|
Advice Beginners BIOS Guide CPUs Links Mainboards Memory Network Storage Video/Sound Cards Contact Forum SiteMap Sponsors WebNews Home |
. | . |
Prices: Mainboards ABIT ASUS Chaintech Shuttle Soyo Tyan CPU Intel P4 2.4C-800 P4 2.6C-800 P4 2.8C-800 P4 3.0-800 P4 3.2-800 AMD AthlonXP XP 1700+ XP 2000+ XP 2400+ XP 2500+ XP 2700+ XP 3000+ XP 3200+ Athlon64 Athlon64 3200+ Athlon64 FX-51 Opteron Opteron 240 Opteron 242 Opteron 244 Opteron 246 Memory Corsair Crucial Kingston Mushkin OCZ |
LOSTCIRCUITS
|
|
| Intel LGA775 SocketT New and (Un) improved? | |
|
(Review by MS, July 28, 2004) |
|
OCZ PC3200 DUAL-CHANNEL EL DDR 512MB(256X2) 400MHz DDR CAS2 - PLATINUM |
Socket 775 LGA, The Pearly Gates To Processor Heaven?
The amount of misconceptions woven around the new processor interface and its land grid array (LGA) contact I/O configuration has been nothing but stunning. Briefly, the new interface was touted to provide a new thermal solution since the surface area of the "bumpless" contacts was greater than the pins and, moreover, the entire processor was not thermally insulated from the mainboard by a rather thick piece of plastic known as the Socket 478. Whether there is any sense in that argument is a matter of perception, truth is that the contacts are far enough from the die to be considered thermally isolated on the substrate with the only connection being the bond wires attaching them to the die itself. Keep in mind that bondwires are typically in the order of some 30 µm diameter and as such would be very poor conductors of heat to begin with. In addition, there is no cooling beyond the socket itself, which means that -maybe- some thermal transients can be buffered over a very short period of time, however, as soon as a thermal equlibrium is reached, there will be no further benefits.
Spring-Loaded
There certainly are merits for the new design, the greatest being the fact that there are no more bendable pins on the CPU. Whoever has tried to rearrange CPU pins into an orderly pattern suitable for reinsertion into a socket will know what I am referring to. I actually tried a comparison between the old and the new design by dropping both CPUs from a standard desk onto a standard office carpet. The result was a socket 478 CPU that was damaged beyond repair whereas the Socket "T" (LGA 775) processor survived about a dozen trials without any harm.
On the surface, the new processors don't look that different other than that "full metal jacket" of the socket enclosing the top edge of the CPU. The CPU itself does not look too alien either, at least as long as we are looking at the top. The bottom view reveals the land grid array (LGA) interface with its 775 contacts that has spawned the name LGA 775 or Socket T (for those with speech impairment or other forms of dyslexia)
Srings for Pins
If there are no pins and the surface is as flat as a land grid array, there has to be another mechanism to warrant contact for all viable pins. The solution in this case is the use of spring-loaded contacts integrated into the socket. The CPU is pressed down by the "metal jacket" onto the pins with enough force to ensure even distribution of contact pressure amongst all pins.
| P4 2.4E (Prescott) At: |
The main concern about the new interface has been from the mainboard manufacturers who observed fatigue of the socket itself after a few changes of processor. There is certainly some truth to it but on the other hand, aside from the elusive overclocker community there is relatively little concern about this becoming a real issue. With regard to the drop test we were not able to establish a significant difference in endurance between Socket 478 and Socket 775 either.
Cooling
As mentioned above, from a technical point of view, it is hardly conceivable that there is any thermal advantage of the new socket over the old one. In absence of real hard measurements we are using inferential evidence in form of the latest Intel creations in thermal management: Enter the helicopter with its unprotected blades waiting for unsuspecting fingers reaching into the case....
The new coolers exceed pretty much everything we have seen so far from Intel and finally provide a semi-intelligent mounting mechanism.
At this point it is difficult to weigh the pro's vs the con's in any but a qualitative fashion, however, it appears as if the initial concerns about pin fatigue may have been exaggerated since we changed CPUs about 1/2 dozen times without noticing any problem. Time will tell, however, our initial impression is more positive than negative.
next page: => DDR2, Serial ATA-2 =>
All advice and educational articles on LostCircuits are free, but if you feel you can, please make a small donation to us!