need your suggestion/expertise on my single loop pissing flow issue

[MadHacker]

well I'm a redundancy kind of person.
i currently have 2 pumps in each of my rad boxes that goes to each machine.
so just thinking about it i would then need to have 2 more pumps.
for a total of 6 pumps.
2 pumps for the radBox to tank, 2 pumps to each of the machines it watercooling.
almost seems kinda ridiculous. but if any 1 pump fails... well nothing dies.

edit: for the extra cost... what does it buy me?

[NaeKuh]

well I'm a redundancy kind of person.
i currently have 2 pumps in each of my rad boxes that goes to each machine.
so just thinking about it i would then need to have 2 more pumps.
for a total of 6 pumps.
2 pumps for the radBox to tank, 2 pumps to each of the machines it watercooling.
almost seems kinda ridiculous. but if any 1 pump fails... well nothing dies.

edit: for the extra cost... what does it buy me?

no no no your looking at the system as 1 big system.

When instead its 2 active systems fighting for equalibirum.

The system which is tank -> rad -> tank... can be kept small, and on bigger diameter tubing.

since flow controls how much heat is carried and dump'd, obviously bigger tubing, using a fish pump to a car rad, or some type of rad of that size is going to have a tiny flow drop.

This is system 1.

System 2, is from the tank to your computers.

Now assume... each of your system is taking 1gpm of flow..
That means 2 systems u got 2 gpm of flow going back into the tank.

Your tank will have way more X gpm of flow to the rad and back, if u designed it right.


As you can see your tank will be kept at close to ambient at all times unless u have an insane massive heat dump going on.

[MadHacker]

no no no your looking at the system as 1 big system.

When instead its 2 active systems fighting for equalibirum.

The system which is tank -> rad -> tank... can be kept small, and on bigger diameter tubing.

since flow controls how much heat is carried and dump'd, obviously bigger tubing, using a fish pump to a car rad, or some type of rad of that size is going to have a tiny flow drop.

This is system 1.

System 2, is from the tank to your computers.

Now assume... each of your system is taking 1gpm of flow..
That means 2 systems u got 2 gpm of flow going back into the tank.

Your tank will have way more X gpm of flow to the rad and back, if u designed it right.


As you can see your tank will be kept at close to ambient at all times unless u have an insane massive heat dump going on.

right but your still snafu if 1 pump fails without redundancy

[NaeKuh]

right but your still snafu if 1 pump fails without redundancy

lol... thats true on any system.

Now u see why in my system.. 5/6 pumps alone with 2 complete loops would have to fail b4 i reach a oh snap situation.

[NapalmV5]

I assumed 3/8 tubing. By looking at that chart you can see that the max head is 15ft -from that you can get an outlet velocity and then a pressure drop across 74ft of tubing.. I didn't take into consideration any of the 90's/blocks. I was just curious if it could make it through the tubing.

lol indeed it does though at low pressure

Well, if you think the CPU block is the culprit, pull it out and make sure it is not clogged. I was having flow issues to the point my RD30 would barely give me 1.2Gpm. After troubleshooting the issue I found that the CPU block was clogged. Fixed my flow issue.

the thing about the ek supreme hf block is that the flow is greater thru the outlet much lower thru the inlet

have you guys used without a jet plate ?

id like to try without the jet plate.. on my system it hinders flow rather than better cool the cpu

ok guys.. when u have large areas u need to cover.. ie.. rad in another room, the best way to handle it is to use a TANK in the room with the computer, and have the tank feed 2 ways.

1 -> computer
2 -> radiator


having 70 feet in tubing in serial is like WOW... u trying to build a space elevator out of tubing?


Seriously... have the tank fed to the room.
Have a large rad cooling the tank constantly... basically feeding the radiator and back to the tank.

Then have your pumps feed from that tank to blocks and back to the tank.


The loop with the radiator will do all the heat exchanging for you without needed restrictions.
Then you can focus on smaller loops from the tank -> PC -> tank.

lol the other day when both rads were in the single loop it was more than 100 ft tubing

thats exactly how i have it now.. the eheim 1250s only serve the 2 rads which cools the tank which feeds the system loop

how big of a tank would you guys recommend for a fast mixture of the different liquid temps ? 1 2 gallon? 5 gallon tank?

[NapalmV5]

before i buy rd30 a few things.. why are you guys selling off your rd30? not good/powerful enough? whats up with rd30? whats the bad?

rd30 has 3/4" fittings so if i go with an rd30 than ill retube the 70-75 ft long current 1/2" tubing to 70-75 ft 3/4" tubing + 2x 3/4" to 5/8" adapters and its done!

or instead of rd30 go with a really big/industrial/automotive pump?

or would rd30 be suffice/offer proper flow?

[SNiiPE_DoGG]

before i buy rd30 a few things.. why are you guys selling off your rd30? not good/powerful enough? whats up with rd30? whats the bad?

rd30 has 3/4" fittings so if i go with an rd30 than ill retube the 70-75 ft long current 1/2" tubing to 70-75 ft 3/4" tubing + 2x 3/4" to 5/8" adapters and its done!

or instead of rd30 go with a really big/industrial/automotive pump?

or would rd30 be suffice/offer proper flow?

Nothing wrong with the RD-30 at all, its just overkill for most users. for you its EXACTLY what you need.

[Rubycon]

Your pump is too powerful. Yes you read that right! 1800 gph is a LOT of flow but you don't need that. In order to get water through your system you simply need more pressure. Those Danner pumps are for circulation - they are essentially a power head on steroids. 180W is a LOT of power as well which gets transferred to the loop!

You need a pump rated for more pressure but much lower flow rates - try 400 gph for starters. If you want a larger pump they make pressure rated pumps for injector (venturi) protein skimmers that will have a 60' head shut off pressure. Beware that you probably want to keep your operating pressure below 30 psig or so. Higher pressures encourage leaks/creeping. If you have a pump capable of developing significant pressure a relief valve plumbed into a bypass loop is a good idea.

Magnetic driven pumps need 120V sinusoidal power to run reliably. Unless it uses and electric motor to spin a cylindrical magnet around a well with the impeller.

Lots of hose/tubing will not offer excessive backpressure if the run is predominately horizontal. (providing the diameter is sufficient) 25mm ID hose for the lazy leg runs would be a good start, then adapt down to common sizes/fittings used for pc water cooling, etc.

[NapalmV5]

thanks sniipe_dogg thanks rubycon i appreciate it guys

yesterday afternoon went out to home depot to get 3/4" tubing ended up getting 1" maybe shouldve stuck to 3/4" idk i will see once i have a push and pull pumps.. hopefully it was all worth it

from system to the 1/2" to 1" compression in the corner ~9 ft x2 1/2" tubing from there to the pump/tank ~24 ft x2 1" tubing..overall shorten by 2-3 ft x2

[NapalmV5]

the 1" tubing the 1800gph pump doesnt like.. barely any pressure at all.. blue screens/reboots/etc. system stability went out the window.. had to go to the 3/4" fitting+tubing

does rd30 perform better on 1" or 3/4" tubing ?? if 3/4" ill have to go all 3/4" more $ ill have to spend

gotta love liquid cooling

[NapalmV5]

got one rd30 (definitely need another) today just finished up and heres how the flow looks like..

while the flow slightly better than the '1800gph' though @ higher pressure now and the noise level difference day and night.. overall way better than before..

when i turn on the flotec transfer pump things get moving.. as suggested i need another rd30

thank you all for all your suggestions much appreciated

[MadHacker]

did you ever check if any of your blocks are clogged?
if you don't want to do a full tear down...
just check your flow near your PC, use a bucket instead and just see how much flow you have going through.

[NapalmV5]

not yet.. i just reversed the flow back and forth..

alright i will try that later today..

thanks MadHacker

[NapalmV5]

got a second rd30 and as you can see the whole problem lies at the godamn cpu block theres no restriction at any of the other blocks

this problem will not go away by simply cleaning off the trapped gunk this block requires weekly maintenance

can i remove the jetplate ? the first time i cleaned off the block i replaced the factory one with the high flow one but the damned thing still clogs up

[wez]

Whats up with those deflated tubes? :O

[zalbard]

Yep, the tubing doesn't look right... Shouldn't affect performance much, I guess, though, but still...

[rge]

Holy crap, my wife would kill me if I did that.... looks like fun though.

Even if you remove the jet plate, whatever non-soluble crap is in your loop will clog the pins. I would use new distilled water only plus Pt nuke/biocide, and maybe 4 Koolance? quick connects and an inline high flow filter, run it for a few days with filter to trap any residual crap and suffer bad flow for few days while filter is in, and then via quick connects remove the filter and snap back together (use male/female for filter and line). If your tubing is not plasticizer free, then I wouldnt waste time and money doing that, since more plasticizer from tubing will be an ongoing problem. Also if you have thermochill rads, or flux problems, I would make sure rads are very clean.

and the compressed tubing does look like an issue, too thin walled?

[Vinas]

Looks like your tubs are being restricted by the suction pressure of your pumps. Try changing to something more rigid, and for the long runs use some PVC. Do some tests to see what works best!

[Annirak]

I think your problem is inlet restriction. A lot of centrifugal pumps can't handle restriction on the inlet. The D30's have just exacerbated the problem; they have more suction, but that just crushes the tubing and makes the inlet MORE restrictive. Looking at your tubing, I'm pretty confident that if you put a reservoir directly before your pumps (with very, very short hose runs, or pipes instead if you can do that) you'll solve the whole problem and, with that much pressure, maybe blow some gaskets)

[Edit:] I don't think the CPU block will cause a problem once you fix your inlet restriction. You might even be able to get away with a single loop.

[Wolf132]

Don't think I've ever seen tubing do what's in those pictures. And if your loop is clogging up weekly, you got other problems.

[UNTAMED]

I have a remote rad set up , not to your scale , 8'down -8'up
-I put 1 655 in the case and 1 at the rads to balance the push\pull pressure most pumps push better than draw.

-can't tell by the pic's if both pumps are at the top[ up stairs] or low ,[downstairs]
-Try moving one pump downstairs and one upstairs -I take from the pic's your upstairs and the tube crush[crushed by air pressure-not sucking] is from pulling the water up to that height maybe.
just saying
-maybe cpu[system]>inlet upstairs pump[1\2"]outlet>[1"]down to rads>[res.?]inlet lower pump >[1\2"]up to system[1"] ?-might make up for the thin wall tube-open for debate for sure.

[Annirak]

they have more suction, but that just crushes the tubing

[crushed by air pressure-not sucking]

The suction, though, is what reduces the internal pressure to the point where the air pressure can crush the tubing. But this is not relevant to the point. The point is that you need to have a low resistance path on the input of the pump. The easiest way to do this is to put a reservoir in front of it.

Also, the output of a pump is definitely low resistance, so cascaded pumps will not have an issue anywhere but on the first pump.

[UNTAMED]

[[The point is that you need to have a low resistance path on the input of the pump.]]
-just saying
-putting a pump up and one down cuts resistance in half\50% does it not ?
-where I,m at loss is that it's greater than 32'[I think] is the max any pump in the world can draw [@sea level?],so how -say a 2psi drop on the block's\tube relate[down-up]- scaled down in a closed loop on these low pressure pumps.

[Annirak]

-putting a pump up and one down cuts resistance in half\50% does it not ?
-where I,m at loss is that it's greater than 32'[I think] is the max any pump in the world can draw [@sea level?],so how -say a 2psi drop on the block's\tube relate[down-up]- scaled down in a closed loop on these low pressure pumps.

The resistance of a loop is the same no matter where you put the pumps. Think of it as pressure gains and pressure losses. The pumps are both pressure gains, any tubing or blocks are pressure losses.

Here's one weird thing:
*Water flowing UP a vertical tube incurs a pressure LOSS of the pressure of a water column the height of the tube PLUS the pressure loss of the tube.
*Water flowing DOWN a vertical tube incurs a pressure GAIN the pressure of a water column the height of the tube MINUS the pressure loss of the tube.

Which means a vertical loop, is exactly the same as a horizontal loop. Remember that absolute pressure usually doesn't matter, provided you don't do what the OP did and get the pressure so low that atmospheric pressure crushes your tubing. This is why siphons work.

Most of this is irrelevant since, from the pics he's posted, the OP's layout is completely flat.

Also remember that you can't have a block so restrictive that NO coolant flows. (well, you can, it's just a completely plugged one). Fluid resistances are dependent on the flow rate, so as the flow rate drops, so does the resistance.

[NapalmV5]

it was a good way to find the clog ^

last night i started the long haul of redoing everything all over again

the guys @ ek must really love this gunk stuff compacted in their cpu block.. i should pack it up and send it over

after thousands of $ wasted sleepless nights frustration to the wahoo i only have myself to blame since i didnt remove the plate right from the beginning as that little voice inside told me.. "idiota remove that crap" no no its called a jet plate and itll jet sheat out to the moon

right now im more concerned about the system not booting up after it booted earlier.. i thought my areca 1231 raid bally up i really thought it was gone.. just found out its still alive and kicking.. but i got no clue about the mobo cpu ram gpus.. if the system doesnt boot by tomorrow im fooked! i mustve missed a wet spot on the mobo idk cause it booted up and started acting up the raid card was saying check the pci bus.. then even my backup system didnt want to boot up only after many attempts.. its been a long 24 hours

well what can i say i got flow and pressure now @ only 1x iwaki rd30.. it cuts thru all the 80 ft tubing/8 blocks+sli bridge/countless 90s like knife thru butter

if no boot tomorrow.. on to giga x58a-oc