Delided Phenom II 940BE

[MrTOOSHORT]

So I was bored tonight and was searching the web for delidded phenoms and came across a couple forums that said it was possible and some tips, so I gave it a shot and was successful. No real difference in temperature though, but looks cool.

anyways check this out.

http://s676.photobucket.com/albums/v...t=IMGP7560.jpg

http://s676.photobucket.com/albums/v...t=IMGP7559.jpg

http://s676.photobucket.com/albums/v...t=IMGP7557.jpg

The phenom II's are glued mixed with solder to the IHS.

[deltatux]

I like to keep the IHS on my processors ... but it's interesting to actually see the die without the IHS on top I guess.

deltatux

[Psychlone]

Yeah, you're f'ing *LUCKY* Bro. Did you heat the solder to the melting point to get it off the cores? Cause I have personal experience with razor blades and solder, and it doesn't work so well...

I was testing the theory on one of my 9 Phenom 9850BE's and discovered that the glue wasn't the only thing holding the IHS on like the older chips (most 939's) - the center of the IHS is literally soldered to the underside of the IHS...so when I pulled the IHS off, it took the core with it.

Fortunately for me, it was only 1 of 9, and it was one of the weaker overclocking steppings, so no loss, but to do that with no alternative would be a catastrophic failure!
All the Phenom and PhenomII CPUs are now put together this way.

DON'T TRY IT for your own sake.



Psychlone

[salanos]

I haven't got money to spend on processors to be dissected, so I won't do it, although I'd like to.

Oh, to have money to spare.

[ST19AG_WGreymon]

Certain steppings have their cores fused with the IHS. Be careful.

[MrTOOSHORT]

Well I took off the IHS spreader off my opteron 165's and got good practice doing it so cutting the black glue around the processor was the easy part. I used the thinest razor blades you can buy, not the ones in the pics. I cut all around the IHS until I felt all the black silicon or whatever it is, was all cut through. Then I followed the intructions from another forum to take a cookie pan and place it on top the stove on a burner going full blast. Then I put some AS5 paste on the centre of the cpu and used my finger to spread it around the area that I thought was the die. When the stove was heated all the way, I put the processor die down for about 30 secs and I lifted the edge of the green pcb of the cpu from both sides with a couple forks and the thing just came off np. Most of the glue/solder was left on the IHS. Then to get the rest of the solder off, I just put electrical tape on the capacitors of the cpu then I just lapped the core very lightly on 2000 grade wet/dry sandpaper with water. That was the longest part and in all for my first time to delid a phenom II, it only took an hour.

[Psychlone]

^ LOL!

Wanna see the IHS from a world-famous Opty 165 (3.2GHz on air, .025V above stock voltage) ??


That's my trusty Opty's IHS right there. It's one of two that are known to reach that speed with that voltage (literally, one-in-a-million) That thing is *still* chugging along at 3.2GHz, running 1.375V under a Big Typhoon at 28*C idle, 46*C full load...been doing that since the 165 came out (and I'm wishing the new Opterons were as cheap as that one was when it came out...holy CRAP - you seen the prices on the new ones lately?!?!!)

Anyway, it's a bit different now - the solder that's used on decent overclocking steppings (like Greymon said) is actually not just solder - you can't get a cookie sheet or butter knife hot enough to melt it without doing damage, and you certainly can't use any direct flame. I once tried with a special iron that used to be used to melt certain alloys for special weights to remove an older 754's IHS that was a good overclocking stepping, but it didn't get hot enough to melt the 'solder' under the IHS...but it did get hot enough to burn the black glue/silicone as well as some of the substrate underneath - that CPU never worked again, even though the IHS was never removed.

So, it stands to reason that the new 'special' steppings have the extra 'solder' points to transfer the heat better to the IHS' underside...where the known standard steppings have either standard solder, or none at all.

In any case, anyone that's reading this thread, I highly discourage de-IHS'ing your CPUs...you may gain 5-7*C from removing the IHS, but you risk damaging the CPU from the point you begin trying to remove it, to the time that you retire the CPU (since every time you remove or reinstall the heatsink/fan that's mounted above it, you risk damaging the cores) And, you can easily drop 5-7*C by adjusting/adding case fans for better internal airflow, and/or getting a better cooling system for your CPU.

@ TooShort: Good job Bro. It's a hard job, and definitely takes some balls to even attempt it - I remember my first time removing the IHS off a precious CPU...and the feeling of accomplishment - but snatching defeat from the jaws of victory (like my one Phenom 9850 above) far outweighs the positive side of removing the IHS for me anymore.

Psychlone

[MrTOOSHORT]

I've like to point out that the bare chip of a delidded phenom is concave where any flat heatsink will not fully lie flat with the cpu. Also there is no temp difference maybe because of this fact or maybe the solder that is between IHS and cpu filled all the gaps so perfect that is was just like a delidded phenom II. In the end I've tried to get that comcave hump more level by lightly lapping the cpu die, but ended up lapping too much and wrecked the chip. Once the chip was installed again, the computer wouldn't post and the chip cracked itself. Too bad I've wrecked that chip because I think it was a golden one ( can do 3.6ghz @ 1.3625 volts ). I've since bought a new Phenom 940 and it requires more volts for the same sets. So be warned that a delidded a Phenom II isn't worth it like it was in the old 939 days. Hope this helps put to rest any risk takers out there LoL.