Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Technology Stocks : Advanced Micro Devices - Moderated (AMD) -- Ignore unavailable to you. Want to Upgrade?

To: hmaly who wrote (78010)	4/23/2002 11:50:20 AM
From: tcmay	Read Replies (3) \| Respond to of 275872

"I believe Mani is saying that, instead of adding another hole as you suggest; if you thicken the walls of the bucket from say 1/8 inch to 1/2 inch, you will slow down the flow. Why would a heat spreader add parallism over a heat sink? " Looking at several designs, including the URL I gave, and also Intel's PIII/P4 descriptions, it's clear that heat spreaders and slugs are used to make contact to the rest of the package. Obviously so, but it bears repeating. As to the "parallelism" and "heat flows in 3 dimensions" point (made by PGerassi in another post recently), of course. Here's an analogy, even one basically isomorphic to the actual package situation: A man is sitting on a wooden bench on a cold winter day. The wood is cold, about 30 F. Being wood, it's not a great conductor of heat. It's not a perfect insulator, but it's not great. The heat flow from the man's butt to the bench is limited by this less-than-great conductivity. To increase the heat flow, the man places a 1-foot by 6-foot metal plate on the bench and then sits down on that. The metal plate is an excellent lateral (and isotropic) conductor of heat. Being that it makes contact with the wooden bench over much more area than the man's butt, it drains heat out of the man with amazing efficiency. It is, of course, a heat spreader. _This_ is the sense in which I meant that the addition of a heat spreader does NOT increase thermal resistance between the die and external world. It _would_ if the area of the heat spreader matched the area of the die. A man sitting on a metal plate precisely as large as where his butt makes contact would see no improvement in heat transfer, in the steady-state situation. But if the heat spreader makes contact with a material of lower thermal conductivity over a larger area, the larger area can (and in the case of the man sitting on the metal plate) make a huge difference. Looking at the many designs for heat spreaders (and also for slugs), it's clear that this is exactly what they are being used for: to make contact with over a large effective surface area. (And, as in the metal plate-wooden bench situation, where the metal plate may have an air gap because of deviations from planarity, using pressure to push them together may be useful. Or using a fluid intermediary, such as a deformable gasket material (like soft copper used in vacuum systems) or thermal paste.) Now it may well be that placing the die directly on a high thermal conductivity leadframe, substrate, metal can, whatever, is an even better thermal situation. That's how it used to be done, of course. (As an aside, I worked on beryllia packages, because of the extremely high thermal conductivity of beryllia as compared to alumina.) But it's clear from the schemes where heat spreaders and slugs are used that this direct attachment is not always possible, or desired for various reasons. And in these cases, adding a heat spreader does NOT increase overall thermal resistance, as Mani claimed, but in fact reduces it. Because of the area effect. Another isomorphic example: electrical grounding. Imagine ordinary soil, or "ground." Suppose one wishes to "ground" an electrical system. According to Mani, placing anything between the system and ground will _add_ to electrical resistance. However, clearly adding an "electrical spreader" will help: a large sheet of copper laid on the ground, or multiple wires laid on the ground, or multiple rods pounded into the ground...all will help. As in the case of using a heat spreader in a package, adding these electrical spreaders improves the transfer. When Mani said that adding a heat spreader cannot help because it's in series between the die and the outside world, this was clearly wrong. If he was arguing that he would prefer solutions where the die is place directly on a leadframe or high thermal conductivity substrate, where a heat spreader could be skipped, this is a completely different issue. The fact that heat spreaders (and slugs) are so common tells us that they are useful. (BTW, there are arguments for using heat spreaders even with traditional lead frame materials. Neither Kovar nor Alloy 42, for example, are as conductive as certain other alloys, so using a heat spreader to increase the overall surface area ("shunt paths," in a sense) could be advantageous.) --Tim May