Thermal Resistivity Table Simplifies Temperature Calculations.

APPLICATION NOTES
Thermal Resistivity Table Simplifies Temperature Calculations
How much power can a microwave device handle before it overheats?
Here is a good approximation of the technique for determining "hot spots" of different structures.
Reprinted with permission from Microwaves/RF
Quick estimates of temperature rises in conduction heat flow situations
can be made easily when it is realized that the steady state Fourier heat
flow equation is the mathematical equivalent of Ohms Law for electrical
circuits with temperature differences replacing voltage differences and
heat flow (expressed in calories, per second, B.T.U. per hour or watts)
replacing current flow. In usual form the one dimensional steady state
flow equation is:
Q = K (A/L) ΔT
(1)
where
Q = heat flow
K = thermal conductivity
A = cross section of heat flow path
L = path length
ΔT = temperature drop along path of length, L.
Inverting the relationship to:
ΔT = (1/K) (L/A) Q
then (1/K) equals thermal resistivity and (1/K) (L/A) equals thermal resistance. (By anology to V = (ρ) (L/A) I and V = R . I). By preparing a table
of thermal resistivity (1/K) in units of °C – inches/W for the materials
encountered in microwave work, temperature drops can be readily estimated in series heat conduction situations. The application of these concepts can best be illustrated by the following two examples.
EXAMPLE PROBLEM 1.
EXAMPLE PROBLEM 2.
Figure 1 shows a thin film resistor in a microwave integrated circuit. How
hot will it get with 2 W dissipated over its 0.1 x 0.2 in. surface (100 W/in.2)?
A miniature rod resistor terminates a coaxial line as shown in Fig. 2.
Estimate the temperature rise at the front end of the resistor relative to a
point on the outer shell when 5 W is being dissipated:
Electrical engineers are usually introduced to heat flow either as studies
in physics or problems to be solved by a mechanical engineer. Heat is
shown to flow via the mechanisms of radiation, convection and conduction—each with their own equations and coefficients.
The subject is made more confusing by the use of c.g.s. units in physics
and British Thermal Units in engineering. Most of the important temperature calculations faced by a microwave engineer result from the need
to determine how hot a particular point in a structure will get when power
is being dissipated. While the ultimate heat sink usually expels its energy to the outside world by free or forced convection to air or liquid, the
internal circuit elements, whose temperature rises usually set a limit on
power handling capability, most often are conduction cooled. Radiation
plays a relatively insignificant role below 300°C.
ΔT1–4 =ΔT1–2 + ΔT2–3+ ΔT3–4
ΔT1–2 (in ceramic) =
(2.13)(0.025)(2)
= 5.3°C
(0.1) (0.2)
ΔT2–3 (in grease) =
ΔT3–4 (in aluminum) =
46 (0.002)(2)
= 9.2°C
(0.1) (0.2)
ΔT1–5 =ΔT1–2+ ΔT2–3+ ΔT3–4 + ΔT4–5
Since the power is uniformly dissipated along the resistor, ΔT1–2 can be
estimated by assuming that the full 5 W travels 1/2 the length of the resistor to the mid plane of the end wall:
ΔT1–2 =
(0.23)(0.125)(2)
= 2.9°C
(0.1) (0.2)
(0.24) (0.020 + 0.031) (5)
= 48.7°C
π(0.020)2
ΔT2–3 (solder joint) =
Therefore,
ΔT1–4 = 5.3 + 9.2 + 2.9 = 17.4°C
Adding to this an estimate of maximum absolute temperature of 80°C for
the aluminum ground plane leads to an estimated 100°C for the maximum resistor temperature. This is a conservative estimate since it
neglects transverse heat spreading, but it serves to illustrate the ease
and value of quick estimates. The major temperature drop occurs at the
grease interface between the ceramic and the metal. Changing from
Aluminum Oxide to Beryllium Oxide with 1/10 the thermal resistivity
wouldn’t help in this situation as much as working to minimize the interface resistance.
Thin Film resistor as heat source in
microwave integrated circuit
RESISTOR
0.1X0.2[2,5X5,1]
SILICONE GREASE
LAYER 0.002[0,05]
COPPER
CIRCUIT
HEAT
ALUMINA CERAMIC
0 025[0 64] THICK
ALUMINUM GROUND PLANE
0.125[3,18] THICK
(0.78) (0.005) (5)
= 3.2°C
π (0.040) (0.062)
For the radial path 3-4, use an average diameter of 0.120 for the area calculation and a path length of 0.080.
ΔT3–4 =
ΔT4–5 =
Then
(0.11) (0.080) (5)
= 1.9°C
π (0.120) (0.062)
(0.11) (0.040 + 0.031) (5)
= 1°C
π (0.062) (0.2)
ΔT1–5 = 48.7 + 3.2 + 1.9 +1 = 54.8°C
A review of the results reveals
that the manner of estimating
the area and length of the metal
paths and the solder joint are relatively unimportant since they
contribute only 11 percent to the
total ΔT. For this configuration
the greatest potential reductions
in temperature drop can be
made by increasing the resistor
diameter or reducing its length.
Resistor as heat source
in coaxial line
0.040[1,02] DIAMETER BeO ROD
T5
0.062
[1,575]
T1
T4
T3
0.2
[5,1]
T2
0.040[1,02]
SOLDER JOINT
0.062
[1,575]
COPPER CHELL
Figure. 2
Figure 1
KEY: Inches [Millimeters] .XX ±.03 .XXX ±.010 [.X ±0.8 .XX ±0.25]
300 Dino Drive, Ann Arbor, MI 48103
888-244-6638 or 734-426-5553 • FAX 734-425-5557
www.aeroflex.com/inmet • [email protected]
REV 3/09
APPLICATION NOTES
Ta ble 1. Thermal Resistivity (R) – Typical Values
R in °C - inches
Watt
conversion from thermal conductivity K to R
R=
R=
22.8
BTU - ft
for K in
K
hr - ft2 - °F
0.094
cal - cm
for K in
K
sec - cm2 - °C
Material
R
Diamond
0.06
Lead
Silver
0.10
Indium
Copper
0.11
Boron nitride
Material
R
1.14
2.1
Material
R
Quartz
27.6
Glass (7740)
34.8
(isotropic)
1.24
Silicon thermal
grease
46
Gold
0.13
Aluminum
0.23
Alumina ceramic
2.13
Water
63
Beryllia ceramic
0.24
Kovar
2.34
Mica (avg)
80
Molybdenum
0.27
Silicon carbide
2.3
Polyethylene
120
Brass
0.34
Steel (300 series)
2.4
Teflon
190
Silicon
0.47
Nichrome
3.00
Nylon
190
Platinum
0.54
Carbon
5.7
Silicone Rubber
190
Tin
0.60
Ferrite
6.3
P.P.O.
205
Nickel
0.61
Pyroceram (9606)
Polystyrene
380
Eutectic lead
tin solder
11.7
Epoxy—high
0.78
conductivity
24
Mylar
1040
Air
2280
KEY: Inches [Millimeters] .XX ±.03 .XXX ±.010 [.X ±0.8 .XX ±0.25]
300 Dino Drive, Ann Arbor, MI 48103
888-244-6638 or 734-426-5553 • FAX 734-425-5557
www.aeroflex.com/inmet • [email protected]
REV 3/09