would adding alcohol in water loop help performance ?

samer.forums · Nov 4, 2017

Hello,

would adding alcohol in water loop help the cooling performance ?

and if yes , what percentage ? 40% ? 20%?

USAFRet · Nov 4, 2017

I can only imagine that it would make thing worse.
Alcohol and water don't mix.

samer.forums · Nov 4, 2017

USAFRet :

should I try it ? would it harm any thing ?

USAFRet · Nov 4, 2017

samer.forums :

Well...alcohol may be more toxic to all the fittings and soft bits (seals) in your loop.
You may come home to find the alcohol has eaten the seals, and your loop has peed all over your hardware.

samer.forums · Nov 4, 2017

USAFRet :

samer.forums :

Well...alcohol may be more toxic to all the fittings and soft bits (seals) in your loop.
You may come home to find the alcohol has eaten the seals, and your loop has peed all over your hardware.

Okay thanks.

13thmonkey · Nov 4, 2017

You need to understand the concept of heat capacity, water has an amazing ability to soak up heat energy for a given rise in temperature of the water, Alcohol is much worse. Helium and Hydrogen are better, lithium is close behind and can cope with higher temps, BUT your CPU can't. Alcohol is roughly 1/2 that of water and starts to boil at a much lower temp.

So water is almost the best possible material, I say almost as there is bound to be better, but they may be toxic, or not function in a useful temp range (like sodium in a reactor).

samer.forums · Nov 4, 2017

13thmonkey :

If it boils faster then it takes heat faster as well ... put water on your hand you wont feel cold like when you put Alcohol on your hands. Alcohol absorbs heat faster than water. while water can keep more heat in (capacity) both work in different ways.

13thmonkey · Nov 4, 2017

It's the latent heat of evaporation that you are feeling at that point. Which is fine when you want your medium to evaporate, like in a fridge cooling circuit when you can re-compress it. Cooling loops aren't like that, alcohol boils at around 78C, so sat over your processor at even 65C you'll be getting vapours forming, and at 78 it'll all be trying to be a gas, the pump isn't designed to pump gas, well done you've got an airlock, and the remaining liquid won't move around the system. It's basic physics really.

13thmonkey · Nov 4, 2017

It takes 2.4 joules of energy to raise 1g of alcohol by 1C, it takes 4.2 for water. So in 100s a 100W (100 J/s) processor will move 200g of water by 100x100/4.2/200 = 11C for alcohol it would be 100*100/2.4/200 = 21C

So water is a much better choice at moving heat from the CPU to the radiator, you'd have to move alcohol nearly twice as fast (assuming it didn't boil, and assuming that they have equal densities, which they don't but accounting for that would just make the situation worse.)

The Original Ralph · Nov 4, 2017

USAFRet :

i believe water and alcohol do mix, - back in the 70s when i mechaniced at a dealership, the rage was to use "Gas-Dri" in the gas tank - it attracted the water in the tank, and helped break it down into smaller molecules or clumps (really technical term there) so it could pass thru the fuel system. A second example is rubbing alcohol, some is 70% and some is 91% - those percentages refer to the amount of alcohol to water

but to your second point, the effect on the seals, i agree entirely

to the OP - if you can determine that the seals and the tubing are compatible with alcohol, it should help cool your system, but i'd really recommend checking with who ever you sourced the cooling system from to see if they'll tell you whether the components are compatible. Alcohol can also be "corrosive" to the plastic in the pump - by corrosive i mean it can make the plastic brittle, cause it to shrink and shrink enough to cause cracks, which would mean leaks - just so you know

samer.forums · Nov 4, 2017

so basically , I will need metal tubing and a special pump. This is going to be very expensive to test.

I was thinking to put the pump After the Radiator , this way No Alcohol Vapor will pass through the PUMP.

but anyways this is way beyond my budget , just for testing.

USAFRet · Nov 4, 2017

samer.forums :

The 'fluid' circulates throughout the whole system.
If some of that 'liquid' is passing into gas anywhere in the piping...bad things will happen.

(in addition to alcohol not being anywhere near as good as water)

13thmonkey · Nov 5, 2017

samer.forums :

Did you not understand the sums? Alcohol is worse, period. A hot CPU will turn it to vapour, and it is less than 50% as good at moving heat by volume (as it is roughly 80% the density of water), and volume is what counts.

This is not a matter of opinion as to what is better, it's just physics. You may find that some additives are used to reduce surface tension to improve whetting, reduce mould formation, but these will do little to change the heat capacity.

samer.forums · Nov 5, 2017

13thmonkey :

I do understand what you said. and I never said any opinions. I just like testing stuff .

your physics is missing huge part .. The Radiator will cool down alcohol faster than Water ... you did not put the fast cooling rate into your calculations , you just calculated the capacity of taking in heat , and missed to add the speed of heat transfer in and out of Alcohol.

if the Alcohol takes heat fast and looses that hat fast , you will not need alot of heat capacity to keep things cooler.

My Thoughts were about faster heat dynamics vs more heat capacity..

Actually I am thinking now about using Ammonia instead of Alcohol...

The Original Ralph · Nov 5, 2017

samer.forums :

i'll try this again - no need for metal tubing, as some plastic tubing is compatible with or impervious to alcohol - you just need to find out from whoever you sourced the kit from to see if the tubing they supplied is good to go. Otherwise, here's a link to a good source for a selection of tubing with specification included with each tubing https://www.mcmaster.com/#standard-plastic-tubing/=1a4ev4e. When you find a tubing that is good to go with alcohol, be sure to research the engineering spec listed for it, if one is - it's doubtful you'll find one that will tolerate more than 20% alcohol. If your tubing is metric, change the search to "metric plastic tubing". Seals - if your system has O rings, they're probably metric as well, as everybody is sourcing their materials from china - Flourosilicone are impervious to everything short of hot lava. If you are tempted to go with silicone tubing, don't - it's real vulnerable to breaking wherever it's scratched - similiar to sheet rock, it will break wherever the surface is cut / scratched.

as to the pump, if there's a mfgr's identity on the pump, contact them re the alcohol issue.

whatever percentage alcohol the tubing shows to be good for, don't get brave and exceed that spec. Back in the 90s, when congress passed the "ethanol" requirement that all gasoline be 10% ethanol, a lot of car engines were destroyed when a refiniery in texas mixed one batch too rich with alcohol.

13thmonkey · Nov 5, 2017

samer.forums :

13thmonkey :

I do understand what you said. and I never said any opinions. I just like testing stuff .

your physics is missing huge part .. The Radiator will cool down alcohol faster than Water ... you did not put the fast cooling rate into your calculations , you just calculated the capacity of taking in heat , and missed to add the speed of heat transfer in and out of Alcohol.

if the Alcohol takes heat fast and looses that hat fast , you will not need alot of heat capacity to keep things cooler.

My Thoughts were about faster heat dynamics vs more heat capacity..

Actually I am thinking now about using Ammonia instead of Alcohol...

Locally, at the block, it'll all be about heat absorption.

Try it and see, why don't you plumb it into your car's radiator?

USAFRet · Nov 5, 2017

Ammonia?

I'm surprised you've not started investigating a mineral oil tank, or a mini-fridge based system.

13thmonkey · Nov 5, 2017

Peltiers FTW!

samer.forums · Nov 5, 2017

USAFRet :

... I dont want noisy stuff , I will try the older ways , when they used Ammonia to make ICE even in the desert , without any noise or any noisy machine..

samer.forums · Nov 5, 2017

13thmonkey :

samer.forums :

Locally, at the block, it'll all be about heat absorption.

Try it and see, why don't you plumb it into your car's radiator?

no it is not about heat absorption alone , you are mistaken.

The cooling loop has 2 parts , heat absorption of liquid , and the cooling of that liquid and how fast it is cooled (Radiator).

you calculated the first part and ignored the second part.

USAFRet · Nov 5, 2017

If alcohol, or ammonia, or liquefied puppy farts were a better fluid medium than some flavor of "water", why is that not a design concept and selling point in mid to high end liquid cooler solutions?

I wonder why that is?
Could it be the result of actual testing? And they found that 'water' is indeed the best solution?

13thmonkey · Nov 5, 2017

the radiator is so remote from from the CPU block that it doesn't matter, all it controls it whether you can get the temp of the fluid entering the block back down to a useful temp, and lets say that you can. Lets go into the calcs in more detail at the block:

The block might hold 2ml of liquid, so that's 2g of water, and about 1.6g of alcohol. And for now assume a flow rate of 2ml/s

On a 65W processor in 1 second that would heat up the water by 65/4.2/2 = 7C, for the Alcohol the rise is 65/2.4/1.6 = 17C, so assuming an ambient of 30C for the liquid (close to case temp) you'll have water leaving the block at 37, and alcohol leaving at 47C. The Water entering the rad is at 37, and the alcohol is 47.

Both will lose some heat in the rad, but the alcohol in the block has much less thermal headroom before it boils, so if the rads don't get the alcohol down to 30 again, then you'll start to get the exit temps from the block climbing much quicker than for the water, and if it hits 78C, it will boil.

You have the additional problem that locally in hotspots in the block with a die temp of 70+C it will get very close to 78C, and it will boil. With water this can only happen if the CPU has hit it's throttle temp.

samer.forums · Nov 5, 2017

13thmonkey :

wrong , Alcohol will lose heat faster in the radiator than water , you need to calculate that right them we can agree or not.

I am not saying you are wrong by the way , I am just saying you are not calculating it right. also , in Alcohol the length of loop is a huge factor in cooling unlike water , because it cools down faster ...

one more thing to add , The alcohol loop should be vertical and not horizontal , so in case there is a Vapor between the CPU and RAD , it will move up ...

anyways I am thinking more about Ammonia now , I will see what I can find online.

samer.forums · Nov 5, 2017

USAFRet :

no one knows it all , and no one tested it all . actually it took us 70 years to move water cooling from cars to PCs .. you assume that the PC world is NASA LABS and that important.

well . it is not that important .. it is Just a PC and some people trying to overclock it.

Industry never OC their PC and will never pay to research anything. this is a hobby work. we all started it in the garage before it hit the market.

USAFRet · Nov 5, 2017

Obviously, you are not the first to consider alocohol instead of water:

http://www.hardwareheaven.com/community/threads/using-alcohol-instead-of-water-in-a-water-cooled-machine.139359/

https://arstechnica.com/civis/viewtopic.php?f=7&t=1037636
- The specific heat capacity of Ethyl Alcohol is about 2420J/kg*K whereas it is 4186J/Kg*K for water.

https://linustechtips.com/main/topic/7897-ethanol-cooling/

https://linustechtips.com/main/topic/590485-running-ethanol-in-a-water-cooling-system/

http://archive.atomicmpc.com.au/forums.asp?s=2&c=6&t=7347

http://www.overclock.net/t/55700/cooling-with-rubbing-alcohol

http://koolance.com/frequently-asked-questions#coolant_choices

would adding alcohol in water loop help performance ?

Notable

Titan

Notable

Titan

Notable

Titan

Notable

Titan

Titan

Distinguished

Notable

Titan

Titan

Notable

Distinguished

Titan

Titan

Titan

Notable

Notable

Titan

Titan

Notable

Notable

Titan

Share this page