AD594/AD595
REV. C
8
OUTLINE DIMENSIONS
Dimensions shown in inches and (mm).
TO-116 (D) Package
0.100
(2.54)
BSC
0.700 ?/SPAN>0.010
(17.78 ?/SPAN>0.25)
0.180 ?/SPAN>0.030
(4.57 ?/SPAN>0.76)
0.085 (2.16)
0.035 ?/SPAN>0.010
(0.89 ?/SPAN>0.25)
0.047 ?/SPAN>0.007
(1.19 ?/SPAN>0.18)
0.017
+0.003
0.002
0.43
+0.08
0.05
(
(
0.125
(3.18)
MIN
0.31 ?/SPAN>0.01
(7.87 ?/SPAN>0.25)
0.01 ?/SPAN>0.002
(0.25 ?/SPAN>0.05)
0.095 (2.41)
0.30 (7.62) REF
0.430 (10.92)
14
1
7
8
0.040
(1.02) R
PIN 1
0.265
(6.73)
0.290 ?/SPAN>0.010
(7.37 ?/SPAN>0.25)
Cerdip (Q) Package
14
1
7
8
PIN 1
0.77 ?/SPAN>0.015
(19.55 ?/SPAN>0.39)
0.310
(7.87)
0.260 ?/SPAN>0.020
(6.6 ?/SPAN>0.51)
SEATING
PLANE
0.100
(2.54)
BSC
0.035 ?/SPAN>0.010
(0.889 ?/SPAN>0.254)
0.180 ?/SPAN>0.030
(4.57 ?/SPAN>0.76)
0.125
3.175)
MIN
0.032
(0.812)
15?/SPAN>
0?/SPAN>
0.300 (7.62)
REF
0.148 ?/SPAN>0.015
(3.76 ?/SPAN>0.38)
0.010 ?/SPAN>0.001
(0.254 ?/SPAN>0.025)
0.018
(0.457)
0.600 (15.24)
BSC
THERMOCOUPLE BASICS
Thermocouples are economical and rugged; they have reason-
ably good long-term stability. Because of their small size, they
respond quickly and are good choices where fast response is im-
portant. They function over temperature ranges from cryogenics
to jet-engine exhaust and have reasonable linearity and accuracy.
Because the number of free electrons in a piece of metal de-
pends on both temperature and composition of the metal, two
pieces of dissimilar metal in isothermal and contact will exhibit
a potential difference that is a repeatable function of tempera-
ture, as shown in Figure 14. The resulting voltage depends on
the temperatures, T1 and T2, in a repeatable way.
CONSTANTAN
IRON
UNKNOWN
TEMPERATURE
ICE POINT
REFERENCE
T1
T2
CONSTANTAN
Cu
Cu
V1
Figure 14.  Thermocouple Voltage with 0癈 Reference
Since the thermocouple is basically a differential rather than
absolute measuring device, a know reference temperature is
required for one of the junctions if the temperature of the other
is to be inferred from the output voltage. Thermocouples made
of specially selected materials have been exhaustively character-
ized in terms of voltage versus temperature compared to primary
temperature standards. Most notably the water-ice point of 0癈
is used for tables of standard thermocouple performance.
An alternative measurement technique, illustrated in Figure 15,
is used in most practical applications where accuracy requirements
do not warrant maintenance of primary standards. The reference
junction temperature is allowed to change with the environment
of the measurement system, but it is carefully measured by some
type of absolute thermometer. A measurement of the thermo-
couple voltage combined with a knowledge of the reference
temperature can be used to calculate the measurement junction
temperature. Usual practice, however, is to use a convenient
thermoelectric method to measure the reference temperature
and to arrange its output voltage so that it corresponds to a ther-
mocouple referred to 0癈. This voltage is simply added to the
thermocouple voltage and the sum then corresponds to the stan-
dard voltage tabulated for an ice-point referenced thermocouple.
CONSTANTAN
IRON
T1
V1
CuNi
Cu
Cu
T3
V
1
' = V
1
FOR PROPERLY
SCALED V
3
' = f(T
3
)
V3'
V1'
V2
Figure 15.  Substitution of Measured Reference
Temperature for Ice Point Reference
The temperature sensitivity of silicon integrated circuit transis-
tors is quite predictable and repeatable. This sensitivity is
exploited in the AD594/AD595 to produce a temperature re-
lated voltage to compensate the reference of cold junction of a
thermocouple as shown in Figure 16.
Cu
CONSTANTAN
IRON
T1
Cu
T3
Figure 16. Connecting Isothermal Junctions
Since the compensation is at the reference junction temperature,
it is often convenient to form the reference junction by connect-
ing directly to the circuit wiring. So long as these connections
and the compensation are at the same temperature no error will
result.