Convection Heat Transfer with Application to Electronics Cooling

1
Convection Heat Transfer with Application to
Electronics Cooling

• Presentation by
• Mark Hoffman
• Steve Laurino
• Matt Weber
• MER050
• 5/29/02

2
Presentation Outline

• Laboratory Introduction
• Procedure
• Experimental Results
• Disscussion of Results
• Conclusions
• Questions

3
Introduction

• Experiment to model the heat transfer from an IC
  chip (aluminum block to simulate Pentium III
  chip)
• Trials performed with and without heat sinks to
  determine effects
• Trials performed with free and forced convection
  to determine effects
• Force convection trials performed with various
  fan velocities to determine effects

4
Introduction- Definitions

• Free Convection- Relative motion between fluid
  and surface due to buoyancy within the fluid,
  i.e. temperature or mass gradient
• Forced Convection- Relative motion between fluid
  and surface is maintained by external means, i.e.
  fan, pump, etc.

5
Introduction- Definitions

• Heat Sink- device attached to a microprocessor
  chip to keep it from overheating by absorbing its
  heat and dissipating it into the air.

6
Introduction-Theory

• Power into the simulated IC module is equal to
  the conductive and convective heat transfer out
  of the module
• This assumes there is no heat generated or heat
  stored within the IC module

• Pin Q(conv) Q(loss)
• Egen, Est 0

7
Introduction-Theory

• Resistance between module and air is equal to the
  temperature difference between the module and
  free stream divided by the convective heat
  transfer

• R case?air
• (T(mod) -T(amb) )/Q(conv)

8
Setup

• Tabletop windtunnel
• Simulated IC module
• Power Supply
• 3 Type k thermocouples
• Thermocouple Junction
• Data acquisition system

9
Setup

• Thermocouple Locations
• T(ambient) Free stream (center of honeycomb
  stabilizer)
• T(Plexi) Underside of windtunnel directly
  beneath aluminum block
• T(module) Center of Aluminum block

10
Procedure-Week 1Forced Convection

• Transient- Fan set to 12V, dT20 C
• Without heat sink- Fan set to varying voltages,
  dT 20 C
• With heat sink- Fan set to varying voltages, dT
  20 C
• Liquid Crystal Imagery Fan set to varying
  voltages, T(mod) 85 C
• (Allowed to reach steady state in all cases)

11
Procedure Week 2Natural (free) Convection

• Varying heater power ? steady state conditions
• Without heat sink- Heater set to varying power,
  time allowed to attain steady state.
• With heat sink - Heater set to varying power,
  time allowed to attain steady state.

12
Results- Week 1
13
Results- Week 1
14
Results- Week 1
15
Results- Week 1
16
Results- Week 1
17
Results- Week 1
18
Results- Week 1
19
Results- Week 1
20
Results- Week 2
21
Results- Week 2
22
Results- Week 2
23
Results- Week 2
24
Conclusions-Week 1

• A fan voltage increase leads to a decrease in
  resistance, an increase in convection heat
  coefficient, an increase in Qconv, and does not
  change Qloss noticeably.
• Similar results are seen with increases in air
  velocity because fan voltage is directly related
  to velocity

25
Conclusions- Week 1

• Observed resistance from module with the heat
  sink to air was extremely close to the given
  resistance curve for the heat sink.
• As the air velocity increased, the distance from
  the edge of module to location where a
  temperature of 33C was observed also increased.

26
Conclusions- Week 2

• The resistance remains approximately constant for
  each case, with and without the heat sink. The
  resistance with the heat sink is lower than
  without the heat sink.
• H is higher with the heat case for each power
  setting

27
Conclusions- Week 2

• Qloss increases as the power increases for both
  cases. The case without the heat sink experiences
  a greater heat loss because in the cases with the
  heat sink more heat is lost through convection so
  there is less heat that needs to be lost through
  conduction to the plexiglas maintain the system
  energy balance.

28
Conclusions- Week 2

• Qconv increases almost linearly with an increase
  in power. The value of Qconv with and without the
  heat sink is approximately the same for each
  power level.

29
Sources of Error

• Lab settings - non-steady environment
  conditions (movement around setup, doors
  and windows opening and closing)
• Outlier data - some calculated data fell so
  far from the rest of the data that it
  radically altered the scale of the plots
• Mislabeled data - confusing and mislabeled data
  submitted by teams to web site

30
Recommendations

• Make a template for team data submissions so that
  all data will be appropriately labeled.
• More time before presentations would facilitate
  more accurate data calculation
• Data should be submitted and finalized in a
  timely manner in close proximity to the
  completion of collection of experimental data

31
Questions

• ?