Do AIO coolers run out of water/liquid?

Discussion in 'Overclocking & Cooling' started by Overblod, Sep 9, 2018.

  1. Overblod

    Overblod [H]Lite

    Messages:
    121
    Joined:
    Mar 28, 2018
    I have a Corsair Hydro series 240mm cooler, its about 1.5 years old, has been running great, cooling performance has not deteriorated/changed, but still, I am wondering, as custom loops use reservoirs or T-links, does AIOs run out of liquid too?
     
  2. Araxie

    Araxie [H]ardness Supreme

    Messages:
    6,143
    Joined:
    Feb 11, 2013
    no, they do not run out of liquid.. there's still a certain amount of overtime evaporation which it's in fact more related to become somewhat "thick and dry" liquid but it take really a lot of years to achieve a point where it really matters..
     
    Overblod likes this.
  3. auntjemima

    auntjemima Hand Jobs Legend

    Messages:
    4,203
    Joined:
    Mar 1, 2014
    As stated, not really. My original H100 is still running since around launch.
     
    Overblod likes this.
  4. Master_shake_

    Master_shake_ Little Bitch

    Messages:
    7,194
    Joined:
    Apr 9, 2012
    my coolit eco c240 did.

    well the o-ring at the cold plate gave up so...

    if no leaks no worries.
     
    Overblod likes this.
  5. NGX

    NGX Gawd

    Messages:
    529
    Joined:
    Oct 6, 2006
    Granted you don't take them apart and loose any coolant, and granted they don't leak, then no they won't run out.
     
    Overblod likes this.
  6. pendragon1

    pendragon1 [H]ardForum Junkie

    Messages:
    11,396
    Joined:
    Oct 7, 2000
    ^ what they said.
     
    Overblod likes this.
  7. polonyc2

    polonyc2 [H]ardForum Junkie

    Messages:
    15,111
    Joined:
    Oct 25, 2004
    how many years do they normally last for?
     
  8. pendragon1

    pendragon1 [H]ardForum Junkie

    Messages:
    11,396
    Joined:
    Oct 7, 2000
    ive got several that are between 3-7 years. the oldest, an h60, clogged before loosing any.
     
    Overblod likes this.
  9. Overblod

    Overblod [H]Lite

    Messages:
    121
    Joined:
    Mar 28, 2018
    cool! thanks everyone
     
    pendragon1 likes this.
  10. Dermen

    Dermen Limp Gawd

    Messages:
    223
    Joined:
    Oct 10, 2017
    I've heard a few stories of AIOs that are 5+ years old losing enough liquid to affect performance. There were a few posts were people added more liquid and got performance back.
     
    N4CR likes this.
  11. warderkeeju

    warderkeeju n00bie

    Messages:
    27
    Joined:
    Apr 8, 2018
    I had an H115i go bad on me -- not sure if it lost liquid or the pump failed. It was also on for nearly 24/7 since purchase across a couple systems. I might've just worn it out.

    With that being said they are designed to be sealed units where you shouldn't need to add liquid to them.
     
  12. cyberguyz

    cyberguyz Limp Gawd

    Messages:
    468
    Joined:
    Aug 28, 2014
    I haven't have it happen to me.
     
  13. IdentityCrisis

    IdentityCrisis [H]Lite

    Messages:
    104
    Joined:
    Oct 29, 2012
    I'm sure there is loss over time, but as stated, it is likely minimal, I haven't ever held onto an AIO long enough to experience that though. I've seen it mentioned as a concern many places before though.
     
  14. Aluminum

    Aluminum Limp Gawd

    Messages:
    490
    Joined:
    Sep 18, 2015
    Technically yes, but not in the useful life of the unit, or not before something else fails first.

    Short version is liquids pass through solids at some rate, you can't beat chemistry-n-physics but you can do your best to engineer around it.

    YMMV for defective or especially shitty products, keep in mind AIO at their core are a budget cost-cutting design (Al + Cu for starters) so there is that to keep in mind.
     
    DigitalGriffin and NGX like this.
  15. MyThLoSt

    MyThLoSt Limp Gawd

    Messages:
    182
    Joined:
    Dec 4, 2007
    I got a h50 from launch basically still works to this day, would be interesting to open up and see though.
     
  16. Master_shake_

    Master_shake_ Little Bitch

    Messages:
    7,194
    Joined:
    Apr 9, 2012
    when they do fail though you can make them in to their own loop this is my coolit 240mm rad and a corsair h55 waterblock pump.

    98398_3158800.jpg
     
    Burticus, FlawleZ, dvsman and 2 others like this.
  17. NGX

    NGX Gawd

    Messages:
    529
    Joined:
    Oct 6, 2006
    Proper use of zip ties right there!
     
  18. doyll

    doyll Kyle's Thermocouple is HOT

    Messages:
    968
    Joined:
    Feb 4, 2012
    Funny how so many posts with so few about how CLCs on very hot CPUs tend to die after a year or two .. because the harder they work / hotter the coolant gets the shorter their lifespan is .. almost always because of coolant loose.

    So yes, there is coolant loss .. and if you work your system really hard (high wattage CPU) with a CLC it has a good chance of going bad in 1.5-2 years. It's the reason most coolant systems have a fill cap for topping up coolant. ;)
     
  19. Overblod

    Overblod [H]Lite

    Messages:
    121
    Joined:
    Mar 28, 2018
    Currently using my aio on i7-6850k, it's overclocked to 4.2 but never exeeds 57 degrees. I was told that it's because it's lid is soldered. But now I am switching to 8086k, delided, and planning to OC it, so was trying to make sure I will be okay with old aio. Seems like it's good for now, but might need replacement in year or so.

    Another reason why custom loops need refilling, susceptible to heat, might be because they generally include GPUs too, and those run much hotter than CPUs
     
  20. Master_shake_

    Master_shake_ Little Bitch

    Messages:
    7,194
    Joined:
    Apr 9, 2012
    how do clc's work harder? all mine were constant rpm pumps so the "work" is the same cold or hot.

    hell even that corsair one is constant rpm.

    sounds like hokum to me.
     
  21. pendragon1

    pendragon1 [H]ardForum Junkie

    Messages:
    11,396
    Joined:
    Oct 7, 2000
    he is kinda right. higher sustained coolant temps, above 60c, can cause premature coolant evap and pumps to wear out. so if you are using an inadequately sized aio it could die early. I think that's part of the enermax issue they are having with their TR aios.
     
    NGX likes this.
  22. Aluminum

    Aluminum Limp Gawd

    Messages:
    490
    Joined:
    Sep 18, 2015
    Enermax problem was improper sealing/pressure of the loop according to their own analysis of failed units, their response was posted on this very forum.

    Considering the Gen1 handles a 2990WX just fine at stock boost etc I wouldn't call it undersized. If anything they are one of the few that has a proper flow rate on their pump, most of the Asetek units I've seen are fucking anemic. The fact that cooler vendors just threw in an adapter plate and relabeled existing designs was shameful IMO, though I get that AMD also did it for TR's launch because it was new and so much bigger than before.

    CLC in generally are definitely overrated as hell, way too much faith in the marketing behind them. It lets you have different case layout options for sure, but if you have the room big air is safer in all situations and even cools better more often than people seem to realize.
     
  23. MacLeod

    MacLeod [H]ardness Supreme

    Messages:
    8,061
    Joined:
    Jul 28, 2009
    I ran my old H100 for like 6 years I think. First on a heavily overclocked FX 8150 then on a heavily overclocked FX 8350. I finally retired it back in November for my H110i but it was working just as well when I took it out as it was when I first installed it around 2011.
     
    auntjemima likes this.
  24. doyll

    doyll Kyle's Thermocouple is HOT

    Messages:
    968
    Joined:
    Feb 4, 2012
    A CLC 'working hard' is the process of transferring high amounts of heat from component (CPU) to coolant and on into airflow. through radiator.

    The more heat a CPU makes the more heat / work the CLC has to do to remove that heat. Higher amounts of heat from CPU means CLC runs hotter moving hotter coolant .. and the higher coolant temp the more the coolant is to 'leak' through the hoses as well as break down the coolant additives like corrosion inhibitors. This also means more heat in the pump and pump motor which increases the wear on pump.

    Ask around with peeps using CLCs on high overclocked CPUs with lots of heat and you will find there is a significant number of them having CLC failures from about 1 year to 18 months .. usually before 2 years of hard use. "Hard use" meaning transfering high amounts of heat.

    Some last longer, some last less time. Problems are a combination of things from coolant loss to coolant additive breakdown and resulting corrosion, coolant thickening, etc. Almost all CLC failures are pump related and many of those are the result of air in the pump .. mostly caused by slow loss of coolant. If you go back and look at when H100 first came out you will find there was an extremely high number of problems with them. So many the market was flooded with 'refurbished' H100s' selling for about half what new ones were priced at. I would say you have been very lucky to have yours last as long as it has. New Enermax Liqtech TR4 started having problems in a few short months with corrosion and other issues.
    Heck, it doesn't take a rocket scientist to understand what I was referring to.
     
  25. FlawleZ

    FlawleZ Gawd

    Messages:
    573
    Joined:
    Oct 20, 2010
    So what you're saying is, we need to have someone test an array of CLCs for a period of time at higher static temperatures? Maybe even one test at say 50C constant, another 60C, and maybe one at 70C constant. Would be interesting to see a cutaway of the CLCs after say 12-18 months at each temp range to compare any differences.

    FWIW, I've been using my H100i for the last 4 years without issue. First on an overclocked X5675 and for the last 1.5 years on an overclocked 3970X. Havent noticed any degradation in performance.
     
  26. doyll

    doyll Kyle's Thermocouple is HOT

    Messages:
    968
    Joined:
    Feb 4, 2012
    That would be nice to see, but I doubt anyone is going to run 3 systems for a couple of years then open up CLCs to see what's there. I don't like CLCs. Don't like the sound of pump running, don't like paying more for no real improvement in cooling over a good air cooler in case setup to flow air properly, don't like how they are not as dependable as a good air cooler. I ran custom loops for many years, but about 7 years ago air coolers had improved to a point they do the job very well at prices way lower than any H2O system can. I know lots of users get 3-6 years, but there are also lots that only last 18-30 months.

    CLC reviews from day one have been seriously skewed to favor them over air coolers. Most cases have very poor fans, and even if they come with decent fans they rarely include enough of them. End result is airflow through case is bad and air temp into coolers when both CPU and GPU are working hard is way warmer than room .. and with air cooling every degree warmer the air into cooler is the component translates to almost the same degrees hotter component will be. So if air inside of case is 8c warmer than room the components are 8c hotter. But with a CLC radiator as intake it's using coolest possible air and give lower temps. Put half of the extra money a CLC costs over a good air cooler into a couple of good case fans and system will be as cool and almost always quieter too.
     
  27. cyberguyz

    cyberguyz Limp Gawd

    Messages:
    468
    Joined:
    Aug 28, 2014
    I'm pretty sure the CLC I am using is 100% better than any air cooler on the market for a Threadripper without being a 5 lb lump of aluminum and copper sitting on top of my cpu right smack in the middle of my case. As for a custom loop, well I know I will need to maintain it every few years so no hardship there. Maintaining my beloved compy is a labor of love.
     
  28. Master_shake_

    Master_shake_ Little Bitch

    Messages:
    7,194
    Joined:
    Apr 9, 2012
    the "hoses" clcs use are not like rubber water cooling hoses.

    they are almost all plastic and they don't allow coolant to evaporate, that is the reason the are closed loop because service is next to impossible.

    the biggest point of failure on these things is the pump.
     
  29. Zepher

    Zepher [H]ipster Replacement

    Messages:
    16,341
    Joined:
    Sep 29, 2001
    My H60 ran 24/7 for 5 years with my slightly OC'd 3770k. I sold the machine 2 years ago to a friend and then my brother acquired it from him a few months later and has been using it 24/7 as well for his media server, so the cooler is over 7 years old now.
     
  30. caddys83

    caddys83 2[H]4U

    Messages:
    3,307
    Joined:
    Feb 14, 2009
  31. doyll

    doyll Kyle's Thermocouple is HOT

    Messages:
    968
    Joined:
    Feb 4, 2012
    If you believe that I've got post holes to sell you.
    All liquid cooling loops need maintenance. Doesn't matter if they are AIO, AIO/CLC or custom loop the coolant in them is all basically the same. They all loose coolant, and after a year or 2 the anti corrosion and gunk inhibitors start to fail and corrosion starts.

    cyberguyz, what unit do you have on your TR4? Some of the TR4 units are not run of the mill CLCs mostly made by Asetek.
     
  32. pendragon1

    pendragon1 [H]ardForum Junkie

    Messages:
    11,396
    Joined:
    Oct 7, 2000
    hose type depends on brand/model and all do have some evap at high temps. low/normal temps has next to none.
     
  33. Aluminum

    Aluminum Limp Gawd

    Messages:
    490
    Joined:
    Sep 18, 2015
    Hi there, take that "plastic" and look up the permeability (water/mix at normal temp & pressure). Unless they have invented some new perfect wonder material it will have fluid loss over time. You won't "see" it because normal evaporation is much faster unless its made from bottom grade chinesium blend.

    Just because that fluid loss is not the first thing that kills most AIOs does not mean it is not happening.
     
  34. bigdogchris

    bigdogchris Wii was a Novelty

    Messages:
    17,474
    Joined:
    Feb 19, 2008
    I have a Nepton 140XL with three years of nearly continuous power on and it's still running well.
     
  35. cyberguyz

    cyberguyz Limp Gawd

    Messages:
    468
    Joined:
    Aug 28, 2014
    The majority of CLCs are asetek rebrands. Those are not designed for a TR and will leave you wanting when you start overclocking anything over 8 cores. AMD provided the asetek bracket for the TR only because there were absolutely 0 coolers available for TR when it was launched. It was intended to be just good enough for a stock TR, but never for use in an overclocked one.

    Mine is an Enermax Liqtech TR4 360 (1st generation - it is there in my sig) until I change it out for a custom loop. It is the only AIO/CLC actually developed to properly fit a TR4 that I know of. It does have a downside in that a lot of these units have been failing due to gunk /corrosion buildup after less than a year. When that happens you can either RMA it, toss it, or tear it down, clean and re-fill it - basically turning it into a custom LC. When mine goes I will tear it down, clean it, refill it using my own brand of alchemy and leave it in my closet as a backup.

    One thing that air coolers do have going for them is lack of maintenance beyond occasionally changing the TIM every few years. But their performance doesn't come near a properly designed CLC.

    To the others in the thread my CLC uses neoprene (synthetic rubber - not plastic) + a braided polyamide (PA or "Nylon") jacket on its hoses. If any osmosis takes place with that, It will take a very, very long time. Neoprene is popular for gas and cooling hoses, both bare and with braided stainless steel jackets, in cars due to their resistance to heat and low osmosis factors (you really don't want your braided stainless gas lines to be weeping gas in your car!!).
     
    Last edited: Sep 17, 2018 at 2:42 PM
  36. criccio

    criccio Fully Equipped

    Messages:
    11,201
    Joined:
    Mar 26, 2008
    Only problem i had was my H100i v2's pump died after 2 years.
     
  37. doyll

    doyll Kyle's Thermocouple is HOT

    Messages:
    968
    Joined:
    Feb 4, 2012
    Your's is an AIO, not a CLC. CLCs are factory sealed in such a way they cannot be disassembled .. well not disassembled and reassembled like yours can. While your AIO is not great it is way better built than CLCs are.

    Sadly the only good AIO I know of is Swiftech. Aphacool make some but there have been some issues and poor customer support.

    Yeah, while yours is the only AIO made for TR4, it is definitely not as good as it should be.

    And you are absolutely correct about the tubing used on CLCs versus what is used in internal combustion engines.

    Much of the time pump failures are a result of low coolant level / air bubbles causing the pump to run dry and fail.
     
  38. criccio

    criccio Fully Equipped

    Messages:
    11,201
    Joined:
    Mar 26, 2008
    Probably. I did dismantle it when my replacement arrived (another H100i v2) and while there was still liquid in it, it didn't necessarily seem like enough though i had nothing to compare it to.
     
  39. pendragon1

    pendragon1 [H]ardForum Junkie

    Messages:
    11,396
    Joined:
    Oct 7, 2000
    or improper installation that causes that bit of air to sit in the pump.
     
  40. Brian_B

    Brian_B [H]ard|Gawd

    Messages:
    1,557
    Joined:
    Mar 23, 2012
    Hmm.

    The amount of “work” done by a CLC is pretty much constant.

    Most don’t vary the pump speed. Fan speed may vary but that doesn’t really affect the work done inside the loop.

    Higher temps could have a sight effect - lower density on the liquid, but the temp deltas are so small the effect is minuscule.

    Given that CLCs don’t have an expansion volume (or rather they just rely on elasticity of the coolant lines) higher temps could also lead to higher internal pressures, and that could lead to premature seal failure. But that doesn’t really have anything to do with work being done by the CTC, the same effect would be noticed with high ambient temps or inefficient/dirty radiator. Again, the temp deltas aren’t that great, but in a sealed system with inadequate expansion that could very well end up being significant. That would be a design flaw rather than a systemic problem shared by all CLCs though - having a semi-elastic line solves the problem neatly enough.