The new 775 HS SUCK!!!!!!!

[Bigeasy2218]

I recently built a home office PC with a asus P5P800 and a P4 640. the HS popped off in the middle of the day while I was gone out of the house and when I got back and found out the problem I took a look at both the HS and the CPU and found the the paste was hard and crusty on the CPU and still kinda plyable on the HS. I have since reapplyed the HS with new paste and I seem to get reboots here and there I was wondering if this could be the problem and if the CPU is damaged I ordered a new one from newegg today just in case. my temlps arent to drastic mostly in the range of 32C - 46C underload but if the CPU is damaged then it wont matter what I do for temps cause the damage is already done. any advice would be helpful.

[batboy]

I suspect you are not getting the heatsink properly locked down.

[3DFlyer]

If it popped off, and was running like that, that CPU is history. I'm surprised it runs at all now. It's a miracle it didn't take the board out with it.

[AdvanS13]

I suspect you are not getting the heatsink properly locked down.

[batboy]

If it popped off, and was running like that, that CPU is history. I'm surprised it runs at all now. It's a miracle it didn't take the board out with it.

Not necessarily. These processors have excellent thermal throttling capabilities. There is a good possibility the CPU is ok.

Use a finger to press down on each of the heatsink mounting "legs" one at a time. If one or two feel "springy" then they might not be locked properly. I have found the stock coolers require a little extra attention to detail to install. The first time I mounted one, I had the same problem.

[3DFlyer]

Not necessarily. These processors have excellent thermal throttling capabilities. There is a good possibility the CPU is ok.

Use a finger to press down on each of the heatsink mounting "legs" one at a time. If one or two feel "springy" then they might not be locked properly. I have found the stock coolers require a little extra attention to detail to install. The first time I mounted one, I had the same problem.

Yeah, I know about the throttling, but that CPU was left for hours throttling up and down with no HSF. That throttling isn't soemthing you want to rely on.

Guys tried to "test" the throttling back when the Prescott's fist came out, and the CPU's failed. They thought that becasue the CPU would throttle they could pull the HSF off and test to see how it woked. All it did was tyorch the core. The CPU doesn't throttle until it overheats, and that heat is what kills them. That thing was overheating and throttling ans cycling like that for hours.

I would never trust that CPU now. He's already lucky it didn't take out his mobo. I wouldn't risk the rest of the system with it. a 200 dollar CPU that's had that happen is not worth the risk. It may still work, but I gurantee it's ruined. It'll never be stable again. I'd be willing to bet that 1/2 the tranbsistorsa have been burned out of that thing from overheating and cycling.

[Bigeasy2218]

I went ahead and ordered a new CPU and a Zalman Heatsink for it I am just going to replace it and be done with it I dont want to take the chance and end up being without a PC for a week while I figure out what else could be causing the problems. thanks for your replys

[Jimbob7]

The CPU is fine, the pc will shut off after a certain temp, i've seen it many times.

Scott.

[wingman99]

Yeah, I know about the throttling, but that CPU was left for hours throttling up and down with no HSF. That throttling isn't soemthing you want to rely on.

Guys tried to "test" the throttling back when the Prescott's fist came out, and the CPU's failed. They thought that becasue the CPU would throttle they could pull the HSF off and test to see how it woked. All it did was tyorch the core. The CPU doesn't throttle until it overheats, and that heat is what kills them. That thing was overheating and throttling ans cycling like that for hours.

I would never trust that CPU now. He's already lucky it didn't take out his mobo. I wouldn't risk the rest of the system with it. a 200 dollar CPU that's had that happen is not worth the risk. It may still work, but I gurantee it's ruined. It'll never be stable again. I'd be willing to bet that 1/2 the tranbsistorsa have been burned out of that thing from overheating and cycling.

I use to use the overheat technique to get a new CPU now it just keeps on shutting down on me at there safe temp cut off. After a 8 hours i gave up, it still prime 95 and ran just fine. Intel did a good job keeping us from doing that anymore to get a new chip. I have a new technique that i use now.

[3DFlyer]

The CPU is fine, the pc will shut off after a certain temp, i've seen it many times.

Scott.

Some motherboarss have a overtemp shutdown, but you have to have it activated and set properly. if it's not, it won't shutdown. It will overheat, throttle back until the temp decreases, then throttle back up, overheat, and will "cycle" like that. The throttleing is not what kills them, it's the overheating cycling that gets them.

[stevenb]

Some motherboarss have a overtemp shutdown, but you have to have it activated and set properly. if it's not, it won't shutdown. It will overheat, throttle back until the temp decreases, then throttle back up, overheat, and will "cycle" like that. The throttleing is not what kills them, it's the overheating cycling that gets them.
No it should shut down, intel has put 2 probes into the presscott, one the motherboard can access and one it cant. The one it cant controls the thermal throttling and shutdown, and it way more accurate than the one that the motherboard reads.

[wingman99]

Some motherboarss have a overtemp shutdown, but you have to have it activated and set properly. if it's not, it won't shutdown. It will overheat, throttle back until the temp decreases, then throttle back up, overheat, and will "cycle" like that. The throttleing is not what kills them, it's the overheating cycling that gets them.

I just checked my bios. Shut down is disabled and i pulled of the fan plug did prime 95 and it shut down fast. but i already new that i have not been able to over heat the CPU since northwood core. You can read all about it on the Intel site it's all internal.

What the mother board reports for Temp's is incorrect my PC shuts off way before the max Tc temp is reached in windows.

[3DFlyer]

Hmmm, mine hit 82C one time when the HSF wasn't seated right, and it did not shut down. I did yank the plug out quickly, and wasn't about to let it do that again. Anyone know how hight his temp is? I know if you keewp letting that thing cycle like that, it will kill it. If the throttling was that good we wouldn't even need heatsinks at all.

[wingman99]

Hmmm, mine hit 82C one time when the HSF wasn't seated right, and it did not shut down. I did yank the plug out quickly, and wasn't about to let it do that again. Anyone know how hight his temp is? I know if you keewp letting that thing cycle like that, it will kill it. If the throttling was that good we wouldn't even need heatsinks at all.

If you don't use a heatsinks you will reduce the 30 year life span

You wont get much work dun with the CPU slowing and shutting down all the time

Throttling is that good
there is two throttles
And 5.2.5 THERMTRIP# Signal
ftp://download.intel.com/design/Pentium4/datashts/30638203.pdf

[Stratcat]

Hmmm, mine hit 82C one time when the HSF wasn't seated right, and it did not shut down. I did yank the plug out quickly, and wasn't about to let it do that again. Anyone know how hight his temp is?
From memory:

1] TM1 Thermal throttling - The one we all see using "Throtttle Watch", where the chip's temp monitoring starts omitting clock cycles internally. This gives a duty cycle ranging from 20% to 50% "off", IIRC. The thresholds temps = ~67 - 72*C for Northies & Pressies, depending on the specific chip. This form of throttling will cycle, as you stated...IOW, the proc keeps functioning.

2] Forgot the name of this 2nd one, but it's a "catastrophic shutdown" - It occurs at either 125 or 135*C die temp (not sure how that's extrapolated from the cpu temp sensing mechanism, and how it relates to the ~72*C TM1 throttling...the 72*C may not indicate the actual die temp, but is a measuring point somewhere on the proc). This one is 100% shutdown. I do not believe it cycles, but have not experimented much with it 'cuz I don't have the guts.

I've seen the 1st type occur many times, since I had a distributed computing farm (seven machines), & was running into all kinds of heat problems when I had to move the farm to the top floor of my house. I can't say it really damaged any of my procs, tho I wouldn't recommend running >60*C on general principal, and personally try to never go over 50*C. I do believe I saw some GNDS on two of my three Northies, due to extended high (1.7V) Vcore.

I've seen the 2nd type occur 3 times:

Once on a PIII w/o a HS, which nearly immediately died (running time = ~30 Seconds).

Once on a Northie w/a mis-seated HS, which went part way thru the boot process, then shutdown. I tried several more "hard" restarts with the same results (running time = ~30 seconds the first time, and maybe 10 - 15 seconds the subsequent times). After noticing the badly seated HS, I cleaned & re-applied AS3, and slapped it back together. The proc seems none-the-worse for the incident, and is now my kid's 2.6 @ 3.2. It's been >2yrs since the incident.

Once on a S478 Prescott, where the fanblades of a Zalman 7000cu got caught on the fan's 3 pin pwr feed/RPM sense lead, before turn on. I powered the machine on, and it actually got thru the POST/Boot sequence, and ran for several minutes, then BSODed, then completely shutdown all video (running time = ~ 3 mins). After I freed the fan, I let it cool for several minutes (the Zalman was sizzlin hot - even out to the very tips of the fins!), & then simply powered it back up, and everything was uneventful. My temps looked fine, so I just immediate deployed it back into my farm running DC. I then ran the 3.2 @ 3.84 for nearly a year, w/o any issues. I didn't removed the HSF immediately after the incident, but later, I did change to a different HSF ass'y, maybe 6 mos afterwards. The AS-5, surprisingly, looked perfect in consistency & distribution, as compared to any other time I removed a P4's HSF ass'y where I used AS-5

All this stuff is off the top of my head, but I'm pretty sure I'm right about the numbers. Anyone really interested, may want to verify the exact catastrophic P4 thermal temp, as to whether it's 125 or 135*C.

FWIW

Strat

[wingman99]

From memory:

1] TM1 Thermal throttling - The one we all see using "Throtttle Watch", where the chip's temp monitoring starts omitting clock cycles internally. This gives a duty cycle ranging from 20% to 50% "off", IIRC. The thresholds temps = ~67 - 72*C for Northies & Pressies, depending on the specific chip. This form of throttling will cycle, as you stated...IOW, the proc keeps functioning.

2] Forgot the name of this 2nd one, but it's a "catastrophic shutdown" - It occurs at either 125 or 135*C die temp (not sure how that's extrapolated from the cpu temp sensing mechanism, and how it relates to the ~72*C TM1 throttling...the 72*C may not indicate the actual die temp, but is a measuring point somewhere on the proc). This one is 100% shutdown. I do not believe it cycles, but have not experimented much with it 'cuz I don't have the guts.

I've seen the 1st type occur many times, since I had a distributed computing farm (seven machines), & was running into all kinds of heat problems when I had to move the farm to the top floor of my house. I can't say it really damaged any of my procs, tho I wouldn't recommend running >60*C on general principal, and personally try to never go over 50*C. I do believe I saw some GNDS on two of my three Northies, due to extended high (1.7V) Vcore.

I've seen the 2nd type occur 3 times:

Once on a PIII w/o a HS, which nearly immediately died (running time = ~30 Seconds).

Once on a Northie w/a mis-seated HS, which went part way thru the boot process, then shutdown. I tried several more "hard" restarts with the same results (running time = ~30 seconds the first time, and maybe 10 - 15 seconds the subsequent times). After noticing the badly seated HS, I cleaned & re-applied AS3, and slapped it back together. The proc seems none-the-worse for the incident, and is now my kid's 2.6 @ 3.2. It's been >2yrs since the incident.

Once on a S478 Prescott, where the fanblades of a Zalman 7000cu got caught on the fan's 3 pin pwr feed/RPM sense lead, before turn on. I powered the machine on, and it actually got thru the POST/Boot sequence, and ran for several minutes, then BSODed, then completely shutdown all video (running time = ~ 3 mins). After I freed the fan, I let it cool for several minutes (the Zalman was sizzlin hot - even out to the very tips of the fins!), & then simply powered it back up, and everything was uneventful. My temps looked fine, so I just immediate deployed it back into my farm running DC. I then ran the 3.2 @ 3.84 for nearly a year, w/o any issues. I didn't removed the HSF immediately after the incident, but later, I did change to a different HSF ass'y, maybe 6 mos afterwards. The AS-5, surprisingly, looked perfect in consistency & distribution, as compared to any other time I removed a P4's HSF ass'y where I used AS-5

All this stuff is off the top of my head, but I'm pretty sure I'm right about the numbers. Anyone really interested, may want to verify the exact catastrophic P4 thermal temp, as to whether it's 125 or 135*C.

FWIW

Strat


That is the old thermal regulation standard. The hot 90nm process is 20c above max Tc of Your CPU
PBR=1 power =115w Max Tc temp=70.8+20=CPU shut down at 90.8

There is thermal monitor and thermal monitor 2

If you relay want the truth spend less time wondering and find out how the CPU is constantly being under clocked and locked at all temperature's if it goes over max Tc according to watts.

Don't miss the link this time we can put this conversation to a end and talk about how many people getting under clocked after overclocking while running prime 95 or super pi or games LOL.

Thermal specifications 5
Intel >ftp://download.intel.com/design/Pentium4/datashts/30638203.pdf

[SolidxSnake]

When I had my 3.0 with the removed IHS, i remember mounting my heatsink incorrectly. I hit the poewr button, not knowing, and then my computer instantly shut off. Maybe a 3 second delay. The computer shuts down automatically, whether the motherboard is set to or not at the CPU's threshold temp.

[Stratcat]

That is the old thermal regulation standard. The hot 90nm process is 20c above max Tc of Your CPU
PBR=1 power =115w Max Tc temp=70.8+20=CPU shut down at 90.8

There is thermal monitor and thermal monitor 2

If you relay want the truth spend less time wondering and find out how the CPU is constantly being under clocked and locked at all temperature's if it goes over max Tc according to watts.

Don't miss the link this time we can put this conversation to a end and talk about how many people getting under clocked after overclocking while running prime 95 or super pi or games LOL.

Thermal specifications 5
Intel >ftp://download.intel.com/design/Pentium4/datashts/30638203.pdfwingman99,

Not sure how, since our earlier posts are time-stamped nearly an hour apart, but somehow we posted nearly concurrently, or at least your post & link wasn't posted yet, as I was composing my post.

So I never got to see your link 'till after I posted, and subsequently didn't read it since my post was already down before I saw it.

Anyway,

Yes, agreed. You are absolutely correct. I was describing the earlier pre-90nM thermal controls, and the numbers I gave from memory reflect that. The TM1 #'s I gave still look good (tho the doc you referred spec'd a 30 - 50% duty cycle vs the 20 - 50 I "recalled"), but I was completely wrong by missing the new TM2/Thermatrip technology.

Thank you for your correction.

And, to add insult to injury, I often admonished peeps to refer to the datasheets, on forums and at work, only to fall prey, myself. :rolleyes:

Thanks again, for pointing the error out.

Strat

[orion456]

The first time I used my Intel 830d a small wire got between the CPU and HS corner. I used the computer for about 5 minutes before I noticed the CPU temp read 90c ... it was still working!! I didn't take time to see how throttled it was, but it must have been close to its shut off temperature. (Of course mobo readings maybe off by 10c so maybe it was only 80c.)

No problems so far with the chip, it still overclocks nicely.