Given my profession I've fixed a lot of computers to date, but this one baffles me.
I have a computer with an AMD Athlon 64 4600+ (2.5 GHz, 65 nm I think). After power on the CPU heats up rapidly coming from room temperature (25 C) to emergency shutdown temp (around 80 C) in under 20 minutes, rising quickly and beginning to level off at the top.
For those of you unfamiliar with this model of CPU, I have another one exactly like it, same age and model in the same model of motherboard, same original AMD heatsink, etc in a home server, which has sat at 2 C above room temperature for it's entire life -- this is not a hot CPU!
The thing is, there is nothing wrong with the heatsink -- as far as I can tell by feeling it is at the same temperature as the display in the BIOS dialog, with maybe about 5 C of lag (50 C feels burning, etc, I do have some experience in feeling the temperature). The fan works full speed and there is no dust in the heatsink. I don't understand.
I can only assume that some sort of malfunction is causing the CPU to generate much more heat than it should. But I don't even know what to try to replace. Some kind of shortcircuit in the CPU? Malfunction in the motherboard causing it to overclock the CPU? PSU? As far as I can tell, the power ratings in the BIOS are spot-on (not perfect but right where they should be).
I have considered maybe it's the outdated BIOS treating the CPU wrong (it has happened before), however this same configuration once worked I am reluctant to think such a problem would spontaneously surface over time.
> Given my profession I've fixed a lot of computers to date, but this one > baffles me.
> I have a computer with an AMD Athlon 64 4600+ (2.5 GHz, 65 nm I think). > After power on the CPU heats up rapidly coming from room temperature > (25 C) to emergency shutdown temp (around 80 C) in under 20 minutes, > rising quickly and beginning to level off at the top.
> For those of you unfamiliar with this model of CPU, I have another one > exactly like it, same age and model in the same model of motherboard, > same original AMD heatsink, etc in a home server, which has sat at 2 C > above room temperature for it's entire life -- this is not a hot CPU!
> The thing is, there is nothing wrong with the heatsink -- as far as I > can tell by feeling it is at the same temperature as the display in the > BIOS dialog, with maybe about 5 C of lag (50 C feels burning, etc, I do > have some experience in feeling the temperature). The fan works full > speed and there is no dust in the heatsink. I don't understand.
> I can only assume that some sort of malfunction is causing the CPU to > generate much more heat than it should. But I don't even know what to > try to replace. Some kind of shortcircuit in the CPU? Malfunction in the > motherboard causing it to overclock the CPU? PSU? As far as I can tell, > the power ratings in the BIOS are spot-on (not perfect but right where > they should be).
> I have considered maybe it's the outdated BIOS treating the CPU wrong > (it has happened before), however this same configuration once worked I > am reluctant to think such a problem would spontaneously surface over time.
> Any ideas?
> LP, > Jure
There are a few possibilities.
1) Thermal diode isn't measuring temperature properly.
2) Silicon die damaged, and Vcore power is shorting to ground.
Stick a clamp-on DC ammeter around the ATX12V yellow wires and measure the current. If you repair a lot of PCs, a clamp-on DC ammeter makes a nice addition, to the regular multimeter for checking voltages. A regular multimeter, is a nuisance for making current readings, while a clamp-on DC ammeter allows the whole PC to be measured in a couple minutes (as long as the main power cable isn't sleeved). If the main cable is sleeved, you can stick an unsleeved ATX extension cable in place, to complete a main cable measurement. A clamp-on DC ammeter works via the magnetic field around the wire, and is also capable of adding the currents of multiple wires. You can take four red wires and stick them all within the jaws of the meter at the same time, to get a total reading.
3) When a processor has a metal lid, it doesn't just lightly touch the top of the silicon die. It can either have some thermal interface material, between the lid and the die. Or in some cases, they use low temperature solder for a mechanical connection.
If the thermal interface solution is not working properly, the silicon die becomes insulated, and the temperature shoots up rapidly until THERMTRIP is triggered.
You should check CPU input frequency, multiplier setting, Vcore setting, as possible reasons for increased power consumption. The 12V current flow measurement, is to confirm the overheat is due to power dissipation. If the 12V measurement is "normal" and under TDP, then it could be a TIM (thermal interface material) issue between the silicon die and the lid.
On Fri, 29 Jul 2011 16:28:23 -0400, Paul <nos...@needed.com> wrote: >Jure Sah wrote: >> Hello,
>> Given my profession I've fixed a lot of computers to date, but this one >> baffles me.
>> I have a computer with an AMD Athlon 64 4600+ (2.5 GHz, 65 nm I think). >> After power on the CPU heats up rapidly coming from room temperature >> (25°C) to emergency shutdown temp (around 80°C) in under 20 minutes, >> rising quickly and beginning to level off at the top.
>> For those of you unfamiliar with this model of CPU, I have another one >> exactly like it, same age and model in the same model of motherboard, >> same original AMD heatsink, etc in a home server, which has sat at 2 °C >> above room temperature for it's entire life -- this is not a hot CPU!
>> The thing is, there is nothing wrong with the heatsink -- as far as I >> can tell by feeling it is at the same temperature as the display in the >> BIOS dialog, with maybe about 5°C of lag (50°C feels burning, etc, I do >> have some experience in feeling the temperature). The fan works full >> speed and there is no dust in the heatsink. I don't understand.
>> I can only assume that some sort of malfunction is causing the CPU to >> generate much more heat than it should. But I don't even know what to >> try to replace. Some kind of shortcircuit in the CPU? Malfunction in the >> motherboard causing it to overclock the CPU? PSU? As far as I can tell, >> the power ratings in the BIOS are spot-on (not perfect but right where >> they should be).
>> I have considered maybe it's the outdated BIOS treating the CPU wrong >> (it has happened before), however this same configuration once worked I >> am reluctant to think such a problem would spontaneously surface over time.
>> Any ideas?
>> LP, >> Jure
>There are a few possibilities.
>1) Thermal diode isn't measuring temperature properly.
>2) Silicon die damaged, and Vcore power is shorting to ground.
> Stick a clamp-on DC ammeter around the ATX12V yellow wires and measure > the current. If you repair a lot of PCs, a clamp-on DC ammeter makes > a nice addition, to the regular multimeter for checking voltages. > A regular multimeter, is a nuisance for making current readings, while > a clamp-on DC ammeter allows the whole PC to be measured in a couple > minutes (as long as the main power cable isn't sleeved). If the main > cable is sleeved, you can stick an unsleeved ATX extension cable in > place, to complete a main cable measurement. A clamp-on DC ammeter works > via the magnetic field around the wire, and is also capable of adding > the currents of multiple wires. You can take four red wires and stick them > all within the jaws of the meter at the same time, to get a total reading.
>3) When a processor has a metal lid, it doesn't just lightly touch > the top of the silicon die. It can either have some thermal interface > material, between the lid and the die. Or in some cases, they use > low temperature solder for a mechanical connection.
> If the thermal interface solution is not working properly, the > silicon die becomes insulated, and the temperature shoots up > rapidly until THERMTRIP is triggered.
>You should check CPU input frequency, multiplier setting, Vcore >setting, as possible reasons for increased power consumption. The 12V >current flow measurement, is to confirm the overheat is due to power >dissipation. If the 12V measurement is "normal" and under TDP, then >it could be a TIM (thermal interface material) issue between the >silicon die and the lid.
>HTH, > Paul
Good job there with your very meticulous explanation and even on what to look for to remedy the issue as well.
However, Jure...
My suggestion, DUMP the CPU and get yourself a newer one as they are cheap enough now a days... better yet, upgrade to a real CPU... Intel. LOL is the way to go.
I stopped supporting AMD machines many moons ago and I am not sorry to turn away clients with their PC issues but once I know they got a AMD? HA! Sorry pal...
Dizz Nutts O'mine wrote: > On Fri, 29 Jul 2011 16:28:23 -0400, Paul <nos...@needed.com> wrote:
>> Jure Sah wrote: >>> Hello,
>>> Given my profession I've fixed a lot of computers to date, but this >>> one baffles me.
>>> I have a computer with an AMD Athlon 64 4600+ (2.5 GHz, 65 nm I >>> think). After power on the CPU heats up rapidly coming from room >>> temperature (25°C) to emergency shutdown temp (around 80°C) in >>> under 20 minutes, rising quickly and beginning to level off at the >>> top.
>>> For those of you unfamiliar with this model of CPU, I have another >>> one exactly like it, same age and model in the same model of >>> motherboard, >>> same original AMD heatsink, etc in a home server, which has sat at >>> 2 °C above room temperature for it's entire life -- this is not a >>> hot CPU!
>>> The thing is, there is nothing wrong with the heatsink -- as far as >>> I >>> can tell by feeling it is at the same temperature as the display in >>> the BIOS dialog, with maybe about 5°C of lag (50°C feels burning, >>> etc, I do have some experience in feeling the temperature). The fan >>> works full >>> speed and there is no dust in the heatsink. I don't understand.
>>> I can only assume that some sort of malfunction is causing the CPU >>> to generate much more heat than it should. But I don't even know >>> what to >>> try to replace. Some kind of shortcircuit in the CPU? Malfunction >>> in the motherboard causing it to overclock the CPU? PSU? As far as >>> I can tell, the power ratings in the BIOS are spot-on (not perfect >>> but right where they should be).
>>> I have considered maybe it's the outdated BIOS treating the CPU >>> wrong (it has happened before), however this same configuration >>> once worked I >>> am reluctant to think such a problem would spontaneously surface >>> over time.
>>> Any ideas?
>>> LP, >>> Jure
>> There are a few possibilities.
>> 1) Thermal diode isn't measuring temperature properly.
>> 2) Silicon die damaged, and Vcore power is shorting to ground.
>> Stick a clamp-on DC ammeter around the ATX12V yellow wires and >> measure the current. If you repair a lot of PCs, a clamp-on DC >> ammeter makes a nice addition, to the regular multimeter for >> checking voltages. A regular multimeter, is a nuisance for making >> current readings, while a clamp-on DC ammeter allows the whole PC >> to be measured in a couple minutes (as long as the main power >> cable isn't sleeved). If the main cable is sleeved, you can stick >> an unsleeved ATX extension cable in place, to complete a main >> cable measurement. A clamp-on DC ammeter works via the magnetic >> field around the wire, and is also capable of adding the currents >> of multiple wires. You can take four red wires and stick them all >> within the jaws of the meter at the same time, to get a total >> reading.
>> 3) When a processor has a metal lid, it doesn't just lightly touch >> the top of the silicon die. It can either have some thermal >> interface material, between the lid and the die. Or in some >> cases, they use low temperature solder for a mechanical >> connection.
>> If the thermal interface solution is not working properly, the >> silicon die becomes insulated, and the temperature shoots up >> rapidly until THERMTRIP is triggered.
>> You should check CPU input frequency, multiplier setting, Vcore >> setting, as possible reasons for increased power consumption. The 12V >> current flow measurement, is to confirm the overheat is due to power >> dissipation. If the 12V measurement is "normal" and under TDP, then >> it could be a TIM (thermal interface material) issue between the >> silicon die and the lid.
>> HTH, >> Paul
> Good job there with your very meticulous explanation and even on what > to look for to remedy the issue as well.
> However, Jure...
> My suggestion, DUMP the CPU and get yourself a newer one as they are > cheap enough now a days... better yet, upgrade to a real CPU... Intel. > LOL is the way to go.
> I stopped supporting AMD machines many moons ago and I am not sorry to > turn away clients with their PC issues but once I know they got a AMD? > HA! Sorry pal...
<steve.houghREM...@THISblueyonder.co.uk> wrote: >Dizz Nutts O'mine wrote: >> On Fri, 29 Jul 2011 16:28:23 -0400, Paul <nos...@needed.com> wrote:
>>> Jure Sah wrote: >>>> Hello,
>>>> Given my profession I've fixed a lot of computers to date, but this >>>> one baffles me.
>>>> I have a computer with an AMD Athlon 64 4600+ (2.5 GHz, 65 nm I >>>> think). After power on the CPU heats up rapidly coming from room >>>> temperature (25°C) to emergency shutdown temp (around 80°C) in >>>> under 20 minutes, rising quickly and beginning to level off at the >>>> top.
>>>> For those of you unfamiliar with this model of CPU, I have another >>>> one exactly like it, same age and model in the same model of >>>> motherboard, >>>> same original AMD heatsink, etc in a home server, which has sat at >>>> 2 °C above room temperature for it's entire life -- this is not a >>>> hot CPU!
>>>> The thing is, there is nothing wrong with the heatsink -- as far as >>>> I >>>> can tell by feeling it is at the same temperature as the display in >>>> the BIOS dialog, with maybe about 5°C of lag (50°C feels burning, >>>> etc, I do have some experience in feeling the temperature). The fan >>>> works full >>>> speed and there is no dust in the heatsink. I don't understand.
>>>> I can only assume that some sort of malfunction is causing the CPU >>>> to generate much more heat than it should. But I don't even know >>>> what to >>>> try to replace. Some kind of shortcircuit in the CPU? Malfunction >>>> in the motherboard causing it to overclock the CPU? PSU? As far as >>>> I can tell, the power ratings in the BIOS are spot-on (not perfect >>>> but right where they should be).
>>>> I have considered maybe it's the outdated BIOS treating the CPU >>>> wrong (it has happened before), however this same configuration >>>> once worked I >>>> am reluctant to think such a problem would spontaneously surface >>>> over time.
>>>> Any ideas?
>>>> LP, >>>> Jure
>>> There are a few possibilities.
>>> 1) Thermal diode isn't measuring temperature properly.
>>> 2) Silicon die damaged, and Vcore power is shorting to ground.
>>> Stick a clamp-on DC ammeter around the ATX12V yellow wires and >>> measure the current. If you repair a lot of PCs, a clamp-on DC >>> ammeter makes a nice addition, to the regular multimeter for >>> checking voltages. A regular multimeter, is a nuisance for making >>> current readings, while a clamp-on DC ammeter allows the whole PC >>> to be measured in a couple minutes (as long as the main power >>> cable isn't sleeved). If the main cable is sleeved, you can stick >>> an unsleeved ATX extension cable in place, to complete a main >>> cable measurement. A clamp-on DC ammeter works via the magnetic >>> field around the wire, and is also capable of adding the currents >>> of multiple wires. You can take four red wires and stick them all >>> within the jaws of the meter at the same time, to get a total >>> reading.
>>> 3) When a processor has a metal lid, it doesn't just lightly touch >>> the top of the silicon die. It can either have some thermal >>> interface material, between the lid and the die. Or in some >>> cases, they use low temperature solder for a mechanical >>> connection.
>>> If the thermal interface solution is not working properly, the >>> silicon die becomes insulated, and the temperature shoots up >>> rapidly until THERMTRIP is triggered.
>>> You should check CPU input frequency, multiplier setting, Vcore >>> setting, as possible reasons for increased power consumption. The 12V >>> current flow measurement, is to confirm the overheat is due to power >>> dissipation. If the 12V measurement is "normal" and under TDP, then >>> it could be a TIM (thermal interface material) issue between the >>> silicon die and the lid.
>>> HTH, >>> Paul
>> Good job there with your very meticulous explanation and even on what >> to look for to remedy the issue as well.
>> However, Jure...
>> My suggestion, DUMP the CPU and get yourself a newer one as they are >> cheap enough now a days... better yet, upgrade to a real CPU... Intel. >> LOL is the way to go.
>> I stopped supporting AMD machines many moons ago and I am not sorry to >> turn away clients with their PC issues but once I know they got a AMD? >> HA! Sorry pal...
On 2011-08-20, Dizz Nutts O'mine <kis...@ss.com> wrote:
> On Fri, 19 Aug 2011 11:31:47 +0100, "SteveH" ><steve.houghREM...@THISblueyonder.co.uk> wrote:
>>Dizz Nutts O'mine wrote:
>>> I stopped supporting AMD machines many moons ago and I am not sorry to >>> turn away clients with their PC issues but once I know they got a AMD? >>> HA! Sorry pal...
>>Idiot
> DUMBASS!
No pro is going to turn away a good proportion of customers and potential income on such flimsy grounds, that essentially amount to advocacy rather than anything of substance. If anyone actually did do that that would be reason enough to give them a wide berth. Support is about getting people's machines to do what they want, not pontificating about what they should have bought.
