LIVE!!! Calling all XM chip, 3930K/3960X and ultrabook owners! Heat-Spring TIMs for testing!

do you got pictures of your cooling system?

 

So, I have a AMD A10-4600M and Radeon 7970M, what are my options

I OC my 7970M to the point where it is tortured and holds at 100C with professional ICDiamond done by GentechPC, and I'd love to keep my processor cooler regardless.

Any way I can order some of this foil? I have an MSI GX60 if it matters. Thanks!

 

can you post me a picture of your cooling system. like find a tear down of it online for me. heat spring will not work unless the heatsink is a solid design so if i can at least rule in or out from looking at a picture i'll help reduce wasted attemps. Thanks

 

This is where the rub comes in. For the manufacturer to get accurate data, or to project accurate data to consumers, the TIM needs to work on ALL HS's. if you hand pick the ones to gather data from you then are skewing the results. That is unless the manufacturer of the TIM is trying to target only systems with over built cooling systems and will list those specific systems accordingly.

This is one area I have to give the people doing ICD tests a nod too. They have recognized that ICD can and does increase cooling transfer to the point of over saturation of the HS capabilities at removing heat. Once you have gotten to this point it does not matter the efficiency of the TIM as the system will overheat, there is no stopping this. Most mobile systems fall into the need for a better HS catagory.................

 

Here is the thing. I got limited samples and it costs 5-20 bucks to mail each one and why would I use my samples on systems we have already ruled out or are highly likely not to work? It would be wasteful testing if we tested systems we already know it doesn't work for. I ruled out a lot of systems from my testing of the netbook and the testing (working in progress) of my G51. Also reborn has shown it works very well. The problem is that uneven or low quality heatsinks will never work well so there is no need to test them. All my other BS is do Friday so i have a free weekend to work on mine. I might re do all of my testing since i have learned a few things about the pressure paper and how H-S works. I have tried two pieces and my pressure on my heat sink is nothing so i am weary if it'll work on my G51 since it is a fairly poor heatsink. Anyways there is no reason to test laptops that we have already ruled out. The point is to find what laptops and types of heatsinks it will work on and it has been shown it is fairly finicky(sp?) if the heatsink is not well made. James D should be getting his samples soon so i hope he will go the extra mile and try what i am doing and get a good reading on the pressures. I am interested if he can verify my results so far on how the pressure lays out on the die and TIM. That will give us a good basis on seeing how the TIM works and how good of a heatsink you actually need. Also i am curious if the die is actually even...do you know Tanware?


Number too on your capacity statements. I have done lots of thinking trying to understand how a heat pipe cna have a capacity and this is the only thing i can come up with.

You have made statements on that a Heat Pipe can have a limited capacity but i honestly don't see how that works with what i know about thermal dynamics. If you are referring to what I think you mean than I agree but if you are not than I have to disagree unless there is some rule or law or something that i don't know.

What I know from thermal dynamics is that:
1. heat transfers until it equals equilibrium
2. heat transfers faster with a greater difference between two temperatures
3. heat transfers faster in a better thermally conductive material

So if you mean this than i agree:

The "capacity" of a heat sink/heat pipe is reached when the temp of the heat sink/heat pipe is the same as the CPU or close enough that no more heat can transfer. With that being said I can see in certain heat sinks/heat pipes that would occur when the fan is not fast enough. As soon as either of these happen ambient temps lower/better fan is used the capacity of the heatsink changes, which is why I believe we do not see a difference from ICD/MX-4 in low power CPUs and low quality heat sinks/heat pipes. Assume we don't have a pressure/evenness issue. Which is also why i believe H-S is showing better results in high out put CPUs with solid heatsinks. It doesn't matter if your using a cheap paste or H-S if your CPU is a low watt and your heat sinks/heat pipes are of low quality. Also another piece of evidence is H-S works better in higher temp results, which would further support these conclusions. H-S has a hard time have low idle temps, which is why i believe it is due to these two reason:

1. heat sinks/heat pipes are already saturated at idle speeds so H-S does nothing
2. H-S actually does worse for this reason: When i tested the pressure of H-S in the heatsink I got a pressure pattern of the H-S design. H-S especially in low pressure does not make 100% contact. It only has ~50-70% contact (my estimate..take with a grain of salt) because the pressure is not great enough to shush it flat so it actually has a negative effect on low temps especially when it has lower pressure. (Can also be contributed to uneven heatsink)

So anyways this is my opinion on heat-sinks and how H-S works. I'll elaborate better after i finish my stuff. Learning a lot of new things from this and don't want to devote a day to trying to explain what i have been trying to figure out since i'll have to rethink it again when i write my opinions on it.

This is why i want to finish my stuff before James gets his and maybe he can replicate my results. TANware you seem to be very interested in this. You want to try to replicate what i am saying and see if it works?

 

What happens is there is liquid in the pipes. This under heat from the CPU vaporizes in a vacuum. It then re liquifies at the heat transfer fin area. The liquid wicks back to the cpu pad to revaporize from heat and so on. A thermal system is only able to do this so fast so oversaturating the heat collection pad will not alow liqud to get back to the pad fast enough. In the end this turns all out to be the rated TDP of the cooling system.

As an example my P79xx. The quad core was equiped with just one heat pipe on the CPU and GPU cooling HS. On the CPU side it was easy to overclock the q9200 where oversaturation would occur. @n 3.2 GHz I could prime95 for 90 seconds and see 85c then once the liquid could not return fast enough the heat would quickly (30-45 seconds) rise from there to shutdown. Now this was with ICD, so with the other heat TIM this would not just dissapear, the HS would still oversaturate.

What would it help then, Well casual usage has the CPU heat spiking to the pad, the faster the removal from the CPU to the pad the better. Also the more of an equalibrium of the chip to the HS for the temp. This should alow the CPU to be in a cooler state just before getting loaded up. This in theory would give the cpu longer from the start of a load to HS oversaturation. Also under general use should alow the CPU to run cooler.

So my interest in testing it out, and why I would have paid for it myself, is just under general usage, not a heavy load. ICD already allows enough heat transfer to oversaturate the HS, MX4 wouldn't alow me to run 3.2 GHz even under general use........................

 

how could mx-4 not allow 3.2GHz? mx-4 is better than ICD

 

No it is not when properly applied. there are a few, in the short term, as paste tims that are about as good as ICD. I had an installed ICD that was 1 year old and as good as the day I applied it. I did a take down with pics of the applied paste after pulling it apart. From what I can tell for the first 2 months Tuniq TX-4 works about as well, again properly applied, I suspect on month 3 though it is baking out. Since I got the Samsung though I haven't been using the P79...............

Edit MX-4 would boot but my intensive tasks at 3.2GHz like batch proccessing raw files failed.

 

HopelesslyFaithful was saying that because technical characteristics of MX-4 are better than IC Diamond, but truly those diamonds do not deteriorate with time... and who knows real characteristics which is different to what is written in spec sheets.

I confirm that HopelesslyFaithful sent me H-S and pressure paper for more than 1 laptop to the address I asked and all that arrived. Now it will take some time to arrive to Ukraine where I will share it with STiHiaL's M15X and perhaps some other laptop.
I hope all info I will provide will be useful.

 

i was referring to tx 4 is better than ICD if you apply every 2 months, which is what reborn does. But yes ICD does last far longer than tx 4. Also i have seen ICD crap out after 6 months of F@H so it's life is limited in F@H ussage. TX 4 is good for about 2 months and thats it.

 

TX4 is great for a liquid paste, it is MX4 from artic cool which is not all that good. ICD can crap out at 6 months if again not installed properly. Now proper installation is not just the user applying properly. If the HS is not properly applying good even preasure to the chip surface it will fail as well. ICD is known to be a pain to install compared to standard tim's as well.

I can't say TX4 is better though it could be. The problem with the thin tims is they all start to boil/bake out right away under high temps. ICD, again installed properly, will not bake out at all..........................

Edit; Again this is why if these were avaailable to the market I would have liked to try it out, as it appears it will not bake out but again it looks like these need care for a good even preasure to the chip from the HS..............

 

ICD is easy to apply i don't get people's issue with applying it. I made a silly little paint image on how to apply it. I apply it perfectly and when i take it off there are no bubbles but its performance after 6 months has dropped a good 3-5C.

it is either mx4 or tx 4 that has a 7wmk compared to ICD 4.5wmk. too lazy to look it up but one of them is better but only good for a couple months.

again ICD does bake and loose efficiency...even when installed right. Have done it twice with ~6 months it crapping out. So believe me or note i could care less i am not going to argue with you any more that under 80-90C load for 6 straight months with proper install it still craps out....better than any other paste but still has very noticeable limits

 

@Hopelesslyfaithful: My cooling system is like in the pic:
The system can actually cool pretty well, but I've been having issues with any TIM other than Prolimatech PK-1. PK-1 used to be able to keep my 920xm at 26/26/26/26 multipliers (3.46Ghz) across all cores at around 97 degrees celsius when gaming. AS5 can barely handle 24/23/21/20 multis reaching the same temps when gaming. Unfortunately I can't seem to find any more PK-1 in my country ATM and was curious about these heatsprings. Sorry I didn't answer faster, for some reason it seems my subscription to the thread didn't stick.

 

no biggy. How are you apply the AS5?

 

I use a line method on both the CPU and GPU cores and use a credit card to make as uniform and well-spread film over the CPU as I can. For the GPU I have to use a LOT of paste, i have very poor direct core contact from the heatsink (i ordered some less thick thermal pads to try and lower the heatsink more, I'll see how it works this week).
For the CPU it appears that core contact is more or less perfect, whenever I lift the heatsink the paste on the CPU is almost transparent and yet my temps are horrible compared to PK-1. I'm hoping the heatsink didn't start leaking gas over time, i have no idea how to tell. I ordered some PK-3 and will check how it works this week, that should tell me if the heatsink is starting to be defective or not (i've read reviews that state that PK-3 is better than PK-1, I hope it's true).

 

if i remember correctly with AS5 you do not spread. You use the same method of ICD where you have a bead if a square die or a line for a rectangle one. Then all you do is apply heatsink and let it spread itself. You may be getting excessive heat bubbles because of spreading. The goal is to apply the heatsink in one motion with as little movement as possible so the pressure spreads then TIM pushing out air so there are no bubbles. I could be wrong if that is the proper way for AS5 but i am pretty sure i am right. I also don't understand why any TIM you would spread because the concept remains the same. I always forget the name TX-4 or MX-4 you are supposed to spread it but it doesn't make sense to me since you will get bubbles.

Maybe tanware can put in his 2 cents. I have only used a few TIMs and last time i used AS5 was ~10-12 years ago on a desktop and i used the bead method but than again...i just did what my brother told me back than because i was like 12-14 and had no clue that i was doing ^^ Well one could still argue that today i guess

 

With it being a bare die core (for both CPU and GPU), it's essential that the TIM covers the entire core. Even if bubbles form, due to the pressure and the heat they should be immediately dissolved and the tim should reach optimal performance after about 200 hours of burn-in. The problem is that it simply sucks. I also can't afford to use the pea method since AS5 is slightly capacitive. Too little and i won't get enough time coverage. Too much and it might leak into things around and short my system. Anyway, do you still send the heatsprings out for testing?
I'm willing to pay shipping/cost if it's not too excessive.

 

You can use line and slush it. It doesn't conduct electricity and the little that pushes off won't matter. It is normally to get some on socket. Typing from phone so sorry for half baked response

 

Arctic Silver recommends the spread method for bare-die (read:laptop) applications. And it is indeed electrically conductive. They recommend not using the line or dot method for that exact reason. Most likely it wouldn't be conductive enough to short out your CPU, but I sure wouldn't chance it.

 

Arctic Silver Incorporated - Céramique 2

it says it is barely conductive...i have seen this crap caked onto a die with no issue so i doubt a little run will do anything. I wouldn't think twice. to each their own but i wouldn't worry. I would worry if you dripped it on the board but whatever.

 

Thanks for the advice, but I just got my pack of PK-3.
Conclusions are:
920xm is now runnable at 26/25/24/24. Stable and below 95 degrees.
7970m max temps dropped from 90 to 77 degrees.
Conclusion: Arctic Silver FAIL.

 

yea i can mail it to you no problem just get the prereqs done and give me your address. I just need 20 bucks to cover shipping since your international. I got to make sure i have enough pressure paper left since i sent a ton to James. I might need to order more so i'll have to check later. I got to see if they can mail me the stupid paper in a box or something instead of a tube. The stuff is a pain to work with or cut because it comes rolled up in a scroll....so stupid.

 

If you don't mind me asking, can you clarify what the prereqs are? My brain doesn't think very well right now and i have to admit the first post has me a little confused.
You need Idle / Load measurements of Intel Burn Test, Prime95 and Folding@Home?

 

So I have finally tested it today. Pressure paper shew almost nothing which means bad pressure but I still used just 1 sheet of HS at least for beginning.
I have feeling that temperature between 40-70 acts worse - slower cooling from 70C. As about higher temps under stress I am not sure now. I put it on CPU only.

I am not sure if I put it in the right way though. 1 side of it is more glossy while other is more matte. I putt matte side to the CPU die.
Perhaps I need 2 sheets. Will make stress tests soon.

 

good god.....james you are totally making these charts....or i am just leaving yours as source data for others to tinker with.....this is going to take me 3 days!!!! Trying to not loose files and work with my file maze is terrible

 

Is there a way to purchase one individually? I don't want to buy the 100-set pack. Or is there another thermal paste solution that still has decent performance?


EDIT: Regarding the experiment, I think we should also include high-end desktops, such as i7 3930k OC'ed to 5 GHz, or even a watercooled and OC'ed GTX Titan. Or maybe de-lidded lower end i7s and i5s.

After all, the TIM is quite expensive, so people who have the funds for a hexacore CPU (or multiple CPUs should they break one during the de-lidding process) and a Titan GPU setup are most likely going to buy the TIM, if it has a decent performance.

And no laptop can ever hope to outdo a proper $1500+ desktop build, in thermal output.

 

i would love too but no one wants to do the experiment. After my testing i honestly am going to cut requirements down. Granted i am doing ~4 times the amount of testing as someone else would do...maybe more but this is going to take forever.

Did the 3930x come with the crap tim or did they solder it like normal? Also you got anything to test this with? I still have like 7 2x2 things left.


I am hoping to get ICD testing done and work on H-S Monday. Sunday i got other stuff to do. I'll be up for another 5-6 hours so PM/post if you want. I am still debating how i want to put my data. I am thinking work file would be best and post all the other data in a zip folder for download. I think work would be easiest to show my work with fairly easy reading.

I also have some reservations that this will even work in my R4....my R4 heatsink seems like crap. Or maybe it is how the mother board is sitting that is cause these weird temps with ICD. I get 10C difference from hottest to coldest core :/ I'll explain later...no point in reading it 2 times



The heatsink seems good so i am wondering if it might be the case mounts or how the motherboard is resting cause one side to have 0 pressure. After i get pressure reading i'll know for sure. Maybe i can file (SP?) down parts of the mounting brackets so it will screw down lower. I'll have to mess around with the case after testing. Maybe i can come up with a way for people to mod cases to create better pressure.....time will tell. Crossing my fingers



EDIT: this is a thought of what might be the issue...just speculating until i got time to take apart the cooling system again and see what the pressure reading is. If this is the issue it might be fixable...assuming i don't botch the whole case -_-



Anyones thoughts?


UPDATE: yea this will take me a week. It'll take 4 days total to finish. I am almost 1/4 through. I think i got a couple hours to go. I am doing this:

Testing
1. High Ambient Temps
a. Low air flow
i. ICD
1. Idle
2. F@H
3. Prime95+Furmark (or equivalent)
4. Intel Burn Test
5. On and Off test (F@H)
ii. Heat-Spring
1. Idle
2. F@H
3. Prime95+Furmark (or equivalent)
4. Intel Burn Test
5. On and Off test (F@H)
b. High air flow (cooling pad/forced air)
i. ICD
1. Idle
2. F@H
3. Prime95+Furmark (or equivalent)
4. Intel Burn Test
5. On and Off test (F@H)
ii. Heat-Spring
1. Idle
2. F@H
3. Prime95+Furmark (or equivalent)
4. Intel Burn Test
5. On and Off test (F@H)
2. Low Ambient Temps
a. Low air flow
i. ICD
1. Idle
2. F@H
3. Prime95+Furmark (or equivalent)
4. Intel Burn Test
5. On and Off test (F@H)
ii. Heat-Spring
1. Idle
2. F@H
3. Prime95+Furmark (or equivalent)
4. Intel Burn Test
5. On and Off test (F@H)
b. High air flow (cooling pad/forced air)
i. ICD
1. Idle
2. F@H
3. Prime95+Furmark (or equivalent)
4. Intel Burn Test
5. On and Off test (F@H)
ii. Heat-Spring
1. Idle
2. F@H
3. Prime95+Furmark (or equivalent)
4. Intel Burn Test
5. On and Off test (F@H)

Plus a few other things on top of that. Each test is ~20mins long but have 10-20 mins run off so anyone can see a longer trend if they want to and than i can just take the prime 20 min junk. I am also trying a few other things and you'll just have to wait and see. I hope to maybe finish testing ICD by tomorrow/monday. I'll have to finish heat-spring stuff Wednesday-Friday before i go on a trip.

 

The 3930K comes with an integrated heat spreader I believe. It's the i7 3770k and lower models such as i5 3750k that have the thermal paste heat spreader sandwich. Maybe we should contact the Desktopreview forum for desktop testing, or some other websites such as Overclock.net? There might be a few people with GTX Titans.


As for the testing, I plan on getting an A10 5750m laptop during the Back to School sales. Not exactly a high thermal chip. Maybe I can increase heat output by OCing the GPU. I will also be modifying the cooling system.

I'm just looking for a location where I can buy the TIM individually.

 

You can't buy it indivuduqlly...well there is one reseller but I don't know if they sell pieces like that. I can tell u I am 90% confident it won't help u. I'll know for sure in a few days.

 

Alright just finish ICD testing....~20 hours of testing and charts. I'll do the pressure paper and my one fun side experiment Wednesday and I'll start on Heat-Spring Thursday but probably won't finish it til next week because i have to go on a trip with the wife. I am ~1/3 done -_-

 

Hi i pmed you and said awhile back id be willing to test.. i currently have a 3920xm if that works please respond to the pm i sent you..

 

got it resposned

 

I'd like to test on my m17x r4 w/3720qm pushed to 4.1ghz. What temps are you guys seeing?
Also, would this work on a GPU like a 7970m? That would be super nice.

 

Hm, the OP is no longer available... That's going to be a slight issue.

 

i am back....can you run a tech running prime and furmark on your R4 technos. make a log with hwinfo so i can take a look at it. I am currently having trouble getting it to work on my R4....pressure issues. I'll be busy/gone until monday and i'll be on and off until Wednesday. So just make a 10 min furmark prime run with overclocks and not what TIM you have and send it to me

 

just an update. I'll have to go to the machine shop Wednesday most likely. Also H-S needs to be cleaned. Not sure if i forgot about that or somehow in the process i wasn't informed...whatever. You can use a few type of cleaners on it. Below is a quote form an email from Jim. Also Dumb Dumb I'll send your's out friday after I finish mine because i still need to test a few things. Also Dumb Dumb how much do you want to do? The minimum or do you want to get in-depth with it? The choice is yours. I just got to figure out what I should tell you to test and what stuff i need to send you. Largely based on how much pressure paper I need to send.

Also I am currently lost on what to do with my R4....The pressure in the R4 is not enough to work and I don't see a wait to get it milled or increase the pressure with ease. The copper plate is not really reach able. Also anyone have a desktop with a high-end chip that we can test alone with a high-end heat sink? It is becoming clear so far that the application in laptops is extremely limited....ONLY because manufactures do not offer smooth heat sinks. Reborns worked and i bet any m18x will work but not even an m17x R4 works :/ Anyone have the highend clevo to test and see how that works? Maybe the clevo has a good heat sink that'll work.

Lastly, H-S also has a cure time i guess. So we need to also do a burn in process with these. it ranges from 1-2 days of burning in I guess. Again, see Jim's quote. If you guy's want more info on burn in process let me know and I can ask him.

I'll also post Wednesday my pressure readings for you guy's can get a look at it.

So basically H-S is appearing to be limited to current market of m18x and maybe a few others. I could have gotten a bad heat sink so I am curious on sending a few samples to a couple other R4 and see if they suffer the same as me. Also we really need to test some desktops. I might have enough for 2-3 desktops....so anyone got a 39xxX/K or whatever.

oh and I have another heat sink for my G51j so i'll be testing that too before i go to machinist to see if their is a difference.



EDIT: sorry missed/forgot about your post. I'll send it to you technos but would you mind buying some pressure paper? I am nearly out after testing a lot of stuff on my 2 computers. It is ~55 bucks on Amazon for a "square foot". I have already gone through 2 square feet for myself and sending to people.

 

I posted a link to this thread from Overclock.net

Check with Desktopreview and see if there are any interested testers.

 

yea i'll post in overclock.net in a week but i rather give it to NBR since i started here if possible.

 

Are you still looking for testers?
Please PM me, I was reading this post, and saw it posted over on overclockers thought I would ask.
Running a delidded I5 with water cooling.
Hope to hear back, thank you.

 

this was posted at overclockers??? I ahven't even gotten a chance to post it over there ^^

I have been busy the last couple days making an XP gaming machine and ended up buying a 3770 on k that over heats...need to get a decent heat sink and test if it is stable....if so i'll be testing it for this.


can you send me link to thread at OC.net

 

Heat-Spring TIMs testing for free (limited samples), with strings attached
He was trying to assist you because you are so busy!
PMed him over there and he linked me back here, so I went ahead and registered here.

 

oops, I guess it was overclock.net ;-)
http://www.overclock.net/t/1382539/...or-free-limited-samples-with-strings-attached

 

yea i found it after googling....odd i never got a subscription notification you replied. Oh well.

I am planning on porting this over this weekend i think to overclock.net too....do you know what section i should make this in? I also need people to buy pressure paper....i r broke as a joke. I bought a desktop to test and have spent ~110-120 on pressure paper already. Plus i am going to get my heat sink milled Monday. Forgot my little baby was getting spayed this week. Wife reminded me so I have been stuck at home watching my dog. I can't leave her alone because she has problems with separation anxiety :/ I don' want her to rip her stitches. Plus i wasted a day or two finding a good XP desktop but ended up just getting a new one since it was a good deal...or i hope it was. Tomorrow heat sink comes in and I'll find out. I really would had to send this stuff out without getting pressure readings. I really need a better look at a larger test sample with pressure readings :/

EDIT: BTW do you know of a good de lidding guide? I want to give that a shot but i want a good guide since i have no idea what i am doing in that process.

 

I hope it went well for you!

I watched a youtube video on it, I then used a replaceable razorblade (the kind handymen use) to slowly pull the lid off.
I would highly suggest taking your time. The razor I used was thicker and had a double edge. I started at a corner (wearing latex gloves) and just kinda got it under then, then carefully pried up, while slowly cutting, be VERY careful on how deep you go under the lid. once you feel that corner is loose, go do the next three. Then slowly edge in toward the middle on each corner while prying lightly!

Was SUPER easy!

Preasure paper can be purchased at a hobby store?

 

http://forum.notebookreview.com/not...est-grit-sandpaper-i-see-2000-12000-grit.html

i made a thread about using high grit sand paper to even out a heat sink and it appears to be working but i have ran out and am waiting for my paper from amazon to come in. When i get more paper in i'll be able to finish and see if it really works. I do have hgh hope for it.

 

Hi, this thread started almost 9 months ago, do we have some results to share as of yet? Again I would like to see how well these work and some way to obtain in the retail market some of these if they do work.......................

 

i have a desktop i got to try but unsure how well it will work since it is only a 40w 3770. I'll give it a shot soon. Been trying to get my laptop to work since i have done all the ICD tests but H-S was a fail. Pressure blows. Reborn posted his results way back and it worked well for me but it was better than stock TIM in my R4 but worse than ICD. Again with my R4 it was because lack of pressure. I'll see if i can upload some of the pressure layouts and make a single graph for you. I haven't had time to see if i can force higher pressure in my R4. The evenness of the heat sink as far as i could remember was good but pressure was super low. I'll see what i can do over the weekend. I really need to scan and organize pressure readings before they get lost.....so many of them. I have use about half a paper so far.


I really want to get the heat sink even on my G51j since that CPU can go up to 80w and higher but i can't get it to go even with sand paper. I think i got a couple ideas on how to do it but have to wait until June 4th i think for my subscribe and save to come in :/


i might see if the machine shop can mill me a piece of copper and de solder my heat sink and resolder the new one on but the solder job on the current heat sink looks really good and i don't know if he could match it. I'll go back after i get more sand paper in if i can't get that to work. I would love if i could figure out a good way for people to use sandpaper to even out their heat sinks so that they can have better cooling.


Also tanware you are real smart and know a lot of random stuff. Are CPU dies uneven? Are they perfectly flat?


BTW TANware I did say that H-S will not work for any device that doesn't have high pressure and even layout. May work better than stock or low end TIMs but compared to ICD it is much worse if you have uneven temps or low pressure. As i said doesn't work with my G51j or R4 with current pressure and unevenness

G51j has great pressure after grinding down mounts but horrible uneven. Might be able to fix unevenness. I made it better but than totally jacked it up. -_-

R4 is fairly even from what i remember. I'll have to double check though but the pressure was a joke but that might be easily fixed.

so both are a fail so far.

my desktop i haven't had a chance to test yet but i don't know how well it'll work. It runs cold as it. 40w desktop runs at 55 C under load so it isn't really a high temp area but might work well. I got to give it a shot but after i exhaust my resources with my g51j

 

Well it is rare to have a perfect match. About the only way to achieve this is with a lapped independent heat sink. This would be like where a desktop has a free standing heat sink attached to the socket and not needing support attachment elsewhere. With laptops this is hard to achieve. Even in a desktop, where weight of a heatsink can torque it from the chip, this may be unobtainable. in a perfect world where no stress is applied maybe H-S would do better?

I was still hoping for success stories. As with my P79xx I had used copper shims to level out the heatsink to get even the best results with ICD. Maybe we just need to wait for a heat spring that can function in a substantially lower pressure environment with a bit more forgiveness of patch consistency............

 

BTW using the sand paper is a homemade way of trying to lap the surface but it is really hard to do it right. I'll let you know in June when i get the new paper and let you know how it goes. If you got any ideas on how to make it work better please let me know.

 

I could test it on my AW M15x with a 920xm in it - if youre still wanting results?