Does a custom loop need to be serviced every year?

x509

2[H]4U
Joined
Sep 20, 2009
Messages
2,630
If I do a custom water-cooling loop, do I need to drain the coolant and replace the hoses every year? I'm planning my next build now, which means I'll actually buy parts problem in December or January. Or maybe February.

Years ago when I thought I would do water-cooling I got a Corsair 800D case. It's a great case for watercooling. But "stuff happened," so I never did install watercooling back then. I want to reuse the case and 850 W PS so I have more $$$ to spend on the CPU and GPU.

My new build will probably be an ASUS board, with either an Intel i9 or an AMD Ryzen 3000 series CPU, I haven't decided yet. and I want overclock the CPU a lot for better Photoshop performance. I'm not a gamer, so GPU overclocking isn't important.

If I overclock the CPU, how important is it to watercool the chipset or VR modules?

If I build a custom loop, how important is annual maintenance. AIO watercooling units for the CPU seem to be sealed units that don't require annual maintenance.

x509
 
there's no general rule for custom loops because by their nature, they are all custom built.

maintenance has to do with many factors.

1. Open to environment.
2. Liquid used
3. Metals and materials in contact with the liquid
4. Loss of liquid even if sealed
5. The longevity of components that may be degrading.


Some systems need to be maintained every few months. Some every year.. Some wont need any servicing until the pump dies.

people dont generally water cool the vrm on the motherboard, they tend to have big beefy heatsinks on them these days for any boards you'd be looking to use with an OC setup.

In general you wont be able to overclock any current generation cpu's much ...at least not enough to really matter in real world tasks. So I'd throw that idea away off the bat.

What water cooling will allow you to do that air cooling will be hard pressed to, is allow you to peg the cpu at it's max clock rate for longer and more consistently. Or allow you to run the cpu at a lower audible volume for a given workload.


Heatsink paste (common to watercooling and air cooling) sometimes does have a lifespan of 1 year before it begins to degrade. That's something you want to research a bit before settling on.
 
Last edited:
Depends on the loop and what all parts you use.
I had a system over 5 years nothing done to it other then clean the dust.
I don't want to go into details. But distilled water should be flushed and replaced 6 months to a year. Or that's my thoughts
 
Last edited:
Do you mean "flushed?"

I thought that distilled water is the "purest" coolant. No? What kind of coolant do you use? How much longer do you expect to run this system before rebuilding your loop, or just replacing the coolant?
 
distilled water that contains no additives is only pure at the moment just before you add it to your system. Afterwards, it begins dissolving ions of copper or aluminum depending on the material in your system and organic material that existed in the system will begin to grow as well.

Additives only inhibit this process. Additives only counter this for as long as there is still un-bound additives left to keep the ions / biologicals from accumulating on their own. So any water based setup is going to have a lifetime that depends on the lifespan and type of additives. Pure water will become bad the fastest, followed by additives that only deal with corrosion but not biological buildup.

AIO and certain commercial coolants are not fully or in some cases, at all, water based. These will last many years in most cases. The expense is that they tend to be thicker than water so are harder to pump, and nothing cools as good as water does. But we're dealing with a difference in cooling that doesn't really make a difference in pc water cooling use cases.

I use 50/50 antifreeze meant for modern GM engines. But my system is fully sealed, so i dont have to deal with the scent of antifreeze or anything. Should last pretty much indefinitely
 
Antifreeze! Bingo!!! How much harder does the pump have to work with the antifreeze compared with distilled water?

Just wondering why I have never, ever heard until now of using antifreeze in a water cooling loop. Seems like an obvious thing to do.
 
Antifreeze! Bingo!!! How much harder does the pump have to work with the antifreeze compared with distilled water?

Just wondering why I have never, ever heard until now of using antifreeze in a water cooling loop. Seems like an obvious thing to do.
Wile antifreeze was used regularly back in the day. I don't recommend it.
I like the stuff koolance makes. It's like anti frezz but made for computers and dose not brake down over time like most coolants. I seen their coolant last 10+ years.

Wile antifreeze and antifreeze like coolants are thicker the difference is very little. And you have the added benefits of keeping the pump lubed and o-rings soft. Water can age o-rings faster and cuse them to crack and stuff sooner then some premixed coolants.
 
Auto coolants or antifreeze can last decades or more, and that is in a damn engine not a pc watercooling loop.
 
whether you use anti-freeze or a stripped down version sold as pc water coolant, the end result wont be very different performance wise. The main difference would be color and smell. Lots of people like to add colors and stuff to their liquid, and lots of people have open systems (such as having a reservoir) ... So that tilts the balance in favor of that stuff over antifreeze ...which tend to already have a color and a noticeable soapy/sweet smell.

Use whatever floats your boat. But in general, the more your liquid is h20, the more often you will have to flush it out.

edit: somewhat good to reference https://www.overclockers.com/pc-water-coolant-chemistry-part-ii/

not a whole lot of data to go by in terms of tests and charts. But it's not like you're going to find any long-term testing being done by anyone over the course of a year or two for this from anyone. You just gotta go by who's word you trust the most.

I will say that the AIO's liquid tends to never be factor in failure...but instead, the cheap chinese made pumps are the source of failure (electrical, not impeller related). And they all use glycol based coolants like antifreeze and most commercial pc coolants are.
 
Last edited:
Antifreeze! Bingo!!! How much harder does the pump have to work with the antifreeze compared with distilled water?

Just wondering why I have never, ever heard until now of using antifreeze in a water cooling loop. Seems like an obvious thing to do.
I use 10-15% af in distilled water.
 
I used 50/50 in my custom loop, same as my car, for years. It ran for 5 years solid with no performance issues. My tubing did cloud up, but that was because I had cheap PVC from Home Depot - it didn't affect the blocks or rad or performance though. That was with a Reserator (big external reservoir) that has a small vent hole, so it was not entirely sealed.

I had an open system running distilled water once that would need flushed out every few months, because stuff would start to grow in it. Even though I used clean water, it was open on the reservoir, so stuff would get in eventually.

I tried some UV dyed Koolance coolant once - that stuff plated out my entire setup after about 9 months and did impact performance. It was a pain in the rear cleaning everything back out. Never again.
 
Silvercoil helps.

Once a year doesn't hurt but isn't always needed. Just check the system once in awhile.
 
When I make a loop for someone whose pc illiterate like my nephew last year, I use auto coolant, import variety. I told him to take care of it and flush it once a year. He obviously didn't in two years lol, so when after much begging, I rebuilt it this time with auto coolant.
 
I just built a new machine and recycled most of my cooling loop, which had been in service with no maintenance (except dusting) for at least 4yrs. I did add a bit of water at least 1 time, but I had a large reservoir and despite observable evaporation top offs were not a regular occurrence.
I use pure distilled with some silver kill-coils, and an all copper/brass loop. When I took apart the block (Swiftech Apogee XL), there was no buildup or gunk but the copper parts exposed to water were black and discolored. Other then that however, there were no ill effects and my cooling performance had remained remarkably stable over the life of the loop.

I cleaned off the block (black staining could not be removed) and put it all back together with some fresh tubing in my new machine. I'm still using my original Lian DDC pump....that thing has to be at least 10yrs old.
 
When my pump died a few months ago and swapped it out I drained my loop. It was pretty damn clean still after 8 months with the EK coolant I brought.
 
I tried some UV dyed Koolance coolant once - that stuff plated out my entire setup after about 9 months and did impact performance. It was a pain in the rear cleaning everything back out. Never again.
what coolant was it ? I recall they had one that reacted with one type of tubbing way back in the day.
But as far as I know koolacne never sold dye.
 
Last edited:
I have had ek pastel white in a system for 2 years. I will let you know how it goes when I flush the loop this weekend.:D
 
what coolant was it ? I recall they had one that reacted with one type of tubbing way back in the day.
But as far as I know koolacne never sold dye.

It was many many years ago and I have had way to many beers to recall. You may be right and it was otherly-branded and I purchased it through their store? I bought all my WC components for that build directly through Koolance. That was maybe 2008-2009?
 
I've been using Koolance Liq-702 coolant for a kajillion years now, and I love the stuff. The only time I ever had any problem with it was way back when you had to mix it with distilled water. I had some growth from a system I mixed in way too much distilled water and that was a mess to clean up. Nowadays it comes premixed in 750ml bottles and in a variety of colors (the only one of which I have never used is red). Heck, I've even mixed different colors of the Koolance fluid without issue.

I have used aluminum, brass, and nickel-plated and gold-plated copper hardware all in the same system with Koolance fluid and have had no issues. Koolance used to sell aluminum radiators exclusively and all of their blocks for use with them were gold- or nickel-plated copper, so their fluid works pretty well with mixed metals. I would still avoid mixing bare copper and aluminum in the same system, though. I rarely flush my system, though I will add fresh coolant from time to time as I swap in and out parts (I have QDCs for this). Sometimes I will disassemble a block (again, QDCs) just to make sure the fluid is working well (no gunking, discoloration, or funky smells) and have never had any problems with a working system. I did have an old GPU block I was selling that still had fluid in it after a year of sitting on a shelf (soft tube kept it sealed) and THAT smelled funky and needed to be cleaned before I let it go, but even so, cleanup was quick and easy - little growth, mostly just stank.

Soft tubing gets a bit cloudy over time, but that happens regardless of what you're using. PETG is known to also do this, but it tends to take a LOT longer. I use glass tubing to hook up the pump/reservoir to the radiators and a custom manifold because glass tubing does not stain. I have QDCs and soft tubing runs to each my GPU and CPU separately from the manifold making for super easy upgrades. Admittedly dusty pictures of my rig (#1 in sig) are in our own Watercooling sticky post here on the [H]. I just replaced the Crosshair VI Hero you see in the picture with a Radeon RX Vega 64 with a full cover block, a Crosshair VI Extreme and a Ryzen R9 3900x. It took me less than 2 hours from shutdown to boot up and all I did for the cooling loop was to top off the reservoir a little bit because the video block started out empty. Some people like the look and other's don't (the 'rigged' parts don't look rigged with the tinted glass door panel closed), but I get the best of both worlds of hard and soft tubing... and I like it. :)
 
Last edited:
  • Like
Reactions: drklu
like this
I haven't modified my system in the last 4 years. I haven't drained it, and there is no sign of anything growing in it. It's sealed for the most part, and I've only had to top it off twice. Silver killcoil in the reservoir, distilled water from the grocery market.
 
there's no general rule for custom loops because by their nature, they are all custom built.
[snipped]

Heatsink paste (common to watercooling and air cooling) sometimes does have a lifespan of 1 year before it begins to degrade. That's something you want to research a bit before settling on.

Does this apply just to the CPU or also to the GPU and VRM and chipset heatsinks?
 
Does this apply just to the CPU or also to the GPU and VRM and chipset heatsinks?

Hmm, it depends. There are some TIMs that dry out and lose performance. There are TIMs that don't dry out, don't lose performance years down the road. Be careful of blanket statements. That said, even if it dried out it doesn't matter much, you'd only lose a few degrees of performance. And regarding MBs, that also depends on what the OEM uses, however most do not use anything that needs to be serviced in a year. In other words, they are not intended to be serviced by the end user regardless.
 
To add to what TheSmokingMan said above, most VRMs and Video RAM chips and use thermal pads which never wear out - the downside to thermal pads is that as a general rule they just aren't as thermally effective as paste. I wouldn't sweat it very much as even dried thermal paste is plenty good unless you are pulling the cooler off and trying to reseat it on dry paste - THAT is VERY bad :) In general, if you pull the cooler off of the paste, you re-paste.
 
Typical antifreeze has half the heat capacity of water. Meaning 1 lbm of water changes 1F is equal to 1 BTU of energy. With 100% antifreeze 1 lbm of solution with 1F change in temperature would only remove .5BTU or half the heat removing performance. A 50/50 solution would be 75% as effective in removing heat per 1F. Using a lower ratio of Antifreeze as in 75% water 25% antifreeze should give better performance per 1F change and still give the corrosion resistance and biological protection. I just use pure distilled water and an additive, change out once a year.
 
50/50 is 80% the specific heat of water at room temp+ , that's not taking into account the other additives that change surface tension etc that it has that you'd have to add to water (like the biological and electrical inhibitors) and alter the water's specific heat you're comparing to.

This is compensated easily in our systems by just circulating the water more than absolutely necessary compared to straight water and/or having larger than necessary radiators. Which pretty much everyone does to the point of overkill anyway. So in practice, most people will not see little to no performance decrease in the temps of a system with antifreeze vs water if you're looking at just temps. But there will be no need to flush the system for half a decade if not more in a sealed setup. Your tradeoff is going to be needing to run a beefier pump or the pump at a higher rpm than absolutely necessary and/or requiring a larger radiator than absolutely necessary.

So unless you're looking to ride the minimal configuration needed, you should be perfectly fine picking whatever floats your boat and not have to worry about running hotter than you potentially could ...at least not anything above the low single digit percentages. Other factors will play a more important role. How easy is it to flush your potential system, is it sealed or not, how lazy are you... etc
 
  • Like
Reactions: x509
like this
If I do a custom water-cooling loop, do I need to drain the coolant and replace the hoses every year? I'm planning my next build now, which means I'll actually buy parts problem in December or January. Or maybe February.

Years ago when I thought I would do water-cooling I got a Corsair 800D case. It's a great case for watercooling. But "stuff happened," so I never did install watercooling back then. I want to reuse the case and 850 W PS so I have more $$$ to spend on the CPU and GPU.

My new build will probably be an ASUS board, with either an Intel i9 or an AMD Ryzen 3000 series CPU, I haven't decided yet. and I want overclock the CPU a lot for better Photoshop performance. I'm not a gamer, so GPU overclocking isn't important.

If I overclock the CPU, how important is it to watercool the chipset or VR modules?

If I build a custom loop, how important is annual maintenance. AIO watercooling units for the CPU seem to be sealed units that don't require annual maintenance.

x509
You should probably replace the coolant every year or two, especially if you use straight water. The tubes should be fine for several years if they're reasonably good quality tubing, unless you get some nasty algae grown in there or something.

Bilogical contamination can be mostly mitigated by putting a few drops of copper sulfate solution or some other biocide. You can also use a little piece of silver in there instead. The same people who sell water cooling hardware generally sell both options. Don't do both.

You shouldn't need to water cool the VRM or chipset unless they're woefully under-cooled from the factory. You should be fine with most reasonably modern boards.

I haven't found the maintenance to be all that onerous. If all you're doing is changing the coolant, which is literally just distilled water in my case, it takes maybe half an hour. Plan your install to make maintenance easier with drain and fill ports and it really ain't no thang.
 
"Nucleoxic," over 8 years on distilled water, 24x7 usage (seriously, this thing is *never* down if the power is up). 3.2 GHz 6-core 3930K Core i7 overclocked full-time to 4 GHz (no SpeedStep BS). No kill coil. No bio treatment or dyes whatsoever. Just add more distilled water as needed. I set out to build a 10-year build and it's still kickin' ass!

P.S. And, yes, people *do* cool their VRMs, to allow higher CPU voltage frequencies to support higher and more stable overclocks. Waterblocks on RAM, NB, and SB, too! 840mm of cooling, baby! - mBrane

nucleoxic (cutout).jpg
 
Last edited:
Back
Top