Athlon heat death

[rapsure]

AMD says in their warranty that the Athlon Processor will die in about 1 third of a second. Therefore heatsink removal voids the warranty.:o

[BBA]

Hmmm...I ran a P3-600 coppermine at 800 for a while once before it started crashing...then I realized the heatsink was sitting on the desk...no where close to the CPU.

It still runs at 800 today.

So you have to be REAL careful with Athlon's huh?:D

[madfish]

Now you know why I like Intel Iron!:D

[DemonKnight]

Theres a big differeance in an overclocked P3-600 and an Athlon 1Ghz+ CPU I would put money on it, that any proc running 1Ghz and up will die in 3 seconds or less with out a heat sink.

[BURNin'HELL]

They took the heatsink off of the 2 GHz P4 and it didn't die. It didn't even crash!!

[LittleKing]

You might should have read the article first DemonKnight. They compared four different processors:

1. P4 2.0 Ghz - Alive and kick'n
2. PIII 1 Ghz - Also, alive and kick'n
3. Athon 1.4 Ghz - 6 feet under
4. AthonMP 1.2 Ghz. - Buried next its brother

If you read the article it explains why.

LK

[krusty the klown]

Yeah, the P4 can throttle down - you can see this in action when the game slows RIGHT down as the CPU tries to conserve power and not overheatl. The Integrated Heat Spreader on the P4 must help, too - it adds thermal mass and will reduce the rate of heat rise.

[fewtch]

Gotta admit, Intel's got one over on AMD here.

However, I'm willing to take the chance with an AMD CPU. For one thing, a properly installed heatsink simply won't come off. Sure, we hear isolated instances of it happening, but I would consider such installations to be incorrect -- either from weakening the socket by removing the HS/fan too often, or not paying attention to weight factors. Among "regular users," I've never once heard a story of a heatsink/fan coming off the CPU.

For another, how do you know if/when the P4 is gonna throttle down? How much heat is "too much?" I'd rather decide that myself than Intel, and I'm willing to take the small chance of a thermal problem rather than deal with the possibility of my CPU throttling down in speed just when that speed is needed.

"The man who trades freedom for security does not deserve nor will he ever receive either." -- Benjamin Franklin

[mudoggy]

Personally, I am a big AMD fan for their performance and price point.

But...

... I was amazed. That feature by the P4 is an amazing piece of engineering. True, if you are paying attention or careful, that situation should never arise anyway. Then again, not too long ago I was swapping systems around, and forgot to plug in the HS Fan (for the second or third time.. hehehe) on a k6-2. If that would have been my Athlon 1200, I'd be placing another order with newegg.com.

The smoke footage was cool!! Was that the 'varnish' on the chip, transistors, or heatsink compound that was doing the smoking?

Also, I had NO IDEA those chips could get that hot!!! Lead melts in the 600's F, doesn't it? Incredible! If you weren't paying attention, and bumped it with a finger or hand.. YEEEEOOOOWWW! :)

[JayMan]

I wonder if they were to try it again with the tbird, but have vcool running (which supports throttling by software) if it would save the tbird? It uses the mobo's heat sensor under the socket to get the temps so maybe this would react fast enuff to save the tbrid???

JayMan

[jad1097]

Oh, no I hope my Heatsink does not fall off!


Why would anyone even bother doing a test such as that? A waste of time IMO.

[ragtop]

Jayman:
The Tbirds died so fast, I don't think any throttling software would have helped. Remember, the sampling rates on your MB heat sensors are typically in the once every 5 seconds range. You can probably set them faster, but not likely fast enough.

Oh well, my Tbird 1.33 and motherboard combo is still half the price of a P4 with equivalent performance. In the unlikely event that a Gremlin detached my heatsink, I could replace the CPU and still be $$ ahead.

[Ruckus]

Strangely enough, I had run my Athlon 1.2 GHz twice without a HS/F for about 5 seconds each time and it survived...

Well, I think the difference here was that the HS/F was on it at one point, but then it broke off since one of the socket lugs was poorly glued on with Epoxy after I broke it when I changed out my old HS/F.

After that I had the impression the Athlons could go a lot longer than what the article had said...guess I was wrong.

[Bob The Great]

768 degrees F That's like hotter then the SUN!!!!! Well not really. But it's pretty hot! I've always known that Intel's chips were built better. I have an Intel Pentium3, an AMD Duron, and a Pentium 133Mhz. The Intel chips are just more rugged! But I still like my Duron better. It's an amazingly fast (700Mhz @ 950Mhz with normal cooling), cheap... and well Fast!

[Spacetrek5]

Seems I remember somthing about heat being the determining factor in the life of solid state components, but then I'm pretty old. It does seem a little surprising to me that no one has mentioned the life expectancy of the AMD chips even with the heat sink functioning properly, given that they run so hot.

[Crassula]

Working for a training establishment, I get to hear the weirdest of tales like this one (from a chap supposedly able to train people to put PCs together). He said "I bought an Athlon 800 CPU and to check it was ok I powered it up without a fan/heatsink. It died but I complained to the supplier that there was no warning that the CPU required a fan/heatsink and there was no fan/heatsink supplied. The supplier replaced the CPU - no charge!"

Later that day, I spotted the same chap fixing motherboards directly to the mini tower backplates

Seriously though, I have had CPU failure in machines where thick grimy dust or even tar from cigarette smoke has been deposited on heatsinks and fans (the remaining system components received a good dose too)

[bob sand]

There's small likelyhood of a heat sink "falling off". However. those mini-fans do have a finite life., and will simply quit turning (lose the ability to start rotation). The attached CPU heat sink will then slow an otherwise rapid temp rise, and reduce the maximum temp slightly, but not prevent ultimate damage.

Most commercial/ industrial/ consumer semi conductors have a max temp rating of 75 C (167 F). Military & some automotive get built to withstand 125C (257 F).

Water boils at 212F; most solders start melting in the 300F range: paper combusts at around 725F: plastics start smoking around 400-600F.

If you wisely check your engine oil & coolant, also make sure the computer fans are rotating.

[bob sand]

There's small likelyhood of a heat sink "falling off". However. those mini-fans do have a finite life., and will simply quit turning (lose the ability to start rotation). The attached CPU heat sink will then slow an otherwise rapid temp rise, and reduce the maximum temp slightly, but not prevent ultimate damage.

Most commercial/ industrial/ consumer semi conductors have a max temp rating of 75 C (167 F). Military & some automotive get built to withstand 125C (257 F).

Water boils at 212F; most solders start melting in the 300F range: paper combusts at around 725F: plastics start smoking around 400-600F.

If you wisely check your engine oil & coolant, also make sure the computer fans are rotating.

[falcompsx]

its kind of ironic, because the p4 has a MUCH more secure HS/fan mounting...i can personally guarantee that a p4 heatsink will NOT fall off once properly installed., it's held on in 6 places...the new socket 478 is even better, just as secure, and easier to install.

[fewtch]

Originally posted by bob sand
There's small likelyhood of a heat sink "falling off". However. those mini-fans do have a finite life., and will simply quit turning (lose the ability to start rotation). The attached CPU heat sink will then slow an otherwise rapid temp rise, and reduce the maximum temp slightly, but not prevent ultimate damage.

Most commercial/ industrial/ consumer semi conductors have a max temp rating of 75 C (167 F). Military & some automotive get built to withstand 125C (257 F).

My (Athlon) motherboard has an option to power off the system automatically if a certain temp is reached... believe me, I use it :) . It would easily catch a fan failure and shut off the system, although probably not if the heatsink fell off or something (a highly unlikely event).

As for temps, the Athlon 1.2GHz+ is rated to 95C core temp, according to AMD's literature. I would assume that's the point where damage starts to occur, but not sure.

[JackSheet]

I agree with RagTop.

The P4 is better in many ways, but AMD offers way more bang for your buck.

Rookie mistake: I put my HSF on an Athlon 1400 with the little plastic tab still on the thermal pad. I booted and it ran for a couple of seconds before I realized I forgot to peel away the tab. I figured I was SOL, but when I re-installed the HSF properly, there was no problem.

[HOT RODDER]

Well If you like Intell so be it but I have never had a customer satisfied with Intell speed and I have had several Intells die of heat stroke and after selling a box set P4 system it shot craps in 2 days and I still have never been repayed for this system by the maufacturer but I have put a 1.1 GHZ Athlon in its place with no failures.
I've had so many problems with Intell and warranty and service but AMD seems to be forever gratful in my buisness because I've been compensated if I have had problems in over 372 systems(most of which I found out I was actually defrauded by supplier selling me bad stuff other folks returned) but after buyin last Intell system I still have over $1000 in box system and its junk till I get either replacement parts or replacement tower and they say that because the build ticket hasnt got one little check mark on it even tho I have Visa card bill showing I was charged for an extended warranty,which has nothing to do with it because it was still covered under regular warranty.
And if anyone wants to use Intell so be it but I'll stick with AMD and Ive never had a heat related failure of an AMD.
Just my $.02 worth.
Lee

[SEALTEAMTHREE]

I'd have to say that I'll suffer through not being able to take my heatsink fan off my T-bird while it's running. I think its a fair trade off considering how badly AMD CPU's are beating out the high-end P-IV's in throughput. Besides, I thought the whole purpose of putting a HSF on a processor was you put it on there to keep your CPU from overheating in the first place.

And just like bob sand said earlier, the odds of a HSF falling off your CPU are almost nil. Unless you have a shoddy ZIF socket, you should have no problems with the HSF falling off. Heck, I bypassed my ZIF socket and got a heatsink that screws into the four holes around the socket on most socket A mobo's. I seriously doubt my heatsink is going to fall off, at least not without someone commiting an act of sabotage on my system.

As for the fan failure bit, I've never had a HS fan die on me before I either upgrade my CPU or replace the HSF for something better. The only fans I've heard of dying on a HS were the crappy non-replaceable fans in a thermaltake orb. My friend's super orb's lower fan siezed two months after he bought the thing, and the top one died a month later (when he found out about the first one). Regardless, he's stepped up to a much better HSF with a replaceable fan.

If anything, this somewhat humorus article from Tom's Hardware reinforces the reasons that the manufacturers (AMD and Intel) specifically state that a "Heat-sink fan is necessary for safe, stable operation...". It doesn't show that the P-IV is better than the T-bird or vise-versa. It just goes to show why we use heatsinks.

[RFTHOR]

OH PLEASE GET A LIFE...GEEK SEEMS TO COME TO MIND...HMMMM I GUESS THE THOUGHT IS "GEE WHAT WOULD HAPPEN IF THE HEAT SINK CAME OFF"....DUHHHHHHHHHHHH! WHAT DO YOU THINK WOULD HAPPEN AND IF A PIN CAME OUT OF A GERNADE WHAT WOULD HAPPEN...DUHHHHHHHHHHHH! IF YOU HAVE PROBLEMS WITH HEAT SINKS AND SUCH THE ANSWER IS SIMPLE, DONT TOUCH THE COMPUTER .YOU DONT KNOW WHAT YOUR DOING!!!!! OK NOW THATS SIMPLE RIGHT?!! GOOD LUCK

[Leaf]

This dramatic little pyro experiment would have made more sense if the CPU's were run w/heatsink on and the fan unplugged first (a more common error in my experince). my BIOS can detect the lack of fan (high CPU temp) on my AMD1.4 no problem. as for using no HSF, it's a good thing I can buy five AMD1.4's for the price of the latest pentium4.

[RFTHOR]

YES,IT IS AN EXPERIMENT I WOULD NOT BE INTRESTED IN,SINCE THE OUTCOME IS ALMOST ALWAYS THE SAME DEATH TO THE CPU.ON THE OTHERHAND THERE ARE COUNTLESS OTHER TEST THAT PROVE A PROCESSORS WORTHINESS.BUT I HAVE NEVER(KNOCK ON WOOD) LEFT A HEAT-SINK OR FAN OFF OR UNPLUGGED BEFORE BOOTING A SYSTEM, FOR GOOD REASON(SEE LAST THREAD)!!!AMD PROCESSORS BEAT INTEL AT THERE OWN GAME FOR LESS, AND THAT IS WHAT BEING AMERICAN IS ALL ABOUT.:cool:

[bassinvader]

**** - wish AMD had put a note with their cpu's detailing this overheating fact.
Just upgraded my trusty AMDK6-2/500 to DURON 750 (No Overclock......Yet) on a KT7A Raid - Pop went the little weasel!!!
Took around 5 seconds with the heatsink off!!!
Took it back to the shop - no replacement and here is the kicker.......Not because the chip overheated but......and this has also happened with the replacement.....The HSF due to the pressure involved in attachin to the crappy lugs on the zif socket cracked a small edge off the actual CPU Square Die - The new one seems ok even with a small bit of die missing - Anyone know how much or where the actual cpu core starts into the little square??? I'm just waiting for the thing to fail......The four supporting roundels on each corner of the CPU didnt do any good with the superorb (Its a very heavy HSF beast too) - Wonder how long the zif socket will last.......It seems to me the faster cpu's get the less reliable everything in the case is becoming due to increased voltages , psu's , fans , stress etc...

[conjh13]

Look guys, there's a reason AMD chips sell cheaper - it's because they are cheaper. I'm not saying I'll never buy an AMD system, but I'm very well contented with my dual P3 setup. Pentium4 chips are looking even more attractive now.

[ukulele]

When your hot your hot. Lets pump some electrons AMD.

[No Limit]

I thought AMD was going to add some kind of throttling that the P4 has to their up and coming Athlon XP CPUs.

[rolltrain]

what temperature should I start being concerned about overheating? also I have a 1.4 that only clocks at 1050 how do I reach full potential?

[fewtch]

Originally posted by bassinvader
The HSF due to the pressure involved in attachin to the crappy lugs on the zif socket cracked a small edge off the actual CPU Square Die - The new one seems ok even with a small bit of die missing - Anyone know how much or where the actual cpu core starts into the little square??? I'm just waiting for the thing to fail......
Don't worry about it, it won't fail. I also chipped the edge of the die on my Athlon 1.2GHz (twice), and it didn't affect it in the slightest. If you put the HSF back on and it's running fine, it will continue to run fine unless you chip it further by removing the heatsink again.

Hearing from others too, a few chips on the edges won't damage the CPU, unless it breaches the core... but actually cracking the die is a possibility (and that will kill the CPU), so be careful.

[lost1]

OK, lemme figger this out with some VERY simple economics...

Pentium 4 2.0GHz CPU (processor only)- $606

AMD T-Bird 1.4GHz CPU- $125

AMD T-Bird 1.4GHz CPU AND Soyo K7VTA Pro motherboard- $221

(Prices from www.mwave.com)

Now...

I could smoke five T-Birds
or
Two (almost three) T-Bird/motherboard combinations

for the price of ONE P4 processor.

A few relevant figures-
Number of years I've owned computers: 23
Number of computers I've personally owned- 16 (4-Intel, 12-AMD)
Number of computers I've built- 150+
Intel-based- 9
AMD-based- all the rest
Number of CPU failures (either brand) due to heatsink/fan falling off- 0
Number of CPU failures I've HEARD of due to heatsink/fan falling off- 0

My point:
Does Intel build good hardware? Yep.
Does AMD? You betcha.
Do the folks at Tom's Hardware have too much time & money on their hands? Without a doubt.

Call me silly if you will, but I'll stick with AMD till Intel gives me more than the bragging rights of "CPU Most Likely to Survive Catastrophic Heatsink/Fan Falloff Failure" for the 5X increase in price...

Personally, I think Tom's next ground-breaking all-revealing project should be: Which CPU could survive the longest when exposed to 934.2 volts of DC power (at 39 amps) while bathed in a Gamma radiation bath, and center-punched with a .308 165gr. FMJ bullet at a muzzle velocity of 2640 fps.
Hopefully they'll videotape the results for all to see...

My .02
;)

[dosmastr]

Originally posted by JayMan
I wonder if they were to try it again with the tbird, but have vcool running (which supports throttling by software) if it would save the tbird? It uses the mobo's heat sensor under the socket to get the temps so maybe this would react fast enuff to save the tbrid???


JayMan

i dunno man, but if you check the video, the chip is fried even before the heatsink is moved so that we can see.
plus a couple little birds told me that very few mobos can react much quicker then the diode built onto the palomino, i don't know how accurate they are but thats what they told me.

i tried running a 750 duron without thermal paste, right after i hit the power button the pc speaker started screaming, yanked the pwr cord, the proc looks ok but i'll have to check it

70C is the cut off point for a 1ghz tbird, higher then that and amd ain't gauranteeing the thing will work.
i got a proc i chipped the corner off, but i heard it chip more when i reinstalled it, i don't even get a POST.
165 down the drain... anyone want a 1.3ghz AXIA y keychain for 20bucks?

[newz2me]

Hi guys- I have a Athlon 1GHz that fried all by itself, and I wasn't even using the computer. It was on, the heat sink was on and from across the room I heard a "pop" and the screen went black. When I took it apart, the cpu fan still worked, heatsink and paste were all instact- cpu was fried. Brown burn marks shooting outward in the paste! :(

I had gotten the cpu about 4 months earlier at Tiger Direct, and they were great. I didn't have to send it back to AMD and they sent me a new one before they even got the old one back!

I now turn my computer off at night, but I am hearing noises that makes me think the cpu fan is now going bad, so I have purchased a new one and am putting it on tomorrow.

I am a big smoker (I saw someone's post earlier about crud in the computer) but I keep my computer pretty clean. I have a cpu fan, a case fan, and a power source fan, does anyone have any ideas as to the why and wherefore of the "poof"?

I just installed a 5" bay fan (It fits around a hard drive if you put one in behind it) since I don't have a second case fan opening in the back and hope that and the new cpu fan will do the trick :)

Thanks!

[RFTHOR]

YOU MIGHT HAVE INSTALLED INCORRECTLY AND IT WORKED FOR FOUR MONTHS , THEN ONE DAY POOF....?!!:eek:

[adelatorre]

" paper combusts at around 725F: plastics start smoking around 400-600F."

Bob,

I guess you never read Farenheit 451 by Ray Bradbury

[outlaw_revolver]

It depends on your investment. Whats the chances of your fan falling off? Has anyone seen intel's prices these days woo? There on up there. Lets say you buy a 1ghz athlon for as little under a 100$. The heatsink falls off. You buy a new athlon and still get off cheaper than the p4.

[-Aar0n-]

There are reasons why I think this is.

1) AMD does not have an entire division dedicated to chip packaging. That would be expensive and raise the cost of their CPUs. They have yet to reel in the Tier 1 OEM despite their volume and performance because of Intel's business practices. This leaves them still in the underdog position in business terms as they do not have a guaranteed consistant profit base. I can guarantee that much like with M$ products, the user pays a a very different price even for OEM Intel stuff than Tier 1/2 OEMs do, and they do so to subsidise the lower prices that the top tier OEMs pay. This leaves AMD going after the lower price market.

2) AMD has a policy against remarking. At the same time they want to let the tweakers overclock. If they wanted a more resilient core, they could bond a copper+nickel or aluminium heat spreader to the top of the chip, even a diode. The problem is that you likely wouldn't have access to the L1 or other bridges (the caps are usualy at least 4x the cpu in surface area to insure good heat dispersion). This is the price that we must pay, is a few CPUs falling by the wayside. Think of it this way: who will pop their cores the most by accident? A: The ones tweaking / modding their cores the most (in theory).

3) AMD is trying to sell volume to gain marketshare. Better packagin is extra $$$... and their stategy is that it is better to let a fraction of shipped product poof!, and keep sales numbers high.

4) Tom's a ******* at times (not always). He is basically an industy insider and should know all this already. He likes to feign ignorance to mislead people into think that AMD is doing a crappy job of CPU packaging. He knows that this is really AMDs way of being good to their customer: not charging us for stuff we don't need. So maybe 1 in 100 CPUs might be saved... is it worth the extra %10 per CPU to protect if from primarily user error.

5) AMD was hoping that more HS makers would start using the mounting holes, it is just a better system (BTW: all major OEMs have always added their own home-brewed additional supports to hold the larger heatsinks in place that ususally involved through MB mounting holes). It think alot of you will agree that that the better portion of cracked / chipped cores come from poorly made / designed clips that did not take multiple removal / application operations into consideration... much less this being done by the user (being really careful with things does not make you as safe as a reckless tech who has done it 100x last week -- no offense).

Don't think I am just a biased jerk.. I have had an Athlon pop on me (well not actually me -- my dad left the HS off too... and the core was pretty torn up anyway).

[ding]

What I heard is that Athlons might burn in flames or even _explode_ in case the cooling fan is halted long enough.. It's more than risky if the athlon computer has to work as a server...
I wouldn't leave the Athlon based computer working overnight without good house insurance...

Aren't there any two-fans CPU cooling units avaliable?

Ding

[frostman]

Ive run my AMD 1.0 Ghz overclocked to 1120Mhz (140FSB x8).

It run at abt 47 - 53 Deg C, but today (ambient temp 30Deg C)
the CPU got to 67 Deg C, so I slowed it back down to 140 x7.5 (1025Mhz.)

Im running UD Cancer research software so the CPU is ALWAYS running 100%, no cooling CPU cycles here.

I have made a dual fan heat sink, you just have to buy
a second fan, unscrew the first one and mount them side by side
the screws into the heatsink are self tapping and will go in anywhere on the heat sink,
Note: My cpu run cooler when one fan is blowing in and the other
is sucking out. (and not both going in the same direction.

There is no solder inside a cpu to melt, the connecting wires are bonded atomically using untrasonics (i think) but no solder.

There used to be an old "rule of thumb" that goes.

"The every 10 Deg C a semiconductor operates, halves the life
of the semiconductor"

As for them smoking if you remover the heatsink, simple DONT DO IT. I know my car engine will blow up if I remove all the oil and rev the hell out of it too, so I dont do that either :)))

[ding]

Yeah - I wouldn't run my car without coolant either!
I was going to buy an Athlon processor and MB, but now I'm hesitating... My computer stays turned on 24hours 7days a week... I can't control it all the time and the fan is a mechanical device - it can fail every time. So imagine what would happen if it fails im my flat during the night with a computer working 100% load all the time... I'm sure it's gonna burn either, perhaps in 10 seconds (I assume that radiator is still in place) not in one...
I'm not that worried about the CPU or MOBO, I'm worried about my apartment !! :(

Sounds like it's just safer to go for Intel P4 / Celeron...

I wonder if the insurance companies are considering raising the the insurance rate for people having Athlon @ home.. :p

All my thoughts for this evening

Best regards

Ding

[J.A.D]

Despite all the talk, I havn't heard of a single case where a failed heatsink on an AMD led to a house fire. There is a lot of chips running 24/7. So the way I see it, the chip melting is a low enough probability that I don't bother worrying about it.

[pyang01]

download VCOOL @ http://www.voodoofiles.com/4711

The superior heat killer for Athlon/Duron unlike other heat cooler (Rain, CPUidle etc)

Proven for overclocking, this small utilities work on VIA KT133A chipset, u also can try on KT266 chipset

i am using Athlon 1.4 ghz overclocked to 1.507 ghz running at 44 - 45 celsius with Asus A7V133 motherboard.

keep try