Semiconductor Thermal Resistance Standards versus Real Life. Bernie Siegal Thermal Engineering Associates, Inc.

Transcription

1 Semiconductor Thermal Resistance Standards versus Real Life Bernie Siegal Thermal Engineering Associates, Inc.

2 Overview Introduction Objective Temperature vs. Thermal Current Standard Thermal Parameters Thermal Resistance Definition Environment Equivalent Circuit Standard Reference Environments Definition of Terms Mounting Surface Environmental Using Thermal Data Standards Conclusion References

3 Introduction Objective To demonstrate that semiconductor thermal measurements made according to standards do not necessarily produce parameter values that are useful in real applications.

4 Temperature vs. Thermal Temperature Usually refers to environmental temperature surrounding the packaged device Relates to temperature-variant device attributes Thermal Refers to heat flow, typically from heat generator(s) in package to outside world Deals with various heat flow paths

5 What is Thermal Resistance? Thermal Resistance - Junction to Specified Reference Point Measure of the ability of the physical structure to remove heat from the semiconductor junction to the reference point Temperature rise of the junction above a reference point per watt of continuous power Metric is C/W

6 Thermal Resistance Definition θ J R = ( T T ) J P H R Where: θ JR = Thermal Resistance, junction-to-reference, C/W T J = junction temperature, C T R = reference temperature, C P H = Power dissipation, W

7 Using Thermal Data Calculating Junction Temperature T J = T R + ( P θ ) H JR Where: T J = junction temperature, C T R = reference temperature, C P H = Power dissipation, W θ JR = Thermal Resistance, junction-to-reference, C/W

8 Using Thermal Data θ = θ + θ JA JC SA Ψ = Ψ + θ JA JT CA (heat sink) (natural convection) θ CA θ JA or Ψ JA θ JC or Ψ JT

9 Standard Test Methods Methods provide a basis for comparison of different devices housed in the same electronic package or similar devices housed in different electronic packages

10 Reference Environments Environment drastically effects the thermal performance of a component System and Board Temperature Mounting Surface k eff Air Flow Other Components Established need for standardization

11 Reference Environments θ JA θ JMA θ JB θ JC θ JL θ JX θ JR θ CA θ SA - junction-to-ambient (natural convection) - junction-to-moving air (forced convection) - junction-to-board (ring cold plate) - junction-to-case (cold plate) - junction-to-lead - junction-to-defined environment - junction-to-defined reference - case-to-ambient - sink-to-ambient

12 Junction-to-Ambient θ JA Natural Convection 1 ft 3 test chamber low emissivity material sample mounted in the center of chamber Chamber can be larger for high power dissipation Measure ambient temperature, chamber temperature

13 Still-Air Chamber - θ JA Low Emissivity Chamber Material ; Plexiglass or Cardboard 1ft 3 Volume DUT Test Board thermocouple

14 Juntion-to-Moving Air θ JMA Forced Convection Laboratory Wind Tunnel Speeds range from 0 to 1000 lfpm Laminar and turbulent air flow Orientation and air disturbances important Measurement of air speed, temperature, and pressure

15 Forced Convection- θjma 3/2/2007 MEPTEC07 15

16 Wind Tunnel Specifications Flow Uniformity Uniform to +/-5% of the mean velocity across the center 90% of the test chamber crosssection Constant within +/-5% along the length of the test section. Swirl Less than 5% of the mean flow velocity Measured with a three axis or cross wire anemometer Turbulence Less than 2% Measured with a hot wire anemometer with a frequency response of 10 khz or higher Unsteadyness Less than 5% over the time period of a typical measurement Chamber Size No more than 5% cross-section blockage Non-ducted flow Package Mounting/Orientation With respect to air flow With respect to test chamber Thermocouple mounting

17 Junction-to-Board θ JB Ring cold plate method Forces heat transfer through leads Top and bottom of package thermally isolated 2S2P PCB only Conduction model validation Board level simulation parameter Simple model for board effects

18 Ring Cold Plate 1 to 5 mm Gap Insulation to Board or Insulation to Package Junction Temperature Board Temperature Thermocouple Soldered to Middle Lead Insulate with tape if necessary Insulation Water Channel 5 mm Minimum

19 Junction-to-Case θ JC Standard being developed Must force a known amount of heat flux through the desired thermal interface Measure case temperature to compute Juction-to-Case temperature difference

20 θ JC Apparatus Heater Backing Plate Foil Heater Heater Spreader Plate Heat Flux Sensor Device Interface Plate Heat Collector Thermal Test Board & Package Flux Column Temp Sensors Cold Plate

21 Test Board Standards 2S and 2S2P SMT Leaded, SMT Area Array, Through-Hole DIL, Through-Hole Area Array Board Material, thickness, geometry Trace thickness, geometry Routing rules Nesting

22 PCB Standard Dimensions A = 76.2 mm C = 25 m m A = mm 1 B = m m A = 60.7 m m B = mm B = m h = mm h = mm 54 mm C = mm C = mm e = mm d = mm f = mm e = mm d = mm f = mm G = mm G = mm

23 Heating Curve Example 20 Heating Curve (Average) Thermal Resistance ( o C/W) m/s 2 m/s Heat Sink Still Air 0 1.0E E E E E E E E E+04 Hetaing Time (th, seconds)

24 Standards versus Real Life Standards Well-defined and simple thermal environment Easy environment temperature measurements Relatively easy junction temperature measurement Power dissipation is forced and known accurately Power dissipation is static Measurement subject is only heat source 2s or 2s2p thermal test board Real Life Not well-defined and complex thermal environment Difficult/impossible environment temperature measurements Relatively difficult junction temperature measurement Power dissipation is usually not know accurately Power dissipation is dynamic Measurement subject is one of many heat sources Package mounted on complex/manylayered printed circuit board

25 Summary Be careful when using standard thermal parametric data in real life applications

26 Application-oriented metrics Not as rigorous nor accurate as θ but useful for quick T J approximations Ψ JT Ψ JB - junction to top of case (natural convection) - junction to board (natural convection) Ψ JR = T P H J P H is the total power dissipated independent of the heat flow path