Battery Problems


Fantastic Member
Microsoft MVP
I think anyone building a high end Rig should consider sealed water cooling.
I have mixed feelings about water cooling for two reasons, one obvious, the other not.

The first is obvious; leaks. Even with so called "sealed" systems, leaks can occur. Hoses can crack or get cut, fittings can come loose, or there can even be a manufacturing defect. The problem is, and I've seen it too many times, users are very conscientious in the beginning. And for the first year or two, they regularly check for leaks and coolant levels (for non-sealed systems). But being new, there rarely is a problem so instead of checking regularly, they slack off. That's were the trouble begins because flexible tubing hardens with age, becomes less flexible and can become brittle. Fittings can work loose and spring a leak. As the equipment ages, regular inspections should increase, not decrease.

The second reason is much less obvious. Motherboard designers and engineers intentionally cluster other heat sensitive/heat generating devices around the CPU socket so they too can take advantage of the "expected" downward firing OEM cooler. These devices include the regulator and voltage divider components and even the chipset (as many don't have their own HSF assembly). This is why I don't really like side firing 3rd party coolers either - they only blow in one sideways direction, and not down and about.

So what happens is these users implement alternative cooling solutions for their CPUs but totally neglect "case cooling" and the cooling requirements of their "system" components. All they really achieve is bragging rights for their CPU temps. In the meantime, their motherboard components are being abused by running at least very warm (if not hot) most of the time. Which, if nothing else, increases aging of those devices.

I am actually a fan (no pun intended) of today's OEM coolers. They are not the same as OEM coolers from years back. As I noted above with TIM, neither AMD or Intel (or NVIDIA) want their processors to overheat, become unstable, or worse, be damaged. So contrary to what many still believe or have been told, the OEM coolers are more than capable of providing adequate (and pretty quiet) cooling for their processors, even with mild to moderate overclocking. And again, I stress it is the case's responsibility to provide an adequate supply of cool air flowing through the case. The processor cooler need only toss that heat into the air flow. And a good case will suppress fan noise too! (Love my Fractal Design R4 :))

So I say, unless doing extreme overclocking (or your CPU does not come with a cooler) try the OEM cooler first. You may be surprised that it does the job quite well.

The popular CM Hyper 212 is only warrantied for 1 year! :( OEM coolers are warrantied for 3 full years - not just in case they fail, but to protect the processor they come with too. In the rare event a failed OEM cooler causes CPU damage, ONLY the OEM makers cover replacement of the CPU too. No aftermarket cooler warranty covers damage to or replacement of the processor!

And to that, it is also important to note that cooler temperatures does NOT automatically mean better! You can tell by the last line in my sig that I take heat very seriously! It is, no doubt, paramount that CPU temps remain comfortably within the "normal" operating range. But there is absolutely NO EVIDENCE or scientific reasoning to suggest that a CPU running at 30°C performs better, is more stable, or will have a longer lifespan than it would if running at 55°C. All that extra 25° really gets you is bragging rights. Nothing wrong with bragging rights - as long as the one bragging really understands what it is they are bragging about - and cooling of other components is not neglected.

So unless my client insists on alternative cooling (or selects a CPU that does not come with OEM cooling) I always stress to use the OEM cooler first. Get a good case and ensure it is equipped with quality (and quiet) fans. Then keep an eye on the temps. An aftermarket cooler can always be added later.

The other exception is with our "silent running" HTPC builds for PCs going into home theater setups. I almost always just go for some form of "passive" (no fan) cooling then.

As far as shipping computers with CPU coolers attached, I just don't do it any more. In fact, I don't ship anymore at all. Like William, we are local only now and either I take it to the client myself and set it up there, or I instruct and caution them how to lay the computer on its side so the HSF is sitting on the board, not hanging off it.


Hi Digerati.
I am very mindful of the points you raised, however with good brands of sealed water coolers
now I am not expecting any issues.
I am running 3 computers now with Water cooling and finding them brilliant.
I watch temps very carefully.
Maybe over time evaporation may take place but yet to see that evident.
Saying these things may happen is really being paranoid.
Of course there are lemons in every product made.
Water cooling is the next step from air cooling and I am amazed how far I can O/Clock and still
stable in tests.
I got sick of being scratched working around the cooler fins and having to remove one cooler fan
when testing different Mem sticks.
They work for me and it's my $$$$.
How no one has come up with a simple support for the large coolers amazes me.
Have you yourself tested water coolers over a long period of time or just scaremongering?

I have watched many videos of guys plumbing their own water cooling systems.
Some of them are quite dangerous and scary.
Just an accident waiting to happen.
Last edited:


Fantastic Member
Microsoft MVP
I am very mindful of the points you raised, however with good brands of sealed water coolers
now I am not expecting any issues.
Being "very mindful" is 90% of the battle. The problem is less experienced and less mindful users, and the aging effects of time.

No doubt good brands of alternative cooler solutions are likely to be less troublesome. But until Man can create perfection 100% of the time, even good brands can have samples (lemons, as you noted) that fail prematurely, or make it past factory quality assurance inspections with a manufacturing defeat. And regardless, even quality hoses and fittings can suffer abuse or become brittle or loose after several years have passed - especially after countless heat-up and cool-down (expansion/contraction) cycles.

It is not likely you, XAOS, would be negligent as clearly you are not naïve about the problems. But you also are not likely the type of user who would come to my shop to have us build you or modify/upgrade your computer either! But not everyone reading is that sort of knowledgeable user. ;)

William B

Senior Member
Personally I have found that idle temps with air are about the same with water, but when overclocking and/or gaming the water cooled CPU will generally stay cooler under extreme loads (prime95 loops or certain games) unless one uses something like a high $$ Noctua air cooler. Both air and water are good.. Both serve a purpose. AIO coolers have come a long way, and I have no issues using a stock CPU HSF with nominal computer operations, or even gaming under stock clock loads. In my opinion and I have been building for 15 years and I also overclock and game, and I believe one only needs to invest in an aftermarket cooler if they overclock. Be it air or water. As Bill mentions the stock heat sink fans do a very good job with normal computer operations.

Thanks for the very informative posts gentlemen I love reading this stuff.
Last edited:


Fantastic Member
Microsoft MVP
but when overclocking and/or gaming the water cooled CPU will generally stay cooler under extreme loads (prime95 loops or certain games)
Sure! But again, cooer does not automatically mean better. You sure don't ever want "hot". But very warm for brief intervals will not harm the CPU.

30°C "sounds" better than 50°C, but the reality is, the 50°C CPU can be expected to be just as stable, perform just as well, and have just as long a lifespan.

And that brings up another point. Prime95 and other benchmark programs are hokey at best. They are designed to abuse hardware by pushing them beyond their design limits. In spite of how they are marketed, they do NOT represent real-world scenarios. They are great in the design labs, but not in consumer scenarios - unless the consumer is solely interested in bragging rights.

Extreme loads while gaming IS a valid, real-world scenario - at least for gamers. But note the load presented by gaming is not a "steady-state" maximum load like benchmark programs present.

A benchmark program is like putting a steady tone through an audio amplifier and cranking the gain (volume) to maximum, then see if the amplifier and speaker can handle it. That is not a fair test because music (voice or instruments) is not a steady tone nor at a constant SPL (sound pressure level - volume). Music has a huge "dynamic range" which presents a rapidly changing load on the amps and speakers. This is the same with gaming.

If everything we see on our display during gaming was in constant and rapid motion with dozens and dozens of 3D "objects" always moving about in 3 dimensions all at once, that would definitely put a huge and constant ("steady-state") demand on our system resources and cooling. But objects move and stop, move and stop. Large chunks of background remain static. So the demand on system resources and cooling rapidly go up and down too. And yes, temps can rapidly rise, but they rapidly drop too. But the point is, the load and cooling demands while playing even the most demanding games is less than that presented during abusive benchmark programs. So I don't like benchmarking programs for that reason.


That Corsair article is an interesting read. I am glad they point out in the beginning the problem of motherboard damage during shipping with big, heavy coolers.

But you can clearly see Corsair's bias in the article. In the 3rd bar chart down, the commentary says, "the Hydro coolers retain their lead while two of the air coolers start to seriously buckle". "Seriously buckle"??? At 60.5°C while overclocking and pushing the CPU "power consumption up another ~110W"??? Not even.

The 4th chart does show two of the "air competitors" having problems. But (1) they slowed the fans down to 40% !!! Why would you do that if you were stress testing the CPU???

To me, all that article is demonstrating is water cooling is quieter. While I am glad to see they say, "This isn't relevant when you're running the fans at full bore...", I feel that article is more marketing "fluff" than substance.

I note EVERY gamer I know enjoys listening to the soundtrack of the game while they are playing. Many have decent computer speakers, some 5.1 surround, they crank up while playing. Others use decent headphones. In either case, fan noise is drowned out by the soundtrack anyway.