Ive made this just this Morning. Delid the 5 2400G with Liquid Metall and on top the last bit of NH-1 paste. Temps with my AIO Alphacool Eisbear LT in Aida 64 COU with GPU are at 52°C =) Better than only on DIE and on top with Kryonat. =P
@@Fincher123 "Delid the 5 2400G with Liquid Metall and on top the last bit of NH-1 paste." So you put both thermal compounds (liquid metal & NH-1) inside the IHS?
@@AmuckPVP No no no! XD And i just replaced the NT-H! with Thermal Grizzly Kryonaut. On the DIE is Liquid Metall and on top between Pump and IHS is Kryonaut. Temps are at OC 1.4250 Volt 4.00GHz and GPU 1600MHz with 1.2500Volt at 52-54°C =P Have a nice Day! =)
@@AmuckPVP if you like to do this... but it will eat in the Copper and get Dark. I Only like to use Kryonaut in contact to Copper. because the process with Nickel is much slower. What Metall is your Colling on top of the IHS?
The results are as expected and I have a few notes for you: 1. Liquid metal between IHS and Cooler is only better by a few degrees versus thermal grease. The higher the temps, the bigger the gap. 2. The x62 has a copper base which dried the liquid metal and needs re-applying. The stain has no impact on performance. Don't sand it! 3. The x62 is not powerful enough for 7700k @4.7GHz. Package temps should be around 60 with core spikes under 70. 4. You sould delid the 7700k and use liquid metal between the DIE and the IHS. Huge impact on temps! Remember to remove the adhesive. 5. Adding another 2 fans for push-pull will help with the noise. 6. Lapping the IHS will gain another 2-3 degrees.
I wouldn't use liquid metal between the IHS and the cooler, the real benefit you can gain with liquid metal is by delidding the cpu and replacing the original TIM with it, not to mention liquid metal fuses with bare copper over time.
I had this happen to my GPU waterblock, I was in a time constraint situation so I had to remove it fast so I used Rouge with a rotary tool and ended up with better temps with a regular paste on the block afterwards lol
Yeah, but it just affects the look of the copper. Go watch Gamers Nexus Liquid metal Tests. The stained copper does not cool any worse. The only thing you should absolutely not do is put liquid metal on Aluminium
@@EVPointMaster Gamers nexus only did a short term test. I used LM for 2 years. It didnt just affect the look it was literally like a thick plating of metal. LM reacts over time with copper, fusing with it and "drying" it out almost as if you plated it. Thats the long term exposure. LM isnt worth it as you have to replace it after a year. I have pictures of the "plating" effect. So this time around I went with a diamond paste and had better temps and wont have to change that paste for years
I used it for about 3 years without maintenance - 0 issues and pretty much same performance as day 1. I only removed it for loop maintenece. Same CPU block now on a 3900X and performing as exppected :)
The silver looking material on the copper cooler is the "intermetallic compound CuGa2". A. I. Ancharov 2008. His research didn't address any changes in thermal conductivity, but your cooler has had a chemical reaction to the liquid metal, aka gallium.
I've been using Conductonaut between die and IHS and between IHS and cold plate for about two years now. It will soak into the copper of the cold plate after awhile and it no longer has perfect contact. Just slap some more on once a year or so and you'll be fine.
Before watching video itself: I see you have copper base, LM leeched in to copper, "dried" and kept copper stained. It's predictable reaction, LM reacts with copper, but slowly, it needs ~3 applications before copper base will stabilize. This is why it's best to use nickel plated base. No need for several reopenings
It's weird to think that this video is a year old already. I use liquid metal on my delidded i9-9900K, which uses a custom CNC machined copper IHS as well. I tried with only liquid metal between the IHS and die, with Arctic MX-4 between the IHS and cooler and saw temps hitting 95 degrees on Prime95 SmallFFT. I replaced the MX4 with liquid metal and did the same test, which yielded maximum temps of 85 degrees. Both of the tests above were done with a Corsair H100i GTX. Once I replaced the H100i GTX with an H150i Pro RGB and used liquid metal all round, the maximum temps were 78 degrees running the same test. Liquid metal does wonders, though the effects of liquid metal are far more noticeable on CPU delids, as well as on laptops.
Your best bet is to delid and use the LM directly on the die and the bottom of the IHS. Between the IHS and your AIO use the standard paste (thermal grizzly!) to prevent the metals from reacting. I'm currently running a 7700k with this setup and i idle at around 28c, maxing out at 60c while gaming.
I've been using Arctic Silver 5 for about 15yrs now, never had a problem with it, great quality paste, rarely need to reapply, and even then taking the HSF unit off looked like it was still fine like a reapplication wasn't necessary.
I’ve always heard that you need as flat of a surface to get the best results with liquid metal. Basically the mating surfaces need to be lapped. I could still see the tooling marks on the copper surface, just a bunch of microscopic cavities that prevent solid contact and restrict heat flow. I wonder if you could have the copper plate taken to a shop that does lapping and could polish it smooth for you.
I did this exact same mod on my 7700k 3 years ago. I delidded and put Conductonaut between the die and IHS and the IHS and Kraken CPU block. About every 6 months it needs a fresh coat as the gallium ions migrate over to the copper in the cold plate. You don't need to scrape the old stuff off, just add a fresh coat like paint.
did you reapply LM to the die? On my 7700k I only applied it on the die but not to my ISH and cooler. I've noticed my temps suffering so I don't know whether to re-do it
@@richcheema I only ever applied it once between the die and IHS. It's the bare copper in the AIO cold plate that soaks up the gallium ions. The nickel coated IHS hasn't seemed to need it, but my temps were fine, so I didn't actually check. You may want to take off the IHS again and add another coat. If that's not helping, it may be your AIO. Over time, permeation reduces the amount of coolant in the loop, leading to higher temps.
You're comparing an 8 month old, degraded liquid metal application to a fresh application of non-metal paste, and even then, they're about the same (AS needs to burn in a bit, so it will probably come down that 1 degree in the next few days). You did say that the fans spin louder than they used to... so what you're seeing now is not the optimal performance that Thermal Grizzly has to offer, but it is close to the best that AS has to offer. They're only about the same after there has been significant migration of the gallium into the copper heat block. That means there is less thermal material between the IHS and the block than there should be, and the mix of metals is off (more tin and indium). That said, I personally use a high performance conventional paste between my desktop's IHS (lapped) and the heat sink (air cooled). The thermals are not so tight that I would consider it worth the effort to mess with the liquid metal. I just wrote a reply to the 2 month of liquid metal video. I'd suggest nickel plating the copper surface before using liquid metal. It's actually very easy to do and can be done in about ten minutes, including the setup time. Look how nice your CPU IHS still looks after all that time... well, the IHS is nickel plated copper! The kit I mentioned in the other post costs under $50 and will last you many, many heat blocks/heat sinks/etc.
Are we not talking about the amount of paste he just put on there? Looked like about 20x as much as is needed, when too thick it can actually insulate the chip from the cooler.
Here's the benefit of liquid metal: its surface contact between chip and heat-sink will definitely not change over time. By virtue of being a liquid it must conform to the surfaces pressing it closely together. If the PS4 had problems with thermal compounds pulling apart or degrading contact over time, liquid metal would certainly solve that.
When you delid your cpu you can get much more higher overclock and system stability. Well worth it placing liquid metal directly on top of the CPU is pretty much pointless
I've been using LM on all of my game consoles (the ones that use a copper heat sink anyway) and have had pretty amazing results. The way I figure it, it doesn't seem as though it dries out any worse than thermal grease and should be good for several years... If by chance it doesn't perform as well years down the road, those consoles etc will be far cheaper to replace in the future. You should check out the audio differences between before and after on a PS4 Pro, it vastly improves how much the fan speeds up to where it's actually tolerable under load in certain areas of games. It seems like when LM bonds with copper it only does it to a certain extent. It would be interesting if someone could come up with a way of doing an absorption test over time to see the rate between say a month and several years. I feel like the evidence will show that the absorption is only a finite amount and won't continue over a certain amount.
Liquid metal doesn't dry out, but like you said, it absorbs into the copper, so you still end up with less liquid over time and have to add more. You are right on the other stuff as well, if you keep adding liquid metal, the copper is "saturated" with LM at some point, so it will basically stop absorbing it. This happens, because the LM doesn't get very deep into the copper, it's basically just on the surface.
Great video! I hope you do one again with that paste in 8 months just to see the how it fairs between the 2. I have a 6700k at 4.6GHz@1.344v, delided. I use the paste Phanteks PH-NDC (This is sourced from Noctua, so it's the same paste as the NT-H1) in the die and the IHS. I ran it at idle, about 28c and at load 70c (Ambien temp about 20c) with a CoolerMaster Master Liquid 240ml RGB. It's one of the best pates I have ever used, even better than the MX-4. No cure time needed. Maybe you can do a comparison with that paste, vs Artic Silver vs MX-4 and whatever else you can think of. I would love to see that.
Don't be afraid to delid your CPU bro. By this time I'm sure you have grown confident enough to try this on your 7700k. On any other CPU (with a few exceptions) I would never even concider deliding a CPU. But a I7 7700k MUST BE DELIDDED! I can not stress this enough. You simply can't really do anything with a 7700k in a non-delidded state. They simply run to hot. Trust me and everyone else who is telling you. please delid this 7700k. I idle at 30 and never break 54 no matter how long i let it set at %100 YOU MUST USE LIQUID METAL between the CPU die and the IHS. This is where the problem is with this processor. You can use paste between the IHS and the cooler and you will see very similar results, so I wouldn't recommend liquid metal here.
No alcohol cleaning of the hsf. Showing any etching of the copper. Didn’t show vacuum or blowing out radiators or fans. 8 months is a long while to build a dust layer
what you should have done is reapply liquid metal to see the differences because it absorbs into copper there for you had less liquid metal. Typically after a second application that's when you get consisted long-term results.
there is a huge advantage of LM... I tried 3 different thermal pastes and multiple times, but all of them were only a bit better than the stock paste on my RTX notebook. With LM, i got much better temps on the CPU and it was the only thermal compound which was good enough so my CPU doesnt throttle. Im only a bit worried about the copper heatsinks but i ordered for this case for 80$ a completely new heatsink, so if the heatsink ever gets damaged bcause of the LM... i can easily replace it
that sounds like to me u did not have enough liquid metal.Readmore how i found liquid metal to work over 2years, 3 cpus and 1 gpu. once the galium defuses with the copper and if there not enough indium left the 2 sufaces will become detached from each other how i apply liquid metal now.... remember there is an oxide layer on both components u must rub ur way through that using firm pressure use the applicator Q-tip to firmly rub in the liquid metal, do not assume that the LM is correctly on the intended area if it covering it! how to know if the LM is properly attached to the surface's: during the rubbing of a surface if an area on the surface begins too squeak or binds noticeably on the intended areas then you'll know once all intended areas are rubbed one u will notice small chunks of contaminants that is the oxide layer u have rubbed through use the applicator Q-tip to remove most of them once removed add additional LM install cooler but then un install the cooler to check the contact area has to much or too little add liquid metal as needed if u follow these instructions u will not need to replace ur application for atleast 2 years i have a saphire radeon r7 260x with a copper base plate that did not have enough LM that after 8 months it stopped the pc from P.O.S.T.ing 2years there wasnt enough indium left over after the galium was gone so the gpu die became detached after the pc shutdown for the night
to get the correct temperatures. you need to saturate the water in the loop close to the Ambient temperature as possible first then you can test tim and at that point, you will see the real difference
I am curious, how many PCs do you have in the room? Would it be crazy for you to do custom loop? I kind of imagine video editing, rendering/compression and uploading happening on a secondary rig with a lower end gpu. Maybe higher core count, value amd cpu. Anyways just curious!
This is weird. Why not use LM between bare cpu die and Ihs? That’s the best way to do it. I have been using LM for 2years on my delidded 6700k, it’s still runny and temps have remained the same as day 1. The copper might seem dry but that’s ok - no need to worry, it’s just gallium doing it’s thing.
Wait isnt Arctic Silver conductive? Also, i believe LM is most effective in a direct die situation. If you coat your other components or something, direct die LM is about the best interface you can get, tho is also situationally advantagous. As far Nocyua fans, Redux fans with chromax corners 4lyfe. I have white corners on those grey housings and it looks dope. Just an idea.
I am curious what happens if you put very tiny dimples in both the CPU cover shield and the coolers contact surface and fill those tiny pores with something better than copper. On a related note, I would pay extra for a cpu that was shipped from the factory with a fancy kind of cover that augmented or improved the performance of a cooler. I would also pay extra for a X570 motherboard that had the cpu soldered in, as I seldom upgrade a processor on the same motherboard.
Well, typical beginners application fault here! Your thermal coupling is only as good as the weakest bridge. That stock i7 7700k has normal thermal compound between the silicone die and its heat spreader. That cpu is now 2 to 2,5 years old, that paste is dried up granting for bad thermal coupling! If you want to do it right, delide that cpu (I recommend tooth silk for that process), clean away that nasty old stuff, put some nice and fresh galinstan on the die, spread it with a cotton tab until there is a super thin film on the die (and on the die ONLY), repeat this at the center of the heat spreader on the side facing the die (again a super thin film!), use some double sided 0,1mm 3M tape to glue processor and heat spreader back together (be aware of alignment, you have only one shot at glueing back together, because the tape glues like crazy and missalignment could smear the galinstan ALL over the processor). After that do your normal cpu cooler mounting!
on a 7700k it only makes sense to use liquid metal if you delidded it. i guess the main problem here is the thermal paste between the die and the ihs. i changed it after 2 years on my 4790k to liquid metal and the difference was 40C (the paste between the die and the ihs was completely dry at that point). at the time i had to use usual thermal paste between the ihs and the cooler but i think the difference would have been higher if i had a waterblock on my ihs. i think if you want to overclock and your cpu is out of warranty its totally worth it to delid the cpu and replace the stock tim with luquid metal. from what i've seen the average temp difference is about 15-20 C. if u have a soldered cpu (like 9900k) i think its worth using liquid metal. (at least on intel). you can cook eggs on this thing so every bit of better heat transfer is worth it. on ryzen i dont think its worth using liquid metal, since the difference in temps is not that high and you cant get a lot more mhz out of it even if you use liquid metal. on ryzen your more bound to the limit of the architecture. often times you cant get more mhz out of it even if the temps are like 60C on full load.
Thanks for the video! Interesting to see, what is happening (or what not) afteer some time when using it. I put it on my GPU a few months ago - maybe I will do a video, too to see how it affects (or not) the GPU DIE.
I would have used thermal grizzly paste but yea I don't know if liquid metal would be useful to most people is for deliders and using it under the IHS and on top might have shown a larger difference. but I also think thermal grizzlys thermal grease would have shown closer results in this test.
That was pretty good. In my case, copper plate against a Ryzen 1800X (Is plated copper) and both surfaces got welded together by the reaction. I had to pull the CPU from the socket, pretty scary stuff but I learned a great lesson and will try LM again now with a 3900X The temperatures were absolutely stellar during the time the LM was there (in liquid or crystal) from its first day till the last. I decided to pull it out because I wanted to see how the LM reacted, not because the temps were bad. The lesson is, the copper is gonna form a new compound, and you need to reapply LM each time and they will not glue together anymore. And if they do, I can always unlock the CPU socket lever ever so slightly that pulling the CPU will not be as aggressive for the pins/socket. How did the LM look after 4 months? Well, the CPU was firmly "soldered" on the copper plate. I needed to use a tool to pry between both surfaces ever so slightly to move it and it popped out. It was very interesting to see both sides had a coat of crystals attached to them, both sides mirroring each other, you could see the bond was great between both surfaces at the time the CPU was installed. I sanded both surfaces and used Noctua Paste afterwards out of fear, I should of leaved the new metal compound formed by the LM on both sides and reapply LM. But I didn't know what would of happened back then. I'm ready to give it another try. About LM application, if it is thin enough it behaves as paint and will not spill at all anywhere.
Haha, yeah I had my Threadripper at 95c trying to losen the cooler and ended up needing a lot of force on a screwdriver to rip them apart, haha, “Threadripper” is now engraved on my cooler which looks cool, haha!
DOG! I'm sure that's a good boy. Even if it's a girl. They're all good boys. Sorry, what were we talking about? Thermal interface material, right. I'm personally using Arctic MX-4 and have been ever since I first had to install a CPU cooler, and it's never disappointed. The delta between different TIMs isn't so massive that I could be bothered to swap for something more premium, especially something as electrically conductive as liquid metal. Sure, it has thermal conductivity out the wazoo, but the downsides are too numerous to ignore. It's vastly more expensive than MX-4, it takes more effort to apply, it can easily short out everything you ever loved (other than the dog) if you do it wrong, it corrodes aluminium so quickly it's not even funny (okay, it's a little funny) meaning it's inherently unsuitable for all heatsinks that eschew the copper heat exchanger and it's a pain in the neck to get rid of once it's somewhere you don't want it. The other day I was idly browsing Ebay for a used 3770K (yes that would be an upgrade stop judging me) and one of the first results was a delidded, liquid metal'd specimen for a fairly reasonable price. Liquid metal under the IHS is, thermally speaking, a significant upgrade over the Colgate Intel uses, of course, but if a certain high-pitched Canadian is to be believed you have to SERVICE that stuff because it seeps. I cannot service this. You can keep that 3770K.
was that mixed with thermal paste? it seem like it on the cpu block. I have delided my 7820X and have liquid metal on the die and the cpu waterblock directly on the cpu die. have had it like that for a long time, its perfoming VERY good low temps and OC to 4.9 GHz all 8 cores with no problems.
Your closing statements are a bit misleading. Liquid metal can definitely improve thermals and give you better performance. It's not for noobs though. You seemed to play it down as if it doesn't make a significant difference, but I know from overclocking my cpu, maintaining lower temperatures definitely keeps the cpu's performance in check, especially at the edge. I'd love to apply some conductonaut to my system, although its outdated and showing it's age and probably isn't worth it, but then again, it might just help it out enough and keep it ticking a bit longer. And, if I did get some conductonaut, I would definitely consider delidding to get the most effective heat transfer. Even if you delid you can still re-attach the IHS if you would rather; using liquid metal between the die, the IHS and the cooler will still be better than not delidding. If you don't have the right tools, delidding is kinda not worth it... And it's definitely not worth it with a soldered IHS. My IHS is definitely not soldered, I've delidded other cpus in the same family.
It is not supposed to be applicated on the Heatspreader. As far as i know, liquid Metal only damages copper when there ist aluminium in contact with it. because liquid Metal contains Gallium which is insanely corrossive against base Metals, and Aluminium reacts corrossive with copper.
u should do a delidding, paint the leads with fingernail polish and straight onto the chip...=) using it with the regular heat spreader doesnt really let the lm shine
Did you delid the 7700K? If not, only between IHS and cooler, that isn't very good. That toothpaste form Intel is really bad. I have mine delid, applied LM between IHS and CPU die, arctic MX-4 between IHS and cooler (Enermax LiqFusion 360). Have stable 5Ghz with max Temp @~75°C
I would freeze in a 18 °C room lol.. 24 is comfortable Btw liquid metal is intended to be placed on the die, not the HIS. Liquid metal easily permeates copper and stains it
liquid metal does not dry out...it reacts to metals or non metals! is use some on my skylake die and on my gtx 1060 i'm curios to see is it will dry out after a few months!
lilman227 yeah, I have my laptop sitting about 2 GHz (so it’s 1070 mobile, at that speed its outbenching other peoples overclocked desktop 1070 thanks to the extra CUDA core of 2048 on the mobile vs 1920 desktop). I mean, it’s literally only power limited on that laptop too but sadly no shunts. I know my Titan Xp hits the max limit of +20% even at stock, like it’s always power limited..., If you don’t shunt mod there is basically no overclocking (pretty much always the same +125 core, maybe +500 or 600 memory) besides undervolting the voltage-frequency curve using CTRL+F in MSI Afterburner and tweaking all the data points, which is just really annoying to me when with AMD cards I just edit the BIOS and set unlimited power so I don’t have to short out electrical components with a corrosive compound cause that just sounds dumb that it’s mandatory to do no matter how you look at it. So if you’re wanting a decent bump, you basically have to shunt mod (I’ll be honest, even my electric bicycle is shunt modded though, haha! Maybe I’m biased but if I’m overclocking, I want real gains in performance, but even with this Pascal will have more steady frame times not hitting power limits so it’s even useful even if you don’t overclock the card since the boost algorithm does some extra by itself).
New subscriber here... Would you recommend repasting every now and then,maybe every month? Well I'm talking about laptops... Good contents and reviews..hooked on watching some of your vids, they are quite interesting...
None all used gallium and it is absorbed by copper just like what you saw in the vid. And as you notice it has less impact on Nickle plated IHS so using it in a Nickle plated cold plate is much more idea than copper, that's why user said if you use copper gold plate you need to reapply after three months. And after reapplication since the 1st application the liquid metal is now absorbed by the copper it is now coating with it so adding another layer will have no mo problem means will not absorb it easily just like the first application.
About Noctua fan pimping: maybe try applying some neon coloured paint to the fan blades and put some UV light in your case? It might even look cooler than RGB... :)
@@TechTalkTobi could just be luck of the draw for me, noticed my temps were getting worse over time and decided to take off and re-apply, only finding what it had become. The paste I still had leftover was in the sealed bag and it was so dried out it was like playdough which was disappointing. Ended up getting some arctic 5 and been pretty pleased, temps haven't gone beyond any higher than 15-20c from ambient. Plus I still have some old arctic silver from when I first started building (like 7 years ago) that still seems like new, so i'd say it's great for minor performance change with longevity
@@akstyle450 maybe not enough ? ... i always put a smal film on the heatspreader that you are not able to see the surface of the heatspreader - and a smal point in the middle. nothing squeeze out and the temps are every time the same and also better than with AS 5
Hands down the best thermal paste, goes in/on everything build or service. Not to say Arctic Silver is not great stuff but if you like to OC Kryonaut all the way.
sorry wont use the liquid metal stuff, as far as i know it all has gallium in it, watch some vids of what gallium does to other metals!! bugger that 2-3 deg difference ill live with
Liquid metal does do preform much better than actual thermal paste. Don't know where you got your results from?!?! if you did to apply it correctly in the first place, you would notice the difference as wel. And about your damaged copper on your cooler heat sink, why not use liquid copper instead boy?!?!
Your application is quite poor You have clumps of liquid metal You have empty spots Will you apply thermal compound this way? There is no way to remove properly applied liquid metal without polishing, which by looks of your CPU you have not done, which is clear indicator that you have not done good job at applying liquid metal
