Custom Loop & Possible Flow and Fluid Temperature Issues

[MaxAvidy]

CPU: 5900X

MB  : X570

GPU: 3090

 

Hi fellow nerds,

 

My question is rather open ended as I'm not sure I actually have a problem, but my OCD will not be denied! I have thermal probes before and after my radiators, plugged into a Corsair Commander for temperature monitoring and fan curves. The question is: Should I be seeing a delta (difference between pre and post radiator probes) of 5 to 6 degrees Celsius? The reason I ask is that in some of LTT's videos on water cooling they talk about the line reaching equilibrium after warming up. For instance when they talk about loop order.

 

The system is in a Corsair 7000D Airflow, and the loop is: EK D5 Pump/Reservoir -> EK GPU cooler (front and rear water block) -> EK CPU water block -> first thermal probe -> EK 120mm Radiator -> EK 360mm Radiator -> EK 420 Radiator -> last thermal probe -> Reservoir. I've also made sure the pump isn't dialed back.

 

I've made the runs as straight as possible, but I'm sure there is some sort of turbulence created by some fittings, not to mention the various radiators and water blocks creating resistance. I also realize the cooling is probably overkill, but I had been wanting to make a custom water cooled system and go all out after saving for awhile. Not to mention watching LTT is a bad influence! 

 

Thanks,

 

The rainbow mess is setup to change the LED color of each component depending on their thermals. CPU by CPU temp, GPU by GPU temp, and reservoir by pre-radiator thermal probe. I was running Cinebench during this picture and getting a multithread score of around 20400 (everything is stock settings)

 

 

[thrasher_565]

you get lower over all temps the faster the flued is but as long as you got abut 1 gpm then your good. each 1 gpm is about 1 c drop. that's how we did thing back in the day but we didn't have things like jet plates that slowed down the flow.  so if your temps are good then i would not worry about it. if you want it lower you need more rads or faster fans. i guess if you had enough rads in between components you could see a bigger difference but i dont no. there are people that got like 5c above ambient witch is nuts... but were are talking like 2000mm+ of rad space...

[MaxAvidy]

Thanks,

 

I'd figure the amount of radiator and fans I'm throwing at this (radiators have push pull setup) would be enough. I'll have to get something to measure the actual flow rate. No idea what a D5 is rated to "push" through as far as how many radiators and waterblocks go.

[thrasher_565]

23 minutes ago, MaxAvidy said:

Thanks,

 

I'd figure the amount of radiator and fans I'm throwing at this (radiators have push pull setup) would be enough. I'll have to get something to measure the actual flow rate. No idea what a D5 is rated to "push" through as far as how many radiators and waterblocks go.

ya i would not worry about it. whats the temps? even if you added another pump thats another say $200 that might only get like 1-2 c drop its not worth it.

 

gallons per minute

liters per minute

liters per hour

 

1 gpm is like 3.79 lpm the d5 can do 1500 ish lph un restricted. so 1500/60? = 25lpm? im bad at math...

Edited November 1, 2022 by thrasher_565

[Applefreak]

4 hours ago, MaxAvidy said:

I also realize the cooling is probably overkill

No such thing... 

 

If you system temps are fine, perfect.

At some point you will hardly notice a difference in input/output temperature under load when measuring the fluid alone.

Something to observe is that the water maybe 30ish or so but the metal fins on the radiator can get hot enough to boil water in some instances. The fluid is only displacing the thermal energy. You should be able observe a difference in temperature under normal loads but at a certain point that difference will be negligible.

The real goal is to have your max temps level out, don't worry about the temperature inside loop. 

Something to keep in mind though when benching or running heavy loads, do not shut off you system right away. Let it cool down first. Why, because unlike a heat sink, you will have less surface area to dissipate the heat of the components and the hot radiator on the intake side will just radiate into the case as well. Generally it does not matter but I know how sensitive those solder joints are so cooler is better. I am guessing you have one radiator on the intake side, that is something that would drive me crazy, blasting hot air into the case, well it does not matter as much but still.

 

Stay away from mechanical flow meters, they only hinder the fluid and can cause leaks in some instances (poor quality). 

The D5 pump is industrial, not like some other pumps that were retrofitted from aquariums.

If you run it at it's max rpm, it is loud and quite powerful but that rarely ever makes any sense at all and is not worth it.

The pump is rated for 1500l/h or 6.6 gpm, typically you will see it putting out around 3 to 5 PSI, depending on rpm and the configuration.

 

As long as you have sufficient return flow as well as no visible bubbles inside your coolers, the system is perfectly fine the way it is.

I guess you have done your fair share of tests but I did observe that in some instances, running fans push/pull did not help matters much or at all. 

 

Also be prepared to change your room thermostat. No need to run 2 room heaters at the same time.

 

 

[thrasher_565]

25 minutes ago, Applefreak said:

Also be prepared to change your room thermostat. No need to run 2 room heaters at the same time.

the heat from the pc will heat the room also and making the room thermostat to shut off sooner so no need to do anything... unless the thermostat in not in the room and the door is closed...

 

ya there so many dam variables.  fans +or- 10%, rads fpi, water additives, pump tops, fittings that bottle neck, dumping the rad heat in the case, using 2 loops, parallel vs series, comparing different sensors/ program read outs, rad thick ness, water blocks, mounting presher, there just so much its not worth the time or money to go after these differences.

 

thing to add is you could add a timer to say your hot water so it shuts off instead of it having hot water 24/7 was the insulation should keep it hot for some time. but there are also more effecnt systems for heating and electricity and the house for that matter but at a cost... you can do cheap diy stuff but most part its illegal... like say a bach heat that heats water with the sun that for example that keep the the water hot and dose not use eletricay and is cheap to build...

 

Edited November 1, 2022 by thrasher_565

[MaxAvidy]

1 hour ago, thrasher_565 said:

ya i would not worry about it. whats the temps? even if you added another pump thats another say $200 that might only get like 1-2 c drop its not worth it.

 

Water temps can get to around 40c if I'm pushing the system maxed out at 2k. GPU won't thermal throttle. The CPU will get to 90 pretty easily in Cinebench, but I think I need to reseat it. Not sure if anyone else knows what sort of temps they're hitting in their loops.

 

1 hour ago, Applefreak said:

No such thing... 

 

If you system temps are fine, perfect.

At some point you will hardly notice a difference in input/output temperature under load when measuring the fluid alone.

Something to observe is that the water maybe 30ish or so but the metal fins on the radiator can get hot enough to boil water in some instances. The fluid is only displacing the thermal energy. You should be able observe a difference in temperature under normal loads but at a certain point that difference will be negligible.

The real goal is to have your max temps level out, don't worry about the temperature inside loop. 

 

I've tried to negate the hot air coming in as much as possible by having that radiator the last in the loop and several fans next to it drawing some air out of the case, hopefully pulling more fresh air up into the case from the bottom. Only so many places to put a radiator in the case. I'll probably play around with finding exactly where temps of various bits sit. I do think there is a restriction somewhere, but that being said, the water is definitely flowing. And yeah, wasn't necisarily worried about the temp itself, but rather what that large difference in temperature might mean. de8auer did a test that making sure the line wasn't running too slow would give you better cooling results. I just don't know if mine is slow enough to affect me there yet. Will check into that given the numbers you folks have shared.

der8auer

 

47 minutes ago, thrasher_565 said:

the heat from the pc will heat the room also and making the room thermostat to shut off sooner so no need to do anything... unless the thermostat in not in the room and the door is closed...

 

You guys may joke, but I have another thick radiator that I'm playing with the idea of mounting to a return vent close to my PC and heating my house with the power of gaming.

 

[Applefreak]

4 minutes ago, MaxAvidy said:

 

Water temps can get to around 40c if I'm pushing the system maxed out at 2k. GPU won't thermal throttle. The CPU will get to 90 pretty easily in Cinebench, but I think I need to reseat it. Not sure if anyone else knows what sort of temps they're hitting in their loops.

 

 

I've tried to negate the hot air coming in as much as possible by having that radiator the last in the loop and several fans next to it drawing some air out of the case, hopefully pulling more fresh air up into the case from the bottom. Only so many places to put a radiator in the case. I'll probably play around with finding exactly where temps of various bits sit. I do think there is a restriction somewhere, but that being said, the water is definitely flowing. And yeah, wasn't necisarily worried about the temp itself, but rather what that large difference in temperature might mean. de8auer did a test that making sure the line wasn't running too slow would give you better cooling results. I just don't know if mine is slow enough to affect me there yet. Will check into that given the numbers you folks have shared.

der8auer

 

 

You guys may joke, but I have another thick radiator that I'm playing with the idea of mounting to a return vent close to my PC and heating my house with the power of gaming.

 

If you have the budget, get a Mo-Ra 3 external radiator. A bit messy with tubes going out of your case but can be done nicely as well, depending on what furniture you have in place. I've built a system for a client using that solution and it does keep her off-grid cabin above freezing in the winter (mining rig, don't ask).

[Applefreak]

[thrasher_565]

3 hours ago, MaxAvidy said:

 

Water temps can get to around 40c if I'm pushing the system maxed out at 2k. GPU won't thermal throttle. The CPU will get to 90 pretty easily in Cinebench, but I think I need to reseat it. Not sure if anyone else knows what sort of temps they're hitting in their loops.

 

 

I've tried to negate the hot air coming in as much as possible by having that radiator the last in the loop and several fans next to it drawing some air out of the case, hopefully pulling more fresh air up into the case from the bottom. Only so many places to put a radiator in the case. I'll probably play around with finding exactly where temps of various bits sit. I do think there is a restriction somewhere, but that being said, the water is definitely flowing. And yeah, wasn't necisarily worried about the temp itself, but rather what that large difference in temperature might mean. de8auer did a test that making sure the line wasn't running too slow would give you better cooling results. I just don't know if mine is slow enough to affect me there yet. Will check into that given the numbers you folks have shared.

der8auer

 

 

You guys may joke, but I have another thick radiator that I'm playing with the idea of mounting to a return vent close to my PC and heating my house with the power of gaming.

 

well in terms of best case for temps it would be open bench > case rads intake and exhaust out side no hot air going thow the rads> case intake thow  all rads and exhaust out with enough fans > 2 intake rads one exhaust >

 

90c on gpu seems hi.. unless its oced? do you have a pic of the build?

 

this is my case and all rads will intake from out side the case and exhaust up and out.

 

 

 

 

[MaxAvidy]

15 hours ago, Applefreak said:

If you have the budget, get a Mo-Ra 3 external radiator. A bit messy with tubes going out of your case but can be done nicely as well, depending on what furniture you have in place. I've built a system for a client using that solution and it does keep her off-grid cabin above freezing in the winter (mining rig, don't ask).

I like that idea. Looks like it's not carried anymore, but I can build an aluminum frame for the extra EK-CoolStream CE 420 I have, and pump that heat right into the house return vents. I need to keep my setup looking somewhat nice since I don't have my own office for the rig.

 

13 hours ago, thrasher_565 said:

90c on gpu seems hi.. unless its oced? do you have a pic of the build?

 

That's on the CPU. GPU seems fine. Again, I think I just need to reseat the block to keep the CPU from thermal throttling. This is more a test in making sure the water is moving quickly enough through the loop to pull as much heat away from the components as possible.

 

I only had this picture at work, but white arrows are flow after the radiators and red is after components and before the last radiator. White rectangles are the radiators. The first is the small one of the left, then to the one on the top, then to the one one right (the front of the case). The rear radiator fan blows out, as do the 6 fans on the top radiator. The six on the front radiator blow in. And those four fans in the back of the picture have no radiator, and they exhaust out to pull some of the hot air coming in front the front radiator. Not perfect, but again, not really about temps, more about difference in temps.

 

13 hours ago, thrasher_565 said:

this is my case and all rads will intake from out side the case and exhaust up and out.

I may try having all of them blow out, and those four in the back pulling in fresh air, but I'll try that after I have the loop acting the way I'd like it.