Leak Testing

Associate
Joined
8 Jan 2011
Posts
372
Location
London, UK
I just got hold of an EK leak tester. Its
a nice piece of kit, comprising a small pump and gauge.

I have been testing a part of a loop with it, and I get a very slow drop of pressure over the period of about 2 hours. Say from 0.25 bars to 0.15.

In terms of my loop, presume this is bad, as it means I have a very slow leak.

But given this is my first time using this kit, I thought I’d ask anyway: is it normal to drop slowly over time? Or should it stay exactly the same? I presume the latter?

Thanks :)
 
Thanks @Cyber-Mav . I guess my question is: if air leaks, will water? Water has surface tension. Air does not. This is a ‘long’ leak, taking a long time to drop.

Probably a stupid question, but I was wondering if it’s ‘normal’ that a water loop would also be 100% airtight.

Anyway I’ll get testing to try to find out where this leak is occurring from. Cheers.
 
This for me is why you shouldn't get too hung up with leak testing with air - it's holding pressure for a period of time, as long as it's not immediately losing pressure. Until you test with water you won't know whether it's really a problem, and at least with water you'll be able to trace a leak fairly easily to allow you to then fix it.
 
Never used one
But 0.25 bar seems a bit low?
Do the instructions say to use that?
I would have expected about 0.5 bar personally

Are you testing individual parts of the loop?
Rather than the whole thing?

Doesn't seem like much of a drop
Over that period of time

Could be the actual leak tester is leaking it?

Edit
DOH! Sorry you said part of the loop

2nd edit
Could do the old school thing
Put some washing up liquid on there
Pump air in look for bubbles
 
Last edited:
I don't think the needle moved much at all when I used mine; although I only mostly left it on for 20-30 minutes as reassurance before moving to wet testing. Except for that one time when I forgot to tighten one of the ports on the top of the res. but then I could hear the leak :)

The one time I ended up leaving it on for longer I think the loop lost closer to 0.025bar (so one mark on the scale).

Just to pick up on @Mcnumpty2323 point and check the obvious - the valve on the body of the tester was completely closed wasn't it?
 
Thanks @Cyber-Mav . I guess my question is: if air leaks, will water? Water has surface tension. Air does not. This is a ‘long’ leak, taking a long time to drop.

Probably a stupid question, but I was wondering if it’s ‘normal’ that a water loop would also be 100% airtight.

Anyway I’ll get testing to try to find out where this leak is occurring from. Cheers.
For that part of the question
Something that's airtight will be watertight
Something that's watertight will not always be airtight
Not a physicist but would guess that's due to
One being a gas one being a liquid
So yes surface tension
Size of molecules involved
Stuff like that
 
Thanks a lot, all. Yes manual says 0.5 bar, I was testing it a bit low.

https://www.ekwb.com/shop/EK-IM/EK-IM-3831109816738.pdf

Did some Googling and it apparently is a thing that systems can fail an air test but then pass a water test, but this is more about plumbing, industrial applications, etc.

But, the very basic physics says air may get out of systems when water may not, mainly due to the difference in viscosity and the differences in pressures between water and air.

Quite how all this applies to, or scales down to, a pc water loop, I don’t know, but bottom line is I think I can’t ignore this very slow air leak.

But, as stated above, a water test will be the only real way to prove it.

The unfortunate part is that this section of the loop has one bit that I spent literally five hours trying to get right… might have to undo it to test it all…!

And yes @VersionMonkey valve was closed. (First few tests it wasn’t…! D’oh!)
 
It could well be an O-ring not sitting right. Better to be safe then sorry. I'm still doing my leak testing with paper tied around the seals of each fitting :cry:
 
Yeah not sure how it scales down
Though plumbing and industrial stuff
Would usually be under higher pressure
Than watercooling loop
Any way to check the accuracy of the leak tester?
Like don't attach the tube to a component
Just blank/stop the tube
No way to liquid test the part in situ?
 
This for me is why you shouldn't get too hung up with leak testing with air - it's holding pressure for a period of time, as long as it's not immediately losing pressure. Until you test with water you won't know whether it's really a problem, and at least with water you'll be able to trace a leak fairly easily to allow you to then fix it.

EK says 15 - 30 minutes. If it doesn't drop then it's fine.

One time recently back in December when I was doing a flush and swapping out parts, I forgot to fully close the torque drain valve and wondering how I had a massive leak from the air tester. :cry:

Never used one
But 0.25 bar seems a bit low?
Do the instructions say to use that?
I would have expected about 0.5 bar personally

Are you testing individual parts of the loop?
Rather than the whole thing?

Doesn't seem like much of a drop
Over that period of time

Could be the actual leak tester is leaking it?

Edit
DOH! Sorry you said part of the loop

2nd edit
Could do the old school thing
Put some washing up liquid on there
Pump air in look for bubbles

0.30 is the recommendation I found and where I marked it with a silver strip. That was where mine held for 2 hours until I let the pressure off.

 
Hmm. Tested to see if it’s the leak tester. That appears fine. Definitely is a problem in this section of the loop. :( . Going to have to rebuild it. Great vid @V F thanks :)
 
Damn that's a pita then
Still at least you had the sense to test
A section at a time. While building it
Rather than do. The whole thing
Then test it
 
Yeah was just in case
The leak tester came out the factory
With a less than perfect seal
Always good to know the instrument you are
Relying on to measure with is itself accurate
 
I’ve been able to narrow it down to one of the fittings. It’s not the rads. It’s not the ‘difficult’ junction I thought it might be. So it was one of the fittings. Will have to explore more tomorrow… need a cold beer now. I’ve aged a little today :cry:
 
I’ve been able to narrow it down to one of the fittings. It’s not the rads. It’s not the ‘difficult’ junction I thought it might be. So it was one of the fittings. Will have to explore more tomorrow… need a cold beer now. I’ve aged a little today :cry:

Also don't rush it. When you think you're tired from working on it, take a break from it. Mistakes happen from rushing and getting very eager pushing through it quickly. Check and test again when you're fresh.

Reddit which is frightening at times is full of mistakes from cheapening out, cutting corners or are trying to speed up the build. Quite frightening the stories that people get up to when not being thorough.

The thing I find frightening is the amount of people getting caught up from Youtube videos saying acrylic is bad and fragile. Then they don't keep an eye on their water temperatures over the long run with PETG.

https://www.reddit.com/r/watercooling/comments/ryk7xg/emergency_leak_pls_help/
https://www.reddit.com/r/watercooling/comments/rzae6k/my_reamer_is_leaving_really_rough_edges_on_my/
https://www.reddit.com/r/watercooling/comments/rz75r3/pop_and_smoke_from_gpu/
https://www.reddit.com/r/watercooling/comments/rw2lqs/when_you_skip_flushing_radiators/
https://www.reddit.com/r/watercooli...this_okay_for_the_rubber_to_be_out_like_that/
https://www.reddit.com/r/watercooli...melted_my_tube_leak_made_my_psu_explode_what/
https://www.reddit.com/r/watercooli...mystic_fog_after_2_monthsgpu_and_8_months_in/
https://www.reddit.com/r/watercooling/comments/rscg4h/pump_stopped_pushing_water_and_gpu_no_longer/
https://www.reddit.com/r/watercooling/comments/rrpt14/happy_cake_day_adios_rog_thor_psu_oof/
 
Oh he's definitely not rushing
I have been following his other post
Takes his time and thinks stuff out
And asks if he's not totally sure
Am looking forward to seeing his finished build
It's been looking great so far
 
I’ve been able to narrow it down to one of the fittings. It’s not the rads. It’s not the ‘difficult’ junction I thought it might be. So it was one of the fittings. Will have to explore more tomorrow… need a cold beer now. I’ve aged a little today :cry:
Yeah you definitely deserve a cold beer mate :)
 
Back
Top Bottom