How safe is watercooling currently?

[Joe]

time for a new poll, so time to ask the obvious. How safe is watercooling a PC currently? I know back in the day it may have been pretty risky, but now... it seem to be getting more and more tried and tested. What you think?

[bigben2k]

Strangely, for all the improvements in performance, I haven't seen a lot of progress in reliability.

-Pressure testing of water blocks, as part of QA (Quality Assurance)

-advice to use hose clamps, even though they're not immediately required.

-Overheating from a failing cooling solution is now covered by built-in thermal protections.

That's about it.

We could start another thread (heck poll) on the topic of "what is curently the most likely point of failure of a water cooled system?"

[BillA]

a reading of the results indicates that experience is part of a successful installation
little different than other activities

for me reliability must include zero maintenance (other than dust removal)
that leaves the G5

[Mr. Tinker]

50% of the safety factor lies in the hands of the the manufacturers, 50% the end user.

[pHaestus]

A better question I think is how reliable is the typical wc pump compared to the typical (stock?) CPU cooler fan.

Leaks are a non-issue if you assemble the system correctly. That just leaves pump failure as a problem. The majority of modern systems will shut down and have a BIOS beep code or error message to tell you when your CPU fan is no longer working. This is still quite rare to find in a watercooled system. You sometimes see 12v pumps with the ability to plug into fan headers (so they COULD be monitored like this) but that's often correllated with poor QC (so they NEED to be monitored like this).

To me that is the only area where a watercooled box is less safe than an air cooled one.

What products fill this gap?

[BillA]

either Laing with the RPM sensor to the bios ?

my difficulty is with the fluid loss/algae potential
6 mos is not hard, 1 yr a bit problematical, don't know of 2 yr DIY systems (my target is 3 yrs)

[Brians256]

If people accepted the performance hit and used 40% ethylene glycol antifreeze, you should be able to run (as I have) a system for 2.5 years with no water change. No biological growth, filming or other deposits as yet noticed*.



* I have not disassembled said system yet since temps are fine and tubes look clear and I don't have the time to just look at a perfectly functioning system.

[BillA]

no periodic refills ?

wow if so, what is the res ? vented or sealed ?

we need a poll on max longevity w/o refilling

[TerraMex]

i went for it.
never had a spill. Test everything outside.
cliché : "problem usually between the chair and the keyboard". patience?

on the pump issue, my mcp600 is still alive (a bit noisy thou),
alot of people on this side of the atlantic have 1048's and 1046's,
I've seen a few Eheim in aquarium systems with over 10 years service, two maintenance stops (replace shaft).

we need a poll on max longevity w/o refilling

seconded. should be interesting.

[Brians256]

Originally Posted by BillA

no periodic refills ?

wow if so, what is the res ? vented or sealed ?

we need a poll on max longevity w/o refilling

Oldest system (2 years):
1QT Tupperware res with a non-airtight lid and gooped brass barbed fittings.
Cheap vinyl tubing.
Epoxied brass barbed fittings on std. chevette heatercore.
No-name pump that is supposed to be the same as a Danner Mag3 (local sales droid said it was an oem model).
Homemade aluminum WB made like XJinn's cross drilled block
Coolant is about 40% ethylene glycol antifreeze (prestone I think) and cheap distilled water

Second oldest system (18months)
1QT tupperware style res (same as above)
tygon tubing
MCP650/D4 pump
Chevette heatercore again
TDX WB
Dangerden additive available 18months ago (forget what they were stocking back then)
This one, I haven't checked as much, although temps still seem fine and tubing seems OK as well.

I don't open up my systems much, since I just like to use them. I don't have as much time to fiddle with this hobby as I did 2+ years ago.

[BillA]

large vented res
only way you can get that maintenance interval (with the crappy tubing we use)

my congrats Brian
lets see how many others figured it out
(gonna be a short list if I know this market)

[TerraMex]

i missed.
well vented res helps ... how?
dust , particles ? he has no filter.
evaporation? -> temp differencial? (way off?)

[BillA]

before I answer, anyone want to say whats going on ?

[Joe]

Why not just answer Bill?

[BillA]

just kinda curious

the culprit is tubing permeability
in a closed system the transpiration losses over time create a negative pressure leading to tubing collapse
- if typical plastic tubing is used then the makeup water from a large vented res will enable the system to run longer

there are other solutions

[Joe]

I know Silicone has that issue (Ive seen LOTS of water transpirate from a silicone based system over time), but what about Vinyl and Tygon, do they have lots of loss associated with them?

[BillA]

yup
thicker slows it, need impermeable tubing

[Joe]

What kind of tuming is impermeable?

[Brians256]

HVAC tubing is impermeable. That's about it. Nothing else really needs flexible impermeable tubing. Hehe, odd that I actually know about this. But, most other applications need either impermeable or flex but not both.

If you want cheap and impermeable, give up flex and use copper tubing.

I already knew about the issues that Bill mentions, which is one of the reasons I have a huge res in both of those system. I thought it was common knowledge. That is one of the reasons why I'm not happy with the bay res used by both DD and Swiftech: it needs to be refilled too often. I was adding more coolant every month or two, and that's just too much work for me. I turned off that system and will restart it when I have a larger res. Plus, that res is hard to keep silent, since it is horizontal. In order to prevent splashing, it has to be kept almost totally full; even minor decreases in fluid volume cause it to splash, foam and gurgle.

[BillA]

some ok polymers (the black in Chinese kits I suspect)
and co-extrusion is possible
$$$, nothing cheap available in the US that I found

hard tubing, thick bonded elastomeric couplings (see the G5) is better

[Joe]

See now this is good info to get in the Wiki!!! Someone go add it!

[Aronj66]

I had my system up for 3 years without touching it. No refills or apearent loss of coolant. Only reason I touched the setup was that the impeller on my pump broke and one day I happened to notice my temps were 15deg. higher than should have been. I looked in my bay res. and saw no movement.. I replaced the pump impeller and refilled the system with new coolant. Lately I keep changing parts too much to try for another long run, but hopefully this whole house buying thing will curb my spending for a while.

[ElMoIsEviL]

Well my server is watercooled and it runs 24/7... so I answered the first.

Uses Little River's WhiteWater with a Black Ice Xtreme and a smallish 300GPH pump. Works like a charm.. keeps is running 4GHz no problem. (P4EE 3.2GHz oc'ed to 4GHz running 24/7).

[FuRi0]

i have a waterchill xtreme kit going and the only probs i had was when i was filling
a fill port would have been nice, when i was filling it up i put presure on the side of one of the tubes and moved it out of the block
part from that runs like a charm