sprung a leak

MikeTimbers said:
Sorry, but a 24-hour leak-test with no power to any components except the pump is not optional. Only very experienced water-coolers can avoid this and even then it's not too smart!
Click to expand...
I'm about to go watercooling, would have put me totally off if he had done.

Were only human, learn from mistakes etc, when up & running make sure you do a 24 hour leak test with just the pump running. :)
 
this isn't the first loop i've put in anyway, so it's not down to poor experience etc... the 2 previous loops i've installed in rigs have been fine with that as a leak test, so i didn't see any reason as to why i'd need to do anything more this time... i got it wrong, and i'm paying for it now.
 
If the barbs are being re-used (or secondhand) and you work on the 'extremely tight' approach you most likely killed that o rings over the three builds. And when your loop heated up it failed :( Had a secondhand barb/o-ring do this to me - but loop was being tested outside the case luckly.

Many use PTFE tape and O rings (belt and braces approach) or follow the guidence that finger tight + quarter turn is enough
Most of use have been their, just glad when it did happen to me it wasn't an expensive leason - you may get lucky when it's all fully dry.
 
Last edited:
these aren't second hand barbs, these came with the mosfet blocks a week ago, and i did the finger tight+1/4 turn as well... when i've drained the loop off i'll inspect the ring and see what the deal is with it. quite tempted just to get compression fittings on everything (got some 45deg ones in the loop for easier hose angling)
 
That just sucks - did you see any damage - or does the block/barb have an over/under sized recess.

Was lots of problems reported with non-compatable barbs a while ago (in connection with the typhoon3 and metal barbs) oversized O'rings solved the issue
up untill then I though all barbs were univerally compatable - they arn't :(

MosFet block sometime have very short internal threads - MIPS blocks come to mind - so spacers are required.
I think koolance one are simular but there barbs are short threaded too.
 
Last edited:
i didn't notice any damage, no. they looked fine.

the place where i bought the blocks from supply the same barb for whichever block you get unless it's a thermochill rad.

there definitely was enough thread on the blocks (ek asus rampage formula mosfets) and i made sure the o-ring was in good nick, and was fully compressed. so i think the o-ring might have been damaged, but not to a noticable degree (or possibly the rubber was old and degraded?)

thanks for the input
 
Don't imagine that compression fittings are any less prone to leakage. If anything they're less safe.

Get a good barb. O-rings on barbs are a gimmick. Get standard barbs and use PTFE. Hook the block up to tubing on mains pipes and see if they leak at the join of barb and block. Mains pressure is way above what any pump is going to do.

Use good tubing that's smaller than the barb. ½" OD barbs and 7/16" ID tubing is perfect. Push the tubing all the way to the base of the barb and use a jubilee clip as well. Not a plastic one. Metal.
 
One thing I did as a kind of heath robinson leak test before I did a proper leak test, was to put all the blocks on the mobo out of the case, with the small interconnecting tubes between each block with the to and from tubes cut to their correct lengths, put your finger over one of the free ends and blow in the other. You can probably generate more pressure doing this than the pump will produce, and if there is any tell tale drop in pressure which you should be able to tell by your mouth, then there is a leak somewhere. You may even be able to hear it. This at least stops any major balls ups with the block barbs before the mobo goes in the case. Did this with the rad too to test its barbs where seated properly, though don't blow too hard with the rad as it can flex quite a bit and you don't want to burst it obviously.
 
A non O-ring barb will usually have a much better machined thread (maybe even knife sharp)
With the addition of tape every part/full lenght of thread is now watertight.

with a 'cheap' O-ring barb - water will pass all the way up the thread (and capilary action is a VERY strong force) with the O being the one and only thing to stop a leak.

It's a good link - but G1/4 is a parralel thread and using tape does help (but I fully accept it's not good enough in a pressurized system - which watercooling isn't)
 
Last edited:
shadowscotland said:
A non O-ring barb will usually have a much better machined thread (maybe even knife sharp)
With the addition of tape every part/full lenght of thread is now watertight.

with a 'cheap' O-ring barb - water will pass all the way up the thread (and capilary action is a VERY strong force) with the O being the one and only thing to stop a leak.
Click to expand...

would you say then as a backup plan use some o-rings as well as ptfe tape?
 
belt and braces - but PiKe is right if it's a quality machined barb (ek/dtek/koolance/bitspower).
in theory a O-ring by it's self is also fine - but you now know different
Only issue with using a o-ring barb without the ring is thread length.
 
Last edited:
so with this ptfe tape, how may winds do you have to put around the barb? is it just the 1 or should you do a couple to make it thicker? i'm presuming you'd only really manage 1 (maybe 2) before it's too chunky to get in the barb hole
 
well, my friend didn't bring the power supply with him over the weekend as he forgot :(

so, i've shelled out for a new psu, which even if it's not that which has blown at least i've got a decent spare now :D

hopefully it'll arrive tomorrow so i'll post up any break throughs i have once i've got it installed...

watch this space!
 
You must log in or register to reply here.
Back
Top Bottom