
1/2 LT1013 0.1
–
TRIM +
+VREG A1 LT1004 22
20k 5 4
+
≈10.7V 1/2 LT1013 2.5V
+
15V +
–15V
I4 20k TYPE K
30k = FLOATING COMMON
VISOL –
8 L1 = PC-SSO-32 (UTC)
7 = 1N4148 VIN K
A3
+ LT398A
A4 = 74C14 LT1025
OUTPUT
0VDC TO 5VDC 1/2 LT1013
– –15V * = 1% FILM RESISTOR GND R–
4.99k
0.02

AN28 F13
20k
F. S. TRIM

Figure 13. 0.25% Thermocouple Isolation Amplifier

an28f
 Application Note 28
demodulated back to DC. I1 generates a clock pulse
A = 50V/DIV (Trace A, Figure 16). This pulse sets the 74C74 flip-flop
B = 50V/DIV
C = 50V/DIV
(Trace B) after a small delay generated by I2, I3 and associ-
D = 50V/DIV ated components. Simultaneously, I4, I5 and Q1 drive L1’s
E = 10V/DIV primary (Trace C). This energy, received by L1’s secondary
F = 10V/DIV (Trace H), is stored in the 47μF capacitor and serves as
the circuit’s isolated supply. L1’s secondary pulse also
G = 10V/DIV
clocks a closed-loop pulse width modulator composed
H = 50mA/DIV
of C1, C2, A3 and A4. A4’s positive input receives A5’s
HORIZ = 50μs/DIV
AN28 F14
LT1025-based thermocouple signal. A4 servo-biases C2
to produce a pulse width each time C1 allows the 0.003μF
Figure 14. Waveforms for Figure 13’s Thermocouple capacitor (Trace E) to receive charge via the 430k resistor.
Isolation Amplifier
C2’s output width is inverted by I6 (Trace F), integrated to
Proper operation of this circuit relies on several con- DC by the 47k-0.68μF filter and fed back to A4’s negative
siderations. Achievable accuracy is primarily limited by input. The 0.68μF capacitor compensates A4’s feedback
transformer characteristics. Current during the clamp loop. A4 servo controls C2 to produce a pulse width that
interval is kept extremely low relative to transformer is a function of A5’s thermocouple related output. I6’s
core capacity. Additionally, the clamp period must also low loss MOS switching characteristics combined with
be short relative to core capacity. The clamping scheme A3’s supply stabilization ensure precise control of pulse
relies on avoiding core saturation. This is why the power width by A4. Operating frequency, set by the I1 oscillator
refresh pulse occurs immediately after data transfer, and on L1’s primary side, is normally a stability concern, but
not before. The transformer must completely reset before ratios out because it is common to the demodulation
the next data transfer. A low clock frequency (350Hz) scheme, as will be shown.
ensures adequate transformer reset time. This low clock I6’s output width’s (Trace F) negative-going edge is dif-
frequency limits bandwidth, but the thermocouple data ferentiated and fed to I7. I7’s output (Trace G) drives Q3.
does not require any speed. Q3 puts a fast spike into L1’s secondary (Trace H). “Sing
Gain slope is trimmed at A5, and will vary depending upon around” behavior by C1 is gated out by the diode at C2’s
the desired maximum temperature and thermocouple type. positive input. Q3’s spike is received at L1’s primary, Pins 7
The “50mV” trim should be adjusted with A5’s output at and 3. Q2 serves as a clocked synchronous demodulator,
50mV. The circuit cannot read A5 outputs below 20mV pulling its collector low (Trace D) only when its base is
(0.5% of scale) due to Q2’s saturation limitations. high and its emitter is low (e.g., when L1 is transferring
data, not power). Q2’s collector spike resets the 74C74
Drift is primarily due to the temperature dependence of flip-flop. The MOS flip-flop is driven from a stable source
L1’s primary winding copper. This effect is swamped by (A1) and it is also clocked at the same frequency as the
the 2.2k series value with the 60ppm/°C residue partially pulse-width modulator. Because of this, the DC average
compensated by I3’s saturation resistance tempco. Over- of its Q output depends on A5’s output. Variations with
all tempco, including the LT1004, is about 100ppm/°C. supply, temperature and I1 oscillator frequency have no
Increased isolation voltages are possible with higher effect. A2 and its associated components extract the DC
transformer breakdown ratings. average by simple filtering. The 100k potentiometer per-
Figure 15’s thermocouple isolation amplifier is somewhat mits desired gain scaling. Because this scheme depends
more complex, but offers 0.01% accuracy and typical on edge timing at the flip-flop, the delay in resetting the
drift of 10ppm/°C. This level of performance is useful 0.003μF capacitor causes a small offset error. This term
in servo systems or high resolution applications. As in is eliminated by matching this delay in the 74C74 “set”
Figure 13, a single transformer provides isolated data and line with the previously mentioned I2-I3 delay network.
power transfer. In this case the thermocouple informa- This delay is set so that the rising edge of the flip-flop
tion is width modulated across the transformer and then output (Trace B) corresponds to I6’s rising edge. No such
an28f

AN28-9
 100k* 68k*

AN28-10
– –
A1 A3
1/2 LT1013 +VREG
10k 1/2 LT1013
3.8k* 100k
15V + + +VISOL
LT1034 LT1034 1.2M*
2.5V 2.5V (TYPICAL)
12k 1k* 430k
+
HP-5082-2810 A4 0.1
270k 33k 10k – 1k*
1/2 LT1013
C2
Application Note 28

+VREG 15V 1/2 LT1018 –

I6 10k +VISOL
330 100pF + 0.68
I7
Q1
47k
–
0.01 I1 +VISO 10k
15k A5
I4
510pF LT1006†
0.68 0.003 +
Q3 +VISOL
I5
I2 – 0.1
L1 +VISOL C1 +
2 1/2 LT1018
510pF • 47k
+ TYPE K
10k
6
•1 + 47k +VISOL
47μF
–
7
+VREG 5 VIN K
I3
3k 3•
+VREG = 1N4148 LT1025
0.015 L1 = PC-SSO-32 (UTC)
S +V R Q2 GND R–
74C74
1k = 2N3904
C D Q
15V AN28 F15

* = 1% METAL FILM TRW MAR-6

–
A2 = 74C14
OUT 1/2 LT1013 200k
+ 100k = FLOATING COMMON
0.68 † = A5 CAN ALSO BE LTC1050

Figure 15. 0.01% Thermocouple Isolation Amplifier

an28f
 Application Note 28
compensation is required for falling edge data because
A = 20V/DIV
circuit elements in this path (I7, Q3, L1 and Q2) are
B = 20V/DIV
wideband. With drift matched LT1034s and the specified
C = 10V/DIV
resistors, overall drift is typically 10ppm/°C with 0.01%
D = 20V/DIV linearity.
E = 2V/DIV
Digital Output Thermocouple Isolator
F = 10V/DIV
G = 10V/DIV Figure 17 shows another isolated thermocouple signal
H = 20V/DIV
AN28 F16
conditioner. This circuit has 0.25% accuracy and features
HORIZ = 50μs/DIV
a digital (pulse width) output. I1 produces a clock pulse
Figure 16. Pulse-Width-Modulation Based Thermocouple (Trace A, Figure 18). I2-I5 buffers this pulse and biases Q1
Isolation Amplifier Waveforms to drive L1. Concurrently, the 680pF-10k values provide a
differentiated spike (Trace B), setting the 74C74 flip-flop
(Trace C). L1’s primary drive is received at the secondary.

TYPE K

+VISOL

+ +VISO I2
A1 0.1
LT1006
K VIN I3
– 7.5k
LT1025 I4
15V 1.5M
0.1
1k* GND R–
330 I5 I1
Q1
1.2M* 15V 0.01

+VISOL 10k
+VISOL
–
I9 I10 I11 0.01 680pF
C1 WIDTH
LT1004 S Q
100k*
+
1/2 LT1017 5 •1 OUTPUT
2.5V

7
Q2 Q3 +VISOL 3
74C74
74C906
I8 I7 I6 33k
0.05 8• 4 15V
301k*
POLYSTYRENE
100pF Q4 2k C
+VISOL
(ALL SECTIONS + 1k R D
PARALLELED) 10μF
10k
L1 = PC-SSO-19 (UTC)
100k
= 1N4148
Q5
* = 1% METAL FILM
150pF
PNP = 2N3809 DUAL
15V AN28 F17
NPN = 2N3904

= 74C14

Figure 17. Digital Output Thermocouple Isolator

an28f

AN28-11
Application Note 28
is low. This condition occurs during data transfer, but not
A = 20V/DIV
during power transfer. The demodulated output (Trace H)
B = 20V/DIV
contains a single negative spike synchronous with C1’s
C = 20V/DIV (e.g., I11’s) output transition. This spike resets the flip-flop,
D = 0.05V/DIV providing the circuit output. The 74C74’s width output
E = 20V/DIV thus varies with thermocouple temperature.
F = 20V/DIV
G = 50V/DIV Linearization Techniques
H = 20V/DIV
AN28 F18
It is often desirable to linearize a thermocouple-based
HORZ = 50μs/DIV
signal. Thermocouples’ significant nonlinear response
Figure 18. Waveforms for Digital-Output Thermocouple Isolator requires design effort to get good accuracy. Four tech-
niques are useful. They include offset addition, breakpoints,
The 10μF capacitor charges to DC, supplying isolated
analog computation, and digital correction. Offset addi-
power. The pulse received at L1’s secondary also resets
tion schemes rely on biasing the nonlinear “bow” with a
the 0.05μF capacitor (Trace D) via the inverters (I6, I7,
constant term. This results in the output being high at
I8) and the 74C906 open-drain buffer. When the received
low scale and low at high scale with decreased errors
pulse ends, the 0.05μF capacitor charges from the Q2-Q3
between these extremes (Figure 19). This compromise
current source. When the resultant ramp crosses C1’s
reduces overall error. Typically, this approach is limited
threshold (A1’s thermocouple related output voltage)
to slightly nonlinear behavior over wide ranges or larger
C1 switches high, tripping the I9-I11 inverter chain. I11
nonlinearity over narrow ranges.
(Trace E) drives L1’s secondary via the 0.01μF capacitor
(Trace F). The 33k-100pF filter prevents regenerative Figure 20 shows a circuit utilizing offset linearization for
“sing around”. The resultant negative-going spike at L1’s a type S thermocouple. The LT1025 provides cold junc-
primary biases Q4, causing its collector (Trace G) to go tion compensation and the LTC1052 chopper-stabilized
low. Q4 and Q5 form a clocked synchronous demodulator amplifier is used for low drift. The type S thermocouple
which can pull the 74C74 reset pin low only when the clock output slope varies greatly with temperature. At 25°C it

R2
100Ω R3
FULL-SCALE 909k
TRIM 1%
R1
1k
1% R4 0.1μF
2.7k
ERROR BEFORE OFFSETTING
VH 15V

V+ 2 – 7
OUTPUT (V)

ERROR AFTER OFFSETTING

5 V+ 6 VOUT
OFFSET AMPLIFIER LTC1052* 10mV/°C
VIN 3 + 800°C TO 1200°C
SIMPLE R,S V– 8
AMPLIFIER – + R4 4
1μF 1.37M 1
LT1025
THERMOCOUPLE 1% 0.1μF 0.1μF
VL TYPE S R5
VO 10k
AN28 F20

GND R– R7 R6
OFFSET
750k 12k *LTC1050 CAN BE USED
0 TL T1/6 TM T5/6 TH TRIM
TEMPERATURE (°C)
LT1009
AN28 F19 –15V 2.5V

Figure 19. Offset Curve Fitting Figure 20. Offset-Based Linearization

an28f

AN28-12
 Application Note 28
is 6μV/°C, with an 11μV/°C slope at 1000°C. This circuit error is 0.75°C max because of the 0.1% gain resistors
gives 3°C accuracy over the indicated output range. The and, to a lesser extent, the output voltage tolerance of the
circuit, similar to Figure 10, is not particularly unusual LT1019A and the gain error of the LTC1091A. It may be
except for the offset term derived from the LT1009 and reduced by trimming the LT1019A or gain resistors. The
applied through R4. To calibrate, trim R5 for VOUT = 1.669 LTC1091A keeps linearity better than 0.15°C. The LTC1052’s
at VIN = 0.000mV. Then, trim R2 for VOUT = 9.998V at 5μV offset contributes negligible error (0.1°C or less).
T = 1000°C or for VIN (+ input) = 9.585mV. Combined errors are typically inside 0.5°C. These errors
Figure 21, an adaption of a configuration shown by She- don’t include the thermocouple itself. In practice, connec-
ingold (reference 3), uses breakpoints to change circuit tion and wire errors of 0.5°C to 1°C are not uncommon.
gain as input varies. This method relies on scaling of With care, these errors can be kept below 0.5°C.
the input and feedback resistors associated with A2-A6 The 20k-10k divider on CH1 of the LTC1091 provides low
and A7’s reference output. Current summation at A8 is supply voltage detection (the LT1019A reference requires
linear with the thermocouple’s temperature. A3-A6 are a minimum supply of 6.5V to maintain accuracy). Remote
the breakpoints, with the diodes providing switching location is possible with data transferred from the MCU
when the respective summing point requires positive to the LTC1091 via the 3-wire serial port.
bias. As shown, typical accuracy of 1°C is possible over a
Figure 24 is a complete software listing* of the code
0°C to 650°C sensed range.
required for the 68HC05 processor. Preparing the circuit
Figure 22, derived from Villanucci (reference 8), yields involves loading the software and applying power. No
similar performance but uses continuous function analog trimming is required.
computing to replace breakpoints, minimizing amplifiers *Including of a software-based circuit was not without attendant conscience
and resistors. The AD538 combines with appropriate scal- searching and pain on the author’s part. Hopefully, the Analog Faithful will
ing to linearize response. The causality of this circuit is tolerate this transgression ...I’m sorry everybody, it just works too well!
similar to Figure 22; the curve fit mechanism (breakpoint
vs continuous function) is the primary difference. References

Digital techniques for thermocouple linearization have 1. Seebeck, Thomas Dr., “Magnetische Polarisation der
become quite popular. Figure 23, developed by Guy M. Metalle und Erze durch Temperatur-Differenz”, Abhaand-
Hoover and William C. Rempfer, uses a microproces- lungen der Preussischen Akademic der Wissenschaften
sor fed from a digitized thermocouple output to achieve (1822-1823), pg. 265-373.
linearization. The great advantage of digital techniques is 2. Williams, J., “Designer’s Guide to Temperature Sensors”,
elimination of trimming. In this scheme a large number EDN, May 5, 1977.
of breakpoints are implemented in software.
3. Sheingold, D.H., “Nonlinear Circuits Handbook”, Analog
The 10-bit LTC1091A A/D gives 0.5°C resolution over a Devices, Inc., pg. 92-97.
0°C to 500°C range. The LTC1052 amplifies and filters the
4. “Omega Temperature Measurement Handbook”, Omega
thermocouple signal, the LT1025A provides cold junction
Engineering, Stamford Connecticut.
compensation and the LT1019A provides an accurate refer-
ence. The J type thermocouple characteristic is linearized 5. “Practical Temperature Measurements”, Hewlett-Pack-
digitally inside the processor. Linear interpolation between ard Applications Note #290, Hewlett-Packard.
known temperature points spaced 30°C apart introduces 6. Thermocouple Reference Tables, NBS Monograph 125,
less than 0.1°C error. The 1024 steps provided by the National Bureau of Standards.
LTC1091 (24 more that the required 1000) ensure 0.5°C
resolution even with the thermocouple curvature. 7. Manual on the Use of Thermocouples in Temperature
Measurement, ASTM Special Publication 470A.
Offset error is dominated by the LT1025 cold junction
compensator which introduces 0.5°C maximum. Gain 8. Villanucci, Robert S., “Calculator and IC Simplify
Linearization”, EDN, January 21, 1991.
an28f

AN28-13
Application Note 28
199k* R R

1k*
0.47μF –
1.0464R 0.65R
A2
+
15V –
R R 20.45R –
VIN A1
+ OUTPUT
E A8
10mV/°C
LT1025 2.174R – +
GND R
A3
TYPE E +

R R 12.74R

3.726R –
A4
+

R R 8.826R

7.918R –
A5
+

R R 11.111R

33.44R –
100k* 16.30R
A6
15V +
100k*
LT1021
10V
A7 AN28 F21

OP AMPS = 2s LT1014 QUAD

* = 0.1% METAL FILM
“R” ANNOTATED VALUES ARE IDEAL TARGET VALUES
R = 10k
= 1N4148

Figure 21. Breakpoint-Based Linearization (See Reference 3)

an28f

AN28-14
 Application Note 28
C1
THERMO- SEEBECK 0.1μF
COUPLE COEFFICIENT IC1
TYPE (μV/°C) PIN
E 60.9 1 R1 R2
J 51.7 8 1k 98.8k
K, T 40.6 7
R, S 5.95 6 V+
VR 2 – 7
R6
+ IC2 6 2 1k
VZ VX
LT1097 3
Tm + CHROMEL COPPER 3 + VT
VB R7 R4
(0°C TO 100Vm 10V 4
4 IC3 1k 16.2Ω
650°C) – CONSTANTAN COPPER V– 2V 5 AD538 14
V+ 6 13
2 V+
1 1μF 1N914
VCC E 12
+ VC
IC1 7 11 1V
LT1025C VC V– R5
8 10 180Ω
GND R– – 1μF VY
NOTES: AN28 F22
1. ALL FIXED RESISTORS 4 5 VY = 1.513V
R3
ARE METAL FILM
2. 150 < (R4 + R5) < 200
5k

10 V
VOUT = m Tm
°C

THERMOCOUPLE AMPLIFIER VOUT = 1.513 VT0.917

Figure 22. Continuous Function Linearization (See Reference 8)

9V
2 6
LT1019A-5
2 0.1μF
+
10μF
VIN
4
8
J LT1025A
–
J TYPE
20k
+ GND COMMON 1N4148
4 5
C0

3 7 LT1091A MC68HC05
+ CS VCC
1μF 6 47Ω
LTC1052* CH0 CLK SCK
2 8 1μF CH1 DOUT MIS0
– 4 GND DIN MOSI
1 10k
5V SS
0.1μF 0.1μF
AN28 F23

1k *LTC1050 CAN BE USED

0.33μF
0.1%
3.4k 178k
1% 0.1%

Figure 23. Processor-Based Linearization

an28f

AN28-15
Application Note 28
* TYPE J THERMOCOUPLE LINEARIZATION PROGRAM
* WRITTEN BY GUY HOOVER LINEAR TECHNOLOGY CORPORATION
* REV 1 10/4/87
* N IS NUMBER OF SEGMENTS THAT THERMOCOUPLE RESPONSE IS DIVIDED INTO
* TEMPERATURE (°C)=M•X+B
* M IS SLOPE OF THERMOCOUPLE RESPONSE FOR A GIVEN SEGMENT
* X IS A/D OUTPUT—SEGMENT END POINT
* B IS SEGMENT START POINT IN DEGREES (°C • 2)
ORG $1000
FDB $00,$39,$74,$B0,$EE,$12B,$193,$262,$330,$397 TABLE FOR X
ORG $1020
FDB $85DD,$823A,$7FB4,$7DD4,$7CAF,$7BC3,$7B8A,$7C24,$7C1F,$7B3A TABLE FOR M
ORG $1040
FDB $00,$3C,$78,$B4,$F0,$12C,$190,$258,$320,$384 TABLE FOR B
ORG $10FF
FCB $13 N•2
ORG $0100
OPT ]
STA $0A LOAD CONFIGURATION DATA INTO $0A
LDA #$00 CONFIGURATION DATA FOR PORT A DDR
STA $04 LOAD CONFIGURATION DATA INTO PORT A
LDA #$FF CONFIGURATION DATA FOR PORT B DDR
STA $05 LOAD CONFIGURATION DATA INTO PORT B
LDA #$F7 CONFIGURATION DATA FOR PORT C DDR
STA $06 LOAD CONFIGURATION DATA INTO PORT C
JSR HOUSEKP INITIALIZE ASSORTED REGISTERS
MES92L NPO
JSR CHECK
LDA #$6F DIN WORD FOR LTC1091 CH0, W/RESPECT TO GND, MSB FIRST
STA $50 STORE IN DIN BUFFER
JSR READ91 READ LTC1091
LINEAR LDX $10FF LOAD SEGMENT COUNTER INTO X
DOAGAIN LDA $1000,X LOAD LSBs OF SEGMENT N
STA $55 STORE LSBs IN $55
DECX DECREMENT X
LDA $1000,X LOAD MSBs OF SEGMENT N
STA $54 STORE MSBs IN $54
JSR SUBTRCT
BPL SEGMENT
JSR ADDB
DECX DECREMENT X
JMP DOAGAIN
SEGMENT LDA $1020,X LOAD MSBs OF SLOPE
STA $54 STORE MSBs IN $54
INCX INCREMENT X
LDA $1020,X LOAD LSBs OF SLOPE
STA $55 STORE LSBs IN $55
JSR TBMULT RETURNS RESULT IN $61 AND $62
LDA $1040,X LOAD LSBs OF BASE TEMP
STA $55 STORE LSBs IN $55
DECX DECREMENT X
LDA $1040,X LOAD MSBs OF BASE TEMP
STA $54
JSR ADDB
CHECK LDA #S7F DIN WORD FOR CH1
STA $50 LOAD DIN WORD INTO $50
JSR READ91 READ BATTERY VOLTAGE
LDA #$02 LOAD MSB OF MIN BATT VOLTAGE
STA $54 PUT IN MSB OF SUBTRACT BUFFER
LDA #$CC LOAD LSB OF MIN BATT VOLTAGE
STA $55 PUT IN LSB OF SUBTRACT BUFFER
JSR SUBTRCT COMPARE BATT VOLTAGE WITH MINIMUM
BPL NOPROB IF BATT OK GOTO NOPROB

Figure 24. Code for Processor-Based Linearization

an28f

AN28-16
 Application Note 28
JSR ADDB
LDA #$01
STA $56 SET BATTERY LOW FLAG
RTS
NOPROB JSR ADDB
CLR $56 CLEAR LOW BATTERY FLAG
RTS
READ91 LDA #$50 CONFIGURATION DATA FOR SPCR
STA $0A LOAD CONFIGURATION DATA
LDA $50
BCLR 2,$02 BIT 0 PORT C GOES LOW (CS GOES LOW)
STA $0C LOAD DIN INTO SP1 DATA REG. START TRANSFER
BACK91 TST $0B TEST STATUS OF SPIF
BPL BACK91 LOOP TO PREVIOUS INSTRUCTION IF NOT DONE
LDA $0C LOAD CONTENTS OF SPI DATA REG. INTO ACC
STA $0C START NEXT CYCLE
AND #$03 CLEAR 6 MSBs OF FIRST DOUT
STA $61 STORE MSBs IN $61
BACK92 TST $0B TEST STATUS OF SPIF
BPL BACK92 LOOP TO PREVIOUS INSTRUCTION IF NOT DONE
BSET 2,$02 SET BIT 0 PORT C (CS GOES HIGH)
LDA $0C LOAD CONTENTS OF SPI DATA INTO ACC
STA $62 STORE LSBs IN $62
RTS

SUBTRCT LDA $62 LOAD LSBs

SUB $55 SUBTRACT LSBs
STA $62 STORE REMAINDER
LDA $61 LOAD MSBs
SBC $54 SUBTRACT W/CARRY MSBs
STA $61 STORE REMAINDER
RTS
ADDB LDA $62 LOAD LSBs
ADD $55 ADD LSBs
STA $62 STORE SUM
LDA $61 LOAD MSBs
ADC $54 ADD W/CARRY MSBs
STA $61 STORE SUM
RTS
TBMULT CLR $68
CLR $69
CLR $6A
CLR $6B
STX $58 STORE CONTENTS OF X IN $58
LSL $62 MULTIPLY LSBs BY 2
ROL $61 MULTIPLY MSBs BY 2
LDA $62 LOAD LSBs OF LTC1091 INTO ACC
LDX $55 LOAD LSBs OF M INTO X
MUL MULTIPLY LSBs
STA $6B STORE LSBs IN $6B
STX $6A STORE IN $6A
LDA $62 LOAD LSBs OF LTC1091 INTO ACC
LDX $54 LOAD MSBs OF M INTO X
MUL
ADD $6A ADD NEXT BYTE
STA $6A STORE BYTE
TXA TRANSFER X TO ACC
ADC $69 ADD NEXT BYTE
STA $69 STORE BYTE
LDA $61 LOAD MSBs OF LTC1091 INTO ACC
LDX $55 LOAD LSBs OF M INTO X

Figure 24. Code for Processor-Based Linearization (Continued)

an28f

AN28-17
Application Note 28
MUL
ADD $6A ADD NEXT BYTE
STA $6A STORE BYTE
TXA TRANSFER X TO ACC
ADC $69 ADD NEXT BYTE
STA $69 STORE BYTE
LDA $61 LOAD MSBs OF LTC1091 INTO ACC
LDX $54 LOAD MSBs OF M INTO X
MUL
ADD $69 ADD NEXT BYTE
STA $69 STORE BYTE
TXA TRANSFER X TO ACC
ADC $68 ADD NEXT BYTE
STA $68 STORE BYTE
LDA $6A LOAD CONTENTS OF $6A INTO ACC
BPL NNN
LDA $69 LOAD CONTENTS OF $69 INTO ACC
ADD #$01 ADD 1 TO ACC
STA $69 STORE IN $69
LDA $68 LOAD CONTENTS OF $68 INTO ACC
ADC #$00 FLOW THROUGH CARRY
STA $68 STORE IN $68
NNN LDA $68 LOAD CONTENTS OF $68 INTO ACC
STA $61 STORE MSBs IN $61
LDA $69 LOAD CONTENTS OF $69 INTO ACC
STA $62 STORE IN $62
LDX $58 RESTORE X REGISTER
RTS RETURN
HOUSEKP BSET 0,$02 SET B0 PORT C
BSET 2,$02 SET B2 PORT C
RTS

Figure 24. Code for Processor-Based Linearization (Continued)

APPENDIX A Thermocouple voltages are generated whenever dis-

similar materials are joined. This includes the leads of
Error Sources in Thermocouple Systems IC packages, which may be kovar in TO-5 cans, alloy 42
Obtain good accuracy in thermocouple systems mandates or copper in dual-in-line packages, and a variety of other
care. The small thermocouple signal voltages require materials in plating finishes and solders. The net effect
careful consideration to avoid error terms when signal of these thermocouples is “zero” if all are at exactly the
processing. In general, thermocouple system accuracy same temperature, but temperature gradients exist within
better than 0.5°C is difficult to achieve. Major error sources IC packages and across PC boards whenever power is
include connection wires, cold junction uncertainties, dissipated. For this reason, extreme care must be used to
amplifier error and sensor placement. ensure that no temperature gradients exist in the vicinity
of the thermocouple terminations, the cold junction com-
Connecting wires between the thermocouple and con- pensator (e.g., LT1025) or the thermocouple amplifier. If a
ditioning circuitry introduce undesired junctions. These gradient cannot be eliminated, leads should be positioned
junctions form unintended thermocouples. The number of isothermally, especially the LT1025 R– and appropriate
junctions and their effects should be minimized, and kept output pins, the amplifier input pins, and the gain setting
isothermal. A variety of connecting wires and accessories resistor leads. An effect to watch for is amplifier offset
are available from manufacturers and their literature should voltage warm-up drift caused by mismatched thermo-
be consulted (reference 4). couple materials in the wire-bond/lead system of the IC

an28f

AN28-18
 Application Note 28
package. This effect can be as high as tens of microvolts A final source of error is thermocouple placement. Remem-
in TO-5 cans with kovar leads. It has nothing to do with ber that the thermocouple measures its own temperature.
the actual offset drift specification of the amplifier and can In flowing or fluid systems, remarkably large errors can
occur in amplifiers with measured “zero” drift. Warm-up be generated due to effects of laminar flow or eddy cur-
drift is directly proportional to amplifier power dissipation. rents around the thermocouple. Even a “simple” surface
It can be minimized by avoiding TO-5 cans, using low measurement can be wildly inaccurate due to thermal
supply current amplifiers, and by using the lowest pos- conductivity problems. Silicone thermal grease can reduce
sible supply voltages. Finally, it can be accommodated by this, but attention to sensor mounting is usually required.
calibrating and specifying the system after a five minute As much of the sensor surface as possible should be
warm-up period. mated to the measured surface. Ideally, the sensor should
A significant error source is the cold junction. The error be tightly mounted in a drilled recess in the surface. Keep
takes two forms. The subtractive voltage produced by the in mind that the thermocouple leads act as heat pipes,
cold junction must be correct. In a true cold junction (e.g., providing a direct thermal path to the sensor. With high
ice point reference) this voltage will vary with inability to thermal capacity surfaces this may not be a problem,
maintain the desired temperature, introducing error. In a but other situations may require some thought. Often,
cold junction compensator like the LT1025, error occurs thermally mating the lead wire to the surface or coiling
with inability to sense and track ambient temperature. Mini- the wire in the environment of interest will minimize heat
mizing sensing error is the manufacturer’s responsibility piping effects.
(we do our best!), but tracking requires user care. Every As a general rule, skepticism is warranted, even in the
effort should be made to keep the LT1025 isothermal with most “obviously simple” situations. Experiment with ser-
the cold junction. Thermal shrouds, high thermal capacity eral sensor positions and mounting options. If measured
blocks and other methods are commonly employed to results agree, you’re probably on the right track. If not,
ensure that the cold junction and the compensation are rethink and try again.
at the same temperature.
Amplifier offset uncertainties and, to a lesser degree,
bias currents and open-loop gain should be considered.
Amplifier selection criteria is discussed in the text under
“Amplifier Selection.”

an28f

Information furnished by Linear Technology Corporation is believed to be accurate and reliable.

However, no responsibility is assumed for its use. Linear Technology Corporation makes no representa-
tion that the interconnection of its circuits as described herein will not infringe on existing patent rights. AN28-19
Application Note 28

an28f

IM/GP 290 2K • PRINTED IN USA

Linear Technology Corporation
AN28-20 1630 McCarthy Blvd., Milpitas, CA 95035-7417
(408) 432-1900 ● FAX: (408) 434-0507 ● www.linear.com © LINEAR TECHNOLOGY CORPORATION 1990