Hardware Exchanging thermal paste?
- Thread starter Clydefrosch
- Start date
- Views 118,493
- Replies 343
- Likes 3
More likely its just gone crusty (Nintys stuff)
If worried about the shield, dont remove it and remove copper shim, clean die, clean shim, new paste and reattach
No it doesn't... a super thin well full spread is always the best method to apply thermal paste when it's direct die contact because it's 100% sure it will cover it... With a desktop CPU it doesnt matter as the die is hidden below the copper protective plate so even if someone put jus a bit less of thermal paste in the middle it won't be too much bad even if it wouldn't cover the whole IC\cooler even though it would be a few degrees higher...
Now a super bad full spread is what can cause bubbles... Everytime I used MX 4 and applied directly to the die it was a super thin full spreaded layer where you couldn't spot anything other than a full perfect coverage and that will work super fine...
ps: What that video shows is exactly a super bad full layer which isnt even thin or evenly spread, there's way many other videos that shows proper full coverage 100% spreaded evenly that always work 100% fine and it's probably still the best for direct die things like GPU's, SOC's and laptop APU's\CPU's\GPU's with the die exposed... With Kryonaut though it might not work very well as I got very disapointed with the way it spreads, but with artic MX4 is a perfect spread... 2 days ago when I opened my GPU which I had put MX4 like a year ago, the die had a full 100% super smooth coverage without any marks, it came super perfect, though this time with the kryonaut it got all messed up as it looks likd god damn silicon glue or some crap lol...
Last edited by guily6669,
I personally believe that it is impossible to spread thinly and evenly, even if the naked eye thinks so. There will be always tiny air pockets, thus the heatsink should do the work of spreading, as it does a better job. This is my experience of using thermal paste for 15 years.
Furthermore, different consistencies in thermal pastes will require different techniques, as such spreading becomes a shore and not so sure (if you get what I mean).
So what works best for me is pea sized/cross for SoC, and cross for desktop cpus. I've managed to get temps down 15 degrees celsius on laptops. Which is where I use most to salvage old/friends laptops. I think Kryonaut is great.
However wether the issues of tiny air are as crucial as it seems, I am not sure. I just prefer to avoid that issue entirely
So if you are getting good results with Arctic silver, that's great!
Furthermore, different consistencies in thermal pastes will require different techniques, as such spreading becomes a shore and not so sure (if you get what I mean).
So what works best for me is pea sized/cross for SoC, and cross for desktop cpus. I've managed to get temps down 15 degrees celsius on laptops. Which is where I use most to salvage old/friends laptops. I think Kryonaut is great.
However wether the issues of tiny air are as crucial as it seems, I am not sure. I just prefer to avoid that issue entirely
So if you are getting good results with Arctic silver, that's great!
Well I have been using it since the 90's and have seen all sorts of videos of too much, too little, spread, X, middle ball, line, whatever and I have seen the well spread thin layer always working fine at least with the thermal paste I have always used...
Maybe Kryonaut will not be good to make a thin layer as it looks like god damn silicon glue lol...
Anyway I always lose a lot of time, but with the MX series the layer how I put it to the naked eye you can't see marks, spots, nothing, it looks like a piece of shiny paper above.
My I7 2600K for now has MX4 for like 2 or 3 years and it's on 4.9ghz on air and its also working super fine at less than 70Cº while gaming and before it had MX3 with the same treatment and basically the same temps...
On my GPU also after removing the MX4 from it, it had a perfect full layer on the die with no marks, nothing at all, it was perfect (But I must say I got a bit disapointed with it vs the MX3 since it had only 1 year and the MX4 really seems to dry faster than the MX3 like I saw some ppl commenting it, I have removed MX3 after like 3 years and was mostly still liquid, the MX4 doesn't really seem to stand up as good though there's a new version of it now which I never tried)... Now with the Kryonaut it's probably not a good idea to spread it as it's a pain in the ass and very thick, very glue like, thought it was much better to use it.
One thing I know for sure is with Artic MX series a nice full thin spread will not have air trapped and will make 100% contact with the die\IC plate all the time while going the ball method or other if you happen to put too little either the die could have fried cores or with IC covers not covering the whole surface it will always lose a bit of efficiency even though it won't matter much, but for die contact most still recommend a full thin layer spread, but I guess it depends on the thermal paste too (like I said with kryonaut I don't recommend a spread of it as its super hard to make an actual full nice layer of it and its a bit glue like so most likely will have air bubbles)...
And if the thermal paste is electrically conductive I would be careful using too much too, but I never used a electrically conductive thermal paste ever, not even before using the MX series when they shown up...
Anyway if the thermal paste is not electrical conductive you can surely just throw as much as you want with the ball method and it will work fine too and the temps will be the exact same as the minimum amount for a full coverage or the spread layer, I just feel safer with the spread method because I use the minimum amount of thermal paste and I'm always sure it has 100% full coverage (I personally use a credit card and start the full spread and in the end I do a full pass from side to side to even it and it comes super perfect, (but sadly doesn't seem to work with the Kryonaut). Also a big plus of a thin spread is it makes less mess and will be easier to clean later without much thermal paste outside the die like when using the ball method and applying a bit more than needed which will make a big mess all around even though it will work 100% fine too.
ps: All I can say is I was expecting Kryonaut to be much better to apply, got very disapointed with it and even thought I had a fake Thermal Grizzly Kryonaut
.
Maybe Kryonaut will not be good to make a thin layer as it looks like god damn silicon glue lol...
Anyway I always lose a lot of time, but with the MX series the layer how I put it to the naked eye you can't see marks, spots, nothing, it looks like a piece of shiny paper above.
My I7 2600K for now has MX4 for like 2 or 3 years and it's on 4.9ghz on air and its also working super fine at less than 70Cº while gaming and before it had MX3 with the same treatment and basically the same temps...
On my GPU also after removing the MX4 from it, it had a perfect full layer on the die with no marks, nothing at all, it was perfect (But I must say I got a bit disapointed with it vs the MX3 since it had only 1 year and the MX4 really seems to dry faster than the MX3 like I saw some ppl commenting it, I have removed MX3 after like 3 years and was mostly still liquid, the MX4 doesn't really seem to stand up as good though there's a new version of it now which I never tried)... Now with the Kryonaut it's probably not a good idea to spread it as it's a pain in the ass and very thick, very glue like, thought it was much better to use it
.One thing I know for sure is with Artic MX series a nice full thin spread will not have air trapped and will make 100% contact with the die\IC plate all the time while going the ball method or other if you happen to put too little either the die could have fried cores or with IC covers not covering the whole surface it will always lose a bit of efficiency even though it won't matter much, but for die contact most still recommend a full thin layer spread, but I guess it depends on the thermal paste too (like I said with kryonaut I don't recommend a spread of it as its super hard to make an actual full nice layer of it and its a bit glue like so most likely will have air bubbles
)...And if the thermal paste is electrically conductive I would be careful using too much too, but I never used a electrically conductive thermal paste ever, not even before using the MX series when they shown up...
Anyway if the thermal paste is not electrical conductive you can surely just throw as much as you want with the ball method and it will work fine too and the temps will be the exact same as the minimum amount for a full coverage or the spread layer, I just feel safer with the spread method because I use the minimum amount of thermal paste and I'm always sure it has 100% full coverage (I personally use a credit card and start the full spread and in the end I do a full pass from side to side to even it and it comes super perfect, (but sadly doesn't seem to work with the Kryonaut
). Also a big plus of a thin spread is it makes less mess and will be easier to clean later without much thermal paste outside the die like when using the ball method and applying a bit more than needed which will make a big mess all around even though it will work 100% fine too.ps: All I can say is I was expecting Kryonaut to be much better to apply, got very disapointed with it and even thought I had a fake Thermal Grizzly Kryonaut
.
Last edited by guily6669,
D
Deleted User
Guest
I wouldnt recommend Artic Silver 5 because its electrically capacitive even though it's not exactly conductive, but if you put a lot of it and it got around some components it may cause problems over time and also for full efficiency it needs cure time of a lot of hours, I'd recommend either Artic MX series which is very close to the Artic Silver 5 in temps but its non capacitive and not conductive and super easy to spread or the best Kryonaut which is quite better than AS5 or MX series in heat transference, though it's a crap to spread it so you should probably use the ball method in the middle and make sure it has plenty enough to cover the whole die so it doesnt have cores left without cooling...
ps: And for the hardcores off course Liquid metal which have been tested on the switch and works fine, the cooler plate isn't aluminum and it's safe to use it, but I still wouldn't recommend it on a portable console that goes shaking everywhere with you, but if you still can't lieve without the best temps out there then go for it and use a protective layer all around the die so it doesnt short anything out.
ps: And for the hardcores off course Liquid metal which have been tested on the switch and works fine, the cooler plate isn't aluminum and it's safe to use it, but I still wouldn't recommend it on a portable console that goes shaking everywhere with you, but if you still can't lieve without the best temps out there then go for it and use a protective layer all around the die so it doesnt short anything out.
Last edited by guily6669,
Arctic Ceramique 2... As good as AS5, completely inert, lasts forever. Ticks all the boxes for me.
Anything good surely will work, but I'd still recommend Kryonaut as it's the best non liquid metal thermal paste out there for now and by quite some difference, I mean the point here is really making the best heat transference and reduce the fan speed as much as possible to also get best battery life so I'd totally recommend Kryonaut, now the problem is the sucker won't probably work good with the spread method, but since it's 100% safe and needs no cure time too, just throw a good ball in the middle of the SOC die...
What I mean is really with Kryonaut being massively better, there's no point going to a lower quality thermal compound. However I don't know is over time how Kryonaut will stand up since its already a thick and bit dry thermal paste out of the box, MX2 and MX3 for example they claim 8 years and I surely have used MX2 for like 4 years and MX3 for 3 or 4 years too and they were mostly still liquid so I'd totally say they are perfect over time though MX4 disappointed me a bit too just after 1 year on my GPU (but anyway the GPU was running at like 87Cº almost all the time on the limit and the MX2 and MX3 I have never used in such high temps so it might be the high temp fault too, it's a Asus Strix RX480 which its badly designed and consumes a lot more power than the other RX480 by it self at same clocks and its flashed to RX580 OC so I run it at 1360 or 1390mhz)...
What I mean is really with Kryonaut being massively better, there's no point going to a lower quality thermal compound. However I don't know is over time how Kryonaut will stand up since its already a thick and bit dry thermal paste out of the box, MX2 and MX3 for example they claim 8 years and I surely have used MX2 for like 4 years and MX3 for 3 or 4 years too and they were mostly still liquid so I'd totally say they are perfect over time though MX4 disappointed me a bit too just after 1 year on my GPU (but anyway the GPU was running at like 87Cº almost all the time on the limit and the MX2 and MX3 I have never used in such high temps so it might be the high temp fault too, it's a Asus Strix RX480 which its badly designed and consumes a lot more power than the other RX480 by it self at same clocks and its flashed to RX580 OC so I run it at 1360 or 1390mhz)...
Last edited by guily6669,
Guys, I know I didnt read the whole thread but increasing the fan speed shouldn' t help to overheat or console lifespan like on PS3 days? Also for thermal paste replecament, I'm also looking for this but looks like we still couldnt decide true answer. Someone should post OC asnwer when decision made
Yes. I don`t have figures to hand. But definitely cooler with Ceramique 2.
--------------------- MERGED ---------------------------
My decision is madeGuys, I know I didnt read the whole thread but increasing the fan speed shouldn' t help to overheat or console lifespan like on PS3 days? Also for thermal paste replecament, I'm also looking for this but looks like we still couldnt decide true answer. Someone should post OC asnwer when decision made
Replacing the paste won’t impact temps much if at all. It’ll cause the fan to not run as fast, thus increasing battery life a bit, and it might make it so it takes longer to get warm, but that’s about it.
Kryonaut seems to be slightly better but from what I've heard any decent thermal paste will do the job (MX4 for example is just fine and performs almost as well for a much lower price)
What does impact temps, (remember - paste is just there to ensure a good thermal connection. The paste doesn`t do any cooling itself.
If you could grind aluminium powder up fine enough(talking microns) and get it to stay in the right places, that would do the job too.(That will ofcourse not happen - stupid idea - don`t try it)
What will affect the temps is CSA(cross sectional area) of copper heatpipe, the air gap in the middle, the size of the heatsink and the speed of the fan.
Its all a compromise, taking into account unit cost, portability and reliability.
Nutshell time: Don`t go replacing your compound just for the sake of it - only do it if you need to. You could be just making work for yourself.
Do it at service time or when the need arises.
I don't think that is any good... I have seen the test of a very expensive one which is supossed to have way higher heat transfer than any non liquid metal thermal paste and the temps were a bit higher on the tests...
There's GPU's that came with that stuff too and had temp problems which to fix was a pain in the ass since the cooler had to be sanded or with normal thermal paste it wouldnt make full nice contact with it, but after the temps went lower quite good...
Anyway the good thing with that is that it's supposed to last like forever so it should never lose performance and never need replacements, but who knows...
Kryonaut = 12.5W/mk
Artic MX4 = 8.5W/mk
With the switch low temps sure I don't think there will be much difference in the temps, but it's up for anyone to judge if it's worth the extra money or not since some people even don't care and go the risky way for liquid metal to get the absolute best results.
I have always used Artic MX series and loved them, I only used Kryonaut for the first time on my PC GPU a few days ago, my experience with both even though I have only a ton with MX series is that MX series are super easy to spread and work well that way, Kryonaut is hard to spread, thick kinda a bit like glue and I have no idea if it will cause air bubbles with the spread method.
Artic MX2 and MX3 also they last for so much time even after like 4 years they were always still a bit moisty\liquid 4 me, MX4 seems to dry a bit faster and Kryonaut I have no clue whatsoever for how much time will it work being effecting as a heat transfer compound...
ps: Anyway if anyone's objective is really getting the most absolute lowest temps possible without any risk like with liquid metal then Kryonaut is probably for you, if you can live with something probably very similar and less expensive I'd go with Artic MX4 and you also get the benefit of being easier to apply with the spread method if you want to go that way... If you are a newbie, stay away from Artic Silver 5 and any capacitive thermal compounds and specially don't even try using anyone which is electrically conductive or you might risk making a mess and short something...
There are quite a few old videos on youtube already with all the mods, new thermal paste direct die contact with the copper sheet removed, pads on rams and from ram metal cover to backplate, a hole in the rear case near the fan for more airflow intake, there's all sorts of stuff on youtube...
I don't think I ever seen was anyone trying to cool the 2 biggest nintendo switch faulty chips, the Video out chip and the charger controller chip, there's a gazillion switches out there which have problems on either of them, maybe I will try cooling them too when I mod mine, one of them I think it's below the main motherboard, if the case below is metal maybe if it touches it thermal paste would work super fine, I still havent opened mine so don't know the clearance between the chip and the switch case and don't even know if it's all metal even though from videos it looks like its all metal...
Yeah, I wish was that there was a cheap custom cooler with the whole backplate soldered to the copper pipe cooler as a all-in-one piece and all made of copper, that would be very handy, I wish we could fully cool the switch without even needing the fan, that would give a more battery life, but then again I guess it would require so much copper that would have been too heavy
and probably would melt on the dock and not good for OC either ...
Last edited by guily6669,
Hey all.
So I repasted my switch last night. I'm using Kryonaut, on everything (no thermal pads on hand).
The back of my console seems to get pretty warm now, but the exhaust air doesn't really feel that warm... wondering if I fucked something up? I'm not new to this kind of stuff (delidded i7 4770k 4.1 GHz stable, liquid metal, repasted quite a few GPU's, etc) but I'm not sure i got it right with the switch.
I did the following:
I play:
Any thoughts? I would've expected the air exiting to be warmer than it is. Not saying it's cool air, per say, but def not what I'd call warm, comparing to the temps I've seen reported in this thread.
So I repasted my switch last night. I'm using Kryonaut, on everything (no thermal pads on hand).
The back of my console seems to get pretty warm now, but the exhaust air doesn't really feel that warm... wondering if I fucked something up? I'm not new to this kind of stuff (delidded i7 4770k 4.1 GHz stable, liquid metal, repasted quite a few GPU's, etc) but I'm not sure i got it right with the switch.
I did the following:
- took the cover apart
- removed the RF shield thing
- removed the heatpipe for the cooler
- removed the shield that is over the SoC and RAM
- Cleaned everything good with 99% isopropyl alcohol.
- Applied thermal grizzly kryonaut to RAM and SoC (about a pea size drop on the SoC)
- reinstalled the plate over SoC and RAM (left the copper shim) hoping that would help spread paste))
- applied more kryonaut
- Installed the heatpipe cooler
- Put a nice line of paste on the heatpipe (probably about 3" - 3.5" long)
- Installed the RF shield thing
- booted.
I play:
- A shit-ton of Warframe
- Dead Cells
- Working through The Legend of Zelda: Links Awakening
- Disgaea 5
- Into the Breach
Any thoughts? I would've expected the air exiting to be warmer than it is. Not saying it's cool air, per say, but def not what I'd call warm, comparing to the temps I've seen reported in this thread.
Framerates won't be affected at all, and even temperatures of the system itself won't be affected much if at all. The Switch targets a specific temperature range and tries to keep it there by adjusting the fan speed. All replacing the paste will do is allow it to keep the target temperatures with a slower fan speed, and increase the amount of time it takes to reach the target temperature.
- No one is chatting at the moment.
-
-
-
-
@
RedColoredStars:
Never even seen a tiger crust pizza in any stores around here. Walmart, Cub, or otherwise. -
-
-
@
RedColoredStars:
Last thing I told her is how much I love her, and that Im not leaving her there forever and I promise to come back and take her back home with me.
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
