second pump in loop?

[(ಠ ͜ ಠ)]

Does a water loop really need 272LPH? I have seen mine down below 70LPH with still very good temps.

No, they don't necessarily, but my original post was in response to the following statement from another forum member:

"Hi,
Yeah the faster the fluid goes through radiators they actually are less efficient at removing the heat from the fluid"


As a blanket statement in response to someone complaining about low flow, this is not accurate or helpful advice and indicates a lack of understanding about thermal transfer. The higher your flow rate in a PC cooling loop (up to its measured point of diminishing return), the better your heat transfer (ie cooling), because you are carrying more heat per unit time (hours, minutes, seconds - whatever you want) away from the block. The same principle applies with the radiator, the only difference being the heat exchange is with ambient air. Use a large enough radiator with sufficient surface area to dissipate heat quickly enough and you can passively cool your system without fans. 48.5L/H may give you reasonable temps (I'm assuming on your CPU + GPU) but increasing the flow to say 170-250 L/H will remove heat more quickly from the water blocks and give you a higher coolant/air delta, which is better for thermal transfer, because heat naturally flows from hot to cold. Generally, more flow = better (up to your system's point of diminishing return), even if it doesn't appear like it's doing much. It may not look like it is making a difference because you are also changing another variable in the system (forcing air through rads with fans, ie. greater air flow at the thermal interface).

Hi,
Only exception is if one uses a chiller and that's into exotic

You can always do like I said and remove all your fans and let us know how great your cooling is at maximum clocks your system can do
You at the arctic circle ? this might be the exception

The principle still holds true, the difference here is that you are using yet another heat exchange interface with the chiller to get colder coolant . You don't understand thermal transfer ;-p

 

[ThrashZone]

Hi,
Yeah I think we need to gang up and file class action lawsuits world wide against those lying fan manufactures we don't need no stinking fans you lying bastards

 

[Vya Domus]

because you are carrying more heat per unit time

That's true but this only works if the cooling power of the radiator was "infinite", which isn't the case, they eventually become saturated in relation to the ambient temperature and once that happens flow rate makes no difference because there wont be enough of a temperature differential once the water enters the radiators.

 

[ThrashZone]

Hi,
Yeah he's leaving out a lot of his personal environment variables like he might have a reservoir buried outside or a in cool basement/...
Otherwise talk about blanket statements miko owns this department so far

 

[thesmokingman]

No, they don't necessarily, but my original post was in response to the following statement from another forum member:

"Hi,
Yeah the faster the fluid goes through radiators they actually are less efficient at removing the heat from the fluid"


As a blanket statement in response to someone complaining about low flow, this is not accurate or helpful advice and indicates a lack of understanding about thermal transfer. The higher your flow rate in a PC cooling loop (up to its measured point of diminishing return), the better your heat transfer (ie cooling), because you are carrying more heat per unit time (hours, minutes, seconds - whatever you want) away from the block. The same principle applies with the radiator, the only difference being the heat exchange is with ambient air. Use a large enough radiator with sufficient surface area to dissipate heat quickly enough and you can passively cool your system without fans. 48.5L/H may give you reasonable temps (I'm assuming on your CPU + GPU) but increasing the flow to say 170-250 L/H will remove heat more quickly from the water blocks and give you a higher coolant/air delta, which is better for thermal transfer, because heat naturally flows from hot to cold. Generally, more flow = better (up to your system's point of diminishing return), even if it doesn't appear like it's doing much. It may not look like it is making a difference because you are also changing another variable in the system (forcing air through rads with fans, ie. greater air flow at the thermal interface).


The principle still holds true, the difference here is that you are using yet another heat exchange interface with the chiller to get colder coolant . You don't understand thermal transfer ;-p

You don't know wtf you're talking about. I even gave you hints on where to look, time frame but nope, you double down on highschool science, throw out Newton's name smh. Stop spreading misinformation.

 

[Bobaganoosh]

this has been a trainwreck of misunderstanding each other lol. This is classic bad communication. I'm 99% sure all of you would agree that airflow and flow of liquid are both necessary and more is better (up to a point of diminishing returns of course), but the arguments have somehow distorted into one assuming the other said "you don't even need fans, airflow is not important" and that one thinks the other said "you don't need the coolant to flow at all, it doesn't matter". Nobody here said either of those things.

 

[kapone32]

Hi,
Only exception is if one uses a chiller and that's into exotic

You can always do like I said and remove all your fans and let us know how great your cooling is at maximum clocks your system can do
You at the arctic circle ? this might be the exception

I remember building an AM4 system for a Customer and forgot to install the fan on the Air cooler before turning the system on. In the BIOS the CPU jumped to 85 C before it shut down within 30 seconds. With a fan installed the same scenario was 33 C. Anyone that thinks that fans do not make a difference do not understand how many of us are not talking about what we read but have already experienced. Newton's laws are stressed when a CPU has 16 billion transistors connected to 1718 pins.

 

[Vya Domus]

I remember building an AM4 system for a Customer and forgot to install the fan on the Air cooler before turning the system on. In the BIOS the CPU jumped to 85 C before it shut down within 30 seconds. With a fan installed the same scenario was 33 C. Anyone that thinks that fans do not make a difference do not understand how many of us are not talking about what we read but have already experienced. Newton's laws are stressed when a CPU has 16 billion transistors connected to 1718 pins.

Water has more thermal capacity than metal, I can run my system with the fans stopped for like 15 minutes at full load before temps reach anything remotely concerning. There are scenarios where I could run the system passive whereas with an air cooler it would throttle into oblivion, it would be stupid but it would work, a loop offers way more thermal mass.

 

[thesmokingman]

That's a whole other ball of wax passive watercooling loops, they can work but again its a whole other thing.

 

[ThrashZone]

I had an aio die it wasn't pretty

Context matters.

 

[kapone32]

Water has more thermal capacity than metal, I can run my system with the fans stopped for like 15 minutes at full load before temps reach anything remotely concerning. There are scenarios where I could run the system passive whereas with an air cooler it would throttle into oblivion, it would be stupid but it would work, a loop offers way more thermal mass.

I understand what you are saying. Nothing in Air compares to the potential of Water. I have one for you too. I built a system for a Mining rig. I used the Arctic 15mm fans for the top rad. Well after an hour the temps overcame those fans. I took a fan from a Corsair rad, once the customer confirmed that noise was not a problem and suddenly that single fan solved the problem. So even if you have fans with a rad the quality of those fans matter.

 

[ShrimpBrime]

Can you move 250w??

That's about 850 btu/hr.

W/mk is watts per meter.

Copper is about 398 w/mk

Decent thermal paste is 8 w/mk

Water is 0.6 w/mk at about 20c. (Sea level)

Air is about 0.025 w/mk.

So moving air and water faster will slightly increase the w/mk, but the copper does most of the work.

From the numbers above, we could calculate the desired amount of CFMs to dissipate 250w or 850 btu/hr.

But I'm not able to sit and do the math cause I'm at lunch at work. But should help understand how moving thermals is actually quite difficult when you look at how poor air and water thermal conductivity properties are even though everyone says water is so good at it, in reality, it is Not.

 

[Vya Domus]

thermal conductivity properties are even though everyone says water is so good at it, in reality, it is Not.

This isn't about thermal conductivity, water has high thermal capacity and poor conductivity, that's why you need to keep recirculating it but once you do it will beat any heatsink you can reasonably construct out of copper/aluminium/whatever.

All of these are eventually cooled by air which has horrible conductivity, the fact that water gives you such a large thermal mass helps the most with that because you have more time to cool it down whereas a metal heatsink immediately becomes saturated with heat and has to be cooled by the air which has terrible conductivity.

The advantage of water is that it can store more heat and you also have more time to cool it down.

 

[kapone32]

This isn't about thermal conductivity, water has high thermal capacity and poor conductivity, that's why you need to keep recirculating it but once you do it will beat any heatsink you can reasonably construct out of copper/aluminium/whatever.

All of these are eventually cooled by air which has horrible conductivity, the fact that water gives you such a large thermal mass helps the most with that because you have more time to cool it down whereas a metal heatsink immediately becomes saturated with heat and has to be cooled by the air which has terrible conductivity.

The advantage of water is that it can store more heat and you also have more time to cool it down.

I can attest to this. The latest Mining rig I built was 2 watercooled 68000XTs and using 2 rads did not just lower the idle to temp to 25 C from 29 C but also meant a maximum temp of 48 C on the hotspot. The GPUs are turned down to 70% max clock but that is still 6 C lower than with 1 pump.

 

[Belfaborac]

I’m really only cooling a 7800x3d and a rx6900xtxh with heavy overlock. The reason for 3 pumps is mainly the mora420 and qdc’s. Don’t wanna strain 1-2 pumps. The mora in itself isn’t restrictive but there is a lot of copper tubing that the coolant has to run through and for redundancy

Seems like a case of serious overkill-kill-kill really. I'm all for extreme projects and whatnot, but only when the need is correspondingly extreme. Your system can easily be cooled with a couple of radiators of the common or garden variety.

Then again, that's none of my business.

I've got 3 pumps myself, as it happens, but only because I wanted to see if they made a difference to anything and was too lazy to remove one or two when it turned out they didn't.

 

[ShrimpBrime]

This isn't about thermal conductivity, water has high thermal capacity and poor conductivity, that's why you need to keep recirculating it but once you do it will beat any heatsink you can reasonably construct out of copper/aluminium/whatever.

All of these are eventually cooled by air which has horrible conductivity, the fact that water gives you such a large thermal mass helps the most with that because you have more time to cool it down whereas a metal heatsink immediately becomes saturated with heat and has to be cooled by the air which has terrible conductivity.

The advantage of water is that it can store more heat and you also have more time to cool it down.

It's all (edit, not all, but also) about thermal conductivity or you could get away with pastes that are lower than 2 w/mk.

The fist weakest link. Thermal past.
Second, distance from source.

On thermal mass, everyone would be happy with the thicker IHS plates on AMD chips. Then you shouldn't need water for extra mass.

In your reality it's extra mass. In mine the water os to relocate the thermals. Otherwise the water does nothing except take longer to absorb and then dissipate thermals.

Waters poor conductivity is why Peltier coolers don't work well in a loop.

The only real restriction is the temperature Limit of the processor. If there was no limit, say 100c, then you could dissipate the heat at any rate you wanted. Wouldn't matter.

The processor thermal limit is the only reason we have these types of conversations.

How many Btu can you move in x time period to keep a cpu under 100c. That's really all there is to it.

So yes, mass counts, but thermal conductivity is equally as important.

Here, try this as an example of coefficients.

Heat Transfer Coefficients in Heat Exchanger Surface Combinations

Average overall heat transmission coefficients for fluid and surface combinations like Water to Air, Water to Water, Air to Air, Steam to Water and more.

www.google.com

 

[Calenhad]

Dual pumps like the EK are designed to work together. It shouldn't matter if one runs at a different rate. This is only an issue with individual pumps. Maybe consider a dual pump?

You might want to read my comment again. This is exactly what I said.

 

[HaKN !]

Seems like a case of serious overkill-kill-kill really. I'm all for extreme projects and whatnot, but only when the need is correspondingly extreme. Your system can easily be cooled with a couple of radiators of the common or garden variety.

Then again, that's none of my business.

I've got 3 pumps myself, as it happens, but only because I wanted to see if they made a difference to anything and was too lazy to remove one or two when it turned out they didn't.

Overkill for what ? 3 pumps or 10 ? I like quiet systems. 1 pump alone will always die faster and spin faster and make more noise. more pumps equal a quieter system. no matter what chart or self beliefs you have. Been doing this forever. will never ever run a single pump in a system again. Try 1 pump with a giant 1260mm rad and some QDC's and come back and tell me again how that 1 pump is doing so well.

 

[Belfaborac]

Overkill to cool your gear. My system is so quiet I can barely hear a whisper at load, with two radiators (2 x HWL 420mm). My single pump spun at exactly the same speed on its own as the three pumps do now, since they're varios set to 3/4 speed. I couldn't hear the single pump, I can't hear the three 'pumps and the performance is the same,with one and three whether I use the vario feature or not.

The argument wasn't about what happens noise-wise with a single pump, a huge radiator and lots of QDCs (of which I have six myself, none of which add any noise whatsoever); it was about a giant set-up being utterly unnecessary to cool the hardware in question.

 

[ThrashZone]

Overkill for what ? 3 pumps or 10 ? I like quiet systems. 1 pump alone will always die faster and spin faster and make more noise. more pumps equal a quieter system. no matter what chart or self beliefs you have. Been doing this forever. will never ever run a single pump in a system again. Try 1 pump with a giant 1260mm rad and some QDC's and come back and tell me again how that 1 pump is doing so well.

Hi,
QDC's don't hurt flow much at all unless you pipe them in the wrong direction
Male is outlet/ female inlet.

I use 4 pairs on my x299 rig "gpu/ cpu/ vrm" with two gtx 240-280mm rads
3 pairs on my z490 rig "gpu only" with a ek pe 280mm rad both with single d5's pwm although I let it run full blast by not connecting the pwm port to the board.
x299 rig has a vario not sure what I set it to 5 or 4 but both are silent
Only time I've ever heard a D5 is if I had a clogged cpu block then it will make noise in a stressful environment.

I have in the past use 2-360mm mora's with both systems and yes they to had QDC's on them but i did use 2 D5's on each rig at that time.

 

[HaKN !]

Hi,
QDC's don't hurt flow much at all unless you pipe them in the wrong direction
Male is outlet/ female inlet.

I use 4 pairs on my x299 rig "gpu/ cpu/ vrm" with two gtx 240-280mm rads
3 pairs on my z490 rig "gpu only" with a ek pe 280mm rad both with single d5's pwm although I let it run full blast by not connecting the pwm port to the board.
x299 rig has a vario not sure what I set it to 5 or 4 but both are silent
Only time I've ever heard a D5 is if I had a clogged cpu block then it will make noise in a stressful environment.

I have in the past use 2-360mm mora's with both systems and yes they to had QDC's on them but i did use 2 D5's on each rig at that time.

Saying a D5 is silent at full blast is just copium. You have your experience and I have mine. I like more pumps with next to no sound instead of having 1 at full blast. Just don’t come and tell me a D5 is “silent” at full blast lol. And saying that QDC’s don’t hurt flow is just a straight up lie. At least know what you’re talking about lol

 

[Vya Domus]

It's the blocks that are most restrictive, rads not so much, people used to use multiple pumps when SLI/Crossfire was a thing, there is no real point in having more than 1 pump nowadays.

 

[ThrashZone]

Hi,
My fans are usually at 700-900 rpm and are the loudest items lol

If your d5's are noisy they are garbage.

 

[Ferrum Master]

Not sure why are you even bothering to quarrel here.

Noise wise, people are different. Some are less deaf than other due to age, surroundings, children, neighbors banging like rabbits. D5 noise range is about 30-45dBA. Here are some older VSGs charts. So if you are in a dead silent room, you will hear even 33dBA. If you have fans on rads, and for some weird reason in 2023 you spin them all, so one FAN on 800-1000RPM does ~25-30dBA, each fan adds two times ie + 3dBA. So yes, if a man runs a lot of fans, he won't hear the pump for sure. And full throttle 47dBA due to the fact it is past 4000RPM, it drives into very audiable for human ear spectrum it should be heard... unless he runs like 6-8 fans all the time even on silent. Simple math. All could be right... so hold yer horses.

Other than that, there is no point running two pumps other than redundancy maybe.

 

[maxfly]

If you've ever had a pump die you understand the value of running dual pumps. Not to mention the increased ease of purging a large loop. There's always value in dual pumps despite what some narrow minded members may profess.

Arguing about pump noise is silly, particularly when most people have never owned more than a couple of pumps at best. I've got 5 d5s currently, 3 from from different manufacturers. They ALL have different noise profiles. From barely audible with your ear next to them to loud af from 5 feet away. I've had 7 or 8 d5s and 10 or 12 ddcs from every manufacturer under the sun and it's been the same story with those. Some are quiet, some are loud. Making all encompassing statements about something as random as pump noise is generally inaccurate.