Sign in with
Sign up | Sign in
Your question

0 ring on GPU water block issues

Tags:
  • Water Cooling
  • GPUs
  • Overclocking
Last response: in Overclocking
Share
July 27, 2012 1:27:31 PM

Just fitted a waterblock to a GPU and the long, snaking O ring came out and now I can't get it back in. I keep trying, but it just will not stay down. Silly question, but can I glue the thing in place? Or is there a particular technique I should know about?

Thanks guys.

More about : ring gpu water block issues

a c 337 K Overclocking
July 27, 2012 1:46:27 PM

Did you take the GPU block apart for some reason? You can also use a bead of silicone or gasket sealer in the groove to effectively make your own gasket, but I would certainly leak-test this first before using.

Edit: the trick is a lot of patience, tedious maneuvering with small tools and luck.
July 27, 2012 1:56:19 PM

Right! I have tried sticking it in the freezer to shrink a little - seems a bit long at the moment - I never stretched it but it seems to long now. Could I glue it in place with a bit of silicone?
Related resources
a c 337 K Overclocking
July 27, 2012 2:04:57 PM

You mean as in to keep it in place in the groove while you insert it around the remainder of the o-ring groove? I suppose, but you wouldn't want to use anything that would deteriorate the rubber of the o-ring itself.

Why did you take the block apart? Did it need to be cleaned?
July 27, 2012 2:14:17 PM

I needed to take it apart to fit the gpu link for the two cards.
July 27, 2012 2:22:21 PM

Seems to me like it has expanded
July 27, 2012 3:32:15 PM

Basically I had heard horror stories of people putting waterblocks in and there being leaks due to already broken o rings. So I wanted to check - I didn't realise it would be so hard to put back in place. It seems to have expanded, so even if I stuck it with something (silicone for example) it would still be too large to fit in to the groove. Looks like I am in trouble with this one.
a c 249 K Overclocking
July 27, 2012 3:52:05 PM

You could use cold water running over it to shrink the O-ring but if it tries to expand before you get it in place use a shallow pan and submerge the water block and the O-ring and use ice in the pan to lower the temperature and refit the O-ring to the groove.

The rubber O-ring will shrink fast the colder it gets, and once it gets to the manageable groove fitting size and you get it back into the groove, reassemble the block immediately.

Use cubed ice around the perimeter of the water block, be patient and careful, get rid of your frustration before you begin, this will work just be patient, you'll have to allow it to get cold. Ryan
July 27, 2012 4:11:05 PM

Thanks a lot - I have pretty much given up on it! Was on the verge of buying a new block. I will try this - why is it some say to heat the o ring? That got me confused - I thought heat expanded rubber, not contracted it. I will try and place block under the water and do it like that. Thanks again.
a c 337 K Overclocking
July 27, 2012 4:27:39 PM

They say to heat it to make it more pliable and easier to maneuver. Heat expands, cold contracts or condenses.
July 27, 2012 5:26:59 PM

Ryan

It worked! Got the thing in by packing it with ice as I went aruond the rim - seemed to shrink before my eyes! There's some remaining ice still in the block, but I am guessing that will evaporate. The only worry is, whether the 0 ring stayed in place whilst I was connecting the rest of the block! Who knows - I guess that's where leak testing is important.

Many thanks again.
July 27, 2012 5:28:03 PM

Hi Rubix - that makes sense. Needed to contract it and ice seems to have done the trick - just hope it stays in place! Thanks for your help (again!)
a c 249 K Overclocking
July 28, 2012 11:30:31 AM

matt77 said:
Ryan

It worked! Got the thing in by packing it with ice as I went aruond the rim - seemed to shrink before my eyes! There's some remaining ice still in the block, but I am guessing that will evaporate. The only worry is, whether the 0 ring stayed in place whilst I was connecting the rest of the block! Who knows - I guess that's where leak testing is important.

Many thanks again.


Congratulations Matt!

How did the leak test go?

Any leaks should reveal themselves on full pump pressure within at least an hour, if you're unsure leak test it longer, however the O-Ring seal properly set back in the groove very rarely leaks unless it gets nicked somewhere.
August 6, 2012 8:29:16 PM

4Ryan6 said:
Congratulations Matt!

How did the leak test go?

Any leaks should reveal themselves on full pump pressure within at least an hour, if you're unsure leak test it longer, however the O-Ring seal properly set back in the groove very rarely leaks unless it gets nicked somewhere.



Hey Ryan!

Have only just now noticed your post...sorry.

Yeah, I have just started leak testing, now that the rig is together. Have been running for 6 hours, and all is dry. I am guessing 24 hours is a true test though, so I'll leave it overnight and see what happens.

I have been reading that just jumping the PSU is not a true test since when the hardware is plugged in the water temp will rise, meaning there could be leaks! That's a bit worrying, since it is precisely not wanting to fry my components that has me using the PSU as a jumper.

a c 337 K Overclocking
August 6, 2012 8:33:51 PM

Water temp will rise because your pump is dumping heat into the loop like it normally would, although no other components should be powered up (CPU, GPU, etc). The only components that would be pulling power are 3.3v and 5v which would be plugged into their respective components via molex or SATA power. If you want to disconnect them, you can as well.
August 6, 2012 8:38:31 PM

Thanks. I think the thing they were saying was that leak testing with just the PSU was one thing, but the true leak test is actually having everything powered up - mobo, cpu, ram, the whole lot. That is when the leaks are likely to appear since the heat being dumped into the system change the composition of the water - i.e. leak testing with a jumped PSU is not a true leak test. Seems weird to me - how will you ever know your system is leak proof if you have to power on your whole system, and kill it in the eventuality that there is a leak?!
a c 337 K Overclocking
August 6, 2012 8:42:37 PM

What you are doing now is going to give you 98-99% of what you're going to find when determining leak testing. Water doesn't really get that warm (depending on what you've calculated your delta to) so you aren't going to see a substantial change in water temperature to cause alarm or any disastrous circumstances. Your O-ring really doesn't have anywhere to go now that you have the block secured, but yes, the real test will take place once you power everything up, check for leaks once again and then put a load on the entire system (GPU and GPU) to completely load to max temps. In this case, it will be helpful to keep a towel and/or several layers of paper towels in place and carefully check all edges and seals of your newly sealed block.
August 6, 2012 8:46:53 PM

rubix_1011 said:
What you are doing now is going to give you 98-99% of what you're going to find when determining leak testing. Water doesn't really get that warm (depending on what you've calculated your delta to) so you aren't going to see a substantial change in water temperature to cause alarm or any disastrous circumstances. Your O-ring really doesn't have anywhere to go now that you have the block secured, but yes, the real test will take place once you power everything up, check for leaks once again and then put a load on the entire system (GPU and GPU) to completely load to max temps. In this case, it will be helpful to keep a towel and/or several layers of paper towels in place and carefully check all edges and seals of your newly sealed block.



Thanks for this - put my mind at ease somewhat. Am looking for leaks all of the place eally. Testing the whole system, not just the GPU. Do you think I could leave it running overnight - getting late here and I can't stay up all night!
a c 337 K Overclocking
August 6, 2012 8:51:50 PM

You can, yes, or you can simply shut it down and pick up leak testing later. This also can give you an idea if you have any issues when shutting down your loop...sometimes very small leaks or loose fittings won't accumulate much until you actually shut down flow.
August 6, 2012 8:58:10 PM

Huh, interesting. Maybe I'll shut it down overnight, run it again for 15 hours tomorrow and then plug everything in and load it. What do you think?
!