Field replaceable module with enhanced thermal interface

Details Field replaceable module with enhanced thermal interface Field replaceable module with enhanced thermal interface

2000-03-24

2002-04-23

Picard, Leo P. (Department: 2835)

Electricity: electrical systems and devices

Housing or mounting assemblies with diverse electrical...

For electronic systems and devices

C361S692000, C361S726000, C257S719000, C174S016300, C165S185000

Reexamination Certificate

active

06377453

ABSTRACT:

FIELD OF THE INVENTION
The present invention pertains generally to cooling of electronic components, and more particularly to thermal interfaces of field replaceable modules.
BACKGROUND OF THE INVENTION
Operation of high speed electronic components produces unwanted heat. For example, high speed computer processor elements such as microprocessors, graphics processors and the like generate unwanted heat that must be removed for efficient operation. Heat removal provides for a lower operating temperature, higher operating speeds and greater computing power. Additional advantages include higher reliability and availability.
To meet ever higher requirements for computing power, processor designs continue to evolve, becoming more complex and operating at ever higher speeds. More complex designs integrate greater and greater numbers of transistors, which each contribute to generation of more heat during operation. As each transistor is operated at higher speeds, heat generation is further increased.
Various cooling schemes are known in the prior art. In general, as cooling schemes become more efficient at removing heat, mechanisms to implement the schemes become larger, heavier, bulkier and more difficult to arrange in computer systems.
The large amount of power required to operate high speed computer systems results in large heat dissipation. Accordingly, cooling techniques are required. In some cooling schemes of the prior art, bulky mechanisms such as heat sinks (including cold plates) for implementing the schemes mechanically interfere with field replacement of modules, including processor modules, board modules, system modules and the like. Additionally, there is a problem of thermal interface resistance between the field replaceable module and the heat sink.
In general, thermal interface resistance is the result of imperfect mechanical contact between two opposing surfaces (i.e., between a lid of an integrated circuit package (or other field replaceable module) and heat sink). This imperfect mechanical contact results from collision of irregular “peak and valley” surface defects on both opposing surfaces.
With a substantial amount of contact occurring only at where the peaks collide, there is a bottleneck restriction for the heat flow, since the space in the valleys is either air or a vacuum. Air is a poor conductor of heat.
Depending on the nature of the irregular surface defects, thermal interface resistance per unit area can be one Inch squared-degree Centigrade per Watt of heat. For a typical processor application, this translates to about a 20° C. to 30° C. rise. Since interface resistance is a function of area, the smaller the interface area, the higher the resistance. For this reason, this phenomena is becoming more and more critical in the electronics cooling area since semiconductor package footprints are getting smaller, with higher heat fluxes. The equation for thermal resistance can be represented with the following equation:
R
interface
=C/A
where C=interface resistance per unit area and A=Area of interface.
A number of different schemes to eliminate such irregular surface defects or to obviate thermal effects of their collision are known in the prior art.
1) Diamond Turned Surfaces—One method for improving the interface resistance between materials is to substantially eliminate irregular surface defects on the mating surfaces. Diamond turned surfaces have mirror-like surfaces that are typically used in optics. Although the interface performance is quite good, it is not very popular because it is extremely expensive. In addition, high pressure is needed for interface contact. This method is only used where longitudinal axis tolerance is tightly controlled (wherein the longitudinal axis is understood to be perpendicular to a major surface of the lid of an integrated circuit package (or other field replaceable module).
2) Thermal Grease—Thermally enhanced greases have been popular for years. They fill the voids created by the valleys in the irregular surface defects. Although the conductivity is still orders of magnitude lower than that of metals, it is still orders of magnitude better than air. Thus, C, the thermal resistance per unit area, decreases. This method is only used where longitudinal axis tolerance is tightly controlled.
The problems associated with grease are:
Grease migrates and is messy.
Grease is hard to control, too much grease will actually decrease thermal resistance since the peaks of the two surfaces may not be touching. For this reason manufacturing does not like to use grease.
In some cases, grease may migrate out of the interface area as a result of temperature cycling. This phenomenon is commonly known as “pumping”.
Grease can age and separate resulting in decreased performance.
Moderate pressure is needed on the interface to ensure contact.
3) Thermal Pads—Pads are currently the most popular interface enhancement method. They are generally thermally enhanced silicone based pads that range in thickness of 3-20 mils. This method is only used where longitudinal axis tolerance is moderately controlled. The popularity of these pads has grown in recent years because they are easy to use during manufacturing with a well controlled attach method. In addition, it can be pre-applied to the interface. The problem with this method is that the thermal resistance is generally more than a factor of two higher than grease. In many applications this is not good enough. In addition, these pads need interface pressures greater than one-hundred pounds-per-square-inch (100 psi) to work.
4) Gap Pads—These are interface pads that fill gaps that are sixty to two-hundred thousandths of an inch (60-200 mils) thick. They provide some limited advantages for interfaces where gap tolerance is not well controlled. However, a major problem is that their resistance is generally ten to twenty times higher than standard thermal pads. Accordingly, in general, they are not suited for high heat flux applications. Moderate interface pressures are needed for optimum performance.
5) Phase Change Materials (PCM)—Another popular interface material are PCMs. The most common are paraffins that come on a thin carrier (2-5 mils) such as aluminum or a screen. These PCMs work on the principle that above a certain temperature such as 51° C., they reflow and fill the voids in the interface. The performance is comparable to that of grease. The carrier makes the use of PCMs easy to implement in manufacturing and easy to use. Moderate pressure is needed on the interface. This method is only used where longitudinal axis tolerance is tightly controlled.
6) Metal Pastes—Metal pastes are not commercially used because they are electrically conductive and poisonous.
In all cases discussed previously herein, some pressure is needed for the interface solution to perform. In each case it is beneficial to minimize the interface material to lower the thermal resistance. In addition, none of the above solutions lend themselves well for a sliding contact.
Liquid cooling methods, in which a liquid is pumped through a cold plate coupled to an integrated finned heat sink, is becoming more popular for use in larger modules such as processor modules, board modules, and system modules. The liquid conduits are typically coupled to the modules themselves and therefore mechanically interfere with the field replacement of modules. In particular, due to the liquid coupling between the modules and chassis, the liquid conduits coupled to the modules must be disconnected prior to exchanging modules. This results in a less efficient field replaceable module exchange method and is prone to leakage of the liquid from the disconnected conduits. In addition, this makes “hot swapping” of modules, which is becoming more and more important in larger systems or in mission-critical systems that require redundancy due to the need to provide “always-on” service, impossible. As used herein, the term “hot swap” refers to the ability of a field replaceable module to be connected to and disconnected from a compute

Affiliated with

Also associated with

Rating

Field replaceable module with enhanced thermal interface does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Field replaceable module with enhanced thermal interface, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Field replaceable module with enhanced thermal interface will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-2930991