# Temperatures of coolant needed.

Discussion in 'Water Cooling' started by Wolfsfriend42, Feb 17, 2018.

1. ### Wolfsfriend42n00b

Messages:
5
Joined:
Feb 17, 2018
Hello Everyone,
I am in need of some information. I do not know if anyone can provide all the temperatures I am seeking. I am trying to figure out if an easy and fairly cheap way of actively cooling the water exists. I need to know the following:
• Ambient Temp
• Case Air Temp
• Liquid Flow Rate
I guess the most important is the temp after the radiator. But I would also like to learn the total heat produced in a given PC. Some may think this simple, but it is not since heat is measured in BTU/h and anything PC related is expressed in Watts.

2. ### SticKx911[H]ard|Gawd

Messages:
2,036
Joined:
Mar 14, 2004
In theory, the temp measured anywhere in the loop will be the same reguardless of Equiptment orientation. Law of thermodynamics if I’m not mistaken.

Are you trying to hook up a compressor to chill the water? I can do some testing on my setup this week at some point.

3. ### Wolfsfriend42n00b

Messages:
5
Joined:
Feb 17, 2018
If what you said was true then cooling would never take place since you claim that the radiator will not lower the temperature at all.

4. ### SticKx911[H]ard|Gawd

Messages:
2,036
Joined:
Mar 14, 2004
not quite. it lowers the overall system temp. (system as in the series of parts entwined by flowing liquid.) Adding more radiators would lower the overall system temp but as in measurable direct difference between 2 points at the same time in a closed continuous loop...not so much. What you would be looking for might be difference in a system without a radiator vs with or etc.

unlike a refrigeration system, there's no pressure differential or phase change.

Reality likes this.
5. ### Zarathustra[H]Official Forum Curmudgeon

Messages:
27,339
Joined:
Oct 29, 2000

For a system to be an efficient water cooling loop you want your flow rate to be high enough that you don't see a noticible increase in coolant temperature across water blocks and you don't see a noticible decrease in temperature across your radiators.

Your flow is then high enough that all the water in your loop regardless of where it is, is most the same temp, and you enter a steady state.

Reality likes this.
6. ### Wolfsfriend42n00b

Messages:
5
Joined:
Feb 17, 2018
Ok, I guess you two are saying that the water that leaves the water block at, let's say the CPU, is no hotter than the water going into it, correct?

7. ### Batboy88Limp Gawd

Messages:
326
Joined:
Dec 25, 2017
around ambient maybe a little above, it all evens out anyways though pretty much. I do think with prior messing around going into the rad first is slightly better than the block. adding another rad may not do much....start getting crazy like that you''ll need even a meaner Pump and flow rates.

a single should be plenty to dissipate a system the various 360mm should even be overkill with that kinda surface area. most applications anyways.

It's the stuff too don't want too low a pump speed/flow or too high...some where in the middle.

I was gonna piece together more of the like those generic china kit's with hard tubing...but glad I didn't the EK quality is wayyyy better. More of what I wanted to do.

Last edited: Feb 18, 2018
8. ### Tsumi[H]ardForum Junkie

Messages:
12,906
Joined:
Mar 18, 2010
With a high enough flow rate, the temperature difference is on the scale of 0.5 C or less. The delta increases as flow rate decreases, and the efficiency of the system at removing heat decreases as well.

Reality and Batboy88 like this.
9. ### Batboy88Limp Gawd

Messages:
326
Joined:
Dec 25, 2017
Temps are slightly worse because we have slightly thicker glycol soloution...but I don't care should be able leave that in there for a longggg time.

10. ### Zarathustra[H]Official Forum Curmudgeon

Messages:
27,339
Joined:
Oct 29, 2000
There is a temperature difference, but if your flow is high enough it is so small that it is not measurable. You want your coolant moving fast enough that you just don't have to worry about this at all.

11. ### VanGoghComplex[H]ard|Gawd

Messages:
1,616
Joined:
Apr 5, 2016
I hate to be that guy, and maybe I'm misunderstanding your question, but every bullet point you asked for is going to be different for every watercooled PC. Are you just trying to get a general idea?

Also, if you're talking about actively cooling the water in your loop to sub-ambient temperatures, be aware that condensation on your components will become a concern.

12. ### Wolfsfriend42n00b

Messages:
5
Joined:
Feb 17, 2018
So far not one person replying has answered my question(s). All I am getting is people telling me that there is no difference in temps before a water block and after it. That water loops essentially do nothing since the temps are the same throughout the whole loop. This is Impossible. The reason for a water loop is to remove heat, plain and simple. If the water leaving a radiator is the same temp as the water going in that would mean that the radiator is not dissipating any heat at all.

Why are you all making something so basic and simple so hard? I do not have the means to get this data myself here at home.

I know the temps will be different, I am just trying to find out what any of them are in any gaming system. The only thing that has been said so far is that they are a bit above ambient. Well from the few pieces of this puzzle I have gathered so far that is wrong. I have been told that at idle it can be 32C-90F and under full load maybe 85C-185F. At idle that is apprx 8C higher than the typical home ambient temp(in winter) of 24C. And under load it is 61C degrees higher than ambient. I am trying to figure out if there is an easy and cheap way to keep the water temp at a constant 20.55C-69F under full load with a CPU & GPU in the loop. The reason for choosing this temp is because condensation can start at 20C-68F in a relative humidity of 100%(which does not usually occur inside homes).

13. ### pendragon1[H]ardForum Junkie

Messages:
11,142
Joined:
Oct 7, 2000
yes there will be a temp difference before and after the rad. eventually there will be a point where the temps stabilize. what are you trying to accomplish?

14. ### Brian_B2[H]4U

Messages:
2,232
Joined:
Mar 23, 2012
1W = 3.412 BTU/hr

Past that, it's all basic thermodynamics:
Q = m * c * dT
Qin = Qout

You can get mass flow rates from pump/fan specs - you'll have to experimentally determine head loss though to get the appropriate point on the flow curve - that will vary installation to installation,
Specific heat capacity is pretty easy. Air and Water are constants at your given temperature range. If your using exotic coolants you'll need to get the number from the manufacturer.

Each block/radiator will also have some amount of efficiency loss, nothing transfers at 100% efficiency. Again, that will either need to be determined experimentally, assumed, or given from the manufacturer. That efficiency isn't lost, it just affects final dTs. You can safely just assume 100% for inital calcs, it will get you pretty close - most equipment is in the high 90%s on liquid applications (Unless your using an EK block - j/k bad joke)

Past that, it's just algebra to solve for the unknown. If you want to calculate what your dT (or inlet/outlet temp given one or the other) should be based on a given TDP of cooled equipment, or if you want to calculate the actual heat rejection based on measured temperatures.

Last edited: Feb 18, 2018
pendragon1 likes this.
15. ### VanGoghComplex[H]ard|Gawd

Messages:
1,616
Joined:
Apr 5, 2016
The fluid moves through the loop too quickly, and the amount of wattage being dissipated is too small. You will not see any significant difference between the inlet and outlet of the radiator, or any block. This is possible, and doesn't mean the loop isn't working.

These tests have been done in the WC community. Lots of people have multiple coolant temperature probes in their systems and will all tell you the same thing: the coolant temp in the entire loop will be about the same.

Technically, yes the water leaving your block will be hotter than it was when it entered, and the water leaving your radiator will be cooler than when it entered, but practically, it is all the same temperature. You might see a delta of a degree Celsius between the coolest and hottest part of the loop.

Lots of people have answered your question. You just didn't like/believe the answer. Go spend some money and investigate it yourself if you like; if several experienced watercooling enthusiasts won't convince you, I don't see what else will.

Reality likes this.
16. ### Zarathustra[H]Official Forum Curmudgeon

Messages:
27,339
Joined:
Oct 29, 2000
You don't seem to be understanding how it works. Let me try to explain again.

Yes, if you have VERY low flow, there will be a measurable increase in water temp going across the CPU or GPU blocks, and again a very measurable decrease in temp going across the radiator.

This is - however - not a recommended setup as it results in inefficient cooling, and your components running hotter than they should.

When you increase the flow rate (by turning the pump to a higher setting) the water passes through the CPU/GPU blocks faster, and thus spends less time in the block and heats up less. At the same time the water passes through the radiator very fast, and thus spends less time in it, and cools less across the radiator.

The goal is to have a pump that is fast enough that you hit a steady state, such that the temperature deltas across the CPU/GPU blocks and across the radiators are so tiny that they are not measurable, and instead your entire water loop approximates the same temperature.

You can lower the flow rate to the point where you can measure a difference before and after, but this is a really silly thing to do, buys you nothing and just hurts your cooling performance.

As far as the numbers you are asking for, they are completely impossible to give you because:

1.) A well running water loop with high enough flow rate will not have a measurable difference in temperature anywhere in the loop, because the water is moving that fast.

2.) A poor loop, with too low flow rates will have a difference, but that difference will depend on so many different variables including radiator size, CPU type, GPU type, overclock/voltage level, flow rate, fan type, fan speed, case layout, load level, etc. etc. etc. Water cooling is by its very nature a custom configuration (unless you go the AIO route) and to get answers like these you'd need to find someone with the exact setup you are interested in and ask them to do measurements. It varies so wildly that there is no such thing as a "typical" or "general" value for these things.

The numbers I can give you are as follows:

1.) Even in the states we typically use Celsius for PC cooling measurements (just for the record) No need to convert to Fahrenheit for us.

2.) Most pumps have a rated max coolant temp of 50-60C If you go above that you will see shortened life, if not immediate failure. You don't want to go this hot anyway though, as your CPU/GPU will never run at the coolant temp. There is always a delta. This delta depends on how good the block is, and how well it is mounted, but I'd say on a G2PU its anywhere from 8C to 15C delta, and the CPU usually much more. At least this has been my experience.

3.) With my setup (see here) I usually adjust my fans such that I get about 30C coolant temp at load. Depending on the title and the ambient temp this can be anywhere from 600rpm up to 1800rpm. When the coolant is at about 30C, my GPU at load will usually stay at 36-38C. I think I got a pretty good mating with my EK full cover block, and I used very good paste, which probably helped. My GPU is an overclocked Pascal Titan though, so it puts out a fair deal more heat than most other GPU's.

4.) Yes, you can go slightly sub ambient without taking extreme insulation precautions, especially if you live in a very dry climate, but it has to be done VERY carefully. At 100% RH, even dropping a fraction of a degree below ambient will immediately result in condensation. The lower the RH goes, the lower the dew point drops. The idea is to keep your coolant temp above the dew point at all times, and not cut it too close, as all it takes is one drop of condensation for you to have a very expensive failure. I'd recommend keeping a very comfortable safety margin. You can use one of many dew point calculators out there. This one is pretty good (but sucks on phones).

drescherjm likes this.
17. ### ricson[H]Lite

Messages:
73
Joined:
Oct 8, 2009
Look up BTU to Watts.......1000 watts is 3412.141633 BTU/hr

18. ### Wolfsfriend42n00b

Messages:
5
Joined:
Feb 17, 2018
I give up. All I am getting here in nonsense that doe not answer my question(s). I will find somewhere where helpful people post not those that just talk about other things.

19. ### pendragon1[H]ardForum Junkie

Messages:
11,142
Joined:
Oct 7, 2000
first off, you didn't actually ask any question(s) you just stated thing you want to know. second, people have given the answer(s) but you don't seem to get it. go hit up google...

20. ### Brian_B2[H]4U

Messages:
2,232
Joined:
Mar 23, 2012

SticKx911 and Reality like this.
21. ### Zarathustra[H]Official Forum Curmudgeon

Messages:
27,339
Joined:
Oct 29, 2000
If you find anyone else out there that gives you a different answer, then they don't know what they are talking about.

We have been trying to help you, and explain to you how these things really work, but you seem to have a predetermined belief that it should work differently and are unwilling to listen to sense. If what we have painstakingly and repeatedly tried to explain does not make sense to you, then I think you have a completely non-existant understanding of thermodynamics, and maybe should try a simpler project.

Maybe an AIO?

By all means, go elsewhere. Feel free to come back when you are willing to actually listen to the people who know what they are talking about.

You can lead a horse to water, but you can't make him drink, and all that.

22. ### VanGoghComplex[H]ard|Gawd

Messages:
1,616
Joined:
Apr 5, 2016
Okay OP. I'll answer what I can for you.

Ambient temperature: 22°c
Coolant temp at idle: 24°c
Coolant temp under load, after saturating: 30°c

This in a system with a 7700K and a 1080, both overclocked, 3x140mm of radiator, and a single DDC pump.

Batboy88 likes this.
23. ### Killerxp100Gawd

Messages:
741
Joined:
May 20, 2011
It's the douche of the week!

For some 1+2 will never = 3, no matter how many times you try to explain it.

Reality likes this.
24. ### Nobu2[H]4U

Messages:
2,591
Joined:
Jun 7, 2007
look at threads documenting sub-ambient cooling projects--they likely have the answers you want. A chiller will likely be the easiest/cheapest means, the method depends on you and your circumstances. I've no first-hand experience, so this is the best advice I can give.

25. ### Tsumi[H]ardForum Junkie

Messages:
12,906
Joined:
Mar 18, 2010
It is not nonsense. What you are failing to comprehend is that although the temperature between inlet and outlet doesn't change much at a high flow rate, the temperature of the coolant rises as load increases, and this is where the heat transfer occurs.

Scenario A with low flow rate: Temperature delta between inlet and outlet of CPU (and thus radiator) is 5 C. Temperature of coolant is thus fluctuating from 25 to 30 C. CPU to CPU block delta is 40 C, thus CPU operates close to 70 C.

Scenario B with high flow rate: Temperature delta between inlets and outlets is very close to 0. Temperature of coolant stays at a fairly constant 27.5 C. CPU therefore operates at 67.5 C.

In reality, scenario A will have a higher operating temperature because heat transfer is not a linear rate.

In any case, listen when multiple people tell you you are wrong. Odds are they are right.

Reality likes this.