Home || News || Reviews || Articles || Contact || Store || Forums || Dumbass Database
News Archive
Dumbass DB

"When they call the roll in the Senate, the Senators do not know whether to answer 'Present' or 'Not guilty.'"
- Theodore Roosevelt

Canada Forums
Explosive Labs
Legion Hardware
MaXit Mag
Overclockers Club
Ultra X
Virtual Hideout
A Tale of Two Sockets
Page 2 - Physical Limitations and Conclusion
Author: Matt Krick (DCFlux)
Editor: Tim Whitcomb (DuffMan)
Date: January 13, 2004
Category: Information
Options: Printable Version 12 pt Times New Roman 10 pt Times New Roman 12 pt Tahoma 10 pt Tahoma

Sockets, like most materials, are also faced with temperature constraints. Too hot, and materials start to break down. Too cold, and the structural integrity can be compromised. Nevertheless, both test sockets were rated to operate between -55° C (-67° F, 218.15 K) and 125° C (257° F, 398.15° K). So, while overheating is definitely within the realm of possibility, damage due to excessive cooling is highly unlikely unless, of course, you light your barbecues with liquid oxygen and cool your computer with liquid nitrogen. Both sockets test about the same here.

We turn now to the construction of the socket itself. While most important, this is also the most easily overlooked. This is also where the socket differences begin to manifest themselves. Company A uses a phosophor-bronze construction with gold plating. Phosophor-bronze is a common material used to make connectors and sockets because it is malleable and can retain its shape. Company B uses a copper-alloy with gold plating. The gold plating is used because it resists corrosion, ensuring a consistently good contact between the pin and the socket. Both companies used a gold plate thickness of 15 microns. This is standard in the industry, and provides just enough not to wear off during normal use. Both companies offer a 30 micron thickness plating at an additional cost.

The socket manufacturers also provide warnings about detrimental conditions. Company A warns that "prolonged exposure to ultraviolet light may deteriorate the chemical composition used in the socket material." While UV light is deadly to most plastics over time, this is of little concern in the safe haven of a computer case. Shielded from the sun, the most dangerous UV source is that lovely cold cathode light you have illuminating your motherboard. Of course, since the socket itself is covered by a fan, a heatsink, and a CPU, you have very little to fear. Company B warns that the lifetime of their ZIF socket is approximately "50 cycles without functional damage". While this number is ridiculously high if you are a grandmother checking your AOL mail, if you do a lot of CPU switching this could spell an early death for your socket. You have your warning.

Socket462 Schematic
Socket462 Schematic

We've seen here that there are some issues with socket construction and design even the difference in the plastic used can yield different electronic properties to hinder your overclocking. Is there any way around this? If you're feeling particularly adventurous, you could solder the CPU directly to the motherboard. This ensures an extremely reliable connection that is, unless you happen to slip while you're soldering any one of 462 pins into an 8 layer board. It appears that we will have to deal with the shortcomings of sockets for some time.

What would I like to see in the future? I envision a new chip-motherboard interface, using a fewer number of larger pins to handle the power (perhaps 10-15 amps per pin). For the low speed bus, handshaking could be accomplished using only 32 regular pins. Optical technology makes high speed I/O possible with 32 or 64 optical transponders linked via light pipes to the motherboard. I call it Socket M/K Ultra. You can call it whatever you like.

Page 1
Main Page

Join Xtreme Tek's Mailing List

Copyright © 1999 - 2019 Xtreme Tek. All rights reserved.