PIII tulatian and Heatsink

[moosedick]

the tulatian's have that heat spreader on them. so how do i install the heatsink and thermal grease? do i have to remve the spreader? or do i install it as normal but just coat the spreader in grease? thanks

[Mr B]

Good question! I've been wondering that myself....

[Ebola]

didn't you guys ever have a ppga socket 370 chip. they have a huge surface like the tualatin. i think you just put a coat of grease over the whole thing.

[6502kid]

I hope Ebola is right, because thats what I did with mine.

The spreader sure didnt look like it was supposed to come
off....

My CPU temp stays at 41c or 42c no matter what I do.

I think the monitor is working, but just reading high.

I had my 800eb in the same board, and it ranged from about
43 to 48c or so depending upon where I had my fans and junk.

Let me know if you find out otherwise, and I will have to look
at mine closer. If I got it wrong, I guess this thing just runs
so cool it dont matter !


:mad: :D :mad:

[Hyperfish]

The same way you would for any chip.

Do NOT remove the IHS!

BTW, you might be able to drop the temps slightly by:

1. Ensuring the venthole on the IHS is clear and free from thermal paste.

2. The overall height of the Tualatin is higher than any previous FCPGA PIII, regardless of stepping.
Some HS clips probably apply too much clamping pressure unless modified because of the height difference.

2. I'm using the same Alpha PEP66 V2 (http://wsphotofews.excite.com/036/CO/nj/RN/tH33530.jpg) (modified for socket use) that I've used on the last 6 FCPGA chips, ranging from early cB0 PIII 700s to a cC0 1G.

My idle temps run 34'-35'C (http://wsphotofews.excite.com/001/Oe/ZG/Pk/Mq13967.jpg) using a Sunon 27cfm fan, full load rarely exceeds 41'C at 1440MHz to 1500MHz, (http://wsphotofews.excite.com/030/qJ/Hx/qV/OA31226.jpg) (160-166FSB) 1.56vCore. (Asus TUSL2-C, 1009.3 Beta BIOS, Week 19 1.2GHz Tualatin (http://wsphotofews.excite.com/031/Gp/Hc/wU/uf34761.jpg))

This (http://wsphotofews.excite.com/037/ve/Mm/T9/zv47904.jpg) temp was after 14 loops of Sandra Burn In, room temp was 75'F. Temp peaked at 41'C after the 50th loop.
Case is a Yeong Yang YY00221 (http://wsphotofews.excite.com/006/GE/YH/j1/Dz96624.jpg) Cube case, which has rather poor airflow in the motherboard chamber side.

[Gabertooth]

Ah Fish,

The vent hole in the IHS is only for the curing process of the last stage manufacturing of the processor. After that its worthless. I've heard that it's no different from a Cumine (on the surface anyway) under the IHS, so getting ever electrically conductive stuff in there shouldn't matter.

-Gabertooth



Originally posted by Hyperfish
The same way you would for any chip.

Do NOT remove the IHS!

BTW, you might be able to drop the temps slightly by:

1. Ensuring the venthole on the IHS is clear and free from thermal paste.

2. The overall height of the Tualatin is higher than any previous FCPGA PIII, regardless of stepping.
Some HS clips probably apply too much clamping pressure unless modified because of the height difference.

2. I'm using the same Alpha PEP66 V2 (http://wsphotofews.excite.com/036/CO/nj/RN/tH33530.jpg) (modified for socket use) that I've used on the last 6 FCPGA chips, ranging from early cB0 PIII 700s to a cC0 1G.

My idle temps run 34'-35'C (http://wsphotofews.excite.com/001/Oe/ZG/Pk/Mq13967.jpg) using a Sunon 27cfm fan, full load rarely exceeds 41'C at 1440MHz to 1500MHz, (http://wsphotofews.excite.com/030/qJ/Hx/qV/OA31226.jpg) (160-166FSB) 1.56vCore. (Asus TUSL2-C, 1009.3 Beta BIOS, Week 19 1.2GHz Tualatin (http://wsphotofews.excite.com/031/Gp/Hc/wU/uf34761.jpg))

This (http://wsphotofews.excite.com/037/ve/Mm/T9/zv47904.jpg) temp was after 14 loops of Sandra Burn In, room temp was 75'F. Temp peaked at 41'C after the 50th loop.
Case is a Yeong Yang YY00221 (http://wsphotofews.excite.com/006/GE/YH/j1/Dz96624.jpg) Cube case, which has rather poor airflow in the motherboard chamber side.

[Hyperfish]

Originally posted by Gabertooth
Ah Fish,

The vent hole in the IHS is only for the curing process of the last stage manufacturing of the processor. After that its worthless. I've heard that it's no different from a Cumine (on the surface anyway) under the IHS, so getting ever electrically conductive stuff in there shouldn't matter.

-Gabertooth





Gabertooth-

I've also read the Intel white papers.

My comment is based on first hand experience of my own Tualatin rig, as well as 4 additional Tualatin rigs I own.

In ALL cases, temps decreased an average of 3'C - 5'C at idle, clear vent hole vs. thermal paste smeared over it.

I've also experienced identical results on 14 P4 rigs I have built or owned, both S423 & S478.

Bottom line, in MY experience, a clear vent hole makes a difference in the real world.

[Gabertooth]

Thanks Fish,


Thanks you for reminding me what it is like "in the real world," I had forgotton that real world performance characteristics can differ. Although I thought I'd get real world numbers from my Iwill DVD266u-rn with both sockets populated..I could be wrong. Also thanks for automatically assuming the your the only one with this blessed chip, I know were all Intel fans here but I also believe that since the "real world" is so dynamic that individual results can be influanced by multiple factors..giving credence to the fact that small variations in a system give rise to differeant outcomes. I appreciate the perspective, although you must concede that since we can never maintain a perfectly closed system, the you not just testing with the vent "open" and "closed".

-Gabertooth





Originally posted by Hyperfish

Bottom line, in MY experience, a clear vent hole makes a difference in the real world.

[Hyperfish]

Originally posted by Gabertooth
[B]Thanks Fish,


Thanks you for reminding me what it is like "in the real world," I had forgotton that real world performance characteristics can differ. Although I thought I'd get real world numbers from my Iwill DVD266u-rn with both sockets populated..I could be wrong. Also thanks for automatically assuming the your the only one with this blessed chip, I know were all Intel fans here but I also believe that since the "real world" is so dynamic that individual results can be influanced by multiple factors..giving credence to the fact that small variations in a system give rise to differeant outcomes. I appreciate the perspective, although you must concede that since we can never maintain a perfectly closed system, the you not just testing with the vent "open" and "closed".

-Gabertooth


Gabertooth-

Just a quick FYI-

I tested ALL the Tualatins within a 3 hour period using my own setup:

Yeong Yang case, 1 x 92mm 53cfm intake, 2 x 80mm 43cfm exhaust fans
Asus TUSL2-C, 1009.3 BETA BIOS
Weeks 19 – 22 1.2GHz Tualatins
Modified Alpha PEP66 V2, 27cfm Sunon, Arctic Silver
2 x 256MB Mushkin High Perf Rev. 2 PC133
2 x 45G IBM 75GXP, RAID0, Promise FastTrak100
1 x 27G IBM 34GXP
Gainward CardExpert GF3
SBL
Intel Pro 100S NIC
Tri-boot Win98SE/Win2K SP2 Pro/Win XP Pro

Ambient temps never varied more than +/- 2.5’F during the tests.
Temps were measured using BIOS, MBM & Asus Probe at idle & 100% sustained CPU load stock speeds.
Temps were again measured Ocing each chip as high as they would go using 1.57vCore as the upper limit.

Testbed for the P4 chips:
Various weeks 1.7GHz P4
Stock Intel HSF, AS
Antec SX1030B, Enermax 431W PSU
2 x 80mm 43cfm intakes, 2 x 80mm 47cfm exhaust fans
Asus P4T, 1005 BETA BIOS
2 x 256MB Samsung PC800 RDRAM
2 x 40G IBM 60GXP, RAID0, Promise FastTrak 100TX2
1 x 60G IBM 60GXP
SIIG SCSI adapter
Elsa 920 GF3
SBL
Intel Pro 100S NIC
Plextor Ultra Plex
Sanyo BP-4 16/10/40
Dual boot Win98SE/Win2KSP2 Pro

Temps were measures similar to above method. (Max OC vCore limit was 1.85V)
ALL S478 chips were tested using an Abit TH7II, same components, same case.

I know, it’s not perfect, but it’s the best I could do using my own equipment and $$$ within a given time limit.
Dunno how you feel, but I certainly feel my testing was as thorough & controlled as real world conditions allow…

:)