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

Okay peeps,

I've got the HS installed in my M18x - specs in my sig below.

A few pics to show initial temps + some temps during full folding@home load. :thumbsup:




CPU running stock clocks.


My initial impressions are that it is quite good. Idle temps aren't the greatest however good enough. High 40s to low 50s for the CPU. Very consistent temps.
The impressive part atm is the temps under load. I noticed around 5 - 7 degrees lower folding temps which has allowed me to up my folding CPU frequency from 2.8ghz to 3.3ghz accross all 4 CPU cores.

My main aim was to keep temps below and from ever reaching 90 degrees as this is when the throttling can start to kick in. So far so good.

The extra benefit of this HS is if I don't ever have to pull the system apart and repaste the thermal paste again.
During heavy 24/7 CPU GPU folding the paste can pretty much disappear lol withing a month or 2. I will see how temps go in a few months time to determine durability and HS effectiveness.

Thanks again to HopelesslyFaithful for letting me test it out.
I hope to have it installed in my other systems when I get some more time.


CPU f@h temps whilst Overclocked to 3.3Ghz.

Cheers.

 

A quick snapshot of how HS looks on the CPU die.



Cheers.

 

BTW you need to Trim the sheet so you don't short it. It is conductive. Also i am uploading the directions for the pressure paper. Can you do the pressure paper still? Sorry for taking so long was going to do it this morning but wife never got up before i left to school and got home not that long ago. Also did you crank that stuff down as much as you could? Also something that might help but hasn't been tried is after you apply it and get it running at 90C try to tighten it some more. The stuff melts at 156C so at 90C it should be far more malleable! Just an idea i had but not standard protocol. If you can do pressure paper and scan those and then try again and see if some more pressure and than tightening again under load will help!


Indium - Wikipedia, the free encyclopedia

uploading to drop box now. It is IMPERATIVE that you follow directions on pressure paper. Temperature and humidity affect it so make sure everything is cold and at room temp and note what room temp is!

safety sheet of H-S
https://www.dropbox.com/s/wh6a4klv22hldyn/Indium Corporation safty sheet 20120913130753.pdf

Prescale Ultra-low Film Documents/Directions
https://www.dropbox.com/s/zup2gdr23eac6g4/Prescale Ultra-low film documents 20121110232058.pdf

This is also added to main page


EDIT: tomorrow i'll find all my videos of H-S and pictures and do an application guide. You want at most 1mm on each side since this is conductive and you don't want to short anything! Just in case you don't see my PM chief you need to trim that ASAP to be safe!

EDIT: geex checked that pic out again Chief and H-S is right next to some stuff on the DIE! I don't want your DIE to DIE!!!

EDIT: Chief is that a thermal pad on your Heat sink? Do you not have a copper heatsink? that blows if you don't what a waste of cooling capacity

EDIT: also don't forget to make csv files as requirements call for and make a PPT show so everyone can see everything together! Can't wait to finish my cooler this week and test out the R4 and G51j!

 

Hey Bud,

Yup all good I managed to trim it to fit nice and snug before installing it.
I remember you saying and reading it was conductive unlike some of the other thermal paste.

I tried some pressure paper testing but it didn't turn out so well.
Basically though on the pressure paper the 1st time it showed most of the die rectangle shaped impression. It wasn't very strong so I tried to screw down the HS a bit more. It made the full die rectangle shaped impression. So it shows that it is making pretty good contact.
I will try again next time when I get the other 2 pieces installed. I'll need to read up on your guide as well. Thanks.

Yeh the cooler is copper. I did clean the CPU and HS before placing the HS. The black you see on the HS is just the black tape that the HS comes preinstalled with.

I'm currently working on your csv logging files. It should be great to see a direct comparison with the files I sent you last time.

Cheers.

 

shouldn't you remove the black tape.......that should interfere right? Also it doesn't look liek copper...the pipes look like copper but the HS does not


The HS looks like galvanized metal....You sure the tap is supposed to be there? and its actual copper?

 

Yeh I can remove it however the die only makes contact with the copper part of the HS. Its only really meant for other CPU die sizes which don't fit the shape of the Heatsink. So I don't think it should interfere to much at this stage. I can take it off on my other 2 installs though to test.

Yeah the part that makes contact between the HS and CPU die is copper. The surrounding bit that makes up the rest of the HS most likely isn't but the die contact point which connects to the 3 pipes is so it should be okay.

Do you know of anyone else having tested HS in say a Desktop environment. Extra HeatSink size and pressure should sure help I reckon.

Cheers.

 

nope no one has tried HS on desktops yet trying to find some people. Extide and jimbo were supposed to but they disappeared :/ NM i looked again and zoomed in...i can just barely see the copper in it nm your good lol....was hard to see that

 

Very impressive results!

You say that the film produce 5-7C less temperatures, but compared to what paste? Would you recommend this to everyone over any paste?

 

Hey Cloud,

Yeh around 5 - 7 degrees Celcius temp drops immediately under load whilst f@h under full load.
I based the temperature results against Dells stock paste on the CPU HS and against MX-4 that I normally use.

With the temp drop and much stable temps I was able to boost my CPU clocks from 2.8GHz folding@home to 3.3Ghz constant on all 4 cores.
So far so good. I have now been folding for over 2 days straight without any throttling.

I used to hit 90 degrees or be extremely close to 90degrees C whilst folding in the past.
Now idle temps aren't the greatest. Other pastes can get the CPU much colder but under load temps especially whilst F@H this HS material seems to work pretty great.

Now recommending it would be quite hard as it costs roughly $250 per sheet or so. HopelesslyFaithful has more info on pricing though. So I would get in touch with him to do some testing of HeatSpring.
I would recommend trying to test it 1st if possible. Its main benefit would be lower load temps depending on usage and CPU Load.

It should also highly benefit F@H users + high constant CPU Benchers. Those whom are constantly pushing the thermal limits of the CPU cooling components.

I am testing durability now. On specs paper it won't dry out or spread to the edges like thermal paste does so in theory it should say intact and worked like close to new for a very long time. Well that is what the specs sheet say and I shall be putting this to the test during a hot hawt summer coming up.

As for myself I'm liking it so far. Running F@H CPU clocked at 3.3Ghz with fans running maxed at 3800RPM.

Cheers.

 

i think he was using MX-4 or whatever something something -4 TX-4? he said that was working better for him than ICD but paste went to crap in a month or two. ICD for me has been doing decent over 6-9 months or more of folding. MX-4 or whatever seems to be better but doesn't last. Also looks like mx-4 or whatever. That white paste and all. If this stuff really lasts essential forever i think that alone is worth it. It costs 15 bucks for 2x2 piece (at 20 pieces for 300 bucks) (could retail that at that price assuming someone bought bulk and got it for 8-10 bucks and added the extra charge for the service) so that can get i think ~8 laptop dies. So thats not a horrible price considering ICD is 25 bucks for 24c tube which gets you maybe 20 apps i think. It might be 2x times more money but hey you get better results and never have to repaste again ^-^


I'll make a new thread in the near future after i post my results and get a little more coverage on this topic and ask everyone if they would want to buy it if i sold it on amazon. I might take 1-2 grand out of saving to sell it on amazon for everyone to buy. I think if buy drop 2gs on this i can get it cheap enough to sell it at a reasonable price and keep my risk of buying 2 grand of this to a minimum.

 

dude only 80-85C? dude overvolt and crack it up to 3.5GHz! and get a cooler to help to run cooler and overvolt to 4 GHz! I am getting 2.8-3.2GHz depending on ambient temps on a G51j! with ICD lol with my jerry rigged computer! Once i finish cooler and try H-S I'll be tring svl7 pin overvolt mod to get to 4GHz and maybe even to 4.5GHz on a 920xm! .....if my 180w PSU can get me there *fingers crossed*

EDIT: reborn you said you were getting 85-90- before right? at 2.8GHz? You still got head room with H-S from old temps get it to 3.5GHz! than you will ahve gained 700mhz within same temps!

 

HAHAAHAh yeh I want to push closer to 4Ghz, that would be awesome to run folding 24/7 ahahahah. Ambient temps will be my worst enemy during summer. Since this is my main machine I need it to be stable as well for work and stuff. So far running it at my current settings allow for the best all round performance and use.

I will try and increase the CPU soon though. TO try and find an even higher stable clock for regular day to day usage and F@h.

I would buy some more in the future.


I just turned of F@h for a few minutes but left the fans running at full tilt 3800RPM.
Here are the temp results - nice GPU temp lol!


Cheers.

 

Does this actually work better vs ICD24? I have a m17x r4 w/ a 3720qm OC'd to 4GHZ and a 7970m that get rather HOT under load, i'm pushing 85-90 on CPU and 78 on GPU

 

I'll be testing my R4 this week I hope.

 

If the other guys who dropped off this thread did it for the reason I did, it's because things kept changing, and requirements kept morphing into different things than we had agreed to do. I guess for me, I didn't want to feel like I agreed to do one thing then get chastised if I didn't want to do the new stuff that kept changing. That and the response and update rate seemed to be a little flaky.

 

reason why things changed was because people complained so i reduced what was required

EDIT: primarily durations....most people didn't want to do the long burn in tests so i hacked everything but the basics

 

If you want to know point of view of noninterested in anything guy I will tell these:
- ekretz was talking about changes made earlier
- You asked like 5 times to confirm that people really, sure,super really and 100% after all new requirements still ready to take a part of it.
- People expected much simpler stuff here.

My friend here who helped you with how to STiHiaL-HWinfo tutorial asked me (as I am the one who told him about this thread) if that guy is going to send anything at all and I said that I don't now. I personally was surprised that you sent smth to someone at all!

Just the way you ask and ask confirmation makes people think that you are either not interested in this or not even going to do it.

The list of agreed people must be automatically filled by the one who already sent you schedules! And you try to send them PM to ask an address instead of updating either firs post which doesn't show up in User PC of members or adding new pots to a thread where more than 50% of members already gave up of waiting for some confirmation of you.

am I right eKretz?

 

alright I updated the thread again and made a few more spoiler sections to make it easier to read and follow. I also made an application form and checklist so there is no reason for anyone to not understand whats required and needed to be done. This should make everyone's life easier. Please for all future people email me everything in a single email...it takes way to much time trying to track down all the info when it is in several PMs and several emails.

If you got an ideas on how to make this easier or if i have made a mistake in one of the documents please let me know so i can fix it. Also please let me know what you think of the new process. Please be understanding that this is the first time i have done anything like this and i am learning as I go. I hope this has fixed all the problems we have had in the past. I know all of the forms and checklists i have made will save me a lot of time and make it much easier for me to keep track of everyone's files and progress.

@JamesD
BTW reason why i asked for replies was because i made changes and i don't want anyone left out. I also didn't want people to feel as if they were being taken advantage of since i changed stuff. I wanted to make sure people agreed and we didn't have conflict with any changes. Also the reason why I want people to post on here that they agree is that i want to make sure i can keep people accountable. I have limited supplies and it costs me money to mail it to people. So give me a break that i don't want my time and money wasted doing this....it is kind of ridiculous that you hold it against me that in my spare time I try to help people out with getting free samples and trying a product for free that has only been available to the server/super computer industry. As i said above this is a new thing for me and i am learning as i go. Also everyone else except Reborn has made it fairly difficult because they email me 1 piece at a time over a week or two and it is incredible hard to keep track of 50 emails from 3 people and what they have sent me.

EDIT: I also add privacy statement and how i protect any sensitive data you send me

EDIT: Also added this to requirements to clarify a few things so testing is accurate.

Now if you want to log the ramp up and ramp down thats fine but at least give it a full 15 mins are max temps. You'll probably run into a good 20-30min log maybe more depending on system. Thanks guys just want these tests to be accurate so no one can say we didn't do our part!

 

Hey HF,

Idk if it works in Hotmail, but in Yahoo you can filter emails depending on content and/or sender. Then make separate folders for each user/email address. That should make it easier for you to track mails.

But for all purposes I'll be sending you everything in a single email as soon as I finish again. Sorry I can't redo the other tests with 15mins on max temps, though.

 

NP also check your email. I mentioned in checklist on thread that Intel burn test and prime95/F@H need to be used. Intel burn test is very hard on CPU but inconsistent. F@H/prime95 are CPU heavy but not the same as intel burn test.

Also the drop box files should be back up. i redid my whole dropbox folder so files needed to be updated

EDIT:nm they are now fixed. I guess if you move the files location in dropbox it kills the link :/ such a pain. So if i move folders or files around link dies. sigh i hope i don't have to reorganize how i have stuff again ^^

 

Resent my mail, HF. Sorry wasn't able to indicate there. I did run the tests while both Prime and Intel Burn Test were running. Pretty frightened here. Dang. Haha.

 

you lost me sorry. You couldn't indicate what? You just need to add it to file name. Also why are you frightened?

 

I wasn't able to indicate which tests I ran. And I was scared because the processor reached an all-time high lol.

 

what do you mean? After you start log and run one test you stop and change name to include which test. I have to know what tests and settings are for each one otherwise i can't use the information because i don't know what it is

 

Sorry. I wasn't clear. Remember you asked me to run both Prime95 and Intel Burn test? I wasn't able to indicate that I ran those tests in the email I sent you, because previously I was only using Prime95.

Anyway that's not too important anymore. were you able to get it?

 

can you either reemail me them renamed with the which tests were used or try to explain which ones which because i don't got a clue

 

Alright. Resent it under the subject "Fw: Heat Spring TIM Application - kisetsu17". I also filled up the application form you provided here and attached it there.

 

How much for a 920xm sized piece? Send us a PM

My current record is 3.33ghz continuous 100% load on my cpu which brings the temp up to 95C in 20C ambient after about 12 hours of video encoding.

If it shaves 5-7C off then I will OV and push to 3.6ghz x 8 and get those jobs done quicker

I read you wanna hit 4-4.5ghz on a 920xm. It ain:t gonna happen lol. Perhaps for a very quick bench but 4ghz plus loaded for hours will require more than water cooling in your laptop haha

 

i'll get back to you tomorrow. I got to look through kisetsu first. Holidays are over so family is gone so i got time to get a bunch of this done. I can get 3.2GHz at 80-85C in a cool room and thats with cooler not maxed.

 

I need you to read the thread for requirements and email me form and prerequisite logs. Shipping should be about....errr USPS site is down. Shipping should be anywhere from 20-25 bucks tops. So far i have averaged a tad under 20 bucks for international shipping.

 

doing my testing right now and i am doing a million different comparisons to give a wide gamut of situations. I installed heat spring and something is wrong. Temps on 3 of 4 cores are a hair higher and one is 10C higher than the others. I am seeing 74-77C on 3 cores and 86-88C on another core so i am not sure what is wrong so i am going to have to take a break and start over this weekend and take a further look to see what is wrong. It might have to do with pressure. My pressure paper layout is horrible. I'll report more when i got some more time. Been up working on this for the last 6-7 hours and it is 6:40 am and i need sleep.

EDIT: yea i am getting 10-20C difference on one core so i got to crack this thing open later and take a look

 

If you re-apply, maybe consider rotating your heat spring 180°C to see if a different core gets hot instead.

Given that heat conductivity is strongly determined by contact & pressure, is there a way you could gauge whether the location of the hot core correlates with the location of the worse contact area indicated by the pressure paper?

 

yea i forgot to keep note on my pressure paper to keep record of top and bottom lol. Can everyone else when you use the pressure paper keep note of top and bottom....it might be beneficial for trouble shooting. Also i used a scrap piece on mine....it was just barely big enough so that might be the issue. I'll use that piece for my IB CPU since it'll fit better there and cut a new piece. Also if the new piece has the same issue i'll email Jim asking him if he can "acquire" me a thinker sample to see if that helps. It could be that laptop heatsinks are too low of quality to work well with H-S. Desktop heatsinks are made much better. I'll keep you guy's posted. got papers and exams so i won't have time til this weekend at best to trouble shoot or maybe not even til next weekend.


@ APU71 and @ Reborn2003 and anyone else i mailed to ^^ (not sure if @Username works here):

Have you guy's received all samples and finished all your testing? I know reborn has most of it done but i can't remember if he finished all 3 and did pressure paper. If you finished please fill out the complete form with after thoughts. I would like an analysis from everyone letting me know if they have noticed any issues or interesting connections. I know reborn said he noticed more consistent temps. Can you try Intel burn test? That test is all over the wall with TDP so if you see improved consistency in temps that means H-S increasing in efficiency with great amounts of heat. Also if you notice large differences in cores due to pressure issue that is VERY important to note. I will be looking very closely at that with mine and compare results. Maybe get a thinker TIM to see if that helps. Thicker the TIM worse it is but if pressure greatly improves that will mitigate the increase in thickness.

Also if you are willing down the road to make those PPT shows that would be awesome. I'll make mine once i figure out what the hell is going on lol. No clue how i am getting 10-20C higher on a single core lol. I know one core in the past always ran +-5C higher with ICD but this is nuts lol. I was able to pull off 70-80TDP with my cooler at 2/3 capacity in both 17 and 25C ambient temps ^^ Cooler was running at maybe +-200 CFM ^^ I am doing a huge analysis between high powered coolers with drilled out cases and high and low ambient temps. Took me 6-8 hours to get 8-12 tests done with all the varied conditions with multiple benchmark programs. I can't remember off hand if i did Intel burn test. I think i did but can't remember. I know F@H and prime95 have very different power requirements. Prime95 uses 5-10w TDP more so temps are worse and clocks can't go as high due to voltage issues. Also i did screw up....i forgot to record idle temps -_- I did record idle temps at beginning of all tests so i'll just use those. I recorded test ramp ups so i can also see how each TIM effects cooling system saturation. This is important for several reason. If you have usage where your not on and off max draw. So let us say you use 100% for 60s and stop for 30s and do that on and off. You may not fully saturate your cooling system and hit max temps. This means for your use you maybe able to squeeze out another 20% overclock while maintaining your ideal thermal envelope. I know some people have usage like this so this maybe useful to them. If it cuts your short crunch time down by 20% due to increase clocks that could save a lot of time. Does anyone know off the top of their head what kind of usage causes this? Rendering?


Email to Jim. It has some interesting ideas i want to try.

EDIT: Also at reborn and APU71. Did I include my contact info for you to get a hold of me if you needed help?

 

For the love of god make graphs/pictures with your temps on one single post when you people are done so that we can examine the results ourselves. I don`t want to read through 100000 words to find the data

 

lol when i finish. I am not done. I got to trouble shoot. The paragraphs are the best part! They have all the nitty gritty details! The real analysis! What do you think of 10-20C higher on a single core lol. Somethings wrong lol.


http://img39.imageshack.us/img39/6231/lolbotchedtest.png

the base 3 cores are 5-10C higher than ICD and the one core is 20+ higher lol. Something got messed up so when i get more time to trouble shoot and figure out what the problem is i'll report back. Also i was trying to use a "scrap" piece. It was a hair to small and a bit messed up so note don't use those type of pieces -_- Also i am stoked to try my idea i mentioned to Jim. I want to try a new pressure piece alone and than try another pressure piece with H-S in there and see how the pressure has changed. I will put money on the pressure will be substantially higher.

 

Yeah that is very odd. Almost like you don`t have good contact in that area but since its so small, is that even possible?

 

i'll elaborate on that later.. It could be a few things. After i finish papers and exams over the next week and a half i'll troubleshoot it and let everyone know what the problem is.

 

alright break is here so Monday or Tuesday i'll start working on this again. Someone emailed me tests and i'll get back to you if i got everything i needed. Also i think it is someone in another country so i'll have to get you to send me a some money for shipping. I'll also be finishing my testing up and reporting back on what the issue i had was and let everyone know what not to do -_-

Also i'll be going through reborns cluster *cough* of a zip file and be posting his data if i got everything i need. Plus this week since i am off i should finish the tests as long as i don't have too many issues with testing again, which means i'll post my complete PowerPoint show so everyone will know what i am looking for. I'll also make a template to make it easier on everyone ^^ We all know we are lazy and hate the "administrative"/paper work BS ^^

I'll also post a full analysis of the pressure stuff too. (refer back a few pages if your out of the loop)

 

All I needed to create that email were just 3 docs on dropbox. This is what is called "instead of 1000 words ". Just make sure those links are pointed out well. I understand you want to do it tip.top but you don.t need to write a book here. We are smart enough to get important info from 3 sentences

.S. I am in a similar situation about writing FAQ about RAM on russia forum. Now I see what to change looking from a side on you.

 

still got to look at yours james and i'll do that in the morning. I was working on mine but have to stop since the power source pulled out of internal fan. I can't find my electrical tape so I'll have to wait until i can get to the store to buy some. Can't do tests without internal case fan ^^. I redid all my pressure tests and have taken 4 samples. You really have to tighten a little on each side rotating like you do when you tighten tires....it complete messed up pressure reading when i tighten one side at a time. My pressure readings are horrid. I only have pressure on 50% of die! I have double stacked H-S because i did pressure test with H-S and it barely reads at max pressure! I'll post them all this weekend when i finish. Really messed up on how terrible my heatsink is. There was no quality control on the heatsinks in the G51j....thanks ASUS for being utter garbage! I hope my dell heatsink fares better when i get a chance. I definitely can see H-S not working for 90% of people due to terrible heatsinks. The fact that i have 50% contact over 18PSI (i think that is the lowest it reads) is abysmal! I have maybe 70% contact above 18 PSI with two H-S. That is .008 in of thickness.....may not sound a lot but for added space in TIM that is huge.

 

hey james everything is good except i need a prime 95 test too. After you send that i can mail it out in 1 day. I am on break and so is my wife so i can make it to the post office that day or the day after. Also i am mailing it to that that in UT right?

 

Will be done, sir
Tomorrow maybe or the day after tomorrow.

You recalled me one joke.

Doctor examines his patient and whispers:
- Gooood... Good... Very good.
- Doc, what exactly is good?

Spoiler

- Good that I don't have these problems which you have.

 

hahaha nice

 

@HopelesslyFaithful Done. Sent everything again. Also told StiHial to do the same.

 

yea i got it today. I'll look over it in the morning and let you know if i got everything ro mailing it....need to go to store...my electrical tape is still missing and need to finish mine lol

 

We all are curious with your laptop's trouble and waiting when you tame that presumptuous core with 10-20 degrees higher temp

 

yea i stacked two and pressure isn't even that good

 

OK, I finally want to get the required testing done for this to see if I can use heatspring to drop temps, and create more room to OC the 980x for F@h.

What exact tests are required? I see 5 min idle, 15 min max temp with Prime95/F@h, and a weeklong log of general usage? Is it a week straight with no restarts? That one I'm a bit confused on.

Also, has there been any results submitted yet with HS? And how did the HS fair against standard TIM?

 

@DR650SE Heat-Spring looks F@H friendly

 

@HopelesslyFaithful: Are you still doing tests on this? Or is there anyway I can buy some of those TIM's. My 920xm likes to flirt with hitting 100 celsius, and my 7970m gpu tends to get significantly hotter than other people's, heading for the 90's