Sign in to follow this  
turboken

De-lidding with AS5 ?

Recommended Posts

Ok so I'm thinking of de-lidding. Currently I'm using AS5 between the CPU and HS. 

Everyone seems to be using the cooler lab pro stuff... Can I just use Artic Silver 5 as thats what I have sitting here on my desk. LOL. 

 

Thanks

Share this post


Link to post
Share on other sites
Help AVSIM continue to serve you!
Please donate today!

I used AS5 in between the IHS and die recently after being dismayed when the CLU I originally used after delidding had either run or hardened (it'd only been on for a month). The results weren't good and essentially had undone my temp gains (er, loss, rather). I was essentially back at stock.

 

I put the CLU back on (both the die and the inside of the IHS to make sure that there was plenty on there) and things are great again.

 

No idea what happened to my first CLU application but, just out of the blue, I decided to run a stress test and the CPU was over 100*C. Glad I ran that test and discovered that...

 

So no, using AS5 doesn't help much which is a little odd considering it has to, at the very least, be better than the stock Intel TIM. Use CLU and watch your temps fall off...

Share this post


Link to post
Share on other sites

The Coolab pro seems to be getting very slight gains (lower temp) over ultra, via TomsHardware. I used Gelid under the IHS and got great results without the worry of vrm shorting. But AS5 should work just fine.

Share this post


Link to post
Share on other sites

Ultra or Pro will produce the best results. There's no doubt about this now, plenty of evidence.

 

If you use AS5 you may experience a reduction in temperature but you'll be lucky if it's 10 degrees. Compare that to Pro which frequently reduces temps by 20 degrees, sometimes more.

 

I watched a video the other day where a guy delidded a 6700K and used Thermal Grisly. A 10 degree reduction was all he managed. And bare in mind that Thermal Grisly is better compound than AS5.

 

So yes, you will probably see a reduction, but don't expect much. And then of course you'll have to decide if a mere 10 degrees [if you're lucky] is worth the trouble. I would say not.

 

There are two factors to consider, type of compound and depth of compound. Intel used a somewhat lesser compound and it was in a thick layer as a result of the adhesive used to attach the IHS.

Share this post


Link to post
Share on other sites

The point is, if you're going to be as stupid as me and the rest of the idiots that have either taken a razor or hammer to a $300+ computer part, get your money's worth and make sure you walk away with the full benefit.

 

I bought some CLU for around $10 on Amazon a month or so ago. Aside from whatever happen to my overclock a week ago, there's no downside. Some complain about applying it but take the brush that comes with it, cut the bristles in half and let your inner Picasso out.

Share this post


Link to post
Share on other sites

Absolutely agree with you Greg. Best to reap the full rewards for all the effort and risk. Yes the risk is small with care, but it still exists.

 

What sort of temp reduction did you get with your delliding Greg?

 

I wonder if anyone has tried delliding with Coollaboratory Liquid Copper? Not as efficient as Pro or Ultra, but in tests it's the best of the conventional pastes.

 

http://www.coollaboratory.com/en/products/liquid-copper/

 

Regarding your issue with Ultra, I have read that it does dry up after a time.

Share this post


Link to post
Share on other sites

 

 


Regarding your issue with Ultra, I have read that it does dry up after a time.

 

Does Pro also have this problem as I would rather not have to replace this stuff on a regular basis.

 

Has anyone used AS5 long term (1 year or more) between the die and heat spreader? Have you encountered any issues?

Share this post


Link to post
Share on other sites

Absolutely agree with you Greg. Best to reap the full rewards for all the effort and risk. Yes the risk is small with care, but it still exists.

 

What sort of temp reduction did you get with your delliding Greg?

 

I wonder if anyone has tried delliding with Coollaboratory Liquid Copper? Not as efficient as Pro or Ultra, but in tests it's the best of the conventional pastes.

 

http://www.coollaboratory.com/en/products/liquid-copper/

 

Regarding your issue with Ultra, I have read that it does dry up after a time.

 

I was pretty OCD about my testing methodolgy. I ran an Intel Extreme Tuning Utility stress test before and after the delid process. The ambient temperature was the same and both runs were preempted by a 5 minute "warm up." At 4.8Ghz, my 3770k cooled off 9.5*C on average across all cores. The average maximums decrease was 10.75*C.

 

Full data here:

 

Before Delid (4.8Ghz HyperThreading On / 25*C Ambient / 5 minute warm up)

                    Core 0         Core 1           Core 2          Core 3            Average:

Average       79.80           83.46             82.77            78.76             81.20

Max                  85                92                90                  86               88.25

 

After Delid (4.8Ghz HyperThreading On / 25*C Ambient / 5 minute warm up)

                    Core 0         Core 1           Core 2          Core 3            Average:

Average       74.54           69.75             74.61            67.85             71.69

Max                 79                76                 81                  74               77.5

 

 

I too have read about CLU drying but it seems like that happens over the course of years, not weeks. I really don't think that's what happened. To be honest, I have no idea what happened but, as the common definition of "insanity" goes, I just reapplied the CLU and am hoping for the same thing not to happen again. Just to reiterate, I applied it on the die--again--but also on the inside of the IHS in case the CLU decides to shift or move around (it really seems to resemble solder and with a melting point of 8*C (found in the MSDS), it is a little viscous at room temperature), there should be more coverage. It's also worthy to note that when you start to clamp the socket brace over the IHS, the IHS will get shoved downward a tad. I wonder if that had anything to do with anything...

Share this post


Link to post
Share on other sites

Does Pro also have this problem as I would rather not have to replace this stuff on a regular basis.

 

Has anyone used AS5 long term (1 year or more) between the die and heat spreader? Have you encountered any issues?

No idea to be honest. I was researching the other day and came across claims of Ultra drying out, but not seen anything re Pro. You have to be careful of these claims I guess, if you look hard enough on the internet you'll find all manner of things, individuals hypothesizing.

 

Best place to look is on the delidding forums for long term use of compounds when delidding....

 

http://www.overclock.net/t/1313179/official-delidded-club-guide

 

As you can see form the chart some of these guys use AS5. Best to ask them what it's like for long tern use between die and IHS. The majority are using Pro or Ultra, would be interesting to ask them if it dries out, given it's a big sample size.

 

The average maximums decrease was 10.75*C.

 

 

Not a massive drop then Greg. Some have achieved minus 20 degrees.

 

Having said that, it could be that your CPU was one of the better examples in terms of the IHS/Die interface.

 

Could be that Intel did a better job of installing the IHS and that's why you achived only a 10 degree reduction in temp. My guess is that those who achieve 20 degrees plus, had CPU's with thicker adhesive and thus paste.

Share this post


Link to post
Share on other sites

Ok so I'm thinking of de-lidding. Currently I'm using AS5 between the CPU and HS. 

Everyone seems to be using the cooler lab pro stuff... Can I just use Artic Silver 5 as thats what I have sitting here on my desk. LOL. 

 

Thanks

 

UPDATE: OK so I had an hour of free time the other day ( I just moved and have a 2 yr old = zero free time) so I performed a delid operation on my cpu. I used AS5 as thats what I had sitting here and if I didnt like the results I would just change it. So I'm not using my sandy bridge 2600k anymore as I built a new computer just before I started this moving thing. So now I have a devils canyon 4690k. 

 

So on to the results....  NOT GOOD, turns out the shin etsu stuff intel uses isn't that bad after all. I'm seeing a 3 degree C rise in temps during Intel Burn Test. So I will order some Coollaboratory Liquid Ultra and report back once reapplied. Also when I get time I would like to do a head to head comparison from my 2600k to 4690k both at 4.8 with build pictures.

Share this post


Link to post
Share on other sites

How thick did you apply it Ken? Most of the results I've seen with AS5 demonstrate at least a few degrees drop in temp. Cooler mounting pressure is another variable.

 

The Intel TIM isn't that bad regarding quality as you say. The issue is more to do with the fact that it's in a thick layer, as a result of the depth of the adhesive.

 

Will be very interesting to see how you get on with Ultra. Keep us posted.

Share this post


Link to post
Share on other sites

How thick did you apply it Ken? 

 

The Intel TIM isn't that bad regarding quality as you say. 

 

 

 

Just a small line across the die, nothing excessive just enough to cover it and a small grain of rice on top between the H110 (same as last time).

 

I suppose its possible it could go down a little with time as it does have a 200 hour cure time and these results were with in 1 hour. But I'm going to change it regardless at this point.

 

I now agree with you, the intel TIM is not that bad after all, but i was expecting a few degrees drop. Maybe mine was applied good from the factory as I didn't see it as overly thick as some have pointed out. Anyway I will order some ultra and try it out next week. Its easy enough now to just clean the AS5 off and apply the Ultra.

 

I will update again in about a week or so when I get this changed.

Share this post


Link to post
Share on other sites

Ok guys and gals I have some results. And the results are very very impressive. 

 

I cleaned up the AS5 and put on the CL-Ultra. Just the smallest drop will do. It will spread for ever and eventually cover the die of the processor. I also put a small amount on the inside of the lid. Then just pressed it together on to the socket. Next I applied a small line of AS5 on the top of the processor and placed the H110 water pump back on. 

 

Both readings were taken with a fresh boot into windows and the quick 10 run of IBT. Intel Burn Test.    I5-4690k @ 4.6 and 1.265v

 

With AS5 for both, the package hit 91C.

 

With Ultra and AS5 on top the package hit only 62C.

 

Thats right a 29C drop. So Yeah I fully recommend this stuff. Time to raise my overclock, 5Gigs here we come!

 

I would post images but It wouldnt let me paste them, so I guess we still cant do that.  :mad:

 

Almost forgot to add, I used clear nail polish over the exposed resisters on top just as a precaution.

 

EDIT:

 

Been a week and everything is working great...  also the comment about 5 gigs, yeah I made it! But I'm thinking 4.8 for 24-7 operation. As it uses far less volts.

Share this post


Link to post
Share on other sites

Thats right a 29C drop.

Good grief, what a fantastic drop in temperature. Makes me feel like reconsidering my decision to not delidd.

 

What was your ambient temperature for the tests though? If your room was cooler for the second test, that would influence the results.

Share this post


Link to post
Share on other sites

It only took about an hour. So I know the temp didn't vary too much. As the hard work of cutting it apart was already done. I was just cleaning the AS5 off and putting new stuff on. Very happy with the results. Just be VERY patient and take your time if you decide to take this on. 

I would still only hope for a 15-20 c drop as that seems more average. Mine seemed to just go back together perfectly. I'm sure there is technique involved but maybe its a little luck too.  :wink:

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
Sign in to follow this