School me on custom loops.

Question for you guys is it worth it to do a push pull fan set up on the top rad? Or just one side and it doesn’t make a difference?
 
Question for you guys is it worth it to do a push pull fan set up on the top rad? Or just one side and it doesn’t make a difference?

P/P nets just under a degree of delta on a 360 on average (data from martins years ago). Ofc, there's the cost of fans, loss of space, fitment etc so that's up to you. It's generally not worth it overall.
 
Question for you guys is it worth it to do a push pull fan set up on the top rad? Or just one side and it doesn’t make a difference?
Unless you have a super thick radiator, only one side would really be needed. And as to which side of the radiator to put it on, I usually put the radiator against the case and the fans below blowing up through the radiator when it's on top.
 
Question for you guys is it worth it to do a push pull fan set up on the top rad? Or just one side and it doesn’t make a difference?
Here’s my reasoning for both fans being intake...

If both top and bottom are set to intake they will cool the radiator with ambient air. In my case that is 24c intake with 28/29c exhaust, that air is forces out of the case because of positive pressure.

If I swap this around the bottom radiator will get nice 24c air and the top unit will be cooled with 27/28c air from within the case.


Overall, I think the loop temp will be 1C different (maybe?) after the water normalizes. But I am currently logging both bottom/top as intake and will change in December to intake bottom/exhaust top.

By the end of the month I will have logs of both setups to show which one offers better cooling. Though, I do not think it will make that much of a difference.

Edit: okay, missed the entire point of the post. I used P/P on the Black Ice Nemesis 560 GTX, think it was 50/60mm thick. For the the O11, just rads against the case and fans pulling through them.
 
Can confirm mo-ra3 is great. That’s what I have in my basement and at one point cooled 6 GPUs and 2 CPUs no problem lol. Way better than screwing around with fitting radiators in a case.
How do you deal with the tube length and pump getting the fluid flowing? I’m assuming the PC isn’t in the basement?
 
Question for you guys is it worth it to do a push pull fan set up on the top rad? Or just one side and it doesn’t make a difference?
My anecdotal experience tells me that if you have enough radiator in your loop and good airflow, push/pull does not net you much and is certainly not worth the extra fan cost. Now if you are in a situation where you have to brute force airflow then it will make a difference. The other way it will can be beneficial is if you have barely enough radiator(a single 360 cooling cpu and gpu) and that extra airflow gets you over the top, so to speak. However, in that instance, IMO, you will get better results adding a single 120mm rad to the loop.
 
Push/pull generally only yields a couple of degrees improvement. The only time its really useful in my experience is if your rad is either really thick or has super high fpi and a single row of fans cant push air thru it effectively. Altho if your trying to push thru a thick rad and a filter, you may benefit from a push pull configuration.
 
How do you deal with the tube length and pump getting the fluid flowing? I’m assuming the PC isn’t in the basement?

The pump is in the basement. A D5 has no issue pumping it ~8ft up and once the line is filled the vertical height has no effect.

Pump/rad/res all in the basement.
 
The pump is in the basement. A D5 has no issue pumping it ~8ft up and once the line is filled the vertical height has no effect.

Pump/rad/res all in the basement.
Is that true? If so that is interesting information. Once the loop is full height traveled has no effect?
 
Is that true? If so that is interesting information. Once the loop is full height traveled has no effect?
It does sound special but for a closed loop it is apparently because when the water fall all the pressure gained compensate the one needed to elevate the water at the same time:
https://www.eng-tips.com/viewthread.cfm?qid=393312
In a closed loop system, like in a chilled water circuit the height of the building is irrelevant because the work done(Pressure) by the pump in lifting the water column is gained when the water column falls by the same height. Hence consider only the dynamic pressure head only, i.e the pressure that the pump has to develop to circulate the fluid overcoming the friction loss and fitting pressure loss. Also add the pressure loss in the most remote AHU or FCU. That will give the Head required by the Pump.

In a closed loop system, the pump discharge head is never selected on the basis of the building height but on the total system pressure loss

The height is important in the sense of adding distance but it would be the same if it was putting the radiator in a far away garage at the same height, not the verticality of it.
 
It does sound special but for a closed loop it is apparently because when the water fall all the pressure gained compensate the one needed to elevate the water at the same time:
https://www.eng-tips.com/viewthread.cfm?qid=393312
In a closed loop system, like in a chilled water circuit the height of the building is irrelevant because the work done(Pressure) by the pump in lifting the water column is gained when the water column falls by the same height. Hence consider only the dynamic pressure head only, i.e the pressure that the pump has to develop to circulate the fluid overcoming the friction loss and fitting pressure loss. Also add the pressure loss in the most remote AHU or FCU. That will give the Head required by the Pump.

In a closed loop system, the pump discharge head is never selected on the basis of the building height but on the total system pressure loss

The height is important in the sense of adding distance but it would be the same if it was putting the radiator in a far away garage at the same height, not the verticality of it.
Right, I just did not want to put words in someones mouth. My understanding is that total loop length does matter, so the 8 feet does matter in terms of flow rate at the end of the loop right? I am not saying a d5 cant handle it but there is an effect.
 
Right, I just did not want to put words in someones mouth. My understanding is that total loop length does matter, so the 8 feet does matter in terms of flow rate at the end of the loop right? I am not saying a d5 cant handle it but there is an effect.
It matter for adding friction with the tubing for 8 feet, it is just that having the 8 feet vertical, instead of horizontal should not be a factor.
 
Right, I just did not want to put words in someones mouth. My understanding is that total loop length does matter, so the 8 feet does matter in terms of flow rate at the end of the loop right? I am not saying a d5 cant handle it but there is an effect.
Length = restriction. There are friction losses in any kind of pipe or tube. So yes, a long loop will be more restrictive than a short one, but D5s and DDCs are lovely in that they're so overpowered you don't really need to worry about that.
 
Another nice thing about push/pull is lower fan speeds, equals lower noise. I've done it on some 60mm rads, a few degrees cooler. That said general consensus is go with the 30mm rads and push configuration. Easier to fit in most cases and performs almost as good.

For many of us it's not the temperatures but the noise we're after
 
Another nice thing about push/pull is lower fan speeds, equals lower noise. I've done it on some 60mm rads, a few degrees cooler. That said general consensus is go with the 30mm rads and push configuration. Easier to fit in most cases and performs almost as good.

For many of us it's not the temperatures but the noise we're after
Plus adding another radiator will always be more effective than adding another row of fans.
 
Plus adding another radiator will always be more effective than adding another row of fans.
That works until you’ve run out of room to add more internal rads and don’t want to go external.

I’m doing push + pull this time around on both my rads. While my last build worked, the air flow with just one set of push GT AP-15s was anemic after dealing with rad, filter, and case perf restrictions. Switching to 38mm deltas resolved the airflow issues, but to get good airflow the system was no longer quiet. I ended up going back to the AP-15s and adding an external mounted push +pull 120mm black ice extreme from my parts pile from way back in the day (2001). That got the temps where I wanted them, but it ended up causing some pet related problems (tail fur caught in fan more than once).
 
I run push/pull, but that's because I use a GTR 480mm rad that would need some very high speed fans to be effective.

So I doubled up to keep the fans slower and lower in db.
 
Push pull really depends on how restrictive the airflow of the rad is.

The BI 240GTSs I had thrived on push pull, but those rads just have really high airflow restriction. The XSPC rx360 I have been using for the last few years does as well with 3 fans in pull, as it does with 6 in push/pull. It just has very low restriction air flow.
 
Honestly it's down to aesthetics. As long as you're buying from a reputable brand you don't really need to worry about quality.

Determine what kind of pump you intend to use, the port arrangement you need, and the size you can fit, and go from there.

I'm personally very fond of Watercool's Heatkiller line. Excellent quality, nice thick glass, modular and flexible. There are other good options out there but that's my favorite.
I'm on my third Watercool res/d5pump. I think the guy that started this thread was just bored. He's laughing at the fan boys.
 
I have always gone AIO's and with COVID being the way it is I have had too much time to look at PC stuff this year. I want to build a system in a Lian Li PC O11 Dynamic Mini (I think its coming out within the next week). I looked at Corsair stuff but that was when I was thinking about doing it in an ITX system and was forced to use the Commander Pro because the lack of fan headers. Now that I am not restricted to that since I can use an ATX board I was wonder what you guys recommend.

System Build:
AMD R9 5900x
AMD RX6800XT or 6900XT
4X 32GB G.Skill Trident Z Neo 3200MHZ CL16
MSI Meg X570 ACE
Corsair SF750

I plan on water cooling both the CPU and GPU. I don't know how much rad I actually need. I wont be overclocking too much.

I already ordered one HWLabs Nemesis 360GTS.
I also ordered the Corsair XD3 pump res combo and XC7 block but now I am unsure if I want to use them.

1.) I was thinking some sort of EK pump res combo or their distro plate. But I don't understand what the different type of pumps that the come with.
2.) Whats D5 vs DDC (3.1 vs 3.2?) etc.
3.) I see that most things use g 1/4 thread. What fittings do you guys recommend. Also I want to learn more about Quick Disconnect fittings (I want to use one for a drain port).
4.) What ID/OD do you guys recommend for PETG tubing?
5.) Coolant and Kill coils etc. I want to go for clear liquid.
6.) I hear about cleaning out fittings and rads before hand what do you guys do to clean it out?
7.) I hear its bad to mix metals so does that mean if I get one copper block I have to get all copper blocks vice versa with nickel blocks?

Any help is greatly appreciated. In terms of budget I was thinking $1k~ for the water cooling stuff. If thats unreasonable I can increase it.
I have had for three years a Watercool Heatkiller cpu TR block/200mm res/D5 pump/EK fittings and 12x16 tubing/EK 45x140x420 rad. NO runs,NO drips and NO errors. On my dinning room table I have two new Watercool 150mm res/D5 pumps, a new EK 420mm Rad and an EK quantum Vector Radeon VII gpu block for a gpu loop. For a cpu run your pump off the cpu fan pwm header and the rad fans off the aux cpu fan header. For a GPU use the GPU's two fan headers for the PWM signal to that loops pump and fans.. Any fan controller only uses the pwm signal, the fan header/pump are powered by a 4pin molex. The new EK fan controller looks great but it boots to Windows 10. I don't see it working on a linux distro. If you wait till the coolant is hot the cpu/gpu will have already throttled back.
 
So I looked at the aquaero since some of you guys really talked it up and I ended up buying a 6 LT. I bought a temp sensor plug to use in the distro plate and a flow sensor what other sensors do you guys recommend?
 
Working on it. Found out it’s a lot harder than I thought and I wasted a whole meter of tubing starting out. I was able to salvage some for the shorter runs not completely happy with the CPU inlet run though. SFF is basically done though.

Leak testing right now with the ek tester and then I will do a flush with primochill sysprep and then put in coolant if everything is ok.

59BDD866-5E84-4166-A642-78BC483E33CD.jpeg
 
Put in the sysprep and so far no issues. Do you you really have to run it for 12 hrs?
 

Attachments

  • 133C62C2-9973-4BFE-ADC4-10F3B58C5A86.jpeg
    133C62C2-9973-4BFE-ADC4-10F3B58C5A86.jpeg
    297.5 KB · Views: 0
Put in the sysprep and so far no issues. Do you you really have to run it for 12 hrs?
Maybe not 12, but the longer it runs the more bacteria and stuff is killed. Same with household disinfectant–it only really works if you leave it on the surface for a while.
 
Ive never used it and never had any nasties in my loop. As long as your not licking your fittings, poking your tim covered pinky into one or coughing into your res a normal flush with clean parts should be plenty. With a proper biocide mix you shouldn't need to run it for more than a few hours unless you're worried something may have gotten contaminated.

Just struck me that your likely asking about the leak test...haha. Its a good idea to run it for a minimum of 8 to 24hrs. Even if you plan on being right next to it while it runs. Its really easy to miss that little drip thats landing on your mb or gpu every half hour.
 
Is that leak tester worthwhile? Seems like you could use it to do leak testing in 15-30 minutes rather than 12+ hours?
 
I love air pressure leak testers. It is one of those things that once you use once, you ask why you never used one before.

I have the Aquacomputer DrDrop version. The EK one may be better, just because it is made of aluminum, whereas the Dr Drop is delrin.

I check individual components with one for 15-20 minutes. Like if you need to take a gpu block apart to clean, you will know immediately if a oring is not seated correctly.

I put mine on pressure for 2 hours, if it holds pressure then just fill with coolant and bleed.
 
Is that leak tester worthwhile? Seems like you could use it to do leak testing in 15-30 minutes rather than 12+ hours?
It seems that every euro/british water cooling build video/guide recommends a leak tester before adding liquid. I have never used one, just paper towels and a few hours time. Just another case of us backwards Americans?:eggface:
 
Thats a good point Epos7. With the leak tester your pretty much gtg.
Ive actually got one of the dr drop testers but ive yet to use it. I either forget i have it or cant find it so i just go old school. I swear i will use it next time tho!
 
  • Like
Reactions: Epos7
like this
It seems that every euro/british water cooling build video/guide recommends a leak tester before adding liquid. I have never used one, just paper towels and a few hours time. Just another case of us backwards Americans?:eggface:
Leak testers are a great tool, but people were making custom loops for a long time without them.

I tend to dislike them because I feel like they are likely to encourage sloppy habits, but I also don't have one so my opinion is likely skewed.
 
Ive never used it and never had any nasties in my loop. As long as your not licking your fittings, poking your tim covered pinky into one or coughing into your res a normal flush with clean parts should be plenty. With a proper biocide mix you shouldn't need to run it for more than a few hours unless you're worried something may have gotten contaminated.

Just struck me that your likely asking about the leak test...haha. Its a good idea to run it for a minimum of 8 to 24hrs. Even if you plan on being right next to it while it runs. Its really easy to miss that little drip thats landing on your mb or gpu every half hour.
My bad, I guess I didn't specify in my post. I actually meant the sysprep. EK says to do the pressure test for 15 mins per component. For me I like the leak testers because it gives me a bit more peace of mind. I don't want to spill water all over my new build that I just spent serious $$$ on and its my first ever custom loop. But EK didn't QC mine correctly and the gauge insert wasn't properly put on so I can only test to about .3 bar before the needle stops moving because the insert touches the needle. Still worked well had no leaks once I put the sysprep in. I am going to drain it soon once the 12 hrs comes up and then wire up the aquaero. I do need to get a fan hub though I am short 2 fan headers on the 6 LT.
 
Crazy how fast the costs add up on a custom loop. I was budgeting $1k then raised it to 1.2k. Then I tallied up all the purchases directly for the loop and I ended up spending almost 1.5k.
 
It definitely adds up quick! The positive is being that wcing tech moves really really slow, your wcing gear will outlast pretty much everything in your next 3 or 4 builds. Ive got a rad thats at least 15yo and a couple of ddcs that are 10 or 12. Much like pc gear youll inevitably end up with a closet full of spare wcing stuffs as you change this n that in the quest for visual and cooling perfection.
For the most part youll only need to buy fittings and tubing for a while. Until the need to buy a new case strikes of course 😉
 
The positive is being that wcing tech moves really really slow, your wcing gear will outlast pretty much everything in your next 3 or 4 builds. Ive got a rad thats at least 15yo and a couple of ddcs that are 10 or 12. Much like pc gear youll inevitably end up with a closet full of spare wcing stuffs as you change this n that in the quest for visual and cooling perfection
That is a positive but, for me at least, not exactly a cost savings. You said it yourself, "visual and cooling perfection." Being into watercooling can save you money if you are into hardware as well by giving you an excuse to upgrade less often. However, that assumes you are into the "visual" part of watercooling. Redoing your loop can scratch the itch of a yearly upgrade and give you an excuse to refresh your rig between cpu and gpu generations. A flashy new res or fans or a cool looking gpu block, while not cjeap, is cheaper than a new system. Trying out new cases and new runs can be as fun as getting a new gpu. My point is that watercooling can be cheaper on your wallet over time if you are into the visual.
 
So I put in the EK Cryofuel Clear coolant a few days ago and I thought it would clear over time and while it did it was still quite a bit cloudy. So I drained it and it seemed to have left quite a bit of residue so I tried to flush it with distilled water and that helped a bit but wasnt quite as I wanted so I bought the primo chill reboot and it helped quite a bit. But it seems like I need to disassemble my distro plate and clean it to get it as clear as I think it should be given how new it is.

My question anyways: do you guys have recommendations of coolant (clear preferably) that doesn't have clouding and particulate issues.
 
So I put in the EK Cryofuel Clear coolant a few days ago and I thought it would clear over time and while it did it was still quite a bit cloudy. So I drained it and it seemed to have left quite a bit of residue so I tried to flush it with distilled water and that helped a bit but wasnt quite as I wanted so I bought the primo chill reboot and it helped quite a bit. But it seems like I need to disassemble my distro plate and clean it to get it as clear as I think it should be given how new it is.

My question anyways: do you guys have recommendations of coolant (clear preferably) that doesn't have clouding and particulate issues.
Best clear coolant: distilled water with kill coil. Running 5 years on the current loop with only annual top offs and everything still looks great.
 
So I put in the EK Cryofuel Clear coolant a few days ago and I thought it would clear over time and while it did it was still quite a bit cloudy. So I drained it and it seemed to have left quite a bit of residue so I tried to flush it with distilled water and that helped a bit but wasnt quite as I wanted so I bought the primo chill reboot and it helped quite a bit. But it seems like I need to disassemble my distro plate and clean it to get it as clear as I think it should be given how new it is.

My question anyways: do you guys have recommendations of coolant (clear preferably) that doesn't have clouding and particulate issues.
I'm a fan of Aquacomputer Double Protect Ultra.
 
So I put in the EK Cryofuel Clear coolant a few days ago and I thought it would clear over time and while it did it was still quite a bit cloudy. So I drained it and it seemed to have left quite a bit of residue so I tried to flush it with distilled water and that helped a bit but wasnt quite as I wanted so I bought the primo chill reboot and it helped quite a bit. But it seems like I need to disassemble my distro plate and clean it to get it as clear as I think it should be given how new it is.

My question anyways: do you guys have recommendations of coolant (clear preferably) that doesn't have clouding and particulate issues.

Did you have any kind of coolant in there before? A clear coolant shouldn't turn cloudy.

Did you use any kind of pre cleaner in you loop? New radiators at least should be flushed before use, they will almost always have residual stuff in them from the manufacturing proces to include random grit and junk.

I honestly wouldn't buy any of EKs coolants. I had issues with thier opaque stuff a couple of years ago, but even their clear coolant shouldn't get cloudy on its own.

I like XSPC ec6 clear and aquacomputers Double protect.

Distilled works as well, but if you have any non nickel plated components, like the copper cpu plate or copper/brass then it really stops being distilled after a while as it strips ions off the copper/metal blocks. That is just reality. I used distilled/silver for years as well, it is definitely the cheapest route but nowhere near the best.

Going distilled is fine, just add in some form of corrosion inhibitor.
 
Last edited:
Back
Top