Intel® 82854 Graphics Memory
Controller Hub (GMCH) for Fanless
Set Top Box Applications
Thermal Design Guide
March 2005
Revision 1.0
Reference Number: D14844-001
Legal Statements
INFORMATION IN THIS DOCUMENT IS PROVIDED IN CONNECTION WITH INTEL® PRODUCTS. NO
LICENSE, EXPRESS OR IMPLIED, BY ESTOPPEL OR OTHERWISE, TO ANY INTELLECTUAL PROPERTY
RIGHTS IS GRANTED BY THIS DOCUMENT. EXCEPT AS PROVIDED IN INTEL'S TERMS AND
CONDITIONS OF SALE FOR SUCH PRODUCTS, INTEL ASSUMES NO LIABILITY WHATSOEVER, AND
INTEL DISCLAIMS ANY EXPRESS OR IMPLIED WARRANTY, RELATING TO SALE AND/OR USE OF
INTEL PRODUCTS INCLUDING LIABILITY OR WARRANTIES RELATING TO FITNESS FOR A
PARTICULAR PURPOSE, MERCHANTABILITY, OR INFRINGEMENT OF ANY PATENT, COPYRIGHT OR
OTHER INTELLECTUAL PROPERTY RIGHT. Intel products are not intended for use in medical, life saving, life
sustaining applications.
Intel may make changes to specifications and product descriptions at any time, without notice.
Designers must not rely on the absence or characteristics of any features or instructions marked “reserved” or
“undefined.” Intel reserves these for future definition and shall have no responsibility whatsoever for conflicts or
incompatibilities arising from future changes to them.
The Thermal Design Guide may contain design defects or errors known as errata which may cause the product to
deviate from published specifications. Current characterized errata are available on request.
The Thermal Design Guide as well as the software described in it are furnished under license and may only be used
or copied in accordance with the terms of the license. The information in this document is furnished for
informational use only, is subject to change without notice, and should not be construed as a commitment by Intel
Corporation. Intel Corporation assumes no responsibility or liability for any errors or inaccuracies that may appear
in this document or any software that may be provided in association with this document.
Except as permitted by such license, no part of this document may be reproduced, stored in a retrieval system, or
transmitted in any form or by any means without the express written consent of Intel Corporation.
Contact your local Intel sales office or your distributor to obtain the latest specifications and before placing your
product order.
Copies of documents which have an ordering number and are referenced in this document, or other Intel literature
may be obtained by calling
1-800-548-4725 or by visiting Intel's website at http://www.intel.com.
AnyPoint, AppChoice, BoardWatch, BunnyPeople, CablePort, Celeron, Chips, CT Media, Dialogic, DM3,
EtherExpress, ETOX, FlashFile, i386, i486, i960, iCOMP, InstantIP, Intel, Intel Centrino, Intel logo, Intel386,
Intel486, Intel740, IntelDX2, IntelDX4, IntelSX2, Intel Create & Share, Intel GigaBlade, Intel InBusiness, Intel
Inside, Intel Inside logo, Intel NetBurst, Intel NetMerge, Intel NetStructure, Intel Play, Intel Play logo, Intel
SingleDriver, Intel SpeedStep, Intel StrataFlash, Intel TeamStation, Intel Xeon, Intel XScale, IPLink, Itanium, MCS,
MMX, MMX logo, Optimizer logo, OverDrive, Paragon, PC Dads, PC Parents, PDCharm, Pentium, Pentium II
Xeon, Pentium III Xeon, Performance at Your Command, RemoteExpress, SmartDie, Solutions960, Sound Mark,
StorageExpress, The Computer Inside., The Journey Inside, TokenExpress, VoiceBrick, VTune, and Xircom are
trademarks or registered trademarks of Intel Corporation or its subsidiaries in the United States and other countries.
*Other names and brands may be claimed as the property of others.
Table 10. Mechanical Drawing List (shown on the following pages). ......................................................... 35
Intel® 82854 GMCH – Thermal Design Guide 5
Revision History
Date Revision Reference # Description
March 2005 1.0 D14844-001 First release.
6 Intel® 82854 GMCH – Thermal Design Guide
11
IInnttrroodduuccttiioonn
1.1 Document Objective
This document is intended to aid system designers to properly implement a thermal management solution to ensure
reliable and efficient operation of the Intel® 82854 Graphics Memory Controller Hub (GMCH). The objective of
thermal management is to ensure that the temperature of the device while operating in a set top box system is
maintained within functional limits. The functional temperature limit is the range within which the electrical circuits
in the silicon can be expected to meet specified performance requirements. Operation outside the functional limit
can degrade system performance, cause logic errors, or cause component and/or system damage. Temperatures
exceeding the maximum operating limits may result in irreversible changes in the operating characteristics of the
components. This document will provide an understanding of the operating limits of the Intel® 82854 GMCH and
suggest proper thermal design techniques based on a particular configuration and system boundary conditions.
1.2 Related Documents
Table 1. Document References
Title Number Location
Intel® 854 Chipset GMCH Product Preview Datasheet 17064 See your local Intel field
ULV Intel® Celeron® M Processor at 600 MHz for Fanless Set Top Box
Applications – Thermal Design Guide, Rev 1.0
Intel® 855GME and Intel® 852GME Chipset Memory Controller Hub (MCH) –
Thermal Design Guide for Embedded Applications, Rev 1.0
Thermal Considerations for Passive Set Top Box Design Guide, Rev 1.0 17114 See your local Intel field
17531 See your local Intel field
273838http://developer.intel.com/
representative
representative
representative
Intel® 82854 GMCH – Thermal Design Guide 7
1.3 Terminology
Table 2. Definitions of Terms
Term Definition
DDR Double Data Rate
FCBGA Flip Chip Ball Grid Array. A package type defined by a plastic substrate on to which a die is
mounted using an underfill C4 (Controlled Collapse Chip Connection) attach style. The
primary electrical interface is an array of solder balls attached to the substrate opposite the die.
LFM Linear Feet per Minute
GMCH Graphics Memory Controller Hub
Natural Convection
Cooling (free convection)
OEM Original Equipment Manufacturer
PCB Printed Circuit Board
TIM Thermal Interface Material: the thermally conductive compound placed in between the heat
TDP Thermal Design Power: a design point for the component. OEMs must design thermal
T
junction, Tj
T
case, Tc
T
sink, Ts
T
amb
T
amb_max
T
air
T
rise
Ψja
Ψ
jc , Ψpackage
Ψ
cs , ΨTIM
Ψ
sa , Ψheatsink
The maximum junction temperature of the die, as measured or specified in the component
The temperature at the geometric center of the top surface of the package. For bare die
The temperature of the heat sink base plate at the center location of the package or die.
The ambient temperature locally surrounding the component. The ambient temperature should
Maximum allowable T
The air temperature external to the chassis enclosure. Also, referred to as the external ambient
The temperature rise of the air as it enters from the chassis till it reaches the region of the
The thermal resistance between the junction and the package case. Also represents the package
The case to sink thermal resistance, which is dependent on the thermal interface material
The sink-to-ambient thermal resistance. A measure of heat sink thermal performance using
The transferring of heat from a surface to a fluid (i.e., air or liquid) where the convection flow
is generated only by fluid buoyancy. No airflow devices (i.e., fans) are used in the system.
sink and die to improve the heat transfer from the die to the heat sink.
solutions that meet TDP and T
specifications as specified in the component datasheet.
junction
datasheet.
package, this is the temperature at the center of the back surface of the die.
be measured approximately one inch (25.4 mm) upstream of a passive heat sink or at the fan
inlet of an active heat sink.
that can be supported by a thermal solution.
amb
temperature, or the chassis ambient temperature.
component: defined as (T
amb
- T
)
air
The thermal resistance between the junction and the ambient air. A measure of global thermal
performance of a component thermal solution module using total package thermal design
- T
power: (T
)/TDP.
j
amb
resistance. A measure of package thermal performance using total package thermal design
power: (T
(TIM). Also referred to as
thermal design power: (T
total package thermal design power: (T
– Tc )/TDP.
j
Ψ
. A measure of TIM thermal performance using total package
TIM
– Ts )/TDP.
c
– T
)/TDP.
s
amb
8 Intel® 82854 GMCH – Thermal Design Guide
22
MMeecchhaanniiccaall GGuuiiddeelliinneess
2.1 Intel® 82854 GMCH Package
The Intel® 82854 GMCH is constructed with a Flip Chip Ball Grid Array (FCBGA) package with a size of 37.5 mm
x 37.5 mm. It includes 732 solder ball lands with a ball pitch of 1.27 mm.
Figure 1 provides details of the package dimensions of the Intel® 82854 GMCH. The drawing is not drawn to scale
and the units shown are in millimeters. The package consists of a silicon die mounted face down on a plastic
substrate populated with solder balls on the bottom side. The Intel® 82854 GMCH also includes capacitors mounted
on the top of the package in the area surrounding the die, as outlined in the top view. Because the die-side capacitors
are electrically conductive, and only slightly shorter than the die thickness, care should be taken when applying a
thermal solution onto the die in order to avoid any accidental electrical shorts.
The thermal solutions enabled for the Intel® 82854 GMCH for set top box applications may have volumetric
constraint zones that will allow for the thermal solution to be assembled to a system board. System designers must
take these zones into account so that the thermal solution will be properly assembled to the board and not interfere
with any other components. An example of such volumetric constraint zones for the reference natural convection
thermal solutions is shown in “Appendix A Mechanical Drawings
The maximum allowable height for a thermal solution is very important in the overall thermal performance and is a
factor in the volumetric constraint for a thermal solution. This height is determined by the form factor in which the
computer system is placed. For the third party reference thermal solutions presented herein, the maximum allowable
height was based on a form factor requirement that is assumed to have the height limitation of 1.5” for the heat sink.
These solutions may apply for other form factors, but it is up to the system integrator to ensure that all thermal and
mechanical requirements are validated in the final intended configuration. Figure 2 shows a generic mechanical
stack-up for the Intel® 82854 GMCH and the geometric parameters that need to be accounted for when determining
the maximum allowable height for a thermal solution.
”.
Figure 2. Typical Mechanical Stack-up
10 Intel® 82854 GMCH – Thermal Design Guide
33
The overall performance requirement of a component thermal solution depends on the following three parameters:
• Thermal Design Power (TDP)
• Maximum junction temperature (T
• Operating ambient temperature (T
The guidelines and recommendations presented in this document are based on specific parameters that are relevant
to designing a natural convection thermal solution. The overall heat dissipation capability of a thermal solution
depends on many parameters, including:
• Package thermal performance or resistance (Ψ
• Thermal performance or resistance of Thermal Interface Material, TIM (Ψ
• Heat sink performance or resistance (Ψ
• Maximum junction temperature, as specified in the datasheet (T
• Operating local ambient temperature to the component (T
To develop a reliable thermal solution all of the appropriate variables must be considered. Thermal simulations and
characterizations must be performed. The solutions presented in this document must be validated in their final
intended system.
TThheerrm
maall GGuuiiddeelliinneess
)
j
)
amb
jc , Ψpackage
ca , Ψheatsink
)
)
)
j
)
amb
cs , ΨTIM
)
Intel® 82854 GMCH – Thermal Design Guide 11
3.1 Heat Sink Design Considerations
There are three fundamental modes of heat transfer to be considered:
1. The conduction from the heat source to the heat sink fins. Providing a direct conduction path from the heat
source to the heat sink fins and selecting materials with higher thermal conductivity improve the heat sink
performance. The cross-sectional area, thickness, and conductivity of the conduction path from the heat source
to the fins directly impact the thermal performance of the heat sink. In particular, the quality of the contact
between the package die and the heat sink base has a greater impact on the overall thermal solution performance
as the cooling requirements become more difficult to satisfy. Thermal Interface Material (TIM) is used to fill in
the gap between the die and the bottom surface of the heat sink which would have been filled otherwise with a
layer of air and microscopic voids. High performance TIMs with good surface wetting characteristics will
thereby improve the overall performance of the stack-up (die-TIM-heat sink). With poor heat sink interface
flatness and/or roughness, TIM may not adequately fill the gap. The TIM thermal performance depends on its
thermal conductivity, surface wetting characteristics, and the pressure applied to it.
2. The convection from the exposed surfaces to the air stream. After conduction carries the heat from the heat
source to the surfaces exposed to air flow, such as the surface of the heat sink fins, the heat is dissipated from
the heat sink by means of either convection or radiation heat transfer. Convection heat transfer occurs to the
airflow from the surfaces exposed to the flow. Convection heat transfer is characterized by the temperature
difference between the exposed surface and the local ambient air and the total area of the exposed surfaces. A
thermal solution with a greater temperature difference and larger exposed area will have better cooling
capability. In convection, the faster the air velocity over the surface and cooler the air, the more efficient the
resulting cooling. In the case of natural convection with no externally driven induced or forced airflow, the
convection flow is created by the buoyancy force acting on the heated hot air, and its velocity depends solely on
the surface geometry and the surface-to-air temperature difference.
3. The radiation from the component to surrounding surfaces. Radiation heat transfer takes place between any two
surfaces facing each other and having different temperatures. In a fanless system, the amount of radiation from
a hot component like a heat sink to the cooler surrounding surfaces like the interior wall of the chassis enclosure
can be a significant portion of the overall heat sink thermal dissipation. It is strongly recommended that the
chassis layout designer take advantage of this naturally existing heat dissipation mechanism by comprehending
the radiation phenomenon and promoting this radiation heat transfer. A bare aluminum heat sink can be simply
anodized or painted to improve the radiation characteristics of the heat sink surface, resulting in a substantial
thermal performance improvement.
3.1.1 Heat Sink Size
The size of the heat sink is dictated by height restrictions in a system and by the foot-print area available on the
motherboard. The height and the base plate size of the heat sink must comply with the requirements and
recommendations published for the motherboard and chassis form factors of interest.
3.1.2 Heat Sink Weight
With the need to push air cooling toward better performance, heat sink solutions tend to grow larger, resulting in
increased weight. The insertion of highly thermally conductive materials like copper to increase heat sink thermal
conduction performance results in even heavier solutions. The heat sink weight must take into consideration the
package and socket load limits, the mechanical capability of the heat sink retention mechanism, and the mechanical
shock and vibration profile targets.
12 Intel® 82854 GMCH – Thermal Design Guide
3.1.3 Thermal Interface Material
A thermal interface material between the die and the heat sink base is generally required to improve thermal
conduction from the die to the heat sink. Many thermal interface materials can be pre-applied to the heat sink base
prior to shipment from the heat sink supplier and allow direct heat sink attach, without the need for a separate
thermal interface material dispense or attach process during final assembly. Common types of interface materials
include elastomers (i.e., Chomerics* T710) and phase change materials (i.e., Thermagon Tpcm). These types of
materials can easily conform to fill small air gaps that are left between the two mating surfaces. These air gaps can
act as insulators and will increase the thermal resistance. An interface material can assist in filling these voids and
reduce the thermal resistance at the interface.
All thermal interface materials should be sized and positioned on the heat sink base in a way that ensures the entire
die area is covered. It is important to compensate for heat sink-to-die attach positional alignment when selecting the
proper thermal interface material size. When pre-applied material is used, it is recommended to have a protective
cover over it prior to shipping. This cover must be removed prior to heat sink installation.
The thickness of the bond line between the heat sink and die is critical to the thermal performance of the TIM. The
bond line thickness is dependent on the pressure between the heat sink and the die. It is imperative that the heat sink
is applied to the die with adequate force. For more information on force required and other important
documentation, refer to the Chomerics website at http://www.chomerics.com
The thermal resistance of a material can be estimated by using the following expression which is applicable for onedimensional conduction heat flow across any conducting materials, such as TIM.
.
where θL = thickness of the material (m)
k = effective thermal conductivity of material (W/m-ºC)
A = cross sectional area of the material (m
= Thermal Resistance through the material (ºC/W)
TIM
θ
TIM
L
=
(1)
kA
2
)
3.1.4 Mechanical Loading
The pressure applied to the surface of the Intel® 82854 package should not exceed 100 psi, equivalent of 12 lbf.
If the pressure on the surface of the package is exceeded, problems may arise. The solder ball joints between the
package and the motherboard may be subjected to fractures that could result in a loss or degradation of electrical
signals from the device. Also, the die may be exposed to warpage or, at unusually high levels of stress, cracking.
If a large compressive load is applied to the die surface precautions should be taken to help alleviate some of the
load. One manner of doing this is to provide some backing support for the motherboard directly underneath the
package. Standoffs can be used between the motherboard and the chassis to add rigidity to the motherboard under
the package and reduce the amount of board flexure under large loads.
The generic clip design may also provide mechanical preload on the package to protect the solder joint against
damage under mechanical shock. The preload within the specified maximum contact pressure serves to compress the
solder ball array between the package and the motherboard. The compression in the solder balls delays the onset of
the tensile load under critical shock conditions, and the magnitude of the maximum tensile load is thereby reduced.
In this manner, the critical solder balls are protected from tensile loading that may cause damage to the solder joint.
Intel® 82854 GMCH – Thermal Design Guide 13
3.1.5 Thermal and Mechanical Reliability
Recommendations for thermal mechanical reliability testing are shown in Table 3. These should be considered as
general guidelines. The user should define validation testing requirements based on anticipated use conditions.
Table 3. Reliability Validation
Test1 Requirement Pass/Fail Criteria2
Mechanical Shock
Random Vibration
Power Cycling (for
active solutions)
Thermal Cycling
Humidity
• Quantity: three drops for + and – directions in each of three
perpendicular axes (i.e., total of 18 drops).
• Profile: 50 G trapezoidal waveform, 11 ms duration, 170 in/s minimum
velocity change.
• Setup: Mount sample board on test fixture
• Duration: 10 min/axis, three axes
• Frequency Range: 5 Hz to 500 Hz
Power Spectral Density (PSD) Profile: 3.13 G RMS
• 7500 on/off cycles with each cycle specified as 3 minutes on, 2
minutes off at 70 °C
• -5 °C to +70 °C, 500 cycles
• 85% relative humidity, 55 °C, 1000 hours
Notes:
1
The above tests are recommended to be performed on a sample size of at least 12 assemblies from 3 different lots
of material
2
Additional Pass/Fail Criteria may be added at the discretion of the user.
Visual Check and Electrical
Functional Test
Visual Check and Electrical
Functional Test
Visual Check and Electrical
Functional Test
Visual Check and Electrical
Functional Test
Visual Check and Electrical
Functional Test
3.2 Natural Convection Cooling
Considerations
Many factors play an important role in the ability to design a natural convection thermal solution that will keep the
die within its maximum operating temperature. Both component attributes (i.e., T
attributes need to be considered. These include:
•Operating local ambient temperature (T
amb
)
• Heat-generating component placement and system orientation
• Location and size of venting
• Available volume for thermal solution
It is very challenging to design one thermal solution that will apply for multiple form factors. Thermal modeling and
analysis needs to be performed in order to optimize the thermal solution for the intended form factor and
environment. This in essence makes most natural convection thermal solutions custom designs.
, TDP, etc.) and the system
junction
14 Intel® 82854 GMCH – Thermal Design Guide
3.2.1 Intel® 82854 GMCH Thermal Specifications
Thermal data for the Intel® 82854 GMCH for set top box applications is presented in Table 4. The data is provided
for informational purposes only. Please refer to the Intel® 82854 GMCH datasheet for the most up to date
information. In the event of conflict, the datasheet supersedes information provided in this document.
Table 4. Thermal Specifications
SKU Core Vcc (V) TDP (W) Tj (°C)
Intel® 82854 GMCH 1.5 5.7 110
3.2.2 Thermal Design Power
Thermal Design Power (TDP) is defined as the worst-case power dissipated by the component while executing
publicly available software under normal operating conditions, at nominal voltages that meet the load line
specifications. The TDP definition is synonymous with the Thermal Design Power (typical) specification referred to
in previous Intel data sheets. The Intel TDP specification is a recommended design point and is not representative of
the absolute maximum power the component may dissipate under worst case conditions. For any excursions beyond
TDP, the Thermal Management features, described in Section 5 of this document, are available to maintain the
component thermal specifications.
Intel® 82854 GMCH – Thermal Design Guide 15
3.2.3 Local Ambient Temperature (T
The local ambient temperature (T
) has a significant influence when developing a thermal solution. The T
amb
amb
)
amb
is
defined as the local temperature at the location approximately one inch upstream of the thermal solution in a passive
system or directly upstream of the fan intake of an active solution. For a natural convection system the T
amb
is
measured at the same point as a passive system, but the exact relative location to the heat sink depends on the
orientation of the heat sink with respect to the local geometry. This location should be chosen so that the measured
temperature is not affected by the air that is exhausting away from the heat sink. In a horizontal configuration, the
measurement should be taken at the sides of the heat sink. In vertical orientation, the measurement should be taken
about one inch below the leading edge. The local ambient temperature includes the external air temperature outside
the enclosure plus any temperature rise due to other components in the system. The recommended measurement
locations for T
are shown in Figure 3.
amb
Figure 3. Local Ambient Temperature Locations for Passive Heat Sink
Air flow
Air flow
T
T
amb
Top View
Top View
amb
C
CLC
L
L
~0.5” to 1.0”
~0.5” to 1.0”
‘=‘‘=‘
‘=‘‘=‘
Air flow
Air flow
T
T
amb
amb
Processor
Processor
Socket
Socket
Baseboard
Baseboard
Side View
Side View
Heat Sink in Horizontal Orientation
Heat Sink in Horizontal Orientation
Heat Sink Height,‘A’
Heat Sink Height , ‘A’
~1/4 to 1/3 of ‘A’
~1/4 to 1/3 of ‘A’
~1/2 ‘A’
~1/2 ‘A’
~0.5” to 1.0”
T
T
amb
amb
Heat Sink Ht.
Heat Sink Ht.
‘A’
‘A’
Side View
Side View
Heat Sink in Vertical Orientation
Heat Sink in Vertical Orientation
~0.5” to 1.0”
‘=‘‘=‘
‘=‘‘=‘
Front View
Front View
C
C
L
LCL
16 Intel® 82854 GMCH – Thermal Design Guide
−
ψ
3.2.4 Thermal Resistance of a Heat Sink
The thermal characterization parameter or Ψ (Psi) is calculated for a given thermal system so that it may be
compared to other thermal systems. In any computer system it is necessary to calculate the required thermal
characterization parameter needed in order to keep the die within its operating temperatures. The thermal solution
must maintain the die at or below the specified junction temperature. The equation for calculating the junction-toambient thermal characterization parameter is shown in Equation 2.
TT
ψ
=
ja
TDP
ambj
(2)
where Ψ
T
T
= junction-to-ambient thermal resistance in °C/W
ja
= maximum junction temperature of die as specified in Table 4
j
= local ambient air temperature in °C
amb
TDP = Thermal Design Power in W
When calculating the required Ψ, it is important to determine the allowable temperature rise from the maximum
operating environment to the component's maximum specification. It is important to know that lower Ψ
values
ja
require better thermal solutions and vice versa.
Typical T
values for natural convection systems depend linearly on T
amb
which in turn depends largely on specific
rise
chassis design and the relative location of the component of interest with respect to other heat generating
components within the system. For typical set top box chasses with good venting designs incorporated at thermally
critical locations, this rise ranges from approximately 10 to 20°C. Considering a typical value for the external air
temperature in set-top-box applications at T
= 35-45°C, T
air
may become as high as 60-65°C. In some worst
amb
cases with thermally challenging circumstances, such as a compact small form-factor box placed in a confined or
enclosed environment, it may be necessary to anticipate T
solution needed to cool an Intel® 82854 GMCH with a T
value of as high as 70°C. As an example, the thermal
amb
of 110 °C and a TDP of 5.7W in a system with a T
j
amb
=
70 °C, would need to have a junction-to-ambient thermal resistance of:
oo
CC
=
ja
70110=−
W
7.5
o
WC
/02.7
(3)
The above case requires a thermal solution with a thermal characterization parameter less than or equal to 7.02°C/W
to keep the component temperatures at or below the specifications.
Intel® 82854 GMCH – Thermal Design Guide 17
j
j
j
Ψ
j
Ψ
j
Ψ
j
j
Ψ
j
Table 5 shows the required thermal performance at various local ambient temperatures for the Intel® 82854 GMCH.
Figure 4, which is essentially a graphical version of Table 5, shows the thermal solution requirements for the Intel®
82854 GMCH at TDP = 5.7W as a function of increasing T
amb
.
Note: Note that as the T
increases, a better Thermal Solution is needed (e.g., Ψja must decrease).
Required Thermal Solution Performance at Various Ambient
40oC
o
Ψ
(
C/W)
a
4oC5oC55oC6
Ψ
a
(oC/W)
a
10.53
(oC/W)
(oC/W)
a
9.68.77.8
o
C6oC
(oC/W)
a
o
Ψ
(
C/W)
a
7oC
a
(oC/W)
7.02
Note: Specifications (TDP, Tj) provided for reference only. Refer to the latest datasheet for the most recent data.
Ψ
= (T
ja
junction
- T
)/ TDP : junction-to-ambient thermal resistance for the thermal solution.
amb
Figure 4. Required Thermal Performance as a Function of Ambient Temperature for Intel® 82854
GMCH at Tj = 110oC and TDP = 5.7W
18 Intel® 82854 GMCH – Thermal Design Guide
3.2.5 Component Placement, System Orientation and
Venting
The placement of heat generating components in a system has an influence in the ability to develop a natural
convection thermal solution. The component placement should be optimized for a number of reasons, these include:
•Location of venting in the chassis: the vents in the chassis will allow for external air that is at a lower
temperature to enter the chassis and for the heated air to escape the system. For critical components with the
highest amount of power generation (usually the CPU and GMCH), it is necessary to place both the inlet and
outlet vents as close to them as possible in the chassis. This will facilitate the movement of air caused by
buoyancy effects hence minimizing the internal temperature rise, T
rising from the component but unable to exhaust out of the chassis.
•Orientation of the system: the orientation of the system influences where the components should be placed and
the ability to develop a natural convection solution. If the system is in a vertical configuration, the component
should be placed at the bottom of the motherboard. This will allow for the air to rise and prevent any
unnecessary pre-heating of the air surrounding the component. When the system is in a horizontal
configuration, the component should be placed on the topside of the motherboard in order to avoid trapping the
air underneath the board. The main goal of the component placement in regards to system orientation is to
minimize the local ambient temperature and avoid placing the component and other critical components in
unfavorable boundary conditions.
•Maximize the effects of radiation from hot components while minimizing the effect of irradiation to the
components: the proximity of the power generating components to each other will affect not only the local
ambient temperature but will reduce the radiation heat dissipation from the components. High power dissipating
parts should be placed as far from each other as motherboard size and electrical routing constraints allow.
, caused by the recirculation of the hot air
rise
The component placement, location of the vents, and the orientation of the motherboard directly influence the value
hence the ability to develop and optimize natural convection thermal solution. A well designed chassis that
of T
amb
provides lower T
will in turn allow smaller and more cost effective thermal solutions to the critical components.
amb
It is highly recommended that thermal simulations and analysis be performed on a system level. A Computational
Fluid Dynamics (CFD) program may be used to study multiple tradeoff scenarios, and system configurations can be
modeled to optimize the thermal solution to meet component's thermal requirements. Proper system level thermal
modeling allows the thermal solution designer to optimize thermal solutions and be confident in their performance
prior to fabricating hardware. This results in better solutions, lower design time, and faster integration. More indepth discussions and guidelines on the system level considerations are provided in the Thermal Considerations for
Passive Set Top Box Design Guide.
.
3.3 Thermal Validations
The performance of a thermal solution is dependent on many parameters including the component’s maximum
junction temperature, T
materials, and the amount of airflow. In this document the designs are targeted for a natural convection environment,
so there is no induced airflow. The guidelines and recommendations presented in this document are referenced to the
parameter, T
, which relates the performance of a local heat sink solution to the local boundary conditions
amb
determined and provided by a given system. It is the responsibility of each product design team to ensure the target
T
is achieved in a given chassis design and to verify that thermal solutions are suitable for their specific use.
amb
, the operating ambient temperature, T
j
, the specific heat sink design, component
amb
Intel® 82854 GMCH – Thermal Design Guide 19
20 Intel® 82854 GMCH – Thermal Design Guide
44
mppeerraattuurree MMeeaassuurree
TTeem
MMeettrroollooggyy
meenntt
m
4.1 Case Temperature Measurements
Intel has established guidelines for the proper techniques to be used when measuring case temperature. Section 5
contains information on running an application program that emulates anticipated TDP.
The top surface temperature at the geometric center of the die corresponds to the maximum Tcase.
4.2 Zero-Degree Angle Attach Methodology
1. Mill a 3.3 mm (0.13”) diameter hole centered on bottom of the heat sink base (see Figure 5). The milled hole
should be approximately 1.5 mm (0.06”) deep.
2. Mill a 1.3 mm (0.05”) wide slot, 0.5 mm (0.02”) deep, from the centered hole to one edge of the heat sink. The
slot should be in the direction parallel to the heat sink fins (see Figure 5 and Figure 6).
3. Attach thermal interface material (TIM) to the bottom of the heat sink base.
4. Cut out portions of the TIM to make room for the thermocouple wire and bead. The cutouts should match the
slot and hole milled into the heat sink base.
5. Attach a 36 gauge or smaller calibrated K-type thermocouple bead or junction to the center of the top surface of
the die using a high thermal conductivity cement. During this step, make sure there is no contact between the
thermocouple cement and the heat sink base because any contact will affect the thermocouple reading. It is
critical that the thermocouple bead makes contact with the die (see Figure 6).
6. Attach heat sink assembly to the GMCH and route the thermocouple wires out through the milled slot.
Figure 5. Zero-Degree Angle Attach Heatsink Modifications (not to scale)
Intel® 82854 GMCH – Thermal Design Guide 21
Figure 6. Zero-Degree Angle Attach Methodology (not to scale)
4.3 Maximum Case Temperature
Specification
Use Table 6 to determine the maximum temperature value when performing thermal laboratory testing with the
Intel® 82854 GMCH using the metrology described in this chapter and the TDP Stress Application. More
information about the TDP stress application may be found in Section 5.
Table 6. Intel® 82854 GMCH Maximum Case Temperature Value
Tcase,max (°C)
105
22 Intel® 82854 GMCH – Thermal Design Guide
55
TThheerrm
aanndd TToooollss
maall MMaannaaggee
meenntt FFeeaattuurreess
m
5.1 Internal Temperature Sensor
The Intel® 82854 GMCH will include an on-die temperature sensor that can be used to protect the die from
exceeding the Tj max specification. Upon detection that the sensor has reached Tj max the Intel® 82854 GMCH
will be capable of initiating a bandwidth throttling event that will reduce the device power and temperature. The
sensor will also prove to be useful in optimizing the thermal design for the package by being able to provide
junction temperature during testing and evaluation of the thermal solution.
5.2 External Temperature Sensor
The Intel® 82854 GMCH is designed to accept an input signal from an external temperature sensor. The external
sensor can be placed in a location close to the DDR memory and upon detecting a “hot” condition the device would
throttle the READ bandwidth. Proper placement of the sensor would have to be determined by the OEM. The OEM
would have to characterize the temperature difference between the sensor and the DDR memory devices to
determine the best placement for the sensor. On detection of a “hot” condition a signal is communicated directly
from the thermal sensor to the device via the ETS# pin as shown in Figure 7. The external thermal sensor can be
programmed via the SMBus.
Figure 7. External Temperature Sensor
Intel® 82854 GMCH – Thermal Design Guide 23
5.3 Thermal Throttling
The Intel 82854 GMCH is available with throttling functionality to protect the device from power virus conditions
that can cause junction temperatures to increase beyond maximum allowable junction temperatures. Two different
methods of thermal throttling are available on the Intel® 82854 GMCH: bandwidth triggered and temperature based
throttling.
There are three important things to remember about throttling:
1. It is only intended to be a safeguard to ensure that junction temperatures do not exceed maximum specified
junction temperatures.
2. The component thermal solutions must still be designed to TDP. Throttling is not recommended as a method of
designing the device cooling capability to levels below TDP.
3. This mechanism was carefully designed to have minimal impact on real applications, while safeguarding
against harmful synthetic applications. However, throttling may affect performance of the device. Performance
of the Intel® 82854 GMCH should be verified by testing with benchmarks.
5.3.1 Bandwidth Triggered Throttling
Bandwidth triggered throttling will limit the maximum bandwidth that can be sustained over long periods as a
safeguard against a thermal virus. This method of thermal management will temporarily decrease bandwidth
performance of the Intel® 82854 GMCH when an application demands large, sustained bandwidth levels that could
cause the device to exceed its maximum junction temperature. However, in order to trigger bandwidth throttling, the
Intel® 82854 GMCH bandwidth must exceed the threshold over an entire sampling window. Most applications use
high bandwidths only in short bursts, and through application analysis, this sampling window has been set large
enough so that these applications that create short bursts in bandwidth will not see any throttling. Only a sustained
high bandwidth for a period longer than the sampling window has the potential of exceeding thermal limits, and the
throttle mechanism is designed to protect the chip against those potentially harmful applications.
Figure 8 provides a theoretical example of how bandwidth throttling would work. In this example, the bandwidth is
set to throttle at 1100 MB/sec. The throttling value would be determined based on the worst case operating
conditions. This throttle setting is enabled upon system boot and only one value can be set for the WRITE
operations of the device. To determine bandwidth, the read/write operations are being monitored continuously by
hardware inside the Intel® 82854 GMCH within a one second window.
24 Intel® 82854 GMCH – Thermal Design Guide
Figure 8. Intel® 82854 GMCH Bandwidth Throttling
Intel® 82854 GMCH Bandwidth
1. The system is operating at an idle workload until an application that requires a large amount of bandwidth is
initiated. The above example shows a case where the peak bandwidth is 1200 MB/sec. for an entire sampling
window interval, and it will be reduced to the bandwidth throttle setting limit of 1100 MB/sec. The throttle
setting of 1100 MB/sec. effectively places a cap on the allowable bandwidth. The peak bandwidth is dependent
on memory populated and can reach up to 1600 MB/sec. The bandwidth throttling limit can be configured via
BIOS or configuration registers.
Note
Applications are still allowed to exceed the bandwidth throttling limit in short bursts that last less than the
sampling window period.
2. The Intel® 82854 GMCH will continue to operate at the throttled amount of 1100 MB/sec. until the application
no longer requires this level of sustained bandwidth. In this case the junction temperature has not increased to a
temperature that is close to the maximum junction temperature limit of 110º C. So it appears that for the brief
period that the large bandwidth level was required the device was unnecessarily throttled. A drawback of using
bandwidth triggered throttling is that under certain conditions when the system is not operating under worse
case conditions the device will be throttled regardless of the junction temperature.
3. Once the application stops the system workload will return to a lower workload.
5.3.2 Temperature Triggered Throttling
Temperature triggered throttling will limit the maximum achievable bandwidth as a safeguard against a thermal
virus only when the junction temperature reaches a specified trip point temperature. This method of thermal
throttling is an improvement over the bandwidth triggered throttling method because the Intel® 82854 GMCH will
only reduce bandwidth performance when it is absolutely necessary under a preset condition.
The temperature throttle trip point is programmed into the Intel® 82854 GMCH at boot. If the temperature of the
device goes beyond the trip point limit, the device will be throttled to a predetermined maximum throttling amount
until the temperature drops below the same temperature limit.
Intel® 82854 GMCH – Thermal Design Guide 25
Figure 9 provides an example of how temperature triggered throttling would optimize throttling under conditions
similar to the scenario that was described in Section 5.3.1. In this scenario the hot trip temperature is set at 100 ºC.
Keep in mind that the Tj,max specification for the Intel® 82854 GMCH is 110 ºC, and the example described in the
section is only intended to illustrate the behavior. The hot trip temperature represents the temperature setpoint at
which the device will initiate throttling.
Figure 9. Intel® 82854 GMCH Temperature Throttling
Intel® 82854 GMCH Temperature
1. The system is operating at an idle workload until an application that requires a large amount of bandwidth is
initiated. The application demands a peak bandwidth of 1200 MB/sec. and the device will sustain this
bandwidth level until the temperature climbs above the hot trip setting of 100ºC.
2. During this test the device operates at a 1200 MB/sec. bandwidth level for a period longer than the sampling
window because the junction temperature has not increased above the hot trip point setting. In this case the
Intel® 82854 GMCH is demonstrating better bandwidth performance while operating under the same
application as in the bandwidth triggering case. This is clearly a preferred method of throttling the device only
when it is absolutely necessary.
3. Once the application stops the system workload will return to its idle level of 200 MB/sec. In this example, the
Intel® 82854 GMCH never required any thermal throttling. The method will potentially allow for large, brief
bursts of bandwidth loading without impeding the performance.
A number of natural convection thermal solutions were analyzed, and a set of baseline heat sink designs optimized
for the Intel® 82854 GMCH in natural convection applications were established for a typical set top box form
factor. As stated in Section 3.2 it is very important to consider all system and component boundary conditions when
designing a natural convection thermal solution.
Based on the optimum heat sink parameters implemented in the baseline heat sink design, a number of passive heat
sinks have been proposed by different suppliers for delivering the thermal performance required to cool the Intel®
82854 GMCH over a range of T
platforms requiring a passive thermal solution if the z-height allowed is similar to the heights typically allowed for
ATX family form factor solutions. The following sections provide details on the different passive heat sink designs
provided by a number of suppliers. A list of heat-sink suppliers who also provide the total thermal solution set,
including TIM and retention mechanisms, is provided in Section 7.1.
The performance predictions of the thermal solutions provided herein are for reference purposes only. These values
were predicted based on a thermal solution stack-up commonly employed for cooling FCBGA packages in a typical
set top box environment. As such, they do not include any variations that may be introduced by employing different
set of solution stack-up with respect to TIM and the retention mechanism, nor do they imply any statistical
significance. It is up to the system integrator to perform validation in the final intended system, including the heat
sink, attachment method, and thermal interface material.
variations as illustrated in Section 3.2.4. These heat sinks are a good fit for
amb
6.1 Aavid Passive Heat Sink
The Aavid heat sink is a black anodized aluminum extrusion, delivered with a low cost spring clip and pre-applied
thermal interface material. Aavid recommends Thermagon T-pcm or T-mate2905 for TIM. These are phase change
materials (PCMs) that are naturally tacky at room temperatures, requiring no adhesives or preheating. PCM means
the material will change properties at elevated temperatures to increase thermal performance. An assembled part
view is provided below. Dimensions are in mm.
Figure 10. Front and top views of Aavid passive heat sink
Intel® 82854 GMCH – Thermal Design Guide 27
The passive heat sink is made of 6063-T5 aluminum extrusion alloy and has been optimized for natural convection
applications while meeting the form factor and retention mechanism constraints of a typical set-top-box system. A
complete set of engineering drawings for the heat sink and the clip is provided in Figure 15 in “Appendix A
Mechanical Drawings”.
Thermal performance, Ψ
, of this heat sink in natural convection cooling is provided in Table 7. Anodizing the heat
sa
sink surface enhances the radiation performance at an incrementally added cost. Typically it would increase the heat
sink part cost by approximately 10%, but the performance gain of anodized heat sink is typically 10-20% over nonanodized parts. In natural convection applications, anodization or even thin surface painting often is a cost effective
performance enhancement method. Using the predicted heat sink performance characteristics, the corresponding
overall thermal performance of the component, Ψ
, is determined using a thermal network analysis. These
ja
parameters are tabulated below:
Table 7. Thermal performance of Aavid passive heat sink
Non-anodized Black Anodized
Ψsa (°C/W)
Ψja (°C/W)
T
amb_max
(°C)
7.5 6.3
6.7 6.0
71 76
The corresponding local maximum ambient air temperature, T
, that can be supported by the heat sink
amb_max
solution is also included in the table. For example, the black anodized heat sink is capable of cooling an Intel®
82854 GMCH at TDP=5.7W with local ambient temperatures up to T
specification of T
=110°C.
j
= 76°C without breaking the thermal
amb
28 Intel® 82854 GMCH – Thermal Design Guide
6.2 CCI Passive Heat Sink
The CCI Technology’s heat sink is an aluminum extrusion, delivered with pre-applied thermal interface material,
Powerstrate 51-SA AF-S-10H-07B or Shinetsu 7762 Thermal Grease, a gasket, and a low cost spring clip. An
isometric and the bottom views are provided below showing the TIM at the center and the peripheral gasket. A
complete set of mechanical drawing of the parts is provided in Figure 16 in the “Appendix A Mechanical
Drawings”. CCI Technology also recommends a spring loaded nylon fastener as an alternative retention mechanism
which may also be used with the heat sinks provided by other suppliers (see Figure 17 in the appendix for details).
Figure 11. Isometric and bottom views of CCI passive heat sink
This passive heat sink is also made of 6063-T5 aluminum extrusion alloy and has been optimized for natural
convection applications while meeting the form factor and retention mechanism constraints of a typical set-top-box
system. Thermal performance, Ψ
, of this heat sink in natural convection cooling is provided in Table 8. Anodizing
sa
the heat sink surface enhances the radiation performance at an incrementally added cost. Typically it would increase
the heat sink part cost by approximately 10%, but the performance gain of anodized heat sink is typically 10-20%
over non-anodized parts. In natural convection applications, it is often the situation that anodization or even thin
surface painting is a cost effective performance enhancement method. Using the predicted heat sink performance
characteristics, the corresponding overall thermal performance of the component, Ψ
, is determined using a thermal
ja
network analysis. These parameters are tabulated below:
Table 8. Thermal performance of CCI passive heat sink
Non-anodized Black Anodized
Ψsa (°C/W)
Ψja (°C/W)
T
amb_max
(°C)
The corresponding local maximum ambient air temperature, T
6.3 5.3
6.0 5.3
76 80
, that can be supported by the heat sink
amb_max
solution is also included in the table. For example, the non-anodized heat sink is capable of cooling an Intel® 82854
GMCH at TDP=5.7W with local ambient temperatures up to T
=110°C.
of T
j
= 76°C without breaking the thermal specification
amb
Intel® 82854 GMCH – Thermal Design Guide 29
6.3 Cooler Master Passive Heat Sinks
The Cooler Master heat sinks are delivered with pre-applied thermal interface material and a nickel plated high
strength steel clip. This TIM, Powerstrate* 51, manufactured by Power Devices, Inc., is a phase-change thermal
interface material. This means the material will change properties at elevated temperatures to increase thermal
performance. These phase change characteristics must be accounted for when testing the Cooler Master heat sinks.
At low temperatures, the heat sink performance will be significantly degraded, but at elevated junction temperatures,
the material will undergo a change phase and improve in performance. For more information, see the Power Devices
website at: http://www.powerdevices.com
place is provided in Figure 12.
Figure 12. Isometric view of Cooler Master passive heat sinks
. An isometric view of Cooler Master heat sinks with the clip inserted in
These passive heat sinks are made of 6063-T5 aluminum extrusion alloy and have been optimized for natural
convection applications while meeting the form factor and retention mechanism constraints of a typical set-top-box
system. A complete set of engineering drawings for the heat sinks, retention clip, and TIM is provided in Figure 18
in “Appendix A Mechanical Drawings
”.
30 Intel® 82854 GMCH – Thermal Design Guide
Thermal performances, Ψ
, of this family of heat sinks in natural convection cooling with and without anodization
sa
are provided inTable 9. Anodizing the heat sink surface enhances the radiation performance at an incrementally
added cost. Typically it would increase the heat sink part cost by approximately 10%, but as can be seen from the
table the performance gain of anodized heat sinks are well over 20% over non-anodized parts. In natural convection
applications, it is often the situation that anodization or even thin surface painting is a cost effective performance
enhancement method. Using the predicted heat sink performance characteristics, the corresponding overall thermal
performance of the component, Ψ
, is determined using a thermal network analysis. These parameters are tabulated
ja
below:
Table 9. Thermal performance of Cooler Master passive heat sinks.
Surface
Clear Chromated Black Anodized
Treatment
Proposal No. 1 2 3 1 2 3
Ψsa (°C/W)
Ψja (°C/W)
T
amb_max
(°C)
8.8 7.5 6.6 7.1 6.1 5.4
7.5 6.7 6.1 6.5 5.8 5.4
67 72 75 73 77 80
The corresponding local maximum ambient air temperature, T
, that can be supported by the heat sink
amb_max
solution is also included in the table. For example, the clear chromated No. 1 heat sink is capable of cooling the
Intel® 82854 GMCH at TDP=5.7W with local ambient temperatures up to T
specification of T
=110°C.
j
= 67°C without breaking the thermal
amb
6.4 Recommended Thermal Interface
Materials
It is important to understand and consider the effect of the interface between the processor and the heat sink base on
the overall thermal solution. Specifically, the bond line thickness, interface material area, and interface material
thermal conductivity must be selected to optimize the thermal solution.
It is important to minimize the thickness of the thermal interface material, commonly referred to as the bond line
thickness. A large gap between the heat sink base and the die yields a greater thermal resistance. The thickness of
the gap is determined by the flatness of both the heat sink base and the die, plus the thickness of the TIM, and the
clamping force applied by the heat sink attachment method. To ensure proper and consistent thermal performance
the TIM and application process must be properly designed.
The heat sink solutions analyzed in this document are assumed to be using a high compliant, low cost thermal
interface materials, such as Chmerics T710, Powerstrate 51-SA AF-S-10H-07B, or Shinetsu 7762 thermal grease.
Alternative materials may be used at user's discretion. Although heat suppliers provide various TIMs as a part of
total solutions, a list of vendors who specialize in TIM is provided in Section 7.2. The entire heat sink assembly,
including the heat sink, attachment method, and thermal interface material, must be validated together in its final
intended use.
Intel® 82854 GMCH – Thermal Design Guide 31
6.5 Recommended Attachment Methods
The thermal solution can be attached to the motherboard in a number of ways. The thermal solutions may be
designed with mounting holes in the heat sink base for a through-hole type of fastening mechanism, or a groove in
between the fins for a spring clip-like retention mechanism. A typical through-hole retention mechanism uses a set
of 4 spring-loaded fasteners to apply an even load on the die. For a spring-clip type mechanism, a set of mating
hooks need to be placed on the motherboard at corresponding locations, as shown in Figure 15 for the Aavid heat
sink. Different dimensions may be employed provided that there are enough clearances with other components on
the motherboard, as well as with the heat sink and a sufficient contact pressure to the TIM can be applied. For more
details and design specific modifications, consult with heat sink suppliers.
32 Intel® 82854 GMCH – Thermal Design Guide
77 VVeennddoorr CCoonnttaacctt IInnffoorrm
maattiioonn
7.1 Heat Sink Suppliers
Aavid Thermalloy LCC
80 Commercial Street
Concord, NH 03301 USA
Tel: 603-224-9988
e-mail:info@aavid.com
web:www.aavidthermalloy.com
CCI – Chaun Choung Technology Corp.
US: 2204 Forbes Drive, Suite 104