Discussion - Less fluid better cooling?

Phrenetical · Nov 7, 2006

I have a questions to ask and hope it will spur some discussion on the mater and no flaming ffs.

Anyway my first WC loop, probably isnt the best setup, fair amount of tubing for convenience, plus using a swiftech microres, parts etc can be seen in sig blackened.

Anyway, im wondering since i was thinking about this, the only place in my loop that cools the fluid is the rad, so the more the fluid passes through it the better the cooling... so redo'in my loop with shorter tubing lengths, maybe a t-line and no res, and repositioning pump and basically have the minimum amount of fluid i can, theoretically have the fluid pass through the rad more often in a set time because there is less of it.

that would mean the fluid passes through the rad more often and hence would help cool and provide stabler temps at max load.

Is this a good idea or should i not bother with it?

Sleepy_Steve · Nov 7, 2006

Not sure...

I would think that the main effect is more rapidly fluctuating temps. Im not sure if that can be construed as less stable max temps, but maybe.

Running a loop that has less tubing, and a smaller res so you get more passes through the rad per hour... seems like it would be along the same effects as a better pump.

However given that you may need to double the flow rate of a pump to drop temps a few C, i would think reducing the volume of the loop is a tedious way of improving preformance.

darkcow · Nov 7, 2006

well, the more water there is, the more heat it takes to heat up that water. but the more time it takes to cool it down. its all relative i'm thinking.

but i think that having a lesser amount of water will probably allow for a much quicker cool down of the components because the water will spend more time in the rad then the pipes.

its always good to have it done correctly, and having less pipes will probably help in pressure from the pump.

i would just wait until that boring day comes along (we all have them) to just take it apart and do it. thats usually when i do big lengthy projects. of course things take longer to complete.

ƒÓÒl · Nov 7, 2006

In my mind, you'd get better temps because of the shorter tubing runs and fewer fittings lessens the restriction to waterflow.

As for more stable temps, there's too many factors to guess.
I use 2.5 gallons of water, and my rad is humongus, and it takes a while for my temps to change. So long in fact that 3-4 minutes of 100% cpu doesn't change my temps at all. Overall it only changes about 3C from idle to burn.
But I'd say that if your rad easily handles the heatload you've currently got, then it should get better just from the increase in flowrate (due to the better heat transfer that the waterflow increase will buy you).

Some of the causes of wide temp changes can also be:
Too much TIM
Too small rad
Too little fannage
Cruddy waterblock (meaning deposits from water/dye/whatever)
Tubing kinks or airbubbles stuck at a fitting
+more

voigts · Nov 7, 2006

darkcow said:
well, the more water there is, the more heat it takes to heat up that water. but the more time it takes to cool it down. its all relative i'm thinking.

but i think that having a lesser amount of water will probably allow for a much quicker cool down of the components because the water will spend more time in the rad then the pipes.

its always good to have it done correctly, and having less pipes will probably help in pressure from the pump.

You guys are thinking about this as if it was an open system, not a closed system. It makes absolutely no difference how much water is in the loop unless the amount of water acts as a way to dissipate heat itself, ie a 25 gallon reservoir or such. Flow rate is all that matters, whether you are talking 1 litre or 10 litres. The more water you have, obviously it will take a bit longer for the water to reach equilibrium, but once it does, it will only take longer for it to cool off once the computer is turned off.

Read this thread from earlier in this year about understanding radiators. Thorilian (no longer on OC unfortunately) explains these concepts far better than I: http://www.ocforums.com/showthread.php?t=456189

Here is an excerpt:

TreeNode said:
Its more like XFlow has higher flow rates while 2-pass (dual row) rads allow the water to stay in the rad longer to cool off more. But, if you consider them side by side with the same pump, an XFlow system would work better with a lower powered pump than a 2-pass rad. With a high powerful pump, a 2-pass would be better.

Thorilian said:
you have to be very carefull here . you are heading the wrong direction and are about to mislead people into popular urban cooling legend

water stays in a rad the same amount of time in a closed loop weather its single or dual pass so flow is always the most important. while technically if you ONLY count a SINGLE revelution in the cycle ONLY keep in mind , THEN AND ONLY THEN is it possible for water to pick up more heat from a dual rad.
( does not include relationship between DT because its only a single cycle)

once you establish that water cooling is a closed loop and will REcycle meaning make more than 1 pass through the loop, you have just established that water will spend the same amount of time in the rad.

how this relates to efficiency is the DELTA T

you want your DT to be as efficient as possible aka as wide as possible.

that means the diference between the hot and cold basically. which will cool a hot frying pan better ? 1 liter per minute of tap water or 1 liter per minute of ice water?

ice water of coarse. so to let water stagnate inside a 2 pass rad is going to be less efficient. your ice water just warmed up and is not making any room for new ice water to come in..

if this is confusing you need to review that this is a closed loop in which the distance the water travels is always the same and the amount of water is the same. .

a race track... say you have a race track and you go 5 miles an hour around the track ( slow flow). you go around a 1 mile track in 12 minutes . its still 1 mile though.

NOW go around a race track at 60 miles per hour . you just went 1 mile in 1 minute. YOU STILL ONLY WENT 1 MILE . you did it 12x as fast ( high flow).

but you picked up more heat ( more than slow flow) because your ice water was FRESHER ( spent less time and kept the Delta T here after called Dt, WIDER). this is where spead aka flow creates efficiency.

TreeNode said:
Though, this is not even considering the flowrates through the blocks and hose length which will alter things. If you have restrictive blocks, you can't possibly get the water circulating the "right way" through an XFlow anyhow so you would go with a 2-pass to balance it out.... etc.

Thorilian said:
now this is just ... i dont know.. i cant see what you are trying to demonstrate.
maybe if you re word it i can figure out what you are trying to say.

pumps are rated to work in efficncy with certain head ratings at certian speeds. so some pumps are like a slow moving wide deep river . they have more force behind them but they are not moving as fast BUT they overcome minor restriction better so that it does not slow their flow FURTHER .

high flow low head pumps are the oposite like a shallow stream that runs at much higher speeds ( or velocity if you like big words). the problem is they are reduced in speed very easily by restriction so thier speed could slow down a LOT more .

so if i have a pump that pumps 10 gallons a minute and its low power and the restriction slows its speed down to 2 gallons per minute it will not be as effective as a higher head pump that pumps max speed of 6 gallons a minute but WITH restriction only pumps 3 gallons . basically the higher head overcame the resitance more effectively to have an over all higher flow to restriction rate..( these are demonstration synthetic numbers only , not based on any particular pump)

this is where choosing your pump is important . some people over do the pump they need in either to high a flow or to high a head for thier system . its a HUGE balancing act where you haev to match your block and rad ( with Fan ) to the right pump for you.

ways to manage the balancing act .
reduce restriction through shorter tubing runs
reduce restriction through less bends in the loop ( this includes going from dua l pass to single pass)
use a pump that is suited to the systems OVERALL restriction require ment and keeps the flow near the water blocks most efficient speed( this also balances the speed vs deminished returns)
balance everything vs noise and stable OC (aka how much bleeding ears are you willing to put up with)

some hints that i dont usually give system designers but might as well.

you can eliminate tubing lenght by using VERTICALLY mounted single pass cores .
keep the pump at the BOTTOM of the loop so that the pump outlet is very close to the core bottom inlet ( this should reduce a large portion of tubing)
the cores outlet should sit just below the water block inlet so that water is still traveling in an upward movement( this bleeds the system as air travels up and with flow not against it)
from your water block you should have a VERTICALLY mounted res or air trap/tline that travels straight down to the pump inlet ( the res acts as both res air trap and tubing emoving even more tubing from the loop)
note that a good res will be tall and skinny with the water INLET slight below the fill opening BUT still higher than the OUTPUT of the water block so that air travels out of the block UPWARDS into the res where it travels to where you fill the res

it literally takes me under 1 minute to completely fill and bleed a system 100%.

your bigest problem with the design is the case. thats why i prefur tall lian li cases for this.

Keep in mind that an Xflow rad is not a true single pass core. It is a dual pass core with different end tanks slapped on. Thorilian is referring to single pass heatercores, or in rads, this would be the PA160 as it is a true single pass rad. The Xflow has been shown by Marci among others to be up to 20% less efficient than its dual pass cousin at the flow rates we normally run at.

Thorilian said:
on the race track .. its 10 miles long( the whole loop) we will say and 1 mile is the rad(pit area) .

if your system flow is 20 miles an hour , in 1 hour you will have completed 2 laps of covered the rad twice or 2 miles of cooling surface. that means you spent 3 minutes per lap in the radiator x2 or a total of 6 minutes

now lets double the speed of the flow . the track distance stays the same and the pit area stays the same . only the flow changes.

if your system flow is 40 miles an hour , in 1 hour you will have completed 4 laps of covered the rad 4 times . that means you spent 1.5 minutes per lap in the radiator x4 or a total of 6 minutes

So in a closed loop, the water spends the same amout of time in the rad regardless of the flow rate or amount of water. The flow allows for a wider Delta T and turbulence within the rad and hence better heat exchange.

freakdiablo · Nov 7, 2006

well, in a sense, it would kind of work. especialy if you have one of those passive heatsink reservoirs. the water spends more time in the res, so it has more time to cool down.

voigts · Nov 7, 2006

freakdiablo said:
well, in a sense, it would kind of work. especialy if you have one of those passive heatsink reservoirs. the water spends more time in the res, so it has more time to cool down.

Immortal_Hero · Nov 7, 2006

How many times do we have to go over this. The water will only get as cool as the rad will cool it with a given heat input to the water. Changing the amount of water is only going to change the heat up and cool down time it will not change the max temperature (in a closed loop). The only way this would work is if you were drawing water from a lake and dumping it into something else on the other side of the loop (open loop). The rad and flow rate determine the max temp for a given heat dump into the water. Increase flow rate and you will see better temps. Flow rate could be improved by decreasing restriction (ie less tubing).

voigts · Nov 7, 2006

Immortal_Hero said:
The rad and flow rate determine the max temp for a given heat dump into the water. Increase flow rate and you will see better temps.

As a general rule, yes, however, increasing flow rate will result in better temps only to a point. The characteristics of the waterblocks themselves also come into play as some waterblocks only benefit from flow up to a certain point, and some benefit more from higher flow than others.

If you make slightly shorter tubing runs and go from a res to a t-line, the increase in flow may be minor enough that you don't even see any noticeable temp changes.

ƒÓÒl · Nov 7, 2006

True that, the Apogee in that Swifty Apex kit (had to look it up) doesn't gain much C/W between 1 & 3 GPH, but it does some from .5 to 1.
I don't know what the Swifty kit actually pushes, but if it's starting out pretty low, an increase can improve it.

In BillA's radiator testing, the BTU's given off didn't really increase between .5 and 1 GPH, some actually dropped, but there aren't any current models in that test save for the heatercore.
Maybe I'll look some more.

Maviryk · Nov 7, 2006

The way I see the question:

Less wire more electricity?

If less fluid means shorter tubing runs, then yes. But if less fluid means smaller ID tubing, then no.

But I regress, it's already been said in previous posts.

ctrl_alt_del_ · Nov 7, 2006

ƒÓÒl

Do you have any pics of your system with the big radiator? I'd like to see it!

ƒÓÒl · Nov 7, 2006

Nah, no pics...wouldn't want to make the children cry

I spend little time on teh purty, and much more time on the function.
I use a Toyota truck main radiator (all copper and brass, pretty much like a Monster Black Ice, but single-pass) and a gallon reservior with 1" tubing running the 8' between the rad and the computer, thus all the water.
I went big because I want to convert to geothermal cooling someday, and the bigger the tubing, the more surface area on a burried pipe.

QuietIce · Nov 8, 2006

Phrenetical - With the MCP655 you have in your loop the tubing change would have a minimal impact - that pump plows right through any resistance from tubing you currently have. However, you might want to experiment a little with the pump settings. If you're running the pump full tilt (P5) chances are you could turn it down some without any meaningful loss of flow and reduce your heat dump (as well as the noise level

) ...

ctrl_alt_del_ · Nov 8, 2006

ƒÓÒl said:
I went big because I want to convert to geothermal cooling someday, and the bigger the tubing, the more surface area on a burried pipe.

You're talking about digging a trench on the backyard and putting copper tubes on the ground and have that cool the water? Now that is a good idea... I bet that water will probably be around 50-60 degrees?

Sleepy_Steve · Nov 8, 2006

If you do that... you have to remember how deep the ground will freeze in the winter.
As well as make sure that the section of the pipes that pass through the frozen layer of soil will be well insulated.
Oh and prolly treat whatever pipes you are droping into the ground to be corosion resistant as well.

But i think it does have the potential to go below ambient room temp. with potential to go below 32F depending on location and setup.

In general, you will see soil temps similar to that of a wine cooler... 58-62 deg F when below the frost layer (i think) but yes ctrl_alt_del_ had the general idea and ballparck numbers.

Maviryk · Nov 8, 2006

Eh... I wouldn't worry about freezing temperatures. Flowing water does not freeze. And especially not heated flowing water.

That's why you only have to leave a drip in your faucets in the winter if you leave your house for an extended period of time, instead of leaving it on full blast.

Graystar · Nov 11, 2006

Phrenetical said:
Is this a good idea or should i not bother with it?

The answer to your question is that you should not bother with it.

The only factor that ever matters is your radiator's ability to cool. That ability usually increases as flow rate increases, but somewhere between .5 and 1 GPM the ability to cool levels off. Yes, there usually is some additional increase as flow continues to increase, but it's minor when compared to total cooling capacity of the radiator. There's also the fact that higher flow rates do bad things elsewhere that are detrimental to cooling.

So if your flow rate is around 1 GPM already, then the only way to improve cooling is with a bigger or second radiator. I would leave the system as is.

nikhsub1 · Nov 11, 2006

Graystar said:
There's also the fact that higher flow rates do bad things elsewhere that are detrimental to cooling.

Care to elaborate on this mystery?

Graystar · Nov 11, 2006

nikhsub1 said:
Care to elaborate on this mystery?

Sure no prob...

First, higher flow rates mean more pump heat dumped into the water. So there's additional heat to get rid of.

Second, higher flow rates mean more resistance to flow. The heat increase from this has already been covered. The point here is that the increased flow is not commiserate with the increased power. That is, the flow rate gained by increasing power 5 watts is not going to be the same as the next 5 watt jump in power. There is a diminishing return. But you still have the 10 watts to deal with.

Third, as flow rate increases the average delta T at the radiator decreases. This reduces the efficiency of the radiator. That is the reason why some radiators don’t do much better with increased flow rates.

(As a side note for the curious, with increased flow, average delta T at the waterblock increases [that’s good] but only as long as the rad can still put out water at the same low temperature as with the slower flow rate. If these detrimental effects that I've described overpower the positive effects, then temp will likely stay the same. In bad cases the CPU temp can rise. Up to 1 GPH the positive effects win. Beyond that...it's iffy)

Discussion - Less fluid better cooling?

Member

Member

Member

Member

Member

Member

Member

Member

Member

Member

Member

Member

Disabled

Member

Member

Member

Member

Unoriginal Macho Moderator

Member

Similar threads