Holy Icarus! Melted / deformed PETG runs in hardline watercooled Node 202

[Smollie1]

So I played CS 1.6 and Halo: Combat Evolved 10 years ago. Then a few months ago I decided to build my first PC and it was going to be a fully hardline watercooled Node 202. To my knowledge this had been done once before by @SonoDanshi here. He used an R9 Nano, which seems sensible. However, I was determined to use a full length GPU and ended up with an EVGA GTX 970 SC. It's currently paired with a X4 860K (don't ask).

The whole ridiculous build log is documented here, I've linked to where it gets interesting i.e. a couple days ago.

 

Anyway, fast forward to Sunday and I finally get it up and running. I play Crysis and the machine gets damn hot, but all seems well. Then today I play some Burnout: Paradise, Portal and Metro: 2033 (which looks awesome) and I notice a leak and that a fitting is very loose. Turns out EVERY fitting is loose and I have a major insane catastrophe leak. Machine is fine although in 1000 pieces again.

 

I've had to solve a lot of problems for this build, but I never expected this one. Looks like I need to accelerate some improvements I was planning to reduce temps i.e. increase airflow with a Dremel to GPU side panel / minimizing 90 fittings to keep flow rate up and maybe adding some 80mm fans.

 

Let the madness continue...

 

 

[Shimejii]

why would you do this tubing for such a crap cpu and mediocre gpu  The custom cooling is going to cost you more then the hardware is worth

 

Also damn those tubes are fucked lol

[Technomancer__]

1 minute ago, Shimejii said:

why would you do this tubing for such a crap cpu and mediocre gpu  The custom cooling is going to cost you more then the hardware is worth

 

Also damn those tubes are fucked lol

Agree

You should have used the water cooling money to get better hardware

[Smollie1]

I'm doing this for the challenge. I could order better parts and have it running in a bigger case tomorrow, but I'd be bored AF instantly.

 

I'm taking suggestions for keeping temps down that don't involve changing core components or getting a new case. That will be considered cheating.

 

And yes, those tubes are fucked real good.

[WkdPaul]

The build is done, it's his choice and money anyway.

 

So now lets try to be helpful instead of criticizing his hardware choice

 

 

Any insight @B NEGATIVE ?

[Smollie1]

@wkdpaul +1

[GzeroD]

2 minutes ago, Ethocreeper said:

Agree

You should have used the water cooling money to get better hardware

I disagree. If the hardware does what you want it to, and you really want to fit a certain form factor you don't need to but the highest end you can afford. Especially if he wanted to try hardline and liked the aesthetic. I like the Idea and love most things SFF so I hope he gets this worked out and pretty again.

[Taswell]

I would water cool a i3 and a 1050ti if i felt like it, i see no problem here 

[RollinLower]

i agree that a couple 80mm fans would help this a lot. just a bit of air flowing past the runs should keep the outer part of the tubes atleast cool enough to prevent deforming. 

also, what kind of pump do you have powering this? if you can get a more powerfull one to increase flow i bet it would help a lot too!

[GzeroD]

I'm also kinda shocked the PETG warped like that.. at Temps below say 140-150C it doesn't even get notably more flexible When I use PETG filament on my 3D printer. ( I have a bunch of PETG parts near my hot end because they hold up better than ABS or PLA and are rigid unlike nylon.

 

Then again I am not shocked the thermals are this tight. I made a ITX rig a while back with a mere 7850 and 3570k and on a single 120mm thin rad temps were hard to control then I overclocked them both; but this build has a significantly higher power draw than that one did. 

 

I kinda wonder if you could pull it off by going with the copper pipe from the build log title and a couple 80mm fans. as the copper would not warp and it would give you more Heat dissipating surfaces.

[Smollie1]

Ah, interesting idea @RollinLower. Right now it's the EK-XTOP SPC-60 PWM - Plexi (incl. pump). I haven't cable manage at all except for the cable combs and if I did, I think I'd find some space for some fans.

[For Science!]

Yeah, my feeling is also too poor airflow. Nothing cooling VRMs, RAM or the heat that irradiates from the GPU backplate

 

I wonder if the PSU can take one for the team by flipping it up and making it into an "exhaust".

 

Also if you can fit a D5 (probably not) then it'll dump less heat into the system

[Smollie1]

Yeah, a D5 would be a really tight squeeze if it fit at all. I think I could get an 80mm fan in the corner with the ram and rear I/O. Also, I'm going chop a big hole along the GPU side (Imagine identical to the grill on GPU side). I could get another 80mm fan somewhere on GPU side.

 

If I set them both to intake, would that improve the situation without creating any extra problems?

 

@GzeroD The copper was an absolute nightmare to work with if you don't have a $100 pipe bender, which I don't. It looks amazing but I move around a lot and PETG is cheaper and easier to work with. One day I will return to the copper (and buy the bender) as it has a bunch of preferable features. I'd probably do a copper / PETG combo.

[SonoDanshi]

Hey @Smollie1 and ... wow... that petg took a beating. Very simple question for you..... Have you tried adding additional rubber feet on the bottom to rise the case off the table as almost all of the airflow going into the node 202 comes from underneath where you have the radiator? I bought some rubber adhesive feet off amazon that were about 3x higher than the ones that come with the case.

[Smollie1]

It was propped up on two pill cases so it was about an inch off the table I'd say. I think you are totally right though. I'm going to figure out some kind of feet situation. I'll probably also prop the back up slightly higher than the front to see the system better and also improve air flow directly under the case.

[SonoDanshi]

What TDP are you dealing with for CPU and GPU? I have the Nano (175W) and a 6600k (91W) and haven't had anywhere close to the kind of heat you are seeing to melt that pipe.

 

EDIT: Just checked and seems to be 145W and 95W respectively, so actually lower than mine. The mystery deepens...

[GzeroD]

22 minutes ago, SonoDanshi said:

What TDP are you dealing with for CPU and GPU? I have the Nano (175W) and a 6600k (91W) and haven't had anywhere close to the kind of heat you are seeing to melt that pipe.

Correct me if i'm wrong but aren't you using a dual radiator vs his single.  

anyway his 970 is 170W power limit (187 with 110% power bump) according to the evga forum  and techpowerup seems to agree (in a review of that specific card ) in it's review of the card and the 870k is 95W

[B NEGATIVE]

3 hours ago, wkdpaul said:

The build is done, it's his choice and money anyway.

 

So now lets try to be helpful instead of criticizing his hardware choice

 

 

Any insight @B NEGATIVE ?

PETG and Glycol coolants = Broken. Glycols are a solvent for PETG.
 

I have been saying till im blue in the face,PETG is not acceptable for our us.

[AnonymousGuy]

56 minutes ago, B NEGATIVE said:

PETG and Glycol coolants = Broken. Glycols are a solvent for PETG.
 

I have been saying till im blue in the face,PETG is not acceptable for our us.

I'm going to go with this as well.  Has to be a chemical reaction not the loop getting so hot it melts PETG.

[Tech22]

4 hours ago, GzeroD said:

I'm also kinda shocked the PETG warped like that.. at Temps below say 140-150C it doesn't even get notably more flexible When I use PETG filament on my 3D printer. ( I have a bunch of PETG parts near my hot end because they hold up better than ABS or PLA and are rigid unlike nylon.

 

Then again I am not shocked the thermals are this tight. I made a ITX rig a while back with a mere 7850 and 3570k and on a single 120mm thin rad temps were hard to control then I overclocked them both; but this build has a significantly higher power draw than that one did. 

 

I kinda wonder if you could pull it off by going with the copper pipe from the build log title and a couple 80mm fans. as the copper would not warp and it would give you more Heat dissipating surfaces.

I have a copper clamp down tube cutter. It is like scissors for REALLY hard tubing. It bent bitspower PETG and deformed it. If you used bitspower PETG, this may be the problem. Granted, this happened back in April-May. 

[Revan654]

Not sure how accurate this data is but PETG malformed at around 82C. It's why allot of companies with reservoirs & such parts says Do not exceed X degree. It's usually around 60 to 70C based on the one manual I have.

[Smollie1]

Ding ding ding. We have a winner i.e. Mayhems UV. The loop is a combination of Bitspower and EKHD tubing. I wouldn't be surprised if the heat also had something to do with accelerating this process. Not sure on the loop temp, plausible that it was in the high 60 /  low 70s. The thermal limit of the 860k is probably 72c and the thermal margin was in the single digits i.e. CPU in high 60s. The GPU temp was in the low 70's.

 

Guess I've learned my lesson. Time to start planning loop v2. I wonder if I can fit 2 more fans, a loop sensor and the lighting I'd planned in there?

 

Why do anything the easy way when you can do it the unreasonable hard expensive way and learn 100x more. I'm not going to stop until I conquer this case for hardline.

 

[LyondellBasell]

*TRIGGER WARNING*

*TRIGGER WARNING*

*TRIGGER WARNING*

 

 

Use acrylic hardline and distilled water.

 

*grabs popcorn*

[Revan654]

2 hours ago, LyondellBasell said:

*TRIGGER WARNING*

*TRIGGER WARNING*

*TRIGGER WARNING*

 

 

Use acrylic hardline and distilled water.

 

*grabs popcorn*

 

 

Or just use Glass.

[For Science!]

5 hours ago, Smollie1 said:

-snip-

Mayhems UV dyes are supposedly water based, no glycol in them.

 

http://mayhems.net/sds/addatives/blue_addative.pdf

 

Or it could be their documentation is inaccurate. which is not anything new with Mayhems

 

-Triggered-