All in one water cpu coolers

[vincedea]

Hey guys. In general how long do these all in one water coolers last? What type of maintenance is required since a regular air cooler you just blow out the dust reapply thermal paste.

 

[DrMrLordX]

MTBF figures may be available from the manufacturer. Failure rates on AiO pumps are a big unknown at this point.

One of the touted features of an AiO unit is that it is zero maintenance. Obviously, you're going to have to get dust out of your rad/rad fans, but that's about it.

 

[BonzaiDuck]

Hey guys. In general how long do these all in one water coolers last? What type of maintenance is required since a regular air cooler you just blow out the dust reapply thermal paste.

In one option, it depends on the cooler selection. Most of them are sealed systems, so you'd only need to clear the radiator fins of dust-bunnies periodically.

With the Swiftech H240X, it offers options for expansion -- thus becoming a "custom-water" rig at that point.

The question about an MTBF ("mean time between failures") spec is only as uncertain as water-cooling veterans and their experience with pumps of the type used in AiO coolers. Somebody would have an idea.

Ultimately, it's your decision to prefer a cooling system that may last a few years or more, over heatpipe coolers which may offer less performance, when heatpipes have an MTBF spec of about a million years.

It would be interesting to hear from the WC enthusiasts about pump longevity. At least with a custom-water rig, you can replace the pump yourself in addition to setting your BIOS to enable thermal monitoring and any other safety features of a similar nature. And of course, with custom-water, you could choose dual pumps or a reservoir with dual pumps, so there's some promise of redundancy.

 

[Blain]

DIY water cooling with large aluminum reservoir tanks buried underground are the coolest.

 

[thilanliyan]

Not sure about AIO coolers but I used a custom system with a Swiftech MCP355 pump for 5+ years with no failure, and I recently sold the pump and rads off. For 1 of those years the pump was running undervolted but the rest of the time it was running at 12v. I did make sure there was a fan blowing on it to keep it cool though because I noticed it got quite hot at 12v if there was no airflow over it.

The MCP355 was replaced with a Swiftech D5 vario and I have it at the lowest setting. Hope this one lasts as long!!

I'm really not sure of the quality of the parts in AIO systems. I prefer a custom loop with high quality parts myself.

 

[BonzaiDuck]

DIY water cooling with large aluminum reservoir tanks buried underground are the coolest.

Mmmm. . . . 55F . . .

Why aluminum? Does it offer the best thermal transfer in the ground with the least corrosion? But then . . . you'd want an aluminum water-block? No?

All of those Rube Goldberg ideas may work great, but you have the complication of routing hoses into the house from the outside. No moving the computer around, either.

So it's still a question of "superb cooling" versus "better-than-adequate cooling." Especially if we're talking about a mini-mid-tower -- even a mid-tower. Or -- long umbilical cords requiring a more powerful pump. Which -- leads to other questions.

 

[Blain]

Diamond is best, silver and copper are good, but aluminum is affordable.
Bury a 12 gallon (or larger), aluminum tank in the crawl space, run lines up through the floor to a pump, then to the PC.
Copper water blocks are fine.
Use distilled water with a cooling/corrosion additive.

 

[BonzaiDuck]

Diamond is best, silver and copper are good, but aluminum is affordable.
Bury a 12 gallon (or larger), aluminum tank in the crawl space, run lines up through the floor to a pump, then to the PC.
Copper water blocks are fine.
Use distilled water with a cooling/corrosion additive.

Big investment of time and materials. On the other hand -- can't remember if it was an idea from Essence-of-War or Guskline -- to use baseboard radiator parts on the room with some number of quick-release connections for computers. It depends on your dedication to liquid-cooling and the mobility of desktops.

I saw some years back where some guy in Arizona had run copper pipe under his lawn, which he watered periodically with the ubiquitous sprinkler system. He had to use a more powerful pump, pushing water through his water-block in the basement computer. [I distinctly remember his computer-equipped office was in the basement, but homes built with basements aren't so prevalent in the southwest . . . ] Of course, where he put the computer wouldn't matter so much, unless he chose the second-floor or the attic. He still needed a more powerful pump . . .

 

[DrMrLordX]

Diamond is best, silver and copper are good, but aluminum is affordable.
Bury a 12 gallon (or larger), aluminum tank in the crawl space, run lines up through the floor to a pump, then to the PC.
Copper water blocks are fine.
Use distilled water with a cooling/corrosion additive.

Wouldn't galvanic corrision still be a potential issue? Or would the additive be enough to stop that?

 

[unixwizzard]

I've been running a Corsair H50 (or whatever they called it) in my 2nd PC pretty much nonstop since 2009 cooling a 1st gen i7 and it shows no sign of slowing down..

fwiw, ymmv of course.

 

[PhIlLy ChEeSe]

24/7?

 

[unixwizzard]

24/7?

me? almost..

SMART on that system's drive reports the Power On Time Count as 44,149 hours.. approx 5.03 years.


it's been off an hour or two here and there, plenty of reboots.. but yeah mostly 24x7

 

[aigomorla]

Not sure about AIO coolers but I used a custom system with a Swiftech MCP355 pump for 5+ years with no failure, and I recently sold the pump and rads off. For 1 of those years the pump was running undervolted but the rest of the time it was running at 12v. I did make sure there was a fan blowing on it to keep it cool though because I noticed it got quite hot at 12v if there was no airflow over it.

The MCP355 was replaced with a Swiftech D5 vario and I have it at the lowest setting. Hope this one lasts as long!!

I'm really not sure of the quality of the parts in AIO systems. I prefer a custom loop with high quality parts myself.

A d5 and even a DDC are industry raited pumps.
These pumps were throw into a vigorous test by Laing before they were released to the industry.

A D5 is used in applications like solar water heater, while DDC's are used in cars.

You can not compare these pumps to a cheap AIO pump.

Its like comparing a tyco kiddies drill to a full boated dewalt power drill.

 

[PhIlLy ChEeSe]

me? almost..

SMART on that system's drive reports the Power On Time Count as 44,149 hours.. approx 5.03 years.


it's been off an hour or two here and there, plenty of reboots.. but yeah mostly 24x7

I would never leave a water cooled unit on all the time, that's me........

 

[BonzaiDuck]

It was never my intention to be a slacker when it comes to Jury Duty, but I discovered the way to get a quick dismissal by either the prosecutors or defense attorneys when they're taking "juror resume's:" When they ask you questions related to the notion of "reasonable doubt," you discuss it in terms of "confidence level" and quantifiable probabilities. They don't like folks on juries who think in terms of probability.

How does this relate to the current topic? The discussion on water-cooling often turns on the probability of things like pump-failure, the probability that a leak will occur, etc. etc.

So, as I still plan to "get my feet wet" with custom-water-cooling during the next year, it is helpful to hear a veteran like AigoMorla give his opinion about water-pumps.

Another forum member had also given his take on the AiO coolers, suggesting the quality of AiO parts has improved -- particularly the waterblocks and pumps.

It's the nature of "uncertainty" or risk that often blows the worries about it out of proportion. And that's why that old NASA white-paper which proposed the MTBF for a heatpipe being a million years seems so profound.

Put it another way. If I could quantify the probabilities about "guilt" associated with a jury trial, and concluded that "I was 95% confident" the defendant was guilty, I wouldn't think the 5% residual is "reasonable doubt." And if there's some 5% chance that a pump will fail after X years of continuous use, that's not so significant either. Unfortunately, the only information we get would be an MTBF spec and an analysis of customer-reviews for Quality-Assurance content.

And -- no -- if you were a criminal defendant in a jury-trial -- you wouldn't want me as a juror . . . I suppose . . . .