Why you shouldn't water cool your PC

[SolarNova]

3 minutes ago, corrado33 said:

 

Says who? Is there a video somewhere that shows a custom loop with a 240 mm rad is better than an AIO with a 360 mm rad? And at what cost? Hundreds of dollars? A decent block ALONE is at least a hundred dollars. Not to mention the radiator(s), fans, fittings, piping, bending tools, time... For what.... looks? Eh, definitely relegated to the realm of RGB. 

Custom loops tend to have better performing blocks, higher flow rates, better rads, and 'can' have better fans, all of which can reduce temps a few degrees for each thing resulting in an end result a good deal better than an AIO of the same size. But yes, far more expensive.

[corrado33]

1 minute ago, SolarNova said:

Custom loops tend to have better performing blocks, higher flow rates, better rads, and 'can' have better fans, all of which can reduce temps a few degrees for each thing resulting in an end result a good deal better than an AIO of the same size. But yes, far more expensive.

All to reach the cooling potential of a $70 air cooler? Seems cosmetic to me. 

[SolarNova]

Just now, corrado33 said:

All to reach the cooling potential of a $70 air cooler? Seems cosmetic to me. 

Well no, at that point a custom loop of the same size as an AIO would be better than an air cooler as well.

 

In regards to cosmetics, yea part of it is down to cosmetics when it comes to custom loops, i dont think any1 would argue that.

 

I mean i didn't just rebuild my system in a new case for performance ..nor did i buy an expensive glass res, or buy a vertical GPU mount for some kind of cooling gains. That was down to cosmetics.

Spoiler

 

[LambdaSystems]

Good video, but sad to see you didn't try using the same fans on all the coolers. Noctuas fans are usually viewed as among the best so I wouldn't be suprised if the Noctua cooler was best because of the difference in fan quality. However changing the AIO's fans would ofc mean that it gets even more expensive.

[Fnige]

3 hours ago, fasauceome said:

Hyper 212 master race

Blowiematron Master race

[Enderman]

47 minutes ago, corrado33 said:

Yes... by a matter of seconds. I sent mine to 5-10 seconds depending on the position of the fan. By contrast, a radiator capable of dissipating 400W of heat continuously will still need to dissipate 4.18 kJ/degree C (assuming 1L of coolant). That translates to roughly 1000 Ws, meaning your radiator will take 2.5 seconds to cool the coolant by 1 degree, assuming your fans are running on full blast. Considering my CPU drops from 75 to 40 in say... 10 seconds. I think it's going to be... much longer for a water cooled machine. To be exact, I'd say it'll take 87.5 seconds, ignoring the inefficiencies gained when the coolant temperature is close to the ambient temperature of course, and assuming your fans are running full blast constantly. 

Lol you think an AIO has 1L of coolant??

 

47 minutes ago, corrado33 said:

Says who? Is there a video somewhere that shows a custom loop with a 240 mm rad is better than an AIO with a 360 mm rad? And at what cost? Hundreds of dollars? A decent block ALONE is at least a hundred dollars. Not to mention the radiator(s), fans, fittings, piping, bending tools, time... For what.... looks? Eh, definitely relegated to the realm of RGB. 

With a custom loop you can put a 360 rad, or more, as much as you want.

That's how you get better performance and lower noise.

Also there is no need to do hardline.

 

Unlike RGB, a custom loop can look good like a work of art.

[Sniperfox47]

I really wish they had included a non-stock control of one of the Corsair AIOs with the fans from the Noctua one.

 

The OEM Corsair fans with the H110i were pretty terrible and I can only imagine the ones on the H100i aren't any better. Noctua crushes in the case of fans so it may be that the AIO is mostly fine and the fans are most of the issue.

[Dissitesuxba11s]

Really interesting how the 360 rad was "beat" (by 2 degrees) by the 240. Kinda curious on what happened there. 

[907rider]

Granted I have a custom loop so maybe this isn't the best example but you're basically saying that the NH-U12A cools better than a 360mm rad? Sorry but something is fishy here... Here's where I'm at: I have 1x 360mm rad, with 3 noctua NF-F12's... This one 360mm rad cools both my i7 3930K at 4.5ghz, pulling on average 147.3 watts, and my RTX 2070, which has a TDP of we'll say 175 watts.  Realistically its much higher as the card is overclocked to 2055mhz but we'll use stock numbers as far as tdp for it now...  

 

So assuming 175w for the GPU and 147w for the CPU, all on one 360mm rad, I should be seeing sky high temps then right?  .....actually no... max cpu temp is ~77C and max gpu temp ~58C.   I can see how an AIO might not cool as well as my custom loop but the fact that I've got 322 watts being pushed into my rad, which is more than double what you guys are putting through the 360mm AIO. Yet my somehow CPU temps are still 3 degrees lower than what was stated for your 360mm AIO?  That makes no sense at all, unless that cold plate is complete and absolute trash.

[Deli]

I don't get why a 360mm AIO performs worse than the 240mm version. I assume the pumps and fans are the same in both models. So the only difference is the size of the rad. How??

[AlexTheGreatish]

1 hour ago, Dissitesuxba11s said:

Really interesting how the 360 rad was "beat" (by 2 degrees) by the 240. Kinda curious on what happened there. 

15 minutes ago, Deli said:

I don't get why a 360mm AIO performs worse than the 240mm version. I assume the pumps and fans are the same in both models. So the only difference is the size of the rad. How??

They have a different CPU block, the one on the 240 is better (although a bit louder)

42 minutes ago, 907rider said:

Granted I have a custom loop so maybe this isn't the best example but you're basically saying that the NH-U12A cools better than a 360mm rad? Sorry but something is fishy here... Here's where I'm at: I have 1x 360mm rad, with 3 noctua NF-F12's... This one 360mm rad cools both my i7 3930K at 4.5ghz, pulling on average 147.3 watts, and my RTX 2070, which has a TDP of we'll say 175 watts.  Realistically its much higher as the card is overclocked to 2055mhz but we'll use stock numbers as far as tdp for it now...  

 

So assuming 175w for the GPU and 147w for the CPU, all on one 360mm rad, I should be seeing sky high temps then right?  .....actually no... max cpu temp is ~77C and max gpu temp ~58C.   I can see how an AIO might not cool as well as my custom loop but the fact that I've got 322 watts being pushed into my rad, which is more than double what you guys are putting through the 360mm AIO. Yet my somehow CPU temps are still 3 degrees lower than what was stated for your 360mm AIO?  That makes no sense at all, unless that cold plate is complete and absolute trash.

The bottleneck is the CPU block, not the rad.

[Benjeh]

5 hours ago, corrado33 said:

Uh oh, all the water cooling fanboys aren't going to like this....

 

Not only is JAY saying that water cooling isn't any better than air cooling, but now Linus too? Oh my goodness. 

 

Air cooling is BETTER and QUIETER. Looks like water cooling just lost the only thing it had going for it. 

 

And they didn't even use the de-facto air cooler that we all say is great (the D15). 

 

Also, what WASN'T mentioned in the video is that while the greater heat capacity of the water cooled system does in fact absorb spikes of heat slightly better, it ALSO means that it'll take longer to cool down, meaning the fans will be running at higher rpms for longer after you shut your game down. 

 

So, in summary. Water cooling is

 

• Worse than air cooling
• Louder than air cooling
• More dangerous than air cooling (dirty water + electronics = not good)
• Introduces more single points of failure than air cooling
• Significantly more expensive than air cooling
• Not as repairable/long lasting as air cooling
• Requires more maintenance than air cooling

 

Can we delegate water cooling to the likes of RGB yet? 

There's more to it than that, if you're really, REALLY pushing your systems like I do, an air cooler just won't do, I have never lost a system on water, but I have lost a motherboard and CPU overclocking on air. Linus said so himself, if you're just using your PC and not tinkering like myself, then custom loops or AIO aren't for you, But i'm yet to see an air cooled pc hit the same voltages and temps as my custom loop.

Anyway, I'd like to point something out, the cooling down takes longer aspect, personally not sure where you got that from as i've filled my rads up with boiling water before and they have cooled down to room temp within a min or so, and thats without fans blowing air over the fins. You can't just focus on the property of water alone as there are many other factors involved.

At the end of the day its a personal choice, I will always watercool, it looks good, sounds good and performs good for my needs, nothing wrong with air, but it can't meet my specific needs.

 

[Shimejii]

51 minutes ago, Deli said:

I don't get why a 360mm AIO performs worse than the 240mm version. I assume the pumps and fans are the same in both models. So the only difference is the size of the rad. How??

Theres many variables at play here that they did a Poor job at dealing with and explaining Fully. Fan speed (RPM) is pretty much the only differing factor between AIO's these days, even between 240/280 and 360 AIOS. the EVGA CLC 280 outperforms the Corsair 360 Platinum and more, but its because its Fan RPM Goes A LOT higher then most fans, and thus are a ton louder.

 

GamersNexus Does a Sound Normalized test, usually at 40 DB, or 45 DB depending on the test case. 

 

Fans can dictate whether or not an aio edges out by a few degrees over the competition. 2-3 Degrees really isnt much of a difference, and is within margin of error. 

[brandishwar]

2 hours ago, Enderman said:

With a custom loop you can put a 360 rad, or more, as much as you want.

That's how you get better performance and lower noise.

Also there is no need to do hardline.

I have three triple-120mm rads in an external radiator box, so the majority of the heat from the CPU and GPU is being taken completely out of the system, with the fans running at 7V - and I could probably turn them down further if I want, but it won't make a difference to the noise level.

[brandishwar]

3 hours ago, SolarNova said:

Well no, at that point a custom loop of the same size as an AIO would be better than an air cooler as well. 

For the simple reason that aluminum radiators are standard on AIOs, whereas copper radiators are standard for custom loops.

[wfmarcum]

The results seem pretty logical overall. I would be really interested to see the same test done in different environments, with different cases. I know someone else commented in a negative light of AIO's for SFF cases earlier, but it might be a use case that provides better results. I don't think you'll be fitting a triple rad in most SFF, but it might be a use case for the single fan. There are also some smaller cases that would accept larger rads but have less airflow over the CPU and might benefit more from an AIO over an air cooler. Would love to see more videos doing similar tests, it's cool to see independent results.

[Drak3]

25 minutes ago, brandishwar said:

For the simple reason that aluminum radiators are standard on AIOs, whereas copper radiators are standard for custom loops.

Aluminum vs copper is irrelevant to real world use. The difference between the two is margin of error.

[Dylanc1500]

4 hours ago, SlimyPython said:

Blowiematron Master race

Pfft, open bath immersion cooling master race.

 

@AlexTheGreatish I'm just kind of curious, but have you guys talked about trying out or testing submersive cooling with something like flourinert or novec?

[Luscious]

When an air cooler beats your $170 premium 360 AIO... YOUR AIO IS A POS

 

Seriously, Corsair is becoming the Walmart of PC components - trying to sell you everything short of the CPU/GPU/mobo to build a complete rig, yet incapable of excelling with any one item. Their cases, RAM, coolers, keyboards, fans etc. are either outperformed by other brands that sell for less, or are such grossly overpriced for the shoddy quality they offer that you are better off shopping elsewhere and saving you from frustration/disappointment down the road.

 

I am not a brand basher, but there's good reasons why I have blacklisted Corsair and refuse to use ANYTHING from them in my builds. My money is something they will never see.

 

BTW: Custom liquid cooling can cool things AIO and air can't, like quad GPU setups, space constrained situations and are particularly good a tackling extreme heat scenarios in high ambient locations. They can also be made sub-ambient quiet REGARDLESS of the heat load or number of fans used.

 

Also, a D5 pump (non-PWM) is rated at 50,000 hours MTBF. Translation = expect to run it 24/7 for FIVE YEARS before any mechanical failure. Run it in pairs like the real professionals do and even a pump failure won't affect your system's up time or risk killing expensive hardware. No AIO pump can touch a D5 for performance or reliability.

[xoftware]

Funny I've always been a "huge fan" of the huge fan chassis as a tradition ever since I built my first PC. (back when you get buy a GPU with only heatsink option! lol.) As far as I'm concerned, I'm still planning on using the Thermaltake V1 for my next build sometime after Ryzen gen 3. (Hard to beat when Fry's had a sale a while back for $25!) There are many sweet contenders like that these days, so maybe us "fanboys" were on to something all along. Something would really have to tempt me to change my mind, a la the Node 202.

 

It's just so satisfying to have a quality air-cooled build! I always loved touching my hand to it and feeling the cool air. Water cooling can't buy that level of peace of mind, and air cooling is usually much cheaper. This is one of those classic examples where buying parts for your PC starts to add up quickly, if not exponentially. Buying a hotter, bleeding-edge, more expensive CPU also implies more cooling budget. As a principle, target a specific tier of performance, and build comfortably to that. If you're in-between performance tiers for gaming or what-have-you, it's a lot more money for nothing honestly. Probably the biggest mistake you can make in PC building. Tech gets better and cheaper everyday, and better CPUs with better thermals/etc will come along, and you can always upgrade incrementally if your needs change.

 

Air-cooling has always been about principle to me. If you can target a CPU that works well with it, that's the ideal IMO, since CPUs are much more powerful than people seem to estimate. Game logic usually doesn't need to break much of a sweat, and GPUs are responsible for a lot of the performance. So, target your tier. It really really helps to understand how game engines/the software works.

 

HUGE FANS ARE BAD ASS TOO ...........

 

Thermaltake Core V1: https://pcpartpicker.com/builds/by_part/M8Jkcf

 

Silverstone Sugo SG05: https://www.newegg.com/Product/Product.aspx?Item=9SIA62V86N8842

Coolermaster Elite 110: https://www.newegg.com/Product/Product.aspx?item=N82E16811119299

 

[Giganthrax]

6 hours ago, LambdaSystems said:

Good video, but sad to see you didn't try using the same fans on all the coolers. Noctuas fans are usually viewed as among the best so I wouldn't be suprised if the Noctua cooler was best because of the difference in fan quality. However changing the AIO's fans would ofc mean that it gets even more expensive.

It's stock vs stock comparison. Why would he upgrade the Corsair AIOs with Noctua fans, and leave Noctua air coolers on stock? Also, then you have to count in the price of Noctua fans for the AIOs.

[BigDamn]

This video is certainly not LTT's best work. It appears they grabbed whatever coolers they had laying around and went to work. Where's the Corsair H115? Or more importantly the Corsair H115 Platinum (using CoolIt's pump design instead of Asetek's)? Correct me if I'm wrong but I believe the H100i V2 has been discontinued, why not more recent hardware?

 

I'm also concerned about the 150i performing WORSE than the H100i V2, which Alex claims is due to the CPU block being worse. Why can't I seem to find information that would prove that? From my research, not only does the H150i perform better than the H100i V2, the H150i should also be able to keep the 9700k under 80c at 5ghz (closer to 70c). This reminds me of when I watched Linus's review of the old H100i (that I've owned for several years) on the 4790k, where he stated it would top out at 90c with a mild OC. On my 4790k? 80c at 4.8ghz (don't remember the voltage). On my Ryzen 1700? Same cooler, 60c at 4ghz 1.4125v. Granted, my H100i is the CoolIt pump.

 

Lastly, why not try exchanging fans? Although Corsair fans are of the best, I would (reluctantly) argue that Noctua is THE best. This is where Noctua handily beats all other CPU cooler manufacturers. With that said, I own a set of a Corsair High Perf 120mm fans that noticeably improve thermals on my H100i (but they're loud!). This obviously wouldn't demonstrate the most cost effective way to cool your CPU, rather, each coolers performance ceiling.

[suchamoneypit]

When I first got into computers I thought AIOs were the coolest thing and I had to get one. I ran several on my PCs I built, and now I find myself actually really liking the look of a good looking air cooler instead. Surpising performance comparison. Didn't expect the 360 rad to perform so poorly.

[dfsdfgfkjsefoiqzemnd]

49 minutes ago, BigDamn said:

Lastly, why not try exchanging fans?

Perhaps because the AIOs already are more expensive and replacing the fans would only increase the price gap?  

 

The Noctua is $100 with 2 Noctua fans. If you take the AIOs and replace the fans (at $25 each), you're looking at around $170 for the H100i and $230 for the H150i (rough calculation based on local prices here in Belgium, so please don't shoot me if I'm $10 off whatever the price is in your local store).  That's not even close to the same price range anymore. 

[strangersound]

The writers did an amazing job on this video. Genius.