Exposed e6600 die and Ultra-120 eXtreme!

[Alexontherocks]

Hi! I am starting this thread to seek some help regarding the use of a TRUE on a IHSless chip (e6600). I have started the same discussion in another thread but in the intel suport section but I really don't know how to move it and it seems that nobody is reading it anymore but it may interest those of you who, like me, have succesfully removed a IHS on a dual core (or other cpu) and tried

I removed the IHS with no hassle and have used a TRUE with liquid metal as TIM (coollaboratory metal liwuid pro). Temperatures are now: orthos priority 10, 3.6 ghz (400x9) @ 1.5v (under load/ 1.52 @ idle) 57°C on both cores and orthos 12 hours stable (Tambient = 23°C).

I would just like to ask a couple of questions.

1) I have lapped my heatsink and it is "suction" flat. I have removed all solder from the cpu die using 2000 grit and obtained a mirror finish on the die itself. Contact using liquid metal is very good I believe but is there anything else I can do? I could try increasing pressure evenly but I fear sudden death for my conroe.

2) Would using rubber insulation (a bit like amd does)on the cpu's pcb increase efficiency of the TRUE? Heat being trapped under the copper base and dissipated?

thanks!

[it_burns_when_i_pee]

what were your temps before?

i read that you can drop 5 degrees removing the IHS and have thought of doing it myself

[Demo]

Prime95 v25.6 heats up the cores better than orthos, 57c on load with orthos looks very good at 1.5xx vcore!

[geisher]

i don't think it's that good Demo, let's me explain, I have a 6600 that does 3.6 at 1.55v( 1.5v without droop) with scythe Infinity and I have the same temp like the OP 57°C at load and 55°C idle. And judging by the performance of the TRUE, he can do better plus his cpu is IHSless.
And my cooler is modded too.

[Alexontherocks]

it_burns_when_i_pee: I have recorded 63°C on both cores before removing the IHS so it definetely seems to a drop plus the system is definetely more stable.

geisher: It all depends on many factors which I am willing to take all into consideration to increase performance but to begin with wht is your ambient temperature? In addition it seems that your idle-load is really minimal. That does not seem right. Download Realtemp and check your temperatures because it is currently the best program to monitor DTS signals form cpus.

If you have considerations, suggestions please tell me. I have lapped my TRUE and it looks very flat (razorblade and light method). I have used coollaboratory liquid pro (might have I used too much? Would temperatures significantly vary if too much TIM was used?)

Would more pressure help?

P.s. If somone needs help with removing the IHS I can give you some hints. The procedure is quite simple (the one envolving a pan).

Cheers

[it_burns_when_i_pee]

why dont you just share, hopefully you have pictures?

[Alexontherocks]

I am embarassed to say that I don't have any pictures. I am currently looking for a intel eXXX series or quad to make a youtube video of the process but the process of decapitating my e6600 will remain a matter between the two of us. :D

Regarding temperatures I tried reseating the cpu with foam and that did not have really any effect. I reapplied liquid metal and tried different fan configurations. I am currently waiting for a second scythe ultra kaze 3000 and will try to keep 2 in push pull configuration (downvolted to 9v). It would seem that 2 fans (90cfm +) will keep my cores at 52°C. This is a very good result and I can't think of anything more to do to improve this...

When I offered help with removing the IHS I simply offered an online (msn) or written response to integrate what the web offers already. Between theory and practice there is a lot of difference and even though it might seem simple to "cut through"the silicon witha a blade for example, I trust that nayone attempting this for the first time will have doubts regarding the distance to the core, how much pressure apply and what to do if the razor is stuck.

The same applies to the procedure of manually "cooking" the cpu on a small pan. I for one was not sure as to how to proceed regarding temperatures and thermal excursion. If anyone were into this method (IMHO the only feasible one) I would suggest keeping the pan lifted from the gas and slowly lower it when the TIM starts to emit fumes (yes there should be TIM on the pan....cheap white goop will suffice) The IHS will actually pop if the razor blades are securely inserted in all 4 corners but if it does not a good 4 minutes after the TIm starts to solidify then you must take the cpu out of the pan, let it cool naturally and start cutting again. Too much exposure to heat (beyon the necessary 98°C) will actually melt the solder keeping the die on the PCB. The critical aspect of the procedure is to calibrate the heat in order to cause melting of the first layer of solder and not the second. Many attempts have failed because either people used force (prying on the IHS) or have exagerated with heat both using too much of it or prolonging exposure beyond necessary.

I hope this helps.

[cegras]

What I find most eye opening is that you dared to sand the die.

[Alexontherocks]

yes well I figured that I had to take out all of the solder and that there was no way of doing without an "aggressive approach". Simple polishing paste might have created inequalities on the surface so I figured:

Silicon is rated 7 on the mohs scale. The die is flat and it should be able to observe the cleaning of the solder before damage is done to the die. Steel is 6 on the mohs scale so if 2000 grit is not enough to even remotely scratch (seriously scratch) steel I figured it could be done safely on the die.

slow, light "8" movements should be enough to polish the surface. Damage should be avoided in this way.:up:

[it_burns_when_i_pee]

for anyone interested in this check this out

http://www.legitreviews.com/article/402/2/

only i read on other sites to not leave the blades in during heating because the pressure could pull the die off the pcb. they say to heat up and then insert the blades to pry it open

doesnt sound too difficult. but many report no temperature drops, some worse most likely to mounting pressure. some report upto 4 degrees. but if their chip isnt lapped then that may well be the difference?

[Alexontherocks]

well it would seem the system used in the review you posted a link to actually destroyed the cpu. It could have been a mistake but most likely the uneven heat from the blowtorch damaged or slightly moved the die from its seating on the pcb.

As for using razor blades it all depends on the type of razor blades. I live in Italy and have found very thin double sided "gillette" blades. These were so thin that they could be (and have been) cut along the centre spacing (for the actual razor to be mounted into) to be used as pressure points. The truth is that the pressure generated from the blades comes directly from the thickness (however thin they may seem to the naked eye they do have thickness and being metal are not compressable) and as such it is a pressur which is achieved delicately, constantly (once inserted) and just enough to sever the solder connection between die and IHS.

Furthermore you must take into account the fact that using creditcards is very risky if done to remove the last traces of solder. The Die is hard but fragile (not unlike many materials used in various sectors of industry, just think of ceramic knives). Using the credit card to scrape the solder over the edges may very well CRACK the die. To accomplish this it may not be a type of damage you will eventually see. It simply takes a small lesion on a microscopic level to impede normal functioning. I am no engineer and certainly not a material expert but it makes a whole lot of sense not to overdo it when cleaning the die by using hard yet flexible scrapers. 2000 grit paper or even less will actually work just fine if the lapping is done properly and without unnecessary pressure. If anything lapping applies pressure on ALL the surface EVENLY thus avoiding (partially) the risk of cracking the die.

As for the results I believe lapping the die did it for me because the solder used and left on the die was irregular (uneven and bubbly once cold) as a thermal interface material and the process of removing the IHS may have changed its adhesion to the die. I figured clean heatsink surface and clean die surface with liquid metal would be best for thermal exchange purposes.

[it_burns_when_i_pee]

thought i had pointed out in the post that theirs didnt work, likely because of the "razor blade" method of removal on the heatsink stage

if you google removing ihs you will find that they all go through the same process. whether its heating up with a soldering iron, frying pan (yes thats right) etc

one or two others report failures too where they put too much pressure prying it off. one guy wrote a good one but has no pictures. he reckons that only 25% of IHS removals are successful and says its only for hardcore gamers. dont know how he got that figure but cant help notice that many say that there is no difference between a removed IHS and a really good lap job. for that reason i think i may give it a miss :(

shame as i would love a few degrees off. im only hitting 55 load as it is though

[Alexontherocks]

obviously it is your call to do it or not. Anyway I have just removed a friend's IHS and untill now I have a perfect score of 100% success. :D

Anyway if you hit 55°C (what software are you using to monitor? I used to get 55°C on everest whilst realtemp reported 58°C) you need not worry but then again you might get the last boost out of your e6600 from remoing the IHS.

Out of curiosity how much vcore are you giving your 3.6ghz cpu? I can't go below 1.5 (maybe 1.475 haven't tested yet) and with these voltages I felt safer to shave off even just a couple of degrees. And yes...I had gone through the lapping process but as you may have read on the forum a guy (apparently working in the business) has pointd out that manual lapping gives a convex surface. (which I can confirm hving lapped a mirror for a 25 cm parabolic telescope some time ago) I believe lapping is perfect as long as you are able to control it perfectly, otherwise IMHO it is just an elaborated yet correct way to remove the nickel plating.

I lapped mine and checked with a razor and it still made partial contact with the cpu's IHS (both lapped very well). I could see it using liquid metal and not using conventional pastes.

I am prepapring a youtube video of the process envolving the pan. Hopefully I can help raise that 25% success rate :D

[it_burns_when_i_pee]

im at 1.5v aswell, had to up my ram to 2.25v though for complete stablility. thing is before upping the ram to 2.25 i could play any game & benchmark fine. but orthos would bring a rounding error after anything from 1 minute to a few hours. at 2.25v orthos runs like a dilluded puppy.

the bios reports my temps at 35, LoL, but coretemp, speedfan and TAT all report within a degree or two of one another. each core hits like 53-55. under orthos its only 40 odd because the 8800 isnt heating things up. i only got 1000rpm fans in my machine too, its dead quiet so i am more than happy with the temps.

been lapping for years and found that slow and steady wins the race, letting the wieght of the object do the work. sounds common sense when you think that gravity weighs down on all the object evenly and your fingers applying pressure can be uneven making it "lean" into the sandpaper.

i wont attack the IHS removal after reading the little or no gains vs a well lapped IHS - considering the risk anyway. as already said, im very happy with my temps, they could go down a few more degrees with louder fans but thats not what i worked at.

[Morais]

I have removed all solder from the cpu die using 2000 grit and obtained a mirror finish on the die itself.

:eek:

[Alexontherocks]

What can I say? You must be the lucky owner of a good chip and an even better cooling setup. what is your ambient temperature? I have had the same problems with orthos myself because it appears it is often the ram that works against stable overclocks. The cpu itself within reason is a stable chip even at high clocks (that is with sufficient voltage) but the RAM at least for me has always been an issue. Voltage has not always helped me. I guess it depends on the memory chips and luck.....

Morais: I did not get your post. I hope you agree with the eXtreme measures I used. This is Xtremesystems after all :D

[Boogerlad]

follow this guide: thetechrepository.com/showthread.php?t=23

[it_burns_when_i_pee]

not that lucky, some people on here have their e6600 @ 3.6 with lower voltage - like 1.3v. i still not got round to pencil'in the vdroop mod on my board.

as for memory, i paid a semi premium for it at the time, but its budget price now in my opinion. even today some report the geil overclocks beyond the 800 whilst keeping same timings but i havent pushed my luck here as i wanted to keep an synchronis overclock, also theres the strap issue. so 400x9 is fast enough for me, its 1.2ghz faster than its intended speed or 50% more for free!

[Alexontherocks]

maybe a different motherboard would improve overclock. More often than not I have the feeling it is my motherboard keeping my cpu @ 3-6ghz. No matter what people say about max fsb....

I have S3+ memory and it is just crap. No other way to put it.

P.s I am running orthos right now with 17.5°C ambient and cores are @ 47°C. What are your ambient temperatures?

[it_burns_when_i_pee]

17 ambient? omg i couldnt cope with that, my thermostat is like 20-21. weather here in the uk is coat or jacket unless its may to august.

after last thread i dropped my DIMM voltage back to 2.15v and despite orthos crying like a crack baby, crysis and (dare i say it) bioshock run fine for a good while. have upp'd it back to 2.25 now tho

i dont experience resets or crashing unless i drop the cpu voltage. i've leave it at its current config. its happy. besides 3.6ghz is fine for me, i dont really want to go further. it's a decent overclock and i think if i pushed my luck something may die?

and until it struggles to run or runs things un-satisfactory will i consider an upgrade.

[Alexontherocks]

Well cpu-wise I have never heard of conroes perishing from too much voltage. Possibily degradation is an issue but that phenomenon occurs over time and normally needs sustained high temperatures to effectively hinder cpu functioning. The day you will no longer be able to keep your current clocks then you may want to start looking for a new core. :D

Seriously though the only thing I can think of, capable of being killed by overvoltage, is a cap or a mosfet.

NB/SB/CPU are unlikely to be damaged by high voltage.

CAPS/MOSFETS/RAM are more likely to be killed by voltage.

By the way 17°C ambient was just an esception to the rule. Usually I have about 23°C ambient. The windows were open and I checked by monitoring temps with low temperatures.

P.s. I would close this thread. If I find anything useful regarding using my naked e6600 with a heatsink I will post once again but for know I am out of ideas and the thread is going off topic :up:

Thanks for your help!