My watercooling days are over...

The alternative to direct air cooling

Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee

chylld
Posts: 1413
Joined: Fri Oct 10, 2003 4:45 pm
Location: Sydney, Australia
Contact:

My watercooling days are over...

Post by chylld » Mon Oct 15, 2007 4:20 am

Well, it's been a good three and a half years, and my watercooling system has finally carked it.

I found it funny that my computer was off when I got home. I turned it back on, but a few seconds into windows it turned itself off again. I pondered for a bit and then checked the water flow - there was none! :shock:

The pump was still vibrating, which meant that either the pump's impeller had broken, or there was a blockage somewhere. closer inspection revealed a disgusting 6-8 inch mass in the tube leading up to the cpu block. I've cleaned bathroom downpipes of 5 years of waste build-up, and this is almost on that level.

I can only surmise that a layer of gunk must have slowly built up on the inside of the tubing, and just recently decided to separate from the tubing and collapse into a wrinkly folded heap and jammed up the water flow path. It is absolutely disgusting to look at, and I am not even going to entertain the thought of clearing it, especially when it's fallen into the cpu block itself.

That it took nearly an hour just to negotiate the tubing and waterblocks enough to leverage them out of the way to put the original heatsinks back on is enough for me to consider watercooling as an interesting hobby that in my case at least is best left as such. Granted 3.5 years is a long time for 24/7 operation, but given that computer parts eventually do fail, I would certainly not like to face the inconvenience I'm facing now of having to remove the tubing and waterblocks from my rig without bathing everything in a fluorescent green glow.

I've built a fair few computers for work recently, all based on the Intel core 2 duo platform, and I have to say they really are quiet enough to not be annoying, even for someone as picky as me. So while as a hobby I don't think the excitement of watercooling will be surpassed, at the end of the day the simplest and most elegant solutions wins, and one that doesn't involve going to fetch some unused towels to prepare to mop up the mess....

Otto69
Posts: 43
Joined: Tue Sep 18, 2007 12:38 pm

Post by Otto69 » Mon Oct 15, 2007 10:08 am

Just out of curiosity, did you drain and refill your system at least once a year in the time you had it running?

peteamer
*Lifetime Patron*
Posts: 1740
Joined: Sun Dec 21, 2003 11:24 am
Location: 'Sunny' Cornwall U.K.

Re: My watercooling days are over...

Post by peteamer » Mon Oct 15, 2007 10:58 am

chylld wrote:My watercooling days are over...
:shock:

.
.
.

:shock:

.
.
.

chylld
Posts: 1413
Joined: Fri Oct 10, 2003 4:45 pm
Location: Sydney, Australia
Contact:

Post by chylld » Mon Oct 15, 2007 1:36 pm

OK I managed to get all the watercooling gear out without spilling any coolant on the computer, the floor or myself. One thing I discovered is that it was not only the tube leading up to the CPU that was blocked; there were several other blocked tubes as well. Given that they obviously didn't all block at the same time, it's fair to assume that the gunk built up in a gradual manner, up to a point where the pump just had enough and stopped moving water.

Here is a picture of the gunk leading up to the CPU:
Image

Here's the tube that was attached to the outlet port on the computer. You can see how the layer of gunk has decided to attach itself to the paper towel plug:
Image

I doubt that any of the waterblocks or the radiator or pump are in any condition to repair, nevertheless I'll check how much they're going for on ebay etc. and make a decision as to whether I'll salvage some parts or just chuck the lot in the bin. I guess I should also call waste disposal services... lots of chemicals here.

Otto69 wrote:Just out of curiosity, did you drain and refill your system at least once a year in the time you had it running?
I didn't drain and refill my system once it got going, and in retrospect it might have been a wise thing to do. However unless you have a tap arrangement that makes this convenient and easy to do, it's not a process one would typically look forward to.

An interesting thing that I found was that after I disconnected the first tube, I turned the pump on however water still failed to flow... it really seemed like a case of gradual gunk build-up everywhere. Had I changed the coolant frequently, I would have noticed this. The best solution I can think of is to blast it with some acidic solution to really clear out the contents...

colin2
Posts: 145
Joined: Wed Jul 11, 2007 2:40 pm
Location: Seattle

Post by colin2 » Mon Oct 15, 2007 5:21 pm

thanks for sharing. kinda like the long-term effects of cholesterol...

jhhoffma
Posts: 2131
Joined: Mon Apr 25, 2005 10:00 am
Location: Grand Rapids, MI

Post by jhhoffma » Tue Oct 16, 2007 6:39 am

Did you add biocide to the water? That looks like algae.

AuraAllan
*Lifetime Patron*
Posts: 713
Joined: Wed Mar 07, 2007 7:49 am
Location: Denmark

Post by AuraAllan » Tue Oct 16, 2007 6:41 am

Wow thats disgusting. :P

Thanks for sharing those nice pics.

desert0mirage
Posts: 18
Joined: Mon Aug 20, 2007 8:54 am

Post by desert0mirage » Tue Oct 16, 2007 7:36 am

WOW, this made my day. Sorry to find your misery amusing, but those pics, and your detailed descriptions made me LOL literally in the office! Makes me glad I bypassed the urge to watercool.

lm
Friend of SPCR
Posts: 1251
Joined: Wed Dec 17, 2003 6:14 am
Location: Finland

Post by lm » Tue Oct 16, 2007 9:14 am

Come on, some little algae scares you? Where's the sport.

chylld
Posts: 1413
Joined: Fri Oct 10, 2003 4:45 pm
Location: Sydney, Australia
Contact:

Post by chylld » Tue Oct 16, 2007 1:41 pm

jhhoffma wrote:Did you add biocide to the water? That looks like algae.
Algae is a good explanation. The coolant additive I used (Nulon Ultra Cool) doesn't tout any anti-algae properties so it may have been a wise thing to add :)

I just realised that Nulon has another product called Radiator Flush and Clean, which may have been a good thing to use had I maintained a regular cleaning cycle.

The noise of aircooling is starting to get to me, hmmm.... stupid little video card fan! :)

Aris
Posts: 2299
Joined: Mon Dec 15, 2003 10:29 am
Location: Bellevue, Nebraska
Contact:

Post by Aris » Tue Oct 16, 2007 2:38 pm

yea thats the biggest reason i stopped watercooling too. most of the people i talked to suggest you clean the entire cooling system once very 6 months, instead of yearly. That is just absurd IMO. Its easily an all day job to take everything out, drain it, clean it, and put it all back in. And other than looking cool, i didnt see any other benefits of using water over air. I've been able to get much quieter systems on air than i ever could with water. Plus now all i need to clean my system out is 10seconds and a can of compressed air.

eXa
Posts: 13
Joined: Mon Mar 27, 2006 1:29 pm
Location: Notodden, Norway

Post by eXa » Fri Oct 26, 2007 3:49 am

this is kinda funny but also kinda sad. Funny because u ran ur system for 3 1/2 years without any maintenance of the loop, and u didnt use anything to protect from algae. On top of that u used some nasty color/uv additive too. hope u atleast used some distilled water... Sad part is that u probably will scare away other people from watercooling because of ur incompetance. But hey, atleast this is nice to show how not to do it. And ur system still

eXa
Posts: 13
Joined: Mon Mar 27, 2006 1:29 pm
Location: Notodden, Norway

Post by eXa » Fri Oct 26, 2007 3:53 am

did ran for 3 1/2 years :p But dont get me totally wrong, im not trying to be so harsh as i sound. ;-) (And sorry about the split post, im surfing on my cellphone...)

chylld
Posts: 1413
Joined: Fri Oct 10, 2003 4:45 pm
Location: Sydney, Australia
Contact:

Post by chylld » Fri Oct 26, 2007 6:14 am

eXa wrote:this is kinda funny but also kinda sad. Funny because u ran ur system for 3 1/2 years without any maintenance of the loop, and u didnt use anything to protect from algae. On top of that u used some nasty color/uv additive too. hope u atleast used some distilled water... Sad part is that u probably will scare away other people from watercooling because of ur incompetance. But hey, atleast this is nice to show how not to do it. And ur system still did ran for 3 1/2 years :p
I guess i should clarify; the coolant i used was 5% (or whatever it said on the bottle) nulon ultra cool, and 95% distilled water. No colour/uv additives were used, the green comes from the ultra cool :)

Incompetence is one word you could use i suppose. However I've yet to find someone who would religiously drain, clean and re-fill their system at least once a year... after all, I figured if it ain't (seem) broke, don't fix it!

YugenM
Posts: 40
Joined: Wed Feb 14, 2007 9:55 pm
Location: Vancouver, BC, Canada

Post by YugenM » Sat Oct 27, 2007 6:13 pm

Only takes <3hr for me to drain my loop, clean components, put it back together and refill.

Don't know why so many people claim it's an "all-day job" :)

Celoth
Posts: 56
Joined: Wed Oct 10, 2007 4:44 pm

Post by Celoth » Sun Oct 28, 2007 3:20 am

Doesn't scare me away from watercooling! :P

I've read some good reviews of TYGON® Silver Antimicrobial Tubing and nothing but distilled water in an all brass/copper loop. The water was still crystal clear after about 1.5 years, according to one guy. As long as the system is leak and airtight (don't want to add oxygen to the loop because of corrosion and algae), pure distilled water should do the trick in any loop, given proper pre-cleaning.

Speaking of pre-cleaning, I have read many use vinegar for their rads, and I suppose I can see how removing an ever so thin layer of the oxidized copper inside can be good, but are there any good alternatives to that? I know it's not gonna eat right through the copper just because of a wash, but if hot water and maybe soap(?) and proper rinsing can do the trick too, I'd rather do that.

eXa
Posts: 13
Joined: Mon Mar 27, 2006 1:29 pm
Location: Notodden, Norway

Post by eXa » Sun Oct 28, 2007 5:57 am

I looked up the nulon ultra cool. Doesnt seemes like someting needed in watercooling. It doesnt really prevent algae from forming. Seemes like its most usefull if u mix copper and alu in the same loop wich u dont wanna do in the first place anyway.

I have ran my loop for about 6-7 months now and im going to clean it out now soon....

MoJo
Posts: 773
Joined: Mon Jan 13, 2003 9:20 am
Location: UK

Post by MoJo » Sun Oct 28, 2007 7:35 am

I have run my system for three years and never cleaned it. I partially drained it once to upgrade graphics card (needed new block) and what came out was 100% clean. Just checked again now and it all looks perfectly clear.

My liquid mix is about 70% de-ionised water and 30% anti-freeze. The anti-freeze will kill any bacteria. Also makes the water a nice blue colour :)

MikeC
Site Admin
Posts: 12285
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

Post by MikeC » Sun Oct 28, 2007 11:25 am

So... is this experience going to result in an update of Chylld's guide to watercooling?

chylld
Posts: 1413
Joined: Fri Oct 10, 2003 4:45 pm
Location: Sydney, Australia
Contact:

Post by chylld » Sun Oct 28, 2007 12:17 pm

MikeC wrote:So... is this experience going to result in an update of Chylld's guide to watercooling?
Yep.. once i finish looking a bit further into the topic I'll update the guide. Everyone's observations here will contribute to the update.

MoJo
Posts: 773
Joined: Mon Jan 13, 2003 9:20 am
Location: UK

Post by MoJo » Sun Oct 28, 2007 3:20 pm

Chylld, can I make a request? :)

Info on selecting radiators would be useful, in particular what size with what fans is required for a given system. The page at http://www.thermochill.com/PATesting/index.php is helpful but only covers a small number of radiators, plus their numbers seem a little contrived. I have done some research but so far have only found anecdotal evidence.

Celoth
Posts: 56
Joined: Wed Oct 10, 2007 4:44 pm

Post by Celoth » Sun Oct 28, 2007 3:20 pm

chylld wrote: Yep.. once i finish looking a bit further into the topic I'll update the guide. Everyone's observations here will contribute to the update.
Basically, after som study and talking to a chemist, all that is needed for a system that stays clean and cools well is distilled water. Make sure everything is sealed tight against both air and water. No additives are needed and may in fact reduce effectiveness over time. As long as no oxygen is added to the system from an open reservoir or T-line or leak, algae and corrosion won't happen. Flushing the radiator and waterblocks with boiling water before use is a good precaution to take though.

Personally I am gonna go for the Tygon silver tubing just to make sure, and I kinda like the look, but it shouldn't be necessary.

Those who like colored/UV water or add other additives, well, chances are they'll gum up the works and reduce effectiveness over time, so they'll have to clean out the system once in a while. That's just the price they must pay for vanity. :D

chylld
Posts: 1413
Joined: Fri Oct 10, 2003 4:45 pm
Location: Sydney, Australia
Contact:

Post by chylld » Mon Oct 29, 2007 12:35 am

Ok I've updated the watercooling guide with some clarifications on the use of additives; thanks for the suggestion Mike.
MoJo wrote:Info on selecting radiators would be useful, in particular what size with what fans is required for a given system.
Radiator size is typically governed by the physical real estate available in any particular application. Given that this is SILENTpcr, it is safe to assume that the primary goal is silence and hence low airflow; and given that low airflow fans have no trouble pulling air through thick (>1") and dense automotive heatercores, it stands to reason that any other alternative would be a compromise.

However I'm not an expert in this area and as such it would be good if someone else can chime in with their thoughts and comments.

eXa
Posts: 13
Joined: Mon Mar 27, 2006 1:29 pm
Location: Notodden, Norway

Post by eXa » Mon Oct 29, 2007 1:59 am

Heatercores are more suited to high cfm(38mm) fans like the thermochill HE series, BIX, and other thick dense radiators. for low speed fans the best u can get is thermochill PA series. not really much competition there... PA>rest. for a less expensive rad there is the swiftec mcr series and the new thin ones from hwlabs. I would actually recommend to stay away from heatercores. They dont have straight fins so they create more turbulence and noise

Marci
Posts: 68
Joined: Thu Oct 27, 2005 5:49 am
Location: West Yorkshire, UK
Contact:

Post by Marci » Mon Oct 29, 2007 3:25 am

The page at http://www.thermochill.com/PATesting/index.php is helpful but only covers a small number of radiators, plus their numbers seem a little contrived
Those numbers were produced by the same independant guy (Bill Adams) that produced Swiftech's numbers, and Black Ice's numbers, and Coolingworks' numbers.... and the data for our older HE Series ( http://www.overclockers.com/articles778/ )

BillA also designed both the Swiftech MCR series and Coolingworks radiators during his periods of employ with each of those two manufacturers.

Swiftech's numbers are on each radiators product page at http://www.swiftnets.com/

Black Ice & Coolingworks data was published by Cooling-masters... http://www.google.com/translate?u=http% ... en&ie=UTF8

Other radiator performance info was covered by BillA in his Assessment of Radiator Performance article, which can be found here: http://www.swiftnets.com/Technical/Asse ... rmance.pdf

On that basis, if our data is "contrived" then so is everyone elses... in which case, may as well ignore the existence of all radiator testing data and just go with whatever floats yer boat, thus the need for "info on selecting radiators" becomes pointless. Before BillA undertook serious radiator testing for the various manufacturers, all there was was anecdotal evidence. Bill's testing produced empyrical reproduceable data via strict methodology.

There is plentiful data out there for how to choose a radiator in the case of our radiators as we made it a mission to ensure such data is publicly available - see http://www.over-clock.com/ivb/index.php?showtopic=20277 for an example.

MoJo
Posts: 773
Joined: Mon Jan 13, 2003 9:20 am
Location: UK

Post by MoJo » Mon Oct 29, 2007 3:52 am

eXa wrote:Heatercores are more suited to high cfm(38mm) fans like the thermochill HE series, BIX, and other thick dense radiators. for low speed fans the best u can get is thermochill PA series. not really much competition there... PA>rest. for a less expensive rad there is the swiftec mcr series and the new thin ones from hwlabs. I would actually recommend to stay away from heatercores. They dont have straight fins so they create more turbulence and noise
I use a heatercore with Yate Loon @ 5V pulling. It seems to work very well, keeping both a 7800GT and A64 X2 4200+ very cool. It would be nice if I could recycle it for a new system, but there is no real way of knowing if it would be up to the task of dissipating 250W of heat (peak).

Marci
Posts: 68
Joined: Thu Oct 27, 2005 5:49 am
Location: West Yorkshire, UK
Contact:

Post by Marci » Mon Oct 29, 2007 4:44 am

Page 9 of the Radiator Assessment article covers the 3x popular Heatercore variants used in watercooling. With a Delta1212 pushing 72cfm, the best the heatercores managed was around 180w at a 10 deg C Air > Coolant delta... more or less exactly the same performance as a PA120.1, `cept the PA120.1 is a smaller footprint...

It would dissipate 250w with 72cfm, but the coolant > air delta would rise by around 5 degrees... so you'd take c/w of block at current flowrate, multiply by heatload, add to [ambient + 15 deg C] for a rough idea of the temp of the item you're cooling.

With a YateLoon at 5v, and a dT of 15 deg C, you'd be looking at 150w of heat dissipation thereabouts. All depends what you class as the highest temperature you'd deem reasonable for your setup vs the noise levels you'll put up with.

MoJo
Posts: 773
Joined: Mon Jan 13, 2003 9:20 am
Location: UK

Post by MoJo » Mon Oct 29, 2007 5:23 am

Marci, if that were true then my current system would overheat, which it does not. In fact my temps are excellent.

Marci
Posts: 68
Joined: Thu Oct 27, 2005 5:49 am
Location: West Yorkshire, UK
Contact:

Post by Marci » Mon Oct 29, 2007 5:41 am

A radiator will rarely overheat, the dT between air and coolant will simply increase until equilibrium is reached.

Note when I say it'll cool 150w at a dT of 15 deg C with a YL at 5v, it'll cool 200w at a dT of 20 deg C, 250w at a dT of 25 deg C etc etc... the delta just increases. There is no heatload that a radiator cannot cool... it's just whether the temps you get at the end of it are what you'd deem reasonable. Bear in mind the figures I'm basing this off are for a DTek ProCore... probably not the exact same heatercore that you have...

Without exact figures for your specific heatercore, all i can do is paint a relational picture of how performance would be expected to work based on a similar radiator.

Also, figures quoted are assuming that all the watercooled items are at 100% load simultaneously... something rarely achieved in a PC... I haven't seen a game yet that puts both GPU and CPU at a continuous 100% load, nor any other software capable of doing so... ATITool can put an ATI GPU and CPU at theoretical full load, but it's still only around 70% of what PSU Calcs would tell you the full load of those items is due to inefficiencies in mobo power regulation etc. S'where it all gets tricky... I think it was Cathar that once told me that best guess would be take 100% power consumption for the items you're cooling... then bring it down to 66% of that for the realistic max load that that system will ever put out. Would have to have a ratch thru a few years of PMs to find it.

I work it out to be 115w full load tops for an X2 4200 & 7800GT according to PSU Calculators... going off 75% TDP. (Selecting the watercooled items here and knocking off the default 38w starting voltage, not including heatdump of pump, which at most would be around 18w if it's a 12v DC pump, which'd give 133w at the very most, thus easily cooled by the ProCore example with a YL @ 5v)

MoJo
Posts: 773
Joined: Mon Jan 13, 2003 9:20 am
Location: UK

Post by MoJo » Mon Oct 29, 2007 8:43 am

Marci wrote:Note when I say it'll cool 150w at a dT of 15 deg C with a YL at 5v, it'll cool 200w at a dT of 20 deg C, 250w at a dT of 25 deg C etc etc... the delta just increases. There is no heatload that a radiator cannot cool... it's just whether the temps you get at the end of it are what you'd deem reasonable. Bear in mind the figures I'm basing this off are for a DTek ProCore... probably not the exact same heatercore that you have...
Ah, that makes a lot more sense. Thanks for explaining it. My figures for total system load including pump were more like 150W, but perhaps you were looking at the 1.35V X2 4200+ (mine is a 1.4V). From your figures, it looks like my heatercore is better than the DTek one.

Post Reply