VOGONS


First post, by Takino-42

User metadata
Rank Newbie
Rank
Newbie

So I got my Athlon 64 X2 5000+ at 65 watts, when I installed it the idle temps were already at 50C. After doing some benchmarks, approximately after 3 minutes it's already at 70-80 and goes higher (theres also other cpu temperature readings that say it's 100-110C). At that point, should I try to delid it? Heard that brisbanes use thermal paste instead of solder under the IHS so I'd might give a try, tho SMDs near the die aren't exactly pleasing... I had 3800+ (windsor) previously and if I remember correctly it's temps were 60-65C at worst with the stock fan, the same 65 watts, but too bad it's got it's pins bent extremely and broken...

Reply 1 of 16, by mockingbird

User metadata
Rank Oldbie
Rank
Oldbie

You need to delid it and apply good thermal paste between the core and the IHS.

mslrlv.png
(Decommissioned:)
7ivtic.png

Reply 2 of 16, by momaka

User metadata
Rank Oldbie
Rank
Oldbie

Yup, probably time to delid and change the thermal paste.

But before doing that, confirm that the temperatures really are what they are. If you have another AM2 CPU, try it out and see what that shows in terms of temperatures. The chances of the mobo misreporting are small, but still there.
Also, what software did you use for checking the temperatures? If just BIOS, then also try some different softwares to see what they all report. There's Speedfan for the oldschool, MSI Afterburner (doesn't always work for AM2 systems, IME), HWinfo, OCCT, AIDA64, and etc.

FWIW, 50C idle really is already too hot. If the CPU heatsink is the "normal" stock one with full-height and relatively dense fin count, your temps shouldn't be anywhere this high.

If this is of any help, the PC I'm posting from right now has a X2 6000+ Windsor with stock cooler. Under max load, it will reach around 60-65C, IIRC.
I'm not running it in stock mode, though. I have it underclocked to 2.4 GHz and running at 1.175V instead of the default 1.4V, which both greatly reduce the heat output, especially under load. I can do full clock (3 GHz) at 1.275V, which again also shave off about 10-12 Watts from the power consumption, which lowers the temperature by 3-4C under load. But I'm only doing that because my motherboard's VRM is starting to look crispy in a few spots (cheapo ECS OEM mobo for a Gateway PC.)

That said, how is your mobo's CPU fan profile configured? (Or is it even possible to configure it?) On my board, I can't do fan control and the CPU fan speed is a bit on the low side until things start cooking... so I have my CPU fan and rear exhaust fans swapped - that is, my rear exhaust fan is on top of the CPU heatsink (since it turns faster) and my CPU fan is doing rear exhaust duty. As a result, I have a system that can cope with really hot non air-conditioned rooms (configured this back in my college days so that this PC would be able to cope with the lack of A/C in the place I was renting.) In any case, you shouldn't need to turn to tactics like this. If the CPU has thermal compound going bad under the IHS, that should be resolved first before anything else.

FWIW, it's not just Brisbane, but most K8 and K10 CPUs that use thermal compound under the IHS. IIRC, only Phenom (I/II) and Athlon X4 don't, since these have higher TDP and would not have lasted with regular thermal compound under the IHS.

Reply 3 of 16, by Takino-42

User metadata
Rank Newbie
Rank
Newbie
momaka wrote on 2026-03-24, 15:39:
Yup, probably time to delid and change the thermal paste. […]
Show full quote

Yup, probably time to delid and change the thermal paste.

But before doing that, confirm that the temperatures really are what they are. If you have another AM2 CPU, try it out and see what that shows in terms of temperatures. The chances of the mobo misreporting are small, but still there.
Also, what software did you use for checking the temperatures? If just BIOS, then also try some different softwares to see what they all report. There's Speedfan for the oldschool, MSI Afterburner (doesn't always work for AM2 systems, IME), HWinfo, OCCT, AIDA64, and etc.

FWIW, 50C idle really is already too hot. If the CPU heatsink is the "normal" stock one with full-height and relatively dense fin count, your temps shouldn't be anywhere this high.

If this is of any help, the PC I'm posting from right now has a X2 6000+ Windsor with stock cooler. Under max load, it will reach around 60-65C, IIRC.
I'm not running it in stock mode, though. I have it underclocked to 2.4 GHz and running at 1.175V instead of the default 1.4V, which both greatly reduce the heat output, especially under load. I can do full clock (3 GHz) at 1.275V, which again also shave off about 10-12 Watts from the power consumption, which lowers the temperature by 3-4C under load. But I'm only doing that because my motherboard's VRM is starting to look crispy in a few spots (cheapo ECS OEM mobo for a Gateway PC.)

That said, how is your mobo's CPU fan profile configured? (Or is it even possible to configure it?) On my board, I can't do fan control and the CPU fan speed is a bit on the low side until things start cooking... so I have my CPU fan and rear exhaust fans swapped - that is, my rear exhaust fan is on top of the CPU heatsink (since it turns faster) and my CPU fan is doing rear exhaust duty. As a result, I have a system that can cope with really hot non air-conditioned rooms (configured this back in my college days so that this PC would be able to cope with the lack of A/C in the place I was renting.) In any case, you shouldn't need to turn to tactics like this. If the CPU has thermal compound going bad under the IHS, that should be resolved first before anything else.

FWIW, it's not just Brisbane, but most K8 and K10 CPUs that use thermal compound under the IHS. IIRC, only Phenom (I/II) and Athlon X4 don't, since these have higher TDP and would not have lasted with regular thermal compound under the IHS.

Runs from min 2400 to 3200 RPM max default, cannot change it, but again, my 3800 was staying cool under it
Forgot to type that the radiator (both the cooler and IHS) is colder than sensors say, so yup I blame it on compound

Last edited by Takino-42 on 2026-03-24, 16:05. Edited 1 time in total.

Reply 4 of 16, by Takino-42

User metadata
Rank Newbie
Rank
Newbie

Good news: successfully delidded my 3800 for practice, didn't touched any SMDs and the die... but I should try to keep it even so the pcb won't scratch

Reply 5 of 16, by Takino-42

User metadata
Rank Newbie
Rank
Newbie
mockingbird wrote on 2026-03-24, 15:38:

You need to delid it and apply good thermal paste between the core and the IHS.

Would be fine if I don't reapply the IHS? (If the cooler contacts with the die of course)

Reply 6 of 16, by AlexZ

User metadata
Rank Oldbie
Rank
Oldbie

Check if you get good die pressure. 3d party heatsinks that don't touch the frame will be fine.

You may also want to read Any love for AM2? . Brisbane is an inferior product to Windsor.

Pentium III 900E,ECS P6BXT-A+,384MB,GeForce FX 5600, Voodoo 2,Yamaha SM718
Athlon 64 3400+,Gigabyte GA-K8NE,2GB,GeForce GTX 275,Audigy 2 ZS
Phenom II X4 955,Gigabyte GA-MA770-UD3,8GB,GeForce GTX 780
Vishera FX-8370,Asus 990FX,32GB,GeForce GTX 980 Ti

Reply 7 of 16, by Takino-42

User metadata
Rank Newbie
Rank
Newbie
AlexZ wrote on 2026-03-24, 16:08:

Check if you get good die pressure. 3d party heatsinks that don't touch the frame will be fine.

You may also want to read Any love for AM2? . Brisbane is an inferior product to Windsor.

Thanks for sending me an AM2 thread here, pretty lazy to check stuff so appreciated.

Last edited by Takino-42 on 2026-03-25, 03:45. Edited 1 time in total.

Reply 8 of 16, by Takino-42

User metadata
Rank Newbie
Rank
Newbie
Takino-42 wrote on 2026-03-25, 00:23:
AlexZ wrote on 2026-03-24, 16:08:

Check if you get good die pressure. 3d party heatsinks that don't touch the frame will be fine.

You may also want to read Any love for AM2? . Brisbane is an inferior product to Windsor.

Wanted to see if stock cooler will, since it lays pretty low (higher when the CPU is there but naturally since it adds the height with it's IHS especially). Gonna check with my recently delidded windsor if it contacts fine(anddd gonna have to pull the broken pins first).
Also thanks for sending me an AM2 thread here, pretty lazy to check stuff so appreciated.

Yup... the frame doesn't let to contact with it... but I do have another cooler that does contact with the die so I'll try it

Reply 9 of 16, by Takino-42

User metadata
Rank Newbie
Rank
Newbie

I have put a 120mm aluminium fan with 95w TDP rating on it for now, will look for 90mm one so it doesn't gonna be a space eater and doesn't block my RAM slots. Doesn't get higher than 65C under full load.
Also tried "glueing" (just putting thermal paste on die and putting IHS on top of it since the thermal paste sticks to it) IHS to the CPU with thermal paste and putting the stock cooler, same story.

Thanks all of you for the info.

Reply 10 of 16, by Takino-42

User metadata
Rank Newbie
Rank
Newbie
momaka wrote on 2026-03-24, 15:39:
Yup, probably time to delid and change the thermal paste. […]
Show full quote

Yup, probably time to delid and change the thermal paste.

But before doing that, confirm that the temperatures really are what they are. If you have another AM2 CPU, try it out and see what that shows in terms of temperatures. The chances of the mobo misreporting are small, but still there.
Also, what software did you use for checking the temperatures? If just BIOS, then also try some different softwares to see what they all report. There's Speedfan for the oldschool, MSI Afterburner (doesn't always work for AM2 systems, IME), HWinfo, OCCT, AIDA64, and etc.

FWIW, 50C idle really is already too hot. If the CPU heatsink is the "normal" stock one with full-height and relatively dense fin count, your temps shouldn't be anywhere this high.

If this is of any help, the PC I'm posting from right now has a X2 6000+ Windsor with stock cooler. Under max load, it will reach around 60-65C, IIRC.
I'm not running it in stock mode, though. I have it underclocked to 2.4 GHz and running at 1.175V instead of the default 1.4V, which both greatly reduce the heat output, especially under load. I can do full clock (3 GHz) at 1.275V, which again also shave off about 10-12 Watts from the power consumption, which lowers the temperature by 3-4C under load. But I'm only doing that because my motherboard's VRM is starting to look crispy in a few spots (cheapo ECS OEM mobo for a Gateway PC.)

That said, how is your mobo's CPU fan profile configured? (Or is it even possible to configure it?) On my board, I can't do fan control and the CPU fan speed is a bit on the low side until things start cooking... so I have my CPU fan and rear exhaust fans swapped - that is, my rear exhaust fan is on top of the CPU heatsink (since it turns faster) and my CPU fan is doing rear exhaust duty. As a result, I have a system that can cope with really hot non air-conditioned rooms (configured this back in my college days so that this PC would be able to cope with the lack of A/C in the place I was renting.) In any case, you shouldn't need to turn to tactics like this. If the CPU has thermal compound going bad under the IHS, that should be resolved first before anything else.

FWIW, it's not just Brisbane, but most K8 and K10 CPUs that use thermal compound under the IHS. IIRC, only Phenom (I/II) and Athlon X4 don't, since these have higher TDP and would not have lasted with regular thermal compound under the IHS.

Oh and about Phenoms... Do lower TDP Phenoms use solder instead of compound? Since I was considering both 95 and 65 watt Phenoms...

Reply 11 of 16, by Kruton 9000

User metadata
Rank Newbie
Rank
Newbie

I recently encountered a similar issue with a 6000+ on the Brisbane core. The website mentioned that the AM2 socket motherboard supports 5000+ and 5600+ Brisbane and 6000+ Windsor (125W and 89W), but there was no mention of support for the 6000+ Brisbane. Nevertheless, it also booted without any issues. However, temperatures for an 89-watt processor were extremely high – just below 50 degrees Celsius without load and over 70 degrees Celsius under load! The cooler with its copper core was barely keeping up. I cleaned and lubricated it, reinstalled it several times, and experimented with thermal paste, but this only helped slightly. And after a while, I started experiencing reboots due to overheating.
A couple of months later, I installed this processor in a different motherboard, which officially supported even Phenoms and, of course, 6000+ processors with the Brisbane core. With a regular aluminum AMD cooler, temperatures were much lower! Even during stress tests, temperatures didn't exceed 60 degrees.
From this, I concluded that on motherboards without official support for processors, they can work, but the motherboard supplies the wrong voltage to them, causing them to run very hot, even though they generally work. I might be wrong, but I'm sharing my experience. This processor has been working flawlessly on a different motherboard for a month now.

Reply 12 of 16, by Kruton 9000

User metadata
Rank Newbie
Rank
Newbie

P.S. I later reused the first cooler with a copper core with an FX-4300 processor, and it handled it perfectly. So the cooler is definitely in good condition.

Reply 13 of 16, by Takino-42

User metadata
Rank Newbie
Rank
Newbie
Kruton 9000 wrote on 2026-03-25, 16:12:

I recently encountered a similar issue with a 6000+ on the Brisbane core. The website mentioned that the AM2 socket motherboard supports 5000+ and 5600+ Brisbane and 6000+ Windsor (125W and 89W), but there was no mention of support for the 6000+ Brisbane. Nevertheless, it also booted without any issues. However, temperatures for an 89-watt processor were extremely high – just below 50 degrees Celsius without load and over 70 degrees Celsius under load! The cooler with its copper core was barely keeping up. I cleaned and lubricated it, reinstalled it several times, and experimented with thermal paste, but this only helped slightly. And after a while, I started experiencing reboots due to overheating.
A couple of months later, I installed this processor in a different motherboard, which officially supported even Phenoms and, of course, 6000+ processors with the Brisbane core. With a regular aluminum AMD cooler, temperatures were much lower! Even during stress tests, temperatures didn't exceed 60 degrees.
From this, I concluded that on motherboards without official support for processors, they can work, but the motherboard supplies the wrong voltage to them, causing them to run very hot, even though they generally work. I might be wrong, but I'm sharing my experience. This processor has been working flawlessly on a different motherboard for a month now.

Did have a bit higher temperatures before BIOS update on my motherboard, the temperatures got better slightly but issue remained...
So yeah motherboard updates matter too

Reply 14 of 16, by shevalier

User metadata
Rank Oldbie
Rank
Oldbie
Takino-42 wrote on 2026-03-26, 00:47:

Did have a bit higher temperatures before BIOS update on my motherboard, the temperatures got better slightly but issue remained...
So yeah motherboard updates matter too

This dont means that the CPU has become cooler.
For example, for the Venice core, the release notes for the BIOS update on some motherboards explicitly stated: ‘Thermal diode readings have been adjusted’.
When the monitoring chip is initialised, a correction factor is explicitly specified there.

Aopen MX3S, PIII-S Tualatin 1133, Radeon 9800Pro@XT BIOS, Audigy 4 SB0610
JetWay K8T8AS, Athlon DH-E6 3000+, Radeon HD2600Pro AGP, Audigy 2 Value SB0400
Gigabyte Ga-k8n51gmf, Turion64 ML-30@2.2GHz , Radeon X800GTO PL16, Diamond monster sound MX300

Reply 15 of 16, by momaka

User metadata
Rank Oldbie
Rank
Oldbie
Takino-42 wrote on 2026-03-25, 09:52:

Oh and about Phenoms... Do lower TDP Phenoms use solder instead of compound? Since I was considering both 95 and 65 watt Phenoms...

I don't remember anymore, but I think they both came with soldered cores, since the lower TDP Phenoms like X2 and X3 (at least Phenom II, IIRC) were typically the same thing as the X4 Phenoms, but with some cores disabled due to possibly being unstable.

I also have a system with a 1st gen Phenom X4. It's the slowest one - an X4 9150e. While only rated for 65W TDP... if I O.C. it to anywhere near close to what the full spec X4 1st gen Phenoms run, my system power (and CPU temperature) quickly shoot up to the same levels as a higher spec Phenom. So I do think they are all based on the same dies.

Takino-42 wrote on 2026-03-25, 03:42:

I have put a 120mm aluminium fan with 95w TDP rating on it for now, will look for 90mm one so it doesn't gonna be a space eater and doesn't block my RAM slots. Doesn't get higher than 65C under full load.
Also tried "glueing" (just putting thermal paste on die and putting IHS on top of it since the thermal paste sticks to it) IHS to the CPU with thermal paste and putting the stock cooler, same story.

65C is still a tad too hot, especially with that cooler.
I wonder what gives here. Perhaps the mobo sensors are not reading this CPU correctly and a BIOS update could correct this?

In any case, at least as a temporary workaround... or just something to try if you are bored:
Install CrystalCPUID and manually set the core voltage to a lower value. Then see if the temperatures decrease. Like I mentioned, I can run my Windsor X2 6000+ (it's the 95W TDP version) at 1.275V instead of the stock 1.4V. And if I lower the core clock by just 200 MHz (2.8 GHz instead of 3), I can drop right down to 1.2V and significantly reduce the temps.

Note that in order to use CrystalCPUID, the OS CPU frequency control must not interfere here. On Windows 7, one easy way to do this is to go into Power Opions and pick the Performance Plan, then go and customize it, and set both the Idle and Active throttle states under CPU to 100%. As for Win XP, I don't remember if anything like that was needed. IIRC, these CPUs didn't throttle speed in XP, so you can just change the CPU frequency (via multiplier) and voltage without doing anything else. But it's been a while, so I don't remember 100%.

Reply 16 of 16, by Garrett W

User metadata
Rank Oldbie
Rank
Oldbie
Takino-42 wrote on 2026-03-24, 16:01:

Good news: successfully delidded my 3800 for practice, didn't touched any SMDs and the die... but I should try to keep it even so the pcb won't scratch

Did you use any specific guide to go ahead with the delidding?