To Water Cool or not to Water Cool

Gamer237

Limp Gawd
Joined
Jun 26, 2017
Messages
144
so heres my dilemma, currently have a kick ass system that's in need of a cooling upgrade. Right now my 8700k is running 4.8 ghz with a Cryorig r1 but to save my life this tower cooler cant keep up with the temperatures if i push the vcore up to 1.35 and the cpu clock at 5ghz. I just finished a stable stress test with OCCT and aida64 w/ avx and the temps im getting are atrocious. Im talking like 90-100 degrees fluctuating. I'm seriously debating on putting this rig under water but im also hesitant that one day im gonna walk in the house and find my PC soaked in coolant goodness.

Have any of you guys ever experienced leaks with a custom loop? If so what are the chances of loops spontaneously leaking? Im also thinking of putting some thread sealant on the compression fittings just for insurance. If you one of you guys have experienced a disastrous leak did your components die or is it still working?

I already have my cart ready and all i need is to do is press that order confirm button but I just have these nagging question in my head lol.

Hopefully some of you guys can put my head to rest haha :whistle:
 
In general leaks are uncommon.

I've been running for two years. I DID have a leak once, but that was because I did something stupid.

I have a fan switch on my case, and I accidentally left my fans off during a gaming session and didn't notice. The coolant got VERY hot (50-60C?) resulting in thermal expansion of the fittings and a slight dripping.

Nothing fried, because my coolant was not very conductive and didn't get inside anything (just dripped on a GPU backplate)

So, short of doing something stupid like this, in general, as long as you do a good leak test when you set it up, it shouldn't leak.

Are you considering doing hard tube or soft tubing? I know hard tube is all the rage these days, but it makes me nervous, as it never sits as secure in the compression fittings as soft tubing does. If you get the right sized tubing and fittings and tighten those things down, you almost cant pull them out with all your muscle. Hard tubes - on the other hand - just kind of slide off. They don't need to be as secure though, as they are typically wedged in between two fittings.

Sealant should not be necessary, at least not with soft tubing, and it may actually do more harm than good. These things are designed to seal tight. If you get other junk in there it could create a leak pathway.
 
Last edited:
I have only experienced leaks when I put something together wrong. And it usually happens within the first 15 minutes. I've had mine on for days at a time without experiencing leaks.
 
In general leaks are uncommon.

I've been running for two years. I DID have a leak once, but that was because I did something stupid.

I have a fan switch on my case, and I accidentally left my fans off during a gaming session and didn't notice. The coolant got VERY hot (50-60C?) resulting in thermal expansion of the fittings and a slight dripping.

Nothing fried, because my coolant was not very conductive and didn't get inside anything (just dripped on a GPU backplate)

So, short of doing something stupid like this, in general, as long as you do a good leak test when you set it up, it shouldn't leak.

Are you considering doing hard tube or soft tubing? I know hard tube is all the rage these days, but it makes me nervous, as it never sits as secure in the compression fittings as soft tubing does. If you get the right sized tubing and fittings and tighten those things down, you almost cant pull them out with all your muscle. Hard tubes - on the other hand - just kind of slide off. They don't need to be as secure though, as they are typically wedged in between two fittings.

Sealant should not be necessary, at least not with soft tubing, and it my actually do more harm than good. These things are designed to seal tight. If you get other junk in there it could create a leak pathway.

Definitely going soft tubing for now. Will go hard tubing maybe some time down the road but from what ive seen PETG hard tubes are prone to cracking and your right they do kind of just slip off.
 
Like the poster above, I've only ever had one leak, and it's because I didn't fully tighten a stopper on the back of my EK res. Totally my fault and could have been easily avoided. Nothing shorted, since I was still running a 24-hour leak test with plenty of paper towel under each fitting - on top of the fact that new coolant is almost nearly non-conductive - coolant will become more conductive the longer it runs in your loop.

As far as tubing choice goes, hard tubing is quite secure, but not nearly as secure as soft tubing. The real risk is whether your chassis has a lot of flex in it, and if you plan on picking it up/dropping it in the future. I've had friends that have picked up their case and the tubing cracked - now, they were running acrylic tubes rather than PETG. The latter has much more flex potential and is difficult to crack. But, PETG is also more porous than acrylic, leading to minor fluid loss over a long enough time.

I am currently running acrylic tubes with a mix of soft tubes in the unseen areas. My current system, which is used for rendering/video editing/gaming, has been going for about 2 weeks now with no issues.

As far as the question of watercooling goes, there are three things I always tell people when they ask me about going this route:

1. Don't forego component upgrades in lieu of a custom loop. A new processor/graphics card will nearly always be a better choice.
2. It's always more expensive than you think it will be. Fittings, fans, and accessories beyond the block/pump/res add up. Add 10% onto whatever budget you come up with. Order more fittings than you need. Order more coolant than you need.
3. Don't mix European hard tube sizes with American hard tube sizes (eg. 12mm vs. 7/16"). Fittings from companies like Alphacool/EKWB don't fit american tubes as well as their own European tubes. Just because something says 12mm ID does not mean it is exactly that. The difference comes down to Imperial vs. Metric and, while minor, can lead to small leaks. Soft tubing doesn't really have this problem since the tubes are stretched over barbs and then compressed.
 
Thanks for the info. I didn’t know that there’s a difference between European tubes and American made ones lol. From the looks of it, it seems like soft tubing is the most reliable route which is perfect for my current situation. Already ordered extra fittings just in case. :) can’t wait to get this project started. Wonder how long it will take alphacool to ship my stuff from Germany lol.
 
Personally I have never suffered a LC loop leak in over 15 years of using liquid cooling. I do know people that have, but those are people who have either cut corners in their leak testing or got pinholes in their soft flexible tubing (thin wall tygon hose was bad for this because it was as soft as silicone). Personally for flexible tubing I use hose with 1/8" thickness ( 3/8 - 5/8 or 10mm - 16mm ID - OD). For me while hard tube is pretty and neat, I much prefer flexible tubing because it is ... well... flexible :)

I'm not sure what your experience is with water cooling but I will be pessimistic and put this here (sorry if I sound like Cap't Obvious).

Working with fittings - compression fittings are popular & cost more than barbs. As well you have to get the ones that fit your hose right both for inside and outside diameter. As for American vs European tube sizes, hard tube is really sensitive to the OD so pick a standard and stick with it. Flexible tube is far more forgiving. Fittings like Alphacool's have ID and OS specs. Pick the ones that match your hose.

Finally test that loop outside your system before installing it!! If you don't know how to rig a power supply to run outside of a system, let us know - it is not hard.
 
As others have said, leaks are generally the result of user error. That you're worried about them is a good indicator to me that you'll be careful when doing this project. As far as leaks go, here is the most important thing you can do, IMO:

After you have assembled your loop and have it all mounted and plumbed, double check your fitting tightness, and then place paper towels under every fitting and block in your rig. When you fill with coolant for the first time, power ONLY your pump. Nothing else! This is easiest with a second power supply on hand, but if you don't have one, take the PSU out of your case and use it physically separated from where the coolant would leak.

Coolant on your components is not a disaster if they're unpowered. If you do have a leak, the spill might wet some things, but all you'll need to do is dry them out thoroughly before powering them on and they'll be fine.

The vast majority of leaks you may experience will happen the instant the loop is filled and circulating. If it doesn't leak in the first 15 minutes of circulating the coolant, it probably won't ever. If you wanna be super thorough, run the pump by itself, circulating coolant for 24hrs. If it doesn't leak in that timeframe, it won't.

Watercooling is very safe as long as you're careful, patient, and pay attention.
 
I've been doing water cooling since 05, 06? I know I had my rig water cooled for at least my last year of college, and I graduated in 07. I had two leaks in my first setup. The first was about a week in, some drops on the back of the gpu from the cpu barb connecting point, and several months later from the reservoir down onto my ide cables. In both cases it was because I did not tighten down the barbs adequately, and I was using the plastic ones that came with the Swiftech kit I had bought into, so I was worried about cracking them. In 09 I built my new computer, with the case and water cooling guts that I still use to this day (same two pumps, reservoir, and two radiators.). It's had the parts and blocks upgraded twice since then, most recently in December of last year. I switched to metal barbs with this build and I haven't had a leak since. I had always used non conductive fluids (Primochill PC ICE) but this time I went with just distilled water & a silver kill coil. Hell I even had one of the clips that tighten down the tube to the barb, the one going from my cpu out to my gpu in, disconnected for MONTHS before I noticed. If you stick with soft tubing and tighten down the barbs all the way (no tape needed) and do a test run with just the pump running, you will likely be fine.

TL;DR:
Tighten down your barns all the way, ensure the tubing is secured to the barbs, and do a test run with only the pump having power and you will be fine.
 
Built over 40 custom rigs with cpu and gpus in the loops...Never had an issue that was not caused by a faulty part (fuck you Alphacool rotary fittings, all 30 of you $15 each bastards that were no good) or user error when attempting to service the loop (only one) or moving the system and shaking the shit out of it...

To put it in perspective, I am so noise sensitive that I used to rotate my case 90 front to back while it was running for 20 mins or so (make sure you are 100% on your leak test before) as it bleeds the system so fast. Those rad/res combination I was using back then was terrible and have never had to do that once I replaced them with different models.


Take your time. Do not cheapen out too much (no you do not need $20 a piece fittings unless you want them and have the budget) and do not be afraid to go used on fittings. They are super easy to clean and its very cheap to change an O-ring..I have saved thousands by buying used good quality compression fittings over the years. I use nothing but distilled water with a silver kill coil.. Have had systems that ran with no change for 15 months and not a single bit of growth. Ask for help if you need it. You will be fine.
 
Built over 40 custom rigs with cpu and gpus in the loops...Never had an issue that was not caused by a faulty part (fuck you Alphacool rotary fittings, all 30 of you $15 each bastards that were no good) or user error when attempting to service the loop (only one) or moving the system and shaking the shit out of it...

To put it in perspective, I am so noise sensitive that I used to rotate my case 90 front to back while it was running for 20 mins or so (make sure you are 100% on your leak test before) as it bleeds the system so fast. Those rad/res combination I was using back then was terrible and have never had to do that once I replaced them with different models.


Take your time. Do not cheapen out too much (no you do not need $20 a piece fittings unless you want them and have the budget) and do not be afraid to go used on fittings. They are super easy to clean and its very cheap to change an O-ring..I have saved thousands by buying used good quality compression fittings over the years. I use nothing but distilled water with a silver kill coil.. Have had systems that ran with no change for 15 months and not a single bit of growth. Ask for help if you need it. You will be fine.
.

FYI, if you are afraid to de-lid your cpu, then spend the $money that Silicon Lottery charges for it. It is a godsend with those crappy intel TIM setups, and its a waste to go with water otherwise..You are just costing yourself 200-400Mhz depending on the quality of your chip and how much Vcore you wish to feed it. I ran a 5Ghz 24/7 3770K de-lidded with direct die water cooling with CL LP TIM and it never broke 63C under a 100% load 24/7 for weeks at a time.
 
Repost from the 2080ti owners thread but here’s the complete build. I’m using alphacool fittings. I had some angled fittings there at first but took them out due to me being uneasy about their long term reliability.
 

Attachments

  • C597A659-1C0D-4E57-A4F0-7A11C38E2103.jpeg
    C597A659-1C0D-4E57-A4F0-7A11C38E2103.jpeg
    468 KB · Views: 0
That's why I go 1/2" ID 3/4" OD tubing. Or 3/8" ID 5/8" OD. The thicker wall really helps with preventing kinks.
 
Repost from the 2080ti owners thread but here’s the complete build. I’m using alphacool fittings. I had some angled fittings there at first but took them out due to me being uneasy about their long term reliability.

I'm a little jelly. I've been wanting to order a 2080ti for some time now, but I insist on getting the FE edition, and I wasn't about to pre-order blind before the H review.

The FE's just can't be found. Best Buy claims to have them, (if you believe nowinstock listings) but then you go to their web page and you can't order them :(
 
Worst part about soft tubing - making sure it doesn’t kink on you.

I think this may have been the case traditionally with lab-style clear tygon tubing, but when I built my build with Primochill Advanced LRT tubing, it was actually quite difficult to get it to kink.

I went with 3/8"x1/2"
 
I'm a little jelly. I've been wanting to order a 2080ti for some time now, but I insist on getting the FE edition, and I wasn't about to pre-order blind before the H review.

The FE's just can't be found. Best Buy claims to have them, (if you believe nowinstock listings) but then you go to their web page and you can't order them :(

It’s okay buddy I’m sure they’ll have plenty in stock before Christmas.
 
Worst part about soft tubing - making sure it doesn’t kink on you.

Worst part about hard tubing - getting it cut to perfect length.

Worst part about WC in general - $$$$$$$$$$

Sure are right about that. Definitely plunged some good $$$$ for my loop but in the end it’s going to be part of my cooling solution for a long time.
 
Sure are right about that. Definitely plunged some good $$$$ for my loop but in the end it’s going to be part of my cooling solution for a long time.

Definitely.

The upfront cost is steep (much steeper than it looks at first once you add it all up, as all those damned little angles and fittings add up) but once that upfront cost is invested, you have a kickass cooling system that will generally last you a long time, and incremental changes (swapping GPU blocks, etc) are generally a much smaller cost.
 
What’s the maintenance cost piece?

Is it a water change every 6 months or year? Full disassembly of the system to clean? I see if you use dye it’s diffetent as well.

I’m looking at water cooling this time around as well but I am finding a lot of differing information.
 
What’s the maintenance cost piece?

Is it a water change every 6 months or year? Full disassembly of the system to clean? I see if you use dye it’s diffetent as well.

I’m looking at water cooling this time around as well but I am finding a lot of differing information.

Depends, if your doing colored coolant your maintenance intervals are going to be more frequent than people who just run distilled water and a kill coil.
 
What’s the maintenance cost piece?

Is it a water change every 6 months or year? Full disassembly of the system to clean? I see if you use dye it’s diffetent as well.

I’m looking at water cooling this time around as well but I am finding a lot of differing information.
I kept the same fluid (that I collected and put back in during upgrades) in my system for 8 years, it sure wasn't nice when it came out! Built my setup in 08, did a full rebuild in December 17 that actually saw me fully remove and clean all the components. My two pumps, two radiators, and reservoir are still in use. I'm now using distilled water and a silver kill coil over the dyed Primochill PC ICE non conductive coolant I was using before.

I think a lot has to do with your personal preference. Certainly my coolant wasn't operating at peak performance, but it still kept my overclocked 8350 and dual 560 ti's cooled pretty well. If you have a dyed system you may want to swap things out more often, and a full cleaning will be needed if you change colors. Just water & a silver coil, I'm planning on going at least another year before I think about opening my system up. You WILL have to top up the system, especially in the first six months, and there will be some evaporation.

*edit* Corrected the bold text from not to now.
 
Last edited:
Depends, if your doing colored coolant your maintenance intervals are going to be more frequent than people who just run distilled water and a kill coil.

Yeah, but adding a kill coil (or any metal ion based biocide, like PT Nuke) can have a corrosive effect. (Probably slight, but still present)

I'll quote myself here:

I ran the numbers. According to my calculations the Silver actually is what exacerbates things. According to engineering guidelines, in wet worst case environments (like I assume the inside of a loop to be) you want your ΔV to be no more than 0.15. While the ΔV between Copper and Nickel is only 0.05V, the ΔV between Silver and Copper is 0.20V and between Nickel and Silver is exactly 0.15, which leaves no safety margin.

This is certainly not as bad as aluminum which would have a ΔV to silver of between 0.55V and 0.80V depending on the type of aluminum, but it is still too close for comfort IMHO.

Higher ΔV values will lead the metal at the anode to slowly corrode and deposit on the cathode. The anode in these cases would be the copper, nickel and aluminum, the cathode, the silver. You could put a sacrificial zinc coil or something like that in there to counteract the effects, but this would likely leave a mess of zinc oxide in your loop, which wouldn't be desirable.

These are certainly marginal and PROBABLY ok, but given the history with Nickel coatings in EK blocks I decided to err on the safe side, skip the silver, and use their coolant with their added corrosion inhibitors. it might even wind up performing better since they also have a surfactant in their coolant, which you wouldn't get in just plain distilled water. With less surface tension, you make better contact with the inside of the blocks, and transfer heat more effectively.

And it's not just my research, WATERCOOL-Jakob (the WATERCOOL rep) said something similar in a different thread:

Other than that, I would advise AGAINST a simple biocide additive. First of all, bio growth is way less of a risk as many internet threads can make you believe. It happens super rarely, and only if the user made some serious mistakes beforehand. Many biocides pose significant secondary threads: silver additives, as silver kill coils, will start a galvanic corrosion against all nickel plated components in your loop (this would, essentially, peel off the nickel plating from your EWK block). ANd copper sulfate based additives can easily be so acidic that they corrode both nickel AND copper components.
Instead, I would recommend using distilled water with a corrosion inhibitor (innovatek protect and derivates, for example). All glycol based corrosion inhibitors double function as biocides, too, so you are really killing two birds with one stone.

So, in general, the best idea is straight up distilled or deionized water with a glycol based corrosion inhibitor (which also acts as biocide) but these additives are tough to find in the U.S. (at least I haven't been able to, but that said I haven't looked very hard, as I haven't been shopping for new coolant lately)

The second choice should be pre-formulated (concentrate or premix) purpose made water-cooling fluid. Avoid the stuff with shimmery microbeads though, as it is typically for show only, and will gunk up your blocks.

And if you don't have a coolant that contains a surfactant you are going to have a tougher time with air getting caught in the system and possibly higher temps.

Now, EK recommends replacing their Ekoolant (or recommended as they have since replaced it with a different line named cryofuel) annually, but I presume this is out of an overabundance of caution in case it starts building up ions and becomes conductive, in case of a leak.

That, and they also have an incentive to want to sell more coolant :p

I haven't replaced mine in two years and it is still fine. I know people make fun of Jayz2cents, but he did a video where he opened a several year old system on the same coolant, and there were no issues at all inside.
 
Last edited:
  • Like
Reactions: Gavv
like this
Nice. Thanks for the information. That makes that decision a whole lot easier.

Appreciate the information and I am sure I will have more questions.
 
I run a silver killcoil with distilled water, and never had a problem. Mostly just copper blocks though.
 
ive been running petras pt nuke- phn biocide for years without a hitch. ive never had an issue with plating coming off or any kind of corrosion... shrug. im still using the same bottle as a matter of fact. from what i understand as long as its mixed properly your fine. im looking forward to seeing some kind of corrosion inhibitor coming from watercool ;) get on the ball Jakob!
 
ive been running petras pt nuke- phn biocide for years without a hitch. ive never had an issue with plating coming off or any kind of corrosion... shrug. im still using the same bottle as a matter of fact. from what i understand as long as its mixed properly your fine. im looking forward to seeing some kind of corrosion inhibitor coming from watercool ;) get on the ball Jakob!
I miss Petras :(
 
As far as tubing choice goes, hard tubing is quite secure, but not nearly as secure as soft tubing. The real risk is whether your chassis has a lot of flex in it, and if you plan on picking it up/dropping it in the future. I've had friends that have picked up their case and the tubing cracked - now, they were running acrylic tubes rather than PETG. The latter has much more flex potential and is difficult to crack. But, PETG is also more porous than acrylic, leading to minor fluid loss over a long enough time.

Any idea how vulnerable glass tubing is to cracking compare to acrylic?
 
Any idea how vulnerable glass tubing is to cracking compare to acrylic?

Technically, acrylic is actually a kind of glass, but less prone to cracking and shattering than your typical glass. I don't even want to think about how difficult it would be just to bend and install glass tubing. I'm sure it's possible, but there would be essentially zero flex and your risk of cracks would be sky high. Unless it is for a show-build or some proof-of-concept, I cannot advise against using glass tubing strongly enough.
 
Any idea how vulnerable glass tubing is to cracking compare to acrylic?

from what ive seen and read(no personal experience) borosilicate glass is actually really durable as long as you dont try to force it. most people ive seen use it didnt try to bend it. rather they used fixed or rotary 45 and 90 fittings or they bought the pre bent pieces from ppcs. altho i have seen loops that were hand bent and they turned out really nice. i cant remember what forum i saw the last one at of course. i would reach out to the professionals and see if they will give you honest sounding answers. i know alphacool and mayhems sells glass tubing it cant hurt to hit em up.
 
Plus you an buy pre-bent glass tubing if you need it. The smoothness of these bends would go some way to reducing build up from using more fittings.

(Don't stress the build up though - it's not a big issue)
 
So, in general, the best idea is straight up distilled or deionized water with a glycol based corrosion inhibitor (which also acts as biocide) but these additives are tough to find in the U.S. (at least I haven't been able to, but that said I haven't looked very hard, as I haven't been shopping for new coolant lately)

I agree with this. Especially distilled water. It’s cheap and bulletproof.

However, glycol-based coolant with corrosion inhibitor is ~extremely~ easy to find. It’s sold everywhere for automobiles, and is in every auto part shop, convenience store, gas station, and Wal-Mart.

I won’t claim it’s universally fine for every system though, it’s not one size fits all, and all the various components in the system need to be selected with respect to compatibility.

I’ve been cheap and just run automotive antifreeze before with good luck, but I would defer to whatever your equipment manufacturers will waranty, because they do have to consider chemical compatibility.
 
I agree with this. Especially distilled water. It’s cheap and bulletproof.

However, glycol-based coolant with corrosion inhibitor is ~extremely~ easy to find. It’s sold everywhere for automobiles, and is in every auto part shop, convenience store, gas station, and Wal-Mart.

I won’t claim it’s universally fine for every system though, it’s not one size fits all, and all the various components in the system need to be selected with respect to compatibility.

I’ve been cheap and just run automotive antifreeze before with good luck, but I would defer to whatever your equipment manufacturers will waranty, because they do have to consider chemical compatibility.

Yeah, I kind of assumed an automotive antifreeze would do the trick, but I didn't know for sure. I thought it safer to look for one purposely marketed towards water cooling loops.

WATERCOOL-Jakob, you are the one who originally recommended the glycol based corrosion inhibitor additive, like that sold by Innovatek. Do you know if regular glycol based automotive antifreeze would do the trick?
 
I agree with this. Especially distilled water. It’s cheap and bulletproof.

However, glycol-based coolant with corrosion inhibitor is ~extremely~ easy to find. It’s sold everywhere for automobiles, and is in every auto part shop, convenience store, gas station, and Wal-Mart.

I won’t claim it’s universally fine for every system though, it’s not one size fits all, and all the various components in the system need to be selected with respect to compatibility.

I’ve been cheap and just run automotive antifreeze before with good luck, but I would defer to whatever your equipment manufacturers will waranty, because they do have to consider chemical compatibility.
that's what I use, the green stuff. mix it like 10% af.
 
swiftechs hyderx- is rebranded Zerex G-93 - 94 ANTIFREEZE COOLANT, ethylene glycol based. aka zerex racing fluid.

they also provide an msds sheet for it.
how to mix it, taken from their website.
For long term protection (3 years), mix (1) bottle to 0.5l (16 oz) of distilled water.
For short term protection (1 year), mix (1) bottle to 1l (32 oz) of distilled water.

http://www.swiftech.com/hydrxcoolant.aspx

koolance uses a premix solution thats propylene glycol based. they also provide an msds sheet.
https://koolance.com/index.php?route=product/category&path=58_191_191
those are the only two i can remember.

i googled "differences between ethylene and propylene gylcol"
taken from veoliawatertech- theres a pdf linked that goes into detail but this is the gist.
Water is probably the most efficent heat-transfer fluid known. ... Between the two, ethylene glycol (C2H6O2) is a better heat transfer fluid than propylene glycol (C3H8O2). Propylene glycol is less toxic and is considered when toxicity is a concern.
 
Propylene glycol is non-toxic. It's often used when you need an anti-freeze near food, It's actually also commonly used in vapes.

It needs to be buffered to not become acidic over time. Proprietary propylene glycol antifreeze blends like Dowfrost have additives to prevent it from degrading and staying anti-corrosive.

In concentrations of greater than 20% propylene glycol inhibits the growth of (but does not kill) microbes; however at very small concentrations like 1% or so both it and ethylene glycol will actually be consumed and cause growth. [link]

Typically in industrial applications a 25% ratio or greater is used.

As far as heat transfer; a 40% solution transfers heat with 90% of the efficiency of pure water, and at a 20% concentration, is 97% as efficient. [link] Ethylene glycol is a slightly better.
 
Back
Top