Lapping A64 IHS and MSI K8N Neo2 Platinum Chipset (DONE, upto 16c Lower) *56k beware*

[FearTec]

Sure would, looks like I have lost more than my warranty

I will keep you posted after further testing.

[Dutch Wink]

hmm, this is the reason i've never followed the 'use an almost transparently thin layer of as5 paste for best results' line for applying goop. if the surface it's going on isnt making full contact then surely thats worse than having a decent layer of goop making complete contact? the surplus will get squished out to the side anyway...

/shrug

[phextwin]

Originally Posted by -ShadowRunner-

Is the reason people get better clocks on phase change because it cools the whole die more effectively?

It's because it's colder, much colder. So to answer your question, yes.

Water and air can never get below ambient temperiture - phase can.
Cooler cpu's clock better.

[aresaid]

Originally Posted by phextwin

It's because it's colder, much colder. So to answer your question, yes.

Water and air can never get below ambient temperiture - phase can.
Cooler cpu's clock better.

Discounting water chillers of course
They are very common in the US especially, and I assume will become here, so dont discount them.

[FearTec]

I have assembled the PC again and all is looking well, time to let the AS5 settle in then I will benchmark.

[DeNs]

Very nicely done FearTec If I weren't planning on selling my Neo2 for the DFI NF3 939 when it comes out, then i'd probably do this tomorrow!

However if I change over to the XP120/DFI board i'll be lapping the IHS down for sure!

Thanks for that, and nice guide.

---dens

[phextwin]

Originally Posted by aresaid

Discounting water chillers of course
They are very common in the US especially, and I assume will become here, so dont discount them.

I consider them phase or TEC cooling :P

[FearTec]

Update: The lapping of the A64 ISH went as planned, well not really.

Mounted the Radiical watercooling block.


Also check out my modded 9800 pro.




After I lapped the processor, chipset heatsink, re jigged some cables, water-cooling and waterblock mounting screws I rebuilt the pc, checked and double checked everything and took a big breath before turning the PC on.

I powered up and nothing. I checked everything again and nothing. The MSI 4 LED's indicated "processor damaged or not installed". After swallowing my heart and taking a few deep breaths while the whole procedure flashed before my eyes I tried again but again nothing.

Considering it was 5AM and I needed sleep I decided to go to bed and sleep on it. The following day I scanned the net for info on this error and found many past experiences that lead me to believe that my processor was dead. I disconnected the water-cooling, cleaned the AS5 and checked the processor, no broken or burnt pins or sign of damage. I brushed the processor just in case there were remains of silicone carbide sandpaper or metal from the lapping procedure. I blew on the processor socket to remove any invisible dust and reinserted the processor, connected the water-cooling and checked everything again.

After I turned on the power supply and turned the PC on, to my surprise the monitor light powered up and the system started to boot. I entered the bios, checked the temps and the temps were amazingly lower.

Below are my results after lapping the AMD Athlon 64 3500+ IHS.


Both tests were with the room temp being a constant 18c.

Before:
Idle with Cool and Quiet On: 34c
Full Load (Folding @ 100%, and many background apps): 51c

After:
Idle with Cool and Quiet On: 22c
Full Load (Folding @ 100%, and many background apps): 35c

The results speak for themselves, the lapping of the (visible) concave Athlon 64 3500+ ISH my temps had fallen 16c. Given that this is a toasty Newcastle core and not a cool Winchester or Venice I think this is quite good.

It would have been nice if AMD lapped it a little better before selling it to me though.

Photo of the system as it stands now.


My live temps can be found here: http://simon.fearby.com/system/A64/

[phextwin]

Originally Posted by FearTec

The results speak for themselves, the lapping of the (visible) concave Athlon 64 3500+ ISH my temps had fallen 16c. Given that this is a toasty Newcastle core and not a cool Winchester or Venice I think this is quite good.

Post hoc, ergo procter hoc eh?

I find a 16 degree drop just for lapping the IHS a little unbeleiviable.
Not saying you are lying or insinuationg and wrong doing on your part - but how do you know that you didn't just have a bad mount beforehand?

[javascripterror]

Just a thought phextwin. If the IHS is so badly concave how would you ever get a good mount?
I have just completed a lapping job for a Prescott with exactly the same problems. Idle 51 degrees becomes idle 34 degrees. I have taken pictures of the amount removed by lapping and it is considerable. Do you want me to post it?

[evilasdeath]

I'd have to agree with phextwin thats too much jsut for lapping your IHS, and those temps u were getting BEFORE lapping are way too high for a watercooled system anyway, i get that with stock fan full load in a warm room.

I'd intending to add some better case fans and a new heatsink probably freezer 64 and i'll prolly lap that before i install that

Ether the first mount of the waterblock didn't make good contact or your IHS was horrendiously bad

[petfoodonly]

Hang on guys, this doesn't make any sense.

Why on earth would AMD/Intel make chips with an IHS designed in such a way that the chip can't be cooled properly. Both of them have/had "problems with heat" and know how bad it is for marketing. It doens't make sense that they would make a conscious decision to have hotter CPUs.

[FearTec]

Originally Posted by phextwin

Post hoc, ergo procter hoc eh?

I find a 16 degree drop just for lapping the IHS a little unbeleiviable.
Not saying you are lying or insinuationg and wrong doing on your part - but how do you know that you didn't just have a bad mount beforehand?

I have mounted and re mounted stock/ after stock and water-cooling at least 20 times and kept getting bad temps/mounts.

This is what I have done before lapping the IHS.

1) I upgraded from stock air to an 80-CFM SUNYON fan (Model: PMD1208PMB1-A (DC 12V @ 9.1W), this fan was so noisy it sounded like a jet taking off.

2) I bought a Zamaln fan mate to reduce the fan but it's wattage max was too low to reduce the SUNON 80CFM fan's power. I then bought a SPIRe fan speed controller from D&K that was rated to slow the SUNYO fan, but after external testing the SPIRE fan controller failed in 10 seconds and fried the fan controller's switching chipset.

3) Bought a WebTemp program so I can monitor my system's temps from work and home.

4) Upgraded to a Thermaltake SLK 984U (based on http://www.overclockers.com/articles698/ aggressive cpu cooling guide but this still topped the temps over 70c in December (with the air con on)

5) Tried combinations of different thermal goup (new and old, I have enough syringes of goup that I look like a drug addict).

6) Upgraded to water-cooling (thanks to radiical.com.au) the temps dropped but I was still having higher than normal temps when compared to others water-cooling.

During and after lapping the processor I can clearly see the raised area around the top of the IHS. I lapped the processor until it was truly flat.

Raised Area around the IHS visible


Before and after temps.


If this was a Venice or Winnie core I would not believe this either but it is a Newcastle core.

Before I lapped the IHS, when I went from load to idle the temp took a good 4 mins to drop the temps of the processor with water-cooling.

After lapping jumps from idle to load are instantaneous, this proves that there is better contact than my stock concave Athlon 64 IHS.

[javascripterror]

petfoodonly,
I do not want to revisit the previous comments about Intel and AMD Cpu's apart from showing that Intel actually have a raised portion of the heat sink base supplied with their CPU's stock coolers that makes contact with inside of the IHS at its centre.


Click to view full-sized image!
Hosted by UGBox Image Store

Water block bases do not have this raised portion so poor contact ensues when a badly concave CPU is used.

[phextwin]

Feartec, here's the million dollar question:

Does it overclock any further than before without the lapped IHS?

Temp sensors lie, frequencies do not. I will gladly eat my words if you get a few more 100Mhz.

Originally Posted by javascripterror

Just a thought phextwin. If the IHS is so badly concave how would you ever get a good mount?

Ahah, another excellent question. (of which i have no answer to )

I have just completed a lapping job for a Prescott with exactly the same problems. Idle 51 degrees becomes idle 34 degrees. I have taken pictures of the amount removed by lapping and it is considerable. Do you want me to post it?

I'll beleive you
But what was load? And why were you idling at 51°C ?