intel's boxed CPU coolers...

sabbath1 · Jul 18, 2003

If I understood it correctly, it was with the heatsink-removal video here at THG that AMD was first alerted with this issue that even the Palomino could smoke. After that I assume that AMD thought it was too much trouble to revamp the architecture and add an integrated protection logic. Cost was surely another factor.

My system: Intel Pentium 4 3.0, 800FSB / TwinMOS 1Gb DDR400 / MSI 875P Neo / Sapphire Radeon 9800Pro / Antec TruePower 550W / 2x Western Digital Raptor / Hercules G.T XP /
SamsungDVD / Lite-On CDRW

sabbath1 · Jul 18, 2003

What kind of thermal protection are we talking about and where did you read up on that?

Actually I knew this before THG wrote their review of Opteron. In that review, you'll find that the Opteron features integrated thermal shutdown. The only way this protection is worse than that of the P4/Xeon/Itanium is that the Opteron/Athlon 64 shuts down abruptly instead of clock-throttling. But it sure seem to work.

But before that of the THG review, I had asked AMD's European support center about this, or actually I complained about their lousy take on thermal protection, and so they answered that Opteron *and* Athlon 64 will feature integrated heat logic. It's about time, I'd say.

There was already a thermal protection mechanism using the on-die diode, ever since Palomino. It was an auto-shutoff mechanism, which only had to signal the mainboard to turn off.

Yes, I know that. But if you have a mainboard that don't support the diode, the processor will smoke. And in some cases, namely ASUS older boards with C.O.P, the cpu could still smoke if you booted it without a properly seated heatsink, because the board needed to be able to reach BIOS before it could activate the protection, I saw evidence of that myself in a german review. Test CPU was a XP1800+.

My system: Intel Pentium 4 3.0, 800FSB / TwinMOS 1Gb DDR400 / MSI 875P Neo / Sapphire Radeon 9800Pro / Antec TruePower 550W / 2x Western Digital Raptor / Hercules G.T XP /
SamsungDVD / Lite-On CDRW

eden · Jul 18, 2003

Cost was surely another factor

The P4 circuitry probably doesn't even take more than 5% of the die. So I doubt cost was an issue. Either they couldn't use clock throttling because Intel might sue, or it was because they had no time.

What I still wonder though, is if the circuitry is on-die for the P4, doesn't it still have to send a message (which could delay the mechanism to prevent the burn)to the mainboard's clock generators? Or are these on the CPU itself and therefore it's a matter of nanoseconds before it responds?

--
<A HREF="http://www.lochel.com/THGC/html/news.html" target="_new">The official Tom's Hardware Guide Forums Photo Album, click here to contribute!</A>

sabbath1 · Jul 18, 2003

The P4 circuitry probably doesn't even take more than 5% of the die. So I doubt cost was an issue. Either they couldn't use clock throttling because Intel might sue, or it was because they had no time.

Both of those alternative could no doubt be right, but they're still bad excuses if they came from AMD

.
There doesn't have to be clock-throttling to protect a CPU from overheating, because it could safely shutdown via the die-integrated protection anyways.
And, if a company doesn't have the time to make quality products maybe they should rethink their wishes to be a respectful and decently performing company.
Don't get upset, Eden, I'm not bashing you here

What I still wonder though, is if the circuitry is on-die for the P4, doesn't it still have to send a message (which could delay the mechanism to prevent the burn)to the mainboard's clock generators? Or are these on the CPU itself and therefore it's a matter of nanoseconds before it responds?

I will shoot myself if I'm wrong, but I'm 99% sure that the P3 and P4's heat shutdown logic is on the CPU die itself, and thus those CPU can never smoke. Unless, of course, the protection of 1 CPU out of 100 is faulty

So the P4 should have a much safer protection which also acts many times faster than the AMD protection could. So in theory, if you had your computer near an oven or something and if you removed your heatsink, it's possible that the AMD processor still couldn't manage to survive due to it's nature of rising in temperatures, very quickly.

My system: Intel Pentium 4 3.0, 800FSB / TwinMOS 1Gb DDR400 / MSI 875P Neo / Sapphire Radeon 9800Pro / Antec TruePower 550W / 2x Western Digital Raptor / Hercules G.T XP /
SamsungDVD / Lite-On CDRW

pIII_Man · Jul 18, 2003

i know for sure that the pIII's were totally on die and need no feedback from the mobo to quit...possibly the new p4's are throtalling via muliplyer instead of fsb...if this were the case since the multi is controlled by the cpu then that would mean that the cpu could adjust it @ will.

There is no smell better than fried silicon :evil:

sabbath1 · Jul 18, 2003

I believe even the P2's had on-die protection. P1's didn't, but they were running quite cold, if my memory serves me.

Don't know about that multiplier thing.

My system: Intel Pentium 4 3.0, 800FSB / TwinMOS 1Gb DDR400 / MSI 875P Neo / Sapphire Radeon 9800Pro / Antec TruePower 550W / 2x Western Digital Raptor / Hercules G.T XP /
SamsungDVD / Lite-On CDRW

pIII_Man · Jul 18, 2003

well i could test it out i got a pII 233...that doesn't see much use...hehe maybe i could live up to my siggy

There is no smell better than fried silicon :evil:

pIII_Man · Jul 18, 2003

in reality though i doubt the pII's had thermal protection...i have run the destchues (sp?) core w/o a heatsink and the die temp did not go above 46C...not that i recomend it..but i just doubt they had em...possibly the katalami or whatever they are called could fry...2.8 is a lot of volts but i still doubt that they had it in those days.

There is no smell better than fried silicon :evil:

sabbath1 · Jul 18, 2003

Haha, you do that and tell us what happened! :lol:

Someone I trust who is very knowledgable about these things told me that even P2's would survive a heatsink removal. I don't know this myself, but since I trust him, he may be right. Then again, he may not be right...hehe
As long as you don't try to sue me, you're welcome to try though

My system: Intel Pentium 4 3.0, 800FSB / TwinMOS 1Gb DDR400 / MSI 875P Neo / Sapphire Radeon 9800Pro / Antec TruePower 550W / 2x Western Digital Raptor / Hercules G.T XP /
SamsungDVD / Lite-On CDRW

pIII_Man · Jul 18, 2003

if you read my last post...i have done it to the 2.0 volt core i think its called the destcues and it works fine...the katalami or whatever the 2.8v core is called i'm not too sure if that could survive a heatsink removal.

There is no smell better than fried silicon :evil:

sabbath1 · Jul 18, 2003

Ah the Deschutes. Was that a PII or a PIII again? Cool to hear that it survived the heatsink removal 😉
It's also totally unbelievable that AMD's first properly working thermal protection comes with Athlon 64/Opteron.

My system: Intel Pentium 4 3.0, 800FSB / TwinMOS 1Gb DDR400 / MSI 875P Neo / Sapphire Radeon 9800Pro / Antec TruePower 550W / 2x Western Digital Raptor / Hercules G.T XP /
SamsungDVD / Lite-On CDRW

co0lguyz · Jul 19, 2003

I'd got my new proc p4 3.06 ghz, and with an asus p4g8x mobo. I used the intel's boxed cpu fan, it made my cpu's temp into 63 (idle) and become 86 (full load). Is it normal?

pIII_Man · Jul 19, 2003

NO...where are you reading these temps from...try using motherboard monitor...

There is no smell better than fried silicon :evil:

pIII_Man · Jul 19, 2003

Both pIIs...i think i may have got em flipflopped...i don't really remember which one used the higher 2.8v...but whatever it was it ran HOT and it would not be smart to remove the heatsink...the pII 400 that i removed the heatsink from was not running all that hot...around 50c...that is just fine for a pII...

After i slapped a heatsink on that pII 400 i found it was a nice overclocker (got it to 4x 133 stable).

Yes it is amazing...but really amd's stock coolers (or coolers within spec) have a very small chance of falling off (especially if installed by an OEM who knows what they are doing...hopefully)...although a good feature...not as terribly important as it may sound IMO there is a much higher chance of the Fan dying than the heatsink falling off....which in that case most motherboards will warn you or autoshutdown no matter what the temp is.

There is no smell better than fried silicon :evil:

co0lguyz · Jul 19, 2003

I got it from Asus Probe 2.20.03. How many temperatures will the temps decrease if i use the thermal paste? Do the new proc already have the thermal paste?

pIII_Man · Jul 19, 2003

try using a program called motherboard monitor (google it) it reads the processors temperature from a thermalcouple inside of the processor...as opposed to asus's very inacurate anal probe.

There is no smell better than fried silicon :evil:

pIII_Man · Jul 19, 2003

and no...they use some sort of thermal pad..greese will help...but as long as you have installed your heatsink right it should be in the 40's.

There is no smell better than fried silicon :evil:

co0lguyz · Jul 19, 2003

I've downloaded it. It's no use, that program reports the same temp as asus's

HardWareBoss · Jul 20, 2003

Nothing like the smell of Napalm.

sabbath1 · Jul 25, 2003

Both pIIs...i think i may have got em flipflopped...i don't really remember which one used the higher 2.8v...but whatever it was it ran HOT and it would not be smart to remove the heatsink...the pII 400 that i removed the heatsink from was not running all that hot...around 50c...that is just fine for a pII...

After i slapped a heatsink on that pII 400 i found it was a nice overclocker (got it to 4x 133 stable).

Yeah PII's had good thermal handling as well.
So for how long were you running that CPU without a cooler?

Yes it is amazing...but really amd's stock coolers (or coolers within spec) have a very small chance of falling off (especially if installed by an OEM who knows what they are doing...hopefully)...although a good feature...not as terribly important as it may sound IMO there is a much higher chance of the Fan dying than the heatsink falling off....which in that case most motherboards will warn you or autoshutdown no matter what the temp is.

Well, no matter if the fan fails or heatsink comes off in any strange way, the Thunderbird CPU's would always start to smoke. And as you know, even the latest Bartons can smoke if there's no proper protection on the motherboard. Thus AMD processors still have a few safety deficits that no one can deny.

But, good times ahead, Athlon 64 will finally bring up a product quality that's on par with Intels.

My system: Intel Pentium 4 3.0, 800FSB / TwinMOS 1Gb DDR400 / MSI 875P Neo / Sapphire Radeon 9800Pro / Antec TruePower 550W / 2x Western Digital Raptor / Hercules G.T XP /
SamsungDVD / Lite-On CDRW

sabbath1 · Jul 25, 2003

Thanks a lot for this info! I took a look at the 4300 cooler, and that is apparently rated much lower in noise than how it actually performs, so the cooler goes back, unopened, to the store.

Now I'm shooting for the upcoming Arctic Cooling Super Silent 4 Pro, that one is supposed to be compatible with my processor.

My system: Intel Pentium 4 3.0, 800FSB / TwinMOS 1Gb DDR400 / MSI 875P Neo / Sapphire Radeon 9800Pro / Antec TruePower 550W / 2x Western Digital Raptor / Hercules G.T XP /
SamsungDVD / Lite-On CDRW

sabbath1 · Jul 25, 2003

Can heatpipe fans can be as good as normal air fans?

My system: Intel Pentium 4 3.0, 800FSB / TwinMOS 1Gb DDR400 / MSI 875P Neo / Sapphire Radeon 9800Pro / Antec TruePower 550W / 2x Western Digital Raptor / Hercules G.T XP /
SamsungDVD / Lite-On CDRW

Syndil · Jul 26, 2003

Can heatpipe fans can be as good as normal air fans?

Absolutely. In fact, if done properly, a HSF with a heatpipe should always outperform a similar HSF without one. By "done properly" I mean that the heatpipe is actually a functioning part of the HSF, and not just added for looks, as may come to be the case if heatpipes start to become in vogue. I expect that once they really catch on, there will be lots of cheap immitators using things that may look like heatpipes, but are non-functioning. Just like there are lots of wow-look-at-me heatsinks out there right now that are fairly lousy coolers, done strictly for looks. So beware of that in the future. But for now there are few HSFs that I am aware of that use a heatpipe, and they all seem to work. Coolermaster's designs are a perfect example of how simple and effective adding a heatpipe to a HSF can be. Their <A HREF="http://www.coolermaster.com/retail/product_detail.asp?lang=eng&at=boutique&category_id=1&product_id=11" target="_new">HHC-001</A> is a fairly standard looking heatsink, excepting the heatpipes. The heatpipes simply provide a low resistance heat path to the far ends of the cooling fins. The thin metal fins provide quite a bit of resistance to heat, so the further away from the base a fin is, the less heat it will be able to absorb and dissipate. The heatpipe provides a low resistance heat path to the far ends of the heatsink's fins, making them more effective.

Also I noticed in THG's recent <A HREF="http://www.tomshardware.com/howto/20030722/index.html" target="_new">review of mini-PCs</A> that two of the systems use heatpipe heatsinks. Makes perfect sense for such a small enclosure -- not much room for a large HSF, so there is a need to provide an alternate escape path for the heat, in the form of heatpipes. Both the models by Shuttle and Soyo use a heatpipe system for CPU cooling.

Personally I have decided on using a <A HREF="http://www.dansdata.com/ncu1000.htm" target="_new">NCU-1000</A> for my next PC. I intend to build an absolutely silent system, and this heatpipe cooler is so effective that I should be able to get away without using any sort of fan at all in the box, except for the PSU fan. In fact, this cooler doesn't even come with a fan, as it is intended to be used without one. I expect the NCU-1000 to reign supreme as a silent CPU cooler for quite a long time. The design is very intelligent, and the heatpipes used are not your average heatpipe either. Instead of the relatively thick round tubes found in other heatsinks, they use wide, flat heatpipes, filled with small capillaries of working fluid. This provides an even better interface with the heat source, making this heatpipe extraordinarily effective. Also the nature of the tiny capillaries make it more resistant to gravity than a normal heatpipe. If you're as big a nerd as I am, you can click <A HREF="http://www.tsheatronics.co.jp/english/technology/index.html" target="_new">here</A> to learn more about the HeatLane heatpipes in this system. Fascinating stuff, if you're interested in thermodynamics.

_________________________________________________
Armadillo[TcC] at Lanwar and MML

traviss187 · Jul 26, 2003

this probably would have fit better near the beginning of the post but just got to it, personal opinion but one reason I think for the high noise level of intel's stock heatsinks is that the northwood core was never intended to run at 3ghz or anywhere near that, but intel delayed release blah blah blah so the heat sinks wern't meant to deal with the heat that the processors are putting out, they just kept increasing fan speeds... I'm hoping to see a new heat sink design for the prescott core, but then again that might just be wishfull thinking

Syndil · Jul 27, 2003

I expect that once they really catch on, there will be lots of cheap immitators using things that may look like heatpipes, but are non-functioning.

It appears to have begun. Zalman (big shock there) has released a so-called <A HREF="http://www.zalman.co.kr/english/product/zm2hc1.htm" target="_new">heatpipe HDD cooler</A>, only the heatpipes don't go anywhere, just from one side of the HDD to the other! Heatpipes do not work if both ends are attached to a heat source. In this configuration, they are likely to be about as effective as empty copper tubes would be (read: not at all). This product is nothing more than a gimmick, much like their "flower" shaped CPU coolers.

_________________________________________________
Armadillo[TcC] at Lanwar and MML

intel's boxed CPU coolers...

Distinguished

Distinguished

Champion

Distinguished

Splendid

Distinguished

Splendid

Splendid

Distinguished

Splendid

Distinguished

Distinguished

Splendid

Splendid

Distinguished

Splendid

Splendid

Distinguished

Splendid

Distinguished

Distinguished

Distinguished

Distinguished

Distinguished

Distinguished

Share this page