New and Improved Watercooling Sticky - Post Your Systems Here

Joined
May 20, 2011
Messages
898
IMG_20201216_171309.jpg
 

lobstar

Limp Gawd
Joined
Nov 24, 2004
Messages
284
Put a block on my 3090, swapped in a Dark Hero/5950x combo, and swapped out the 3700x/Turn Gaming combo in my server for my old C8H/3950x. I have both systems on the same loop. This is the main system.
 

Attachments

  • PXL_20210420_044528512.jpg
    PXL_20210420_044528512.jpg
    449.1 KB · Views: 0

ATLPIMP

[H]ard|Gawd
Joined
Jun 27, 2002
Messages
1,124
Might as well put these [H]ere too... Winter One First Edition.
 

Attachments

  • IMG_4786.png
    IMG_4786.png
    2.5 MB · Views: 0
  • IMG_4803.png
    IMG_4803.png
    2.5 MB · Views: 0
  • IMG_4804.png
    IMG_4804.png
    2.3 MB · Views: 0
  • IMG_4806.png
    IMG_4806.png
    2.5 MB · Views: 0
  • IMG_4816.png
    IMG_4816.png
    2.6 MB · Views: 0
  • IMG_4818.png
    IMG_4818.png
    2.5 MB · Views: 0
  • IMG_4820.png
    IMG_4820.png
    2.3 MB · Views: 0

Jamie Marsala

Limp Gawd
Joined
Mar 9, 2016
Messages
309
I just redid my system in a new case, the Corsair 5000X RGB. I finally found a water block for my video card and it is on its way, MSI GeForce RTX 2080 Super Ventus XC OC.
 

Attachments

  • Photo May 13, 11 16 49 AM.jpg
    Photo May 13, 11 16 49 AM.jpg
    440.1 KB · Views: 0
  • Photo May 13, 11 17 06 AM.jpg
    Photo May 13, 11 17 06 AM.jpg
    539.1 KB · Views: 0
  • Photo May 13, 11 17 19 AM.jpg
    Photo May 13, 11 17 19 AM.jpg
    551.1 KB · Views: 0
  • Photo May 13, 11 18 19 AM.jpg
    Photo May 13, 11 18 19 AM.jpg
    436 KB · Views: 0

motqalden

[H]ard|DCOTM x4
Joined
Jun 22, 2009
Messages
2,317
Well I decided to WC my Radeon VII DC / Mining rig to try to make it less noisy and run cooler. Ended up ordering everything from Ali Express which was a bit of a risk but worked out. I would have bought from a US company but there was nowhere I could find the waterblocks for my GPU's except China. Everything was in pretty good shape except they didn't include any thermal pads and one of my radiators has a bit of a dent on one end but its not a problem.

Got everything put together and it all powered up but my temps were absolutely terrible and it was throttling... Figured out I wasn't supposed to put the plaster washers on both side of the PCB and because I did it wasn't getting good pressure or spreading out my TIM at all. I fixed that on one of the cards and did a test mount and it's looking way way better so I think that was the problem.

Last time i did "custom" water it was a swifttech h20 220 kit and it was CPU only so this is my first foray into GPU cooling.

RadeonVII_Loop.jpg
 

bal3wolf

Limp Gawd
Joined
May 14, 2008
Messages
459
Got to say these compression fitting are so much easier to work with then the old school barb + hose clamp system from when i last did this.
for sure and much safer only issue tight spaces they are hard to twist and with me i have carpel tunel so they can be a pain i use a mechanics glove to give myself grip to turn them easier.
 

Deimos

Gawd
Joined
Aug 10, 2004
Messages
1,000
You probably should run those cards in serial instead of parallel. You will get a small drop in overall flow rate but your temps will be better over all.

The problem with running the cards in parallel is that the water has a lot more time to heat up inside the block, which means parts of the card will end up much hotter and cooling is less efficient. It will be worse with three cards (for every card you have in parallel, the flow rate drops by 50%). I noticed a significant difference between parallel and serial when I had two cards.

If you can feel a difference in temps at the inlet (before the cards) and outlet (after the cards) with your hand, you don't have enough flow through the blocks to cool it efficiently.

At the very least, test it and let us know what you find.
 

guitarslingerchris

Supreme [H]ardness
Joined
Oct 29, 2004
Messages
7,151
You probably should run those cards in serial instead of parallel. You will get a small drop in overall flow rate but your temps will be better over all.

The problem with running the cards in parallel is that the water has a lot more time to heat up inside the block, which means parts of the card will end up much hotter and cooling is less efficient. It will be worse with three cards (for every card you have in parallel, the flow rate drops by 50%). I noticed a significant difference between parallel and serial when I had two cards.

If you can feel a difference in temps at the inlet (before the cards) and outlet (after the cards) with your hand, you don't have enough flow through the blocks to cool it efficiently.

At the very least, test it and let us know what you find.
This is not accurate.
 

cdabc123

2[H]4U
Joined
Jun 21, 2016
Messages
3,523
This is not accurate.
I'm going to have to agree with you on this. Good cooling can still be easily obtained with minimal flow. Sure you are heating up any components downstream of the core but not to any degree that matters. Most of my builds run all components in parallel with relatively low flow.
 

motqalden

[H]ard|DCOTM x4
Joined
Jun 22, 2009
Messages
2,317
You probably should run those cards in serial instead of parallel. You will get a small drop in overall flow rate but your temps will be better over all.

The problem with running the cards in parallel is that the water has a lot more time to heat up inside the block, which means parts of the card will end up much hotter and cooling is less efficient. It will be worse with three cards (for every card you have in parallel, the flow rate drops by 50%). I noticed a significant difference between parallel and serial when I had two cards.

If you can feel a difference in temps at the inlet (before the cards) and outlet (after the cards) with your hand, you don't have enough flow through the blocks to cool it efficiently.

At the very least, test it and let us know what you find.
This is not accurate.

Yeah there seems to be a lot of back and forth about this on various articles and video I've seen and a lot of it depends on how restrictive your blocks are but in general with 3 blocks or more I was reading that you want to at least go partially parallel so that you are not too restricted I have not measured my flow rate but water is returning to the reservoir pretty fast from an "eyeball" test.
But I have also read that at the end of the day the difference is only a couple degrees in most cases and its a pain to tear things apart. My temps so far have only been tested mining ETH at 960mv but my core temps are low 40's and both my junction and memory temps have stayed around or just under 50c. To me this is already good enough that I am not going to sweat the specifics and just run with it. When these were on air I needed to run the fans at 60-70 just to keep my JT from hitting 110 and my memory was usually hitting 90c. I did this build mostly to keep the noise down since the stock coolers are absolute garbage and really annoying and i was getting tired of it especially when it was hotter outside.
 

cdabc123

2[H]4U
Joined
Jun 21, 2016
Messages
3,523
Yeah there seems to be a lot of back and forth about this on various articles and video I've seen and a lot of it depends on how restrictive your blocks are but in general with 3 blocks or more I was reading that you want to at least go partially parallel so that you are not too restricted I have not measured my flow rate but water is returning to the reservoir pretty fast from an "eyeball" test.
But I have also read that at the end of the day the difference is only a couple degrees in most cases and its a pain to tear things apart. My temps so far have only been tested mining ETH at 960mv but my core temps are low 40's and both my junction and memory temps have stayed around or just under 50c. To me this is already good enough that I am not going to sweat the specifics and just run with it. When these were on air I needed to run the fans at 60-70 just to keep my JT from hitting 110 and my memory was usually hitting 90c. I did this build mostly to keep the noise down since the stock coolers are absolute garbage and really annoying and i was getting tired of it especially when it was hotter outside.
How quiet is it? With those large radiators oriented like that I would almost be tempted to let the water temp go up abit and just let the heat rise out of the build (with minimal airflow)
 

motqalden

[H]ard|DCOTM x4
Joined
Jun 22, 2009
Messages
2,317
How quiet is it? With those large radiators oriented like that I would almost be tempted to let the water temp go up abit and just let the heat rise out of the build (with minimal airflow)

I have not gotten far enough yet where I have gotten to play around with fan speeds but from how cool the air is coming off my rads it feels like they are barely breaking a sweat and I think i will be able to run them at around 1000RPM (instead of the stock 1800) and still be able to keep things quite comfortable. The Artic P12's are not really that loud even full blast to me, but its also hard to tell how loud it all is since the rig is currently in a small hallway with 4 tower systems all running similar fans and a bunch of GPU's humming along. I still need to tear everything apart again as the last time i put it together i must have mucked up one of the cards or forgotten to re-past it since two are running great and one is hitting 110 JT and throttling like mad almost instantly. maybe I missed a washer removal somewhere and I was pretty tired a the time :D
 

motqalden

[H]ard|DCOTM x4
Joined
Jun 22, 2009
Messages
2,317
I think once the temps cool off a bit more outside I will be able to do exactly what you are saying and turn off everything but the pump. I usually keep the front door and a window cracked all winter as temps are usually not below freezing here in Vancouver.
 

Deimos

Gawd
Joined
Aug 10, 2004
Messages
1,000
its a pain to tear things apart.
Not really. The way you have it set up, you could just remove three tubes and plug the ports, you'd be done fairly quickly, its the draining and refilling that would take longest (if you need to drain, that is).

This is not accurate.
Based on what? Flow rates make a difference to cooling performance. I tested parallel vs serial running 1080ti in SLI with two D5 pumps + 3x140, 3x120 and 4x120 rads and there was a significant difference. What testing have you done?
 

guitarslingerchris

Supreme [H]ardness
Joined
Oct 29, 2004
Messages
7,151
Not really. The way you have it set up, you could just remove three tubes and plug the ports, you'd be done fairly quickly, its the draining and refilling that would take longest (if you need to drain, that is).


Based on what? Flow rates make a difference to cooling performance. I tested parallel vs serial running 1080ti in SLI with two D5 pumps + 3x140, 3x120 and 4x120 rads and there was a significant difference. What testing have you done?
How about 1080ti SLI with a single D5, 2x 360s and a 480 and the same with two 980 Tis and several other SLI variations over the almost 20 years now of watercooling.

The water spending longer in the waterblocks separately is almost guaranteed to be a net gain if the flow isn't abysmal and you have enough rad/fans in the loop versus one card getting hotter water that doesn't allow it to shed as much heat.
 

hititnquitit

[H]ard|Gawd
Joined
Apr 16, 2018
Messages
1,294
Martinsliquidlab put this to rest 12 or 13 years ago if i remember correctly. I cant be bothered to find the exact date or the link.
There's no temperature benefit. If you maintain 1gph(or right around that) you wont see a difference.
Just do you. If you prefer parallel, rock it. If not, don't.
 

lobstar

Limp Gawd
Joined
Nov 24, 2004
Messages
284
The water spending longer in the waterblocks separately is almost guaranteed to be a net gain if the flow isn't abysmal and you have enough rad/fans in the loop versus one card getting hotter water that doesn't allow it to shed as much heat.
It may seem counter intuitive but this is not the way the physics of the system works. You need water volume over time and surface area. If you assume a system is balanced (the heat plate can transmit energy at a sufficient rate to the coolant and the rads have enough surface area, etc) these are the only things that can make a difference. Rads only increase the dissipation of the heat from the closed loop, again, by available surface area and flow. So all things equal we're left with flow.

I recently took my loop apart because of a rubbery growth afflicting all of my blocks.

PXL_20210821_192448362.jpg

I was down to .2gpm after about a week of it dropping pretty steadily so I got to see this first hand. My water temps skyrocketed from an average of 28c while gaming to 42c. My GPU was hitting 60c at times (took me a few days to get pads). And that's with 2 huge Monsta 480s with 12 fans and a MORA3 with 4 200mm fans and 4 D5 pumps.

Long story short: your water is one 'thing' in your loop. Over time it's temperature will equalize the entire system up to it's total energy capacity. Test this yourself with a temp sensor at the intake side of the reservoir and one at the inlet of your rads. My pair are only 1c difference even with that low flow described above.
 

alxlwson

You Know Where I Live
Joined
Aug 25, 2013
Messages
8,173
It may seem counter intuitive but this is not the way the physics of the system works. You need water volume over time and surface area. If you assume a system is balanced (the heat plate can transmit energy at a sufficient rate to the coolant and the rads have enough surface area, etc) these are the only things that can make a difference. Rads only increase the dissipation of the heat from the closed loop, again, by available surface area and flow. So all things equal we're left with flow.

I recently took my loop apart because of a rubbery growth afflicting all of my blocks.

View attachment 390430
I was down to .2gpm after about a week of it dropping pretty steadily so I got to see this first hand. My water temps skyrocketed from an average of 28c while gaming to 42c. My GPU was hitting 60c at times (took me a few days to get pads). And that's with 2 huge Monsta 480s with 12 fans and a MORA3 with 4 200mm fans and 4 D5 pumps.

Long story short: your water is one 'thing' in your loop. Over time it's temperature will equalize the entire system up to it's total energy capacity. Test this yourself with a temp sensor at the intake side of the reservoir and one at the inlet of your rads. My pair are only 1c difference even with that low flow described above.

Biologic or plasticizer?
 

lobstar

Limp Gawd
Joined
Nov 24, 2004
Messages
284
Biologic or plasticizer?
I have Tygon rubber tubing and was using Mayhem's Nuke XT-1 concentrate. The only thing I can think of is it was biological OR something related to my Little Giant pump since it has a white housing and compressor wheel and was making some noise before I pulled it. But I seriously doubt a pump design for lifting chemicals had a reaction with propylene glycol and whatever else is in Nuke. I'm running just the D5s for now.

PXL_20210821_205506972.jpg
 

hititnquitit

[H]ard|Gawd
Joined
Apr 16, 2018
Messages
1,294
I have Tygon rubber tubing and was using Mayhem's Nuke XT-1 concentrate. The only thing I can think of is it was biological OR something related to my Little Giant pump since it has a white housing and compressor wheel and was making some noise before I pulled it. But I seriously doubt a pump design for lifting chemicals had a reaction with propylene glycol and whatever else is in Nuke. I'm running just the D5s for now.

View attachment 390535
It was probably the tygon, or the tygon, nuke combination. Im not certain which. I stopped using tygon because of white buildup in my loops. I tried 3 different kinds and ended up having the same white crap in my tubes and blocks. I tried 1/2inx3/4in clear, silver anti-microbial and black uv resistant. The black took the longest to cloud up, the clear the fastest. I switched to primochill lrt and pt-nuke and never had it again(6-7yrs).
Switched to hardline and koolance fluid after that and haven't seen it in this loop either.
So my advice is flush your rads and blocks with clr (5 mins sit, then shake and rinse really well). Scrub your pump and res with distilled and dish soap and replace your tubing with something other than tygon.

Edit- from your description in the other thread that isnt any kind of build up ive ever seen or heard of. If its spongy and bounces back it may be part of your old pumps gasket or a part of something that was glued/caulked or possibly excess. If you can take it apart thats where i would start.
 
Last edited:

guitarslingerchris

Supreme [H]ardness
Joined
Oct 29, 2004
Messages
7,151
It was probably the tygon, or the tygon, nuke combination. Im not certain which. I stopped using tygon because of white buildup in my loops. I tried 3 different kinds and ended up having the same white crap in my tubes and blocks. I tried 1/2inx3/4in clear, silver anti-microbial and black uv resistant. The black took the longest to cloud up, the clear the fastest. I switched to primochill lrt and pt-nuke and never had it again(6-7yrs).
Switched to hardline and koolance fluid after that and haven't seen it in this loop either.
So my advice is flush your rads and blocks with clr (5 mins sit, then shake and rinse really well). Scrub your pump and res with distilled and dish soap and replace your tubing with something other than tygon.
This, I dropped Tygon back in the 2000s due to various issues over time and Primochill LRT has been my favorite to use by far.
 

DWolvin

2[H]4U
Joined
Nov 13, 2002
Messages
2,755
Oh, I internally autotranslated to ZMT... :ROFLMAO: And and I get it if it's a clear tubing issue, that is surprisingly common from what I understand. Now I need to go look up what the clear tube I ended up with is... (Duraclear I think)
 

FearTheCow

Supreme [H]ardness
Joined
May 2, 2006
Messages
5,769
Replaced my single xspc pump/res with 2 ultitubes, they pull from the central reservoir, one does cpu the other gpu, then each runs to its own 560mm rad, they then return to the central resevoir.

I was able to reduce my fan and pump speeds to inaudible and haven't hit above 35c while gaming. I was having to run the single pump at full speed and the fans at higher speeds and had higher temps.

The next step is to do hard tubing and an aquaero.
 

Attachments

  • 20210909_175706.jpg
    20210909_175706.jpg
    455.9 KB · Views: 0

Solhokuten

[H]ard|Gawd
Joined
Dec 9, 2009
Messages
1,391
Here is the finished product for the time being. While I still intend to go with metal tubing (post 1744), I had issues chamfering the tubes correctly and ended up destroying a few inner o-rings. Additionally, I only have one cross fitting on hand (used to straighten out the runs). While packages generally take 4-6 weeks to get here, I figure why not use the ZMT I had laying around.

I will tackle the metal tubes again once the next GPU comes in.

O11 Dynamic XL
Asus X570 Tuf Gaming
3900X
32GB Trident Z
EVGA 1080 Ti FE w/EKWB water block
Rads - EKWB Coolstream PE/XE w/NF A12x25 (both set to intake)


Idle temps are
CPU 34C
GPU 24C

25 min full load w/Prime 95 (small ffts) & MSI Kombustor (burn-in)
CPU 71C
GPU 41C






View attachment 276321View attachment 276322View attachment 276323
Nice rig and temps. I have my top rad set as exhaust. Now I'm thinking I should try it as an intake as well.
 
Top