Eye opener! Retail Heatsink Thermal Pad.....

[Mucker]

Yesterday I purchased an Antec 1080AMG case and decided to transfer my entire system (core=Epox 8kha+, 1800XP Palimino) over to the new case from an Antec sx840 case (both are excellent cases by the way! ). Once everything was moved over I fired it up and to my suprise I received a BIOS NO POST, black screen, death! Everything was hooked up identically. I played with everything I could think of....card pulling, CMOS clearing, power switch pins, etc.......NO GO. I was pretty sure I had a processor problem, so I decided to pull it off the board. When I removed the heatsink, I instantly saw the problem.....the thermal pad that is standard on retail AMD heatsinks had melted entirely away leaving a very unacceptable contact layer between the processor and sink. I've been running it for about a year, and looking back at it my temperatures had risen over that period to a max of about 52 deg C. It really opened my eyes to the importance of periodically checking the thermal layer to ensure optimum heat transfer. Needles to say, I cleaned up the processor and heat sink surface, applied thermal grease, and voilla!, the machine booted normally. My CPU temp dropped to a cool 38 deg C. I stepped up the FSB to 142 and raised the Vcore and Vdimm voltages slightly to achieve XP 2000 status. Running stable and cool at 41 deg C. Just thought I'd pass my experience on, I know I will be checking the thermal layer more frequently now that I know...........

GPJ

 

[Actaeon]

Thanks for the story, very interesting indeed.

Kinda strange, knowing the fact AMD supplies the thermal pad with there retail heatsinks, that somehow the contact layer would over time, manage to not perform according to specifications. I would think Thermal Grease would do this sorta thing first before a thermal pad.

 

[Mucker]

When you really think about it, it's a wonder any of those compounds manage to stay set in between surfaces without melting away or at least causing an unacceptable contact layer. It was a shocker to me, I was not running O'clocked that high, 140 fsb, Vcore 1.80, during that time. Yes it would seem that the more vicous pad would be around longer than grease, I was suprised to see what I saw. I'll be checking it periodically now....

GPJ

 

[zephyrprime]

I don't think the thermal pad what what caused your problem. After all, you still had a heatsink and even with a very poor thermal interface, the processor should start up since the processor is only room temperature when you first start it up. It would take a little while to heat up. Also, there are plenty of hardware sites on the web that have tested using not thermal interface at all and they could start their computers.

 

[Mucker]

The only thing different that I know I did was re-seat the processor during the cleaning process (but that was when I ran out of options). I was really careful during the moving process with the board, that is why it was so strange to get a no post because of the processor. It could have been only 2 things that I see, either the re-seating or re-establishing the thermal layer. I am glad I saw what I saw...learn something new everyday

GPJ

 

[Wind]

Just to remind everyone. The thermal pad can only be used ONCE (1 time). If u reinstalled the HSF on the subsequent time, u'll hv to scrap off the melted pad & replace it w/ a: -
1. New thermal pad
2. Other thermal compound...ASII, AS3 & etc.

 

[mechBgon]

Ahem... the stock phase-change thermal pad on an AMD retail heatsink is ONE USE ONLY. It is designed to "melt to fit" the gap between the CPU core and the heatsink, one time. If you remove the heatsink after firing up the system, you must remove all of the stock pad and replace it with a new one or else thermal grease.

Just FYI.

 

[ProviaFan]

Originally posted by: mechBgon
Ahem... the stock phase-change thermal pad on an AMD retail heatsink is ONE USE ONLY. It is designed to "melt to fit" the gap between the CPU core and the heatsink, one time. If you remove the heatsink after firing up the system, you must remove all of the stock pad and replace it with a new one or else thermal grease.

Just FYI.

Some anecdotal evidence: I've been trying to get an AXP 1700+ Tbred-B to work in an old KT133A board which I recently learned to be incompatible with said processor. Anyway, I put the 800MHz Duron back in, after cleaning the thermal pad from the heatsink with Goo Gone, and then cleaning both the processor core and the heatsink with alcohol. In times past, the processor would run at around 46°C; now, after putting some Arctic Silver on the HS and processor core, I get temps from 39 to 41°C at full load. Keep in mind that AS3 requires a few days to "settle" before it starts to work at maximum efficiency, so the temperature should drop even a few °C more in the next week.

 

[Grimace]

you know whats funny (or not so funny) when the mb went bad on our compaq, they sent out a tech to change it out. it only took him about an hour to swap everything over, but he just stuck the hsf back on the cpu, didnt clean the cpu/hs and put a new pad on or anything, just stuck it back on...

 

[TerryMathews]

Originally posted by: Grimace
you know whats funny (or not so funny) when the mb went bad on our compaq, they sent out a tech to change it out. it only took him about an hour to swap everything over, but he just stuck the hsf back on the cpu, didnt clean the cpu/hs and put a new pad on or anything, just stuck it back on...

That should be ok, as the thermal contact material would still be on both the processor and HS.

This stuff isn't some magical elixir. It's just a conductive material that is designed to fill in holes/imperfections in the CPU and HSF. What happens is, the pressure of contact forces the material down into the pores, which increases the surface area of contact of both the HSF and CPU, which causes more heat to be transfered over a given period of time leading to lower temperatures IF you aren't saturating your HSF.