Video Guide: Changing thermal paste(AS5) on Studio XPS 1645

GROOVYFA,

Here ya go, at idle.

GC-Extreme.

 

Thanks for the replies guys!

So it appears that the individual cores run at different temperatures for everyone?! I find this odd, although I haven't messed with thermal pastes and cooling and temperature monitoring since my desktop gaming days so my memory isn't helping me with this recent event.

 

Different loads on different cores will vary the temps.

 

Will not void the warranty. Only time the warranty is voided when opening the laptop is if we can prove "customer-caused damage."

 

HALLELUJAH ITS OFFICIAL! screw drivers raised you can hear the rush to the nearest electronic store to pick up as5

 

I just ordered some Tuniq TX-3 for my new 1647 that is coming. Anyone familiar with it ?

 

I have read a lot about the TX-3 and heard it is really difficult to work with, but currently offers some of the best results. Its has a much thicker consistency then the others, that's why its more difficult to work with. Very good choice though, you cant go wrong at all.

 

Ya I read how hard it was to apply as well but from this video posted by tuniq it seems really simple so i took the plunge.

http://www.youtube.com/watch?v=xTrSOywFn_4

 

AC5 used to be the best you could get, but there are way better solutions now. Even the stock stuff i used for my desktop works quite well (came with my Hyper 212 Plus) im guessing it is Coolermasters Thermal Fusion 400 Compound which btw is also very very good.

 

I use Noctua HT-N1 on my GPU and CPU, it keeps it so cool that the fan eventually just completely turns off and the temps ramp up and then the cycle repeats itself. Idle temps linked below, I've yet to truly test out the load temps yet but don't know how I should go about testing that, maybe running a couple of gaming benchmarks like Crysis might stress it enough.

https://dl.dropbox.com/u/3226360/TEMPS.jpg

 

Here you go everyone: http://benchmarkreviews.com/index.php?option=com_content&task=view&id=150&Itemid=62

This is a comparison review of 80 different types of Thermal Paste/Grease.

AS5 was tied for first place (once the 200 hours of cure time was achieved), and as the most readily available aftermarket TIM, I do believe it's a waste of time and money to go another route.

 

I am going to apply AS5 to my 1645 next week, too.

But I think I will spread it over the core. Why? Because Arctic Silver tells to do so: http://www.arcticsilver.com/intel_application_method.html

 

I am new to the forums here but i have been reading them ever since i ordered my studios xps. Mine currently idles at about 50c. So i decided im going to reapply the thermal paste. Are people having better results spreading it?

 

Thats interesting I have seen other benchmarks where AS5 while still is good is not as good as some others so IDK illl post links , Also alot of ppl like to spread but it depends on how you should do it like for me with Tuniq TX-3 i have to spread it because its so darn thick
http://www.dansdata.com/goop.htm
http://www.overclock.net/air-cooling/676467-thermal-paste-roundup-8-way.html

http://www.overclock3d.net/reviews/cases_cooling/battle_of_the_goop_thermal_paste_comparison/1
http://www.rbmods.com/content/16600/mx-3thermalfusion400chillaramic.aspx

http://benchmarkreviews.com/index.php?option=com_content&task=view&id=138&Itemid=62&limit=1&limitstart=6

http://www.youtube.com/watch?v=12ce9FpKO3I

 

There isn't one fixed perfect method. Do whatever that you think is best at your own discretion. The information on the net is so broad and its for you to choose what suits you best. Cheers! and I hope you get good results.

 

well i jus finished reapplying the thermal paste about 20 minutes ago. I have seen no difference in temps yet. Mine still idles at around 50c or so. Could i have done it incorrectly? Or maybe 20 minutes is not enough time? Any feedback would be greatly appreciated. I did spread the paste over the cpu. If you think that i should be seeing results then mayb i should just try the line method.

Edit: Ok since i am possibly the most impatient person out there i opened her back up and redid it this time with the line method(horizontal). It is still about the same lol, mayb a degree cooler at times. Guess i just need to let it run for a while and break it in

 

Don't worry it will get lower over time, also the magic starts on load temperatures(thats where it shines).

 

yea i figured it would get better overtime. I will keep you guys posted once it starts to really make a difference

 

I replaced my stock paste with AS5 about a month or so ago, and really haven't noticed much difference. (i820) I think I'm going to re-do it, only this time spreading it instead of letting the heatsinks spread the paste.

 

Okay! So I re-did the application of AS5, this time using the "Spread" method, and it's definitely made a difference. Temps are much better. Only bad thing is, it's a lot more difficult to apply.

 

... And you can leave a lot of air bubbles, which are not good..

 

Just wanna thank the OP, I followed your guide using my "old" 1640 laptop, while applying AS5 on the new 1647..Worked out great..Just also wanted to update the thread to show those what the new arrandale chips look like so they aren't surprised :

prior to thermal paste, showing the on chip (NOT on die) gpu adjacent to little ole arrandale...she's cute




for fun hehe:

 

Honestly, I'm up in the air about this one on which side to believe. Some people swear by letting the heatsink handle the spreading, some people swear by using a credit card. If you look at Arctic Silver's instructions, they use the credit card method. I'm at the point where It really doesn't seem to matter. I've done both (properly), and I've had better results with the manual spread method. Say what you will, whatever floats your boat applies here. I'm not that picky I'm afraid.

 

Featherbeard said he achieved better temperatures with the spread method, so your statement is really pointless to him (no offense intended and if the statement was intended for him). Whether the air bubbles are there or not, they seem to not be influencing his temps that much and still give him temps lower than the line/dot method he used before. It makes you wonder whether the whole "air bubble" thing has as much an influence on temperatures as is claimed. I used the dot and the line method as shown in Xmd's video, and my temps are good enough (CPU idle 48-50/load ~65, GPU idle 47-54/heavy load 80), but I wonder if the spread would yield better results? I don't want to open up my system again, as reconnecting the wireless card (the 3 wires) and remounting the exhaust vent are massive pains . I guess it's whatever works for you...maybe in a couple weeks I'll re-do it using the spread method...

 

i'm also tempted to redo my AS5 appplication with the spread method... It makes more sense to apply it to 100% of the surface as opposed to crossing your fingers with the line method that you put enough on and it covers all of the chip.

I've heard the argument that the majority of heat is transferred through the center etc... BUT the entire thing heats up to one extent or another so it simply makes sense to have a well applied spread.

I wish Featherbeard would dish the temps for his line and spread method temps so we can see what kind of differences we're talking here. TBH, I don't want to do this all over again for 1-2 degrees drop, anything more and I'd gladly do it.

 

The thing is that some people don't put enough of the dab to get proper temps. If you apply the same amount of AS5 as you do in the spread method in a dab you should have the same kind of result... but whatever works for you guys. I got good results from it and I'm happy. Do what works best for you.

 

Obviously.
You dont really have to translate what he posted.
Im just saying that with the manual spread technique it is harder to achieve better results as it may leave some air bubbles, my statement was not intended to him directly as he achieved better results with that method. My statement was intended to everyone else who is looking for a method to apply.
Generally the best method is a little dot on the middle of the chip, this method will allow the paste to spread when pressured and that way you are assuring no air bubbles. Which I repeat and maintain, they are not good, just think about it.

 

I put as close to exactly the same amount during the spread method that I used during the 'smoosh' method. (Using multiple photos to get as close as I could.) The one thing I noticed is that the heat sink didn't do a very good job of spreading out the paste very evenly, and I think it's because it can be difficult to push the heat sink down in a 'level' method. It's much easier to use the fan as a pivot point as shown in the video, which didn't spread the AS5 very well across the i820's core. It's a little clumsy to put the heat sink back on these puppies. :\ The downside to the spread method is it's a lot more tricky obviously, to spread the compound, so beware if you try it!

 

Hi Guys,

I just got a replacement for the overheated and throttle infested 1640 (see signature) and now I have a horribly overheating, less throttling CPU (1645 with i7-720) with the same overall configuration.

I know a bit about opening the laptop and getting my hands dirty but don't have ANY background XP in doing this Thermal Paste operation.

Anyways, my questions are -

1. Which paste to use? AS5 or Sunbeamtech Tuniq TX-3?
Both are available readily here... But please advise me by your actual experience and numbers (wrt 1645 if possible).

2. Can this same paste be applied to both the processor and the VGA chip?

3. Applying this paste would be different for a quad core? If yes, how?

4. And how much amount is OK (neither less nor more / A near about 'exact ml' would be nice)

And if any one of you people have actually tried this on a i7-720 based 1645, it would really help if you share your experience!

Thanks,

CK

 

Hey Ckhurana. Ok, so I've actually performed this procedure 3 times on my 1645 with an i7-720. I'll provide answers to your questions first before I detail my experiences.

1. I suggest AS5. This has been the most "trustworthy" thermal paste (for lack of a better word) for a while now. I've heard of some other good ones but I've only used AS5 and it's worked well for me. I think Tuniq is alright too, but I don't have any XP with that.

2. Yes the same paste can be used for both the processor and the video card (4670).

3. Sort of different. The quad core chip is a little bigger and thus has a more rectangular shape. If you use the line method (as XMD showed in his video) you'd have to place the line parallel with the longer sides of the rectangle-shape of the processor (the dot method is best for the square-shaped dual core processors, not these quads). You can use the dot method for the VGA chip though, as it's square shaped.

4. They usually say about the size of one piece of rice. So if you do the line method, try and keep to the same rice volume but extend it out into a line (the line shouldn't extend across the whole chip though, stop short of the edges)

Ok now for my experiences. I applied my thermal paste for the first time in about mid-March. Before that first application the Dell-applied paste gave me average GPU idle temps that were about 52 C (I never checked CPU temps as the GPU was experiencing the overheating). So with the first application I used the dot method for the VGA chip and the line method for the processor (letting the pressure of the heatsink spread the paste for me). My average idle temps after the first application were 50C. My GPU didn't overheat-throttle under load anymore with games that it was overheating with before the application (COD:WAW in particular was now abt 80C from 84C). So then I was stoked and decided to get myself Battlefield Bad Company 2 early mid-April. Well the overheat-throttling came back again at 84C, most likely because BC2 is a more demanding game than COD:WAW. So I decided to try and reapply again.

The 2nd application was a total bust because I wasn't careful in making sure that the black wireless card wire wasn't under the heatsink, thus preventing contact between the heatsink and the chips . I achieved temps of 90C from just watching a video and I was like "What!!" . So I opened it up again, noticed the problem and had to reapply again. So this time I decided to use the spread method (with a small dot placed in the middle of the spread at the end). I made sure the spread wasn't too thick and the dot would at least minimize any bubbles in that crucial core center of the chips, which is where most heat is generated anyway. I noticed that my idle temp average was still at 50C but the minimum temp. actually touched 46C, which it had never done with previous applications. Now you'd think this would help keep BC2 load temps below 84C....but sadly it didn't help. As angry as I was, I tried contacting Bill (dell rep on NBR), though no response there .

So I reluctantly decided to go the notebook cooler route, since I really liked the laptop. Got myself a Zalman nc2000 (since the NZXT CRYO LX is sold out everywhere for some reason). On first playthrough of BC2 my temps never went above 80C. 4-5C drop under load, which is not too bad really. There was no more overheat-throttling, and the game ran smooth as butter, which made me happy. Also this cooler is really quiet and surprisingly light. I still need to test the cooler a little more with BC2 and some other games to fully see all the temperature gains it has allowed me. Though it keeps the laptop really cool to the touch.

So long story short, these thermal paste applications might or might not help your overheating problems. Either way, I'd suggest using the spread method with a little dot in the center of the chips, though, like me, you might just keep overheating. I don't want to discourage you but I'm just telling you that it's a possibility. Good luck though and if you need anymore info, hit me back.

Goodness I write too much...my bad .

 

EDU, nice summary, you put a lot of effort I'll rep you for it. Like he said if the AS5 application doesn't suffice you should buy a notebook cooler, I use the NC2000 and I never get GPU throttling after hours of gaming.

 

is it really important that i get the new paste on? is the stock paste that bad??? would this help reduce throttling issue??

 

Hi EDU,

Thanks for such elaborate steps, details and listing out your personal XP with the Thermal paste... (Yes, you deserve some more rep!)

I actually use the machine at my workplace too, that's the reason I am more inclined towards the paste method rather than getting a cooler (as I physically chug around with this beast on my back and don;t want to carry that extra load.)

Let me follow your advice along with the XMD's excellentvideo and then get back with my experience.

I will try to get hold of the AS5 (but TX3 is more readily available).

Thanks again!

CK

 

AS5 is not as noticeable on idle temps but Load temps is where it truly shines. Yes it does a way better job than the stock, it definitely helps reduce the throttling.

 

Thank you guys for the reps. Glad my long paragraph wasn't completely useless . @ Ckhurana, I totally understand your feelings about not wanting to lug around the extra weight of a cooler (these things are quite big too, not necessarily heavy). Either way, the paste application will definitely (hopefully) help your temperatures by a few degrees. Good luck and keep us posted.

 

Just applied it to my SXPS1640.

In the middle there was what I believe to be the Northbridge, which Dell had put paste on despite it having a thermal pad on the heatsink as well. I put paste on anyway just to be safe, but I'm worrying that this could cause problems and I should have left it with just the pad.

Hoping for the best, will report back with temperatures after further testing.

 

Is it recommended to apply this to a new xps? Cause i was thinking since its new it should get the job done. but if stock paste sucks i should apply my own to it.

 

Reporting back, temps seem a bit lower but it's hard to tell if the fan's kicking in at the same points..

Idling now at about 42 CPU (C2D P8600) and 52 GPU (Radeon 4670).

TF2 gets the CPU into the 60's and the GPU into the 70's. It's also pretty hot out.

Might try reapplying with less of it, might not.

 

I recommend it because not only do Dell apply the paste horribly (honestly mine was like a blob of mashed potatoes thrown over a USB drive or something ), but the paste they use is not of the best quality. I think they reserve the better paste and application for their Alienware line. I would recommend you do it now, as best as you can (use Xmd's video as a guide and take your time), and then forget about it for a while. I can almost assure you that your temps will go down, especially under load. Use Arctic Silver 5 or the newer MX ones (I'm not sure of the name of that one exactly, but they're supposedly good). Good luck.

 

My brother just ordered the Artic Cooling MX-2 and I was gonna take some from him to apply it. I will definitely replace the paste now. They should treat all their customers equally! The XPS isnt a cheap machine.

It was stated that opening it up to put in paste doesnt void my warranty? Is this true before I do it?

 

Reapplied with less of it, definitely a tad better now. I'm wondering about the northbridge though: is there some way to monitor its temperature?

 

I can't get the plastic vent cover back on the right side. The video made it appear he just got lucky. Can someone please tell me how?

 

In the video you can see that he cut out his likely multiple attempts.

I've taken it off twice, and both times it took several tries, probably around 5 minutes. Basically, keep trying, try more angles and distances from the corner of the laptop. Kind of hard to describe- it goes in, you turn it to make it parallel with the laptop, and you slide it another inch or so while it is parallel.

 

I've run into this too, you just have to keep trying different spots -- it should slide right in.

 

I can get it in, but sliding in the final inch never works... I've looked at the wires trying to push some of them and I'm not having any luck.

EDIT: Just got it. I had to move the white wire covered in black more inside the laptop than letting it bulge.

 

Well I'm not noticing a real difference with IC7.... even though it's rated above AS5. I wonder if I should apply AS5 instead

 

Just decided to change the thermal paste on the CPU/GPU to AS5 using the video as a guide (used spread method though). There was basically no Dell paste on the metal CPU/GPU parts, but rather a big goop AROUND them, and basically no dust in in my heatsink fins, so I'm hoping this improves my temps a bit.

Should I wait awhile before running something like Prime or Furmark?

 

Just applied AS5 for the first time ever last night, noticed an IMMEDIATE drop at idle of 7c!! I have my lappy going at least 16 hours a day, if not more, so I can't wait to see how much further they drop at 200hrs!

Props to OP for helping to give me confidence to pop everything open (fyi I used horizontal spread CPU pea GPU)

 

i dont know how much to put. does it have to be perfect? when i think of a pea its pretty big compared to what i put on which is like a round booger.

 

i don't know what kind of pea's you're eating but the ones i'm thinking of are pretty small. take a retractable "click" pen and look at the circle where the pen comes out at the bottom. that is about the right size.