What temperature is the water in a wc system?

JonJ678

JonJ678

JonJ678

Soldato
Joined
22 Dec 2008
Posts
10,369
Location
England
I know it's going to be a range from ambient to very hot depending on whats emptying heat into it and what is removing heat, but roughly what is the range?

If you empty 400W into the system and remove 400W from it, its presumably going to sit, on average, at ambient temperatures. The radiator efficiency increases with temperature, and it takes time for heat to diffuse. I take it this accounts for temperature being above ambient.

But what values are normal?
 
This depends on many things not least of which is how you control your fans. On my old 939 system I ran my fans for lowest noise when the system was idle and varied their speed with reference to rad temp. under load. Like this the max summer water temperature I remember seeing was around 27c, this was for a rad temp. of around 23c and a cpu load temp around 38c (may have been 36c?) iirc. All these temperatures are approximate as none of the sensors are calibrated but from this I could say the rough rules for that system were:

ambient < water temp < (ambient + 10c)

rad temp approx = ambient + (( water temp - ambient ) / 2)

cpu @ load temp < (water temp + 10c)

Looks close. This was with an Opti180 @ 3GHz + 7800GT in the loop.

Very Hot should never come into play, I'd worry if I ever suspected a water temperature much above 30c.
 
Last edited:
JonJ678 said:
If you empty 400W into the system and remove 400W from it, its presumably going to sit, on average, at ambient temperatures.
Click to expand...

No, it's not - for a constant heat output it'll reach a state of stable, dynamic equilibrium at a temperature where the rate of heat transfered to the air from the radiators/other loop components equals the rate your components and pump produce it. For the radiator in a closed loop to work (transfer the heat from water to air) there must be a temperature difference between water temperature and ambient air temperature. Since there's a difference in water in and water out of the radiator, the average water temperature will depend on the amount of heat removed by the radiator for the given airflow, water flow rate to replace cooled water with warmer water and maintain air-water delta T, efficiency in transferring heat energy to fin/pipe cooling surface and from those surfaces to the air passing through the radiator. These combine to give a C/W value - the temperature difference between ambient air and water temperature for a given set-up given the wattage going into the loop from compaonents and pump heat output.

I think that's right, if not I'm sure others can explain better. :)
 
Last edited:
Hmmm, it was only an estimate. I really have no idea. Keeps my i7 cool which is the main thing.

EDIT:
HowHotIsTooHotGuide said:
WATER AT 43 DEGREES CAN SCALD!
Click to expand...

Jaysus - definitely not at anything approaching that then!!
 
Monkey Puzzle said:
No, it's not - for a constant heat output it'll reach a state of stable, dynamic equilibrium at a temperature where the rate of heat transfered to the air from the radiators/other loop components equals the rate your components and pump produce it. For the radiator in a closed loop to work (transfer the heat from water to air) there must be a temperature difference between water temperature and ambient air temperature. Since there's a difference in water in and water out of the radiator, the average water temperature will depend on the amount of heat removed by the radiator for the given airflow, water flow rate to replace cooled water with warmer water and maintain air-water delta T, efficiency in transferring heat energy to fin/pipe cooling surface and from those surfaces to the air passing through the radiator. These combine to give a C/W value - the temperature difference between ambient air and water temperature for a given set-up given the wattage going into the loop from compaonents and pump heat output.

I think that's right, if not I'm sure others can explain better. :)
Click to expand...

From reading the above a couple of times, I think its what I said but in considerably more detail and preceded by 'no, it's not'. I didn't make myself very clear at all though.

The efficiency of a radiator where water temperature equals ambient is zero, it moves no heat at all. Hence higher temp water than ambient required, how much higher depends on thermal coefficients/air flow rate/water flow rate/conductivities of the fluids and of course on the wattage dumped into the flow. <--clarification of first post

I think I prefer your description as it is much clearer, thank you very much for posting it. Particularly appreciate the description of C/W.

I didn't realise temperatures were maintained that close to ambient. Good news. I take it that more radiators stands a good chance of lowering temperatures, but the improvement will fall off rapidly with the number introduced?

If water is maintained at roughly ten above ambient (say) then it looks like the thermal conductivity of the waterblock for the cpu is very important, as it doesn't get to eject heat into the ambient air. Cheers guys, shall continue researching this.
 
I know after a few hours gaming I can feel the front of my bay res and its just warm, probably mid 20's or so... but thats just my personal estimate, no sensors or anything. Thats with my CPU running mid 50's so about half the temp of my CPU I'd say
 
More radiators = greater pressure drop, this is bad :( Assuming two rads the obvious solution is to run the rads in parallel. Yay, lower pressure drop, this also halves the flow rate of the water through each rad. Common sense says this is good as the water spends more time in the cooling zone, but iirc the reality of the situation differs from what one imagines. At higher flow rates the water releases proportionately more heat to the rad then it does at lower flow rates, much the same way that water through a block removes more heat at higher velocity than it does at lower velocity.

More radiators = greater surface area, this is good :) also greater potential for adding more fans meaning the fans turn more slowly for a given water temperature. Given a large surface area even the smallest air flow has a significant effect on the cooling ability of the system.
 
Must be flow rate against wet wall area (shows if rad tubing is furred up) if you put a fast flowrate thoughrads you will get the worst poss situation as to high a rate does not give the water time to give up its heat.
I run complete passive and a reasonable overclock and can run 12-14hrs video encoding and temp's don't top 38-40c so it will work passive depends how you want to run the system.
 
Jay Bee said:
if you put a fast flowrate thoughrads you will get the worst poss situation as to high a rate does not give the water time to give up its heat.
Click to expand...

Common sense :)

The problem with water cooling is the results often seem counter intuitive :( At low velocity the water tends towards laminar flow, this creates a relatively slow moving layer of water nearest the surface of the block or rad pipe. This slow moving layer effectively insulates the faster flowing inner volume of water. A high velocity tends towards turbulent flow stripping away the insulating layer allowing more water to come into contact with the surface of the block or rad pipe. I came across some nice test results quite recently, unfortunately they were on the internet and I forgot to bookmark - probably on xtreme if it ever comes back on line.


Grats on your passive results by the way, I compromised. It started slowly with just one fan, then another until I finally went back to a full rad system. Oh well, it was great fun at the time.
 
Jay Bee said:
Must be flow rate against wet wall area (shows if rad tubing is furred up) if you put a fast flowrate thoughrads you will get the worst poss situation as to high a rate does not give the water time to give up its heat.
Click to expand...

ecat said:
More radiators = greater pressure drop, this is bad :( Assuming two rads the obvious solution is to run the rads in parallel
Click to expand...

The obvious solution is to get a single bigger radiator if feasible, as it avoids extra restrction/pressure loss from the 90 degree bends of going into the radiator plenum and the pressure drop from going to the larger cross-setional area of the plenum compared to tubing.

Running in parallel causes pressure drop from halving the flow resistance when the flow splits (assuming the split lines are same diameter as the original), and also means the flow through the radiators is half as fast - the water spends twice as long in the radiator. This is a bad thing in terms of heat transfer because the given amount of water gives up more of its heat, lowering its temperature, and lowering the temperature difference between the water and air going through the radiator means less heat transfer. It's better to have higher flow, because the water is replenished with warmer water, with a higher delta T to the air. Faster flow, as mentioned, also promotes turbulence rather than laminar flow, though I suspect that flow will still be laminar in the very thin flattened radiator tubes (turbulence is more easily achieved in higher diameter tubing). You have to remember that it doesn't matter how long a given 'unit' of water spends in the radiator per pass through the loop if you like, since that 'unit' will always spend the same amount of time in the radiator regardless of whether it goes through the loop (and therefore radiator) once every 20 seconds or once per hour. Higher flow is better, though the gains in higher flow get less the faster the flow - it's a case of diminishing returns (asymptotic).

Cathar (of Storm waterblock fame) tested different diameter tubing for a standard length loop typically used in a pc (6ft ish iirc) and with typical component blocks and found the difference to temperatures was minimal between 1/4" at one extreme and 1/2" at the other - it was something like 0.5C or less difference. The Germans quite happily use low-bore, low flow systems.

It's generally recommended that your loop have at least 1gallon/minute flowrate - which is about 90% ish of the maximum it could achieve iirc. You also have to remember that to get higher flowrates you need a more powerful pump that dumps more heat into the loop to get rid of, so while some people invest in crazily powerful Iwaki pumps they probably don't see much benefit, unless they have lots of very restrictive blocks and the powerful pump boosts their flow from very low to high.
 
Monkey Puzzle said:
snip
Click to expand...

All good stuff with the possible exception of:

Monkey Puzzle said:
This is a bad thing in terms of heat transfer because the given amount of water gives up more of its heat, lowering its temperature, and lowering the temperature difference between the water and air going through the radiator means less heat transfer. It's better to have higher flow, because the water is replenished with warmer water, with a higher delta T to the air.
Click to expand...

By the same token does the resulting greater delta T between the water and block not result in a greater transfer of heat at the block ? Common sense tells me the equilibrium point would remain about the same ? But we already know how far common sense get us :D
 
Wow. Cheers all, I've learnt a lot from the above. Decided to try water sooner rather than later, but put it in a new thread for fear of stopping any further useful information emerging here :)
 
ecat said:
All good stuff with the possible exception of:



By the same token does the resulting greater delta T between the water and block not result in a greater transfer of heat at the block ? Common sense tells me the equilibrium point would remain about the same ? But we already know how far common sense get us :D
Click to expand...

Not quite sure what you mean - faster flow means greater overall transfer of heat from water to air (up to a point), since it minimises the decrease in heat transfer due to the water getting closer to ambient air temperature as it leaves the radiator - more heat is transferred from water to air from warm water than from a radiator with warm water at the inlet and near ambient temperature at the outlet.

To use the unit of water analogy -

Loop A: with 10 'units' of water in the loop, and it takes a minute for a unit to pass through the radiator

Loop B: with 10 'units' of water in the loop, and it takes 6 seconds for a unit to pass through the radiator. Identical to loop A except flow rate 10x higher.

Loop B transfers more heat overall, because while in the first loop that 'unit' will be colder when it leaves the radiator, the 9 units behind it are not transferring any heat (sat in pvc tubing). In the second loop the individual 'unit' of water spends only 6 seconds in the radiator, but the 9 other 'units' have all been through the radiator as well. The heat transferred per unit is more than 1/10th of the unit in loop A, so the total heat transferred is higher.
 
Last edited:
You must log in or register to reply here.
Back
Top Bottom