Air makes its way from the reservoir into the radiator.

[Vya Domus]

I got this strange issue where over the course of a couple of weeks some of the air from the reservoir slowly makes it's way to one of my radiators. How that happens is a mystery to me, there should be no way for the air to end up there as far as I can tell.

The air gets stuck at the top of the rad in the second picture, it wouldn't really matter to me but once that happens it makes a pretty loud gurgling sound when the pump speed goes up. Anyone has any ideas ?

 

[thesmokingman]

Evaporation... there's a formula for it.

If air was getting in you'd have sprung a leak. If there's trapped air, that means you don't have enough pump to push the air pockets out. If that's the case you're gonna have to play the tilt the whole rig till you get the trapped air out game.

Btw, I absolutely despise tiny reservoirs man!

 

[Vya Domus]

The water that evaporates should just end up back in the reservoir, this already happened twice, every time I flip the case to get the air out of that one radiator a few weeks later it somehow makes it way back again.

 

[ir_cow]

It's normal for water levels to go down over time from evaporation. I lose a little every month as well.

If you just filled the loop, after about a week all the micro bubbles come out and you'll notice a dip as well. So if your not leaking, top it off and your good to go.

 

[Vya Domus]

No but that's the thing, the water level goes up, this has been happening for like a couple of months.

Air gets trapped into the rad, I flip the case around, water level comes back down at the same level as before, then after a few weeks it goes up again. I haven't done anything to the loop since I last put it together. I don't see any bubbles when in use, obviously at some point the air is somehow sucked into the pump but how I can't understand.

 

[thesmokingman]

No but that's the thing, the water level goes up, this has been happening for like a couple of months.

Air gets trapped into the rad, I flip the case around, water level comes back down at the same level as before, then after a few weeks it goes up again. I haven't done anything to the loop since I last put it together. I don't see any bubbles when in use, obviously at some point the air is somehow sucked into the pump but how I can't understand.

It's cuz you're using a tiny res, any disturbance and it sucks air into the loop that is in the tiny ass res.

 

[Vya Domus]

It's cuz you're using a tiny res, any disturbance and it sucks air into the loop that is in the tiny ass res.

Man I thought this as well but I had to tilt the case like 45 degrees with the pump full speed for the air to make it's way to the pump inlet inside the reservoir, I can't see how this would happen in normal use.

 

[BoggledBeagle]

You can make the 240 rad on the back wall a cross flow rad by putting Y splitter in front of both lower connections a draw the water from the top. The air will obviously have no chance to accumulate there.

You can also have 2 outputs from the rad, one on the top, while slightly lowering the rad performance, because a part of the water will leave it early.

 

[Steevo]

Oxygen dissolves in your coolant, the turbulence and pressure change in the radiator causes it to come out and forms air pockets.

 

[Shrek]

Heating drives the dissolved oxygen out.

 

[Vya Domus]

You can make the 240 rad on the back wall a cross flow rad by putting Y splitter in front of both lower connections a draw the water from the top. The air will obviously have no chance to accumulate there.

This rad doesn't have a top port unfortunately, it would have complicated things too much regardless.

Heating drives the dissolved oxygen out.

Oxygen dissolves in your coolant, the turbulence and pressure change in the radiator causes it to come out and forms air pockets.

Maybe this is what's happening but seeing as the coolant is constantly moving how come it can accumulate like that this quickly ?

 

[Noci]

It is indeed one of the mysteries of water cooling. I don't buy the Oxygen theory as in heating/cooling of the liquid it is very unlikely, it just doesn't make sense in chemistry.

I would rather go with the trapped air combined with the evaporation theories, as they seem more plausible. For example after running my loop for almost a year, sometimes I still see bubbles entering the reservoir. I also have to regulary adjust the level, but as time goes by less and less, if it makes you feel better . But I never kept track of how much volume over which period of time (a nice long term project for one who wants to research this).

I have approximately 2 liters of cooling liquid in the system and after a couple of weeks the level is lowered around 1 centimeter in a 50mm diameter tube reservoir (see picture).

You have soft tubing, which are known for some miniscule evaporation just as with AIO coolers. It's by far not enough to explain that bubble in your resevoir, but it surely can contribute to it.

My explanation (feel free to disagree, but then please subtantiate your dsissemination):

Eventhoug you don't see air being recirculated by your pump, micro-bubbles are circulated all the time but are too small to be visually noticed. Once they get 'cornered' in a low flow velocity section in your rad they accumulate untill they form bigger bubbles that eventually will be caught again by the flow and end up in the reservoir like a visous circle. So if you bleed the air in the reservoir and fill it up with cooling liquid, after a while it should stabilize.
So unless you have a leak, there's nothing to be worried about.

 

[A Computer Guy]

It's cuz you're using a tiny res, any disturbance and it sucks air into the loop that is in the tiny ass res.

I'm about to use an FLT80 pump/res combo. I might have to look forward to this same issue.

Man I thought this as well but I had to tilt the case like 45 degrees with the pump full speed for the air to make it's way to the pump inlet inside the reservoir, I can't see how this would happen in normal use.

You also have a lot of radiator. It will take some for any air trapped there to make it back to the reservoir. Just keep topping off the res at the times you get the most air. If the fluid level gets too low in the reservoir the D5 might suck the air back into the loop. I have a picture of this when I was testing my new pump last year for noise. In the picture below on the left my distroplate was only about 1/2 filled and the suction was so strong without any of the other components in the loop the water vortexed and sucked in liquid/air mixture almost instantly as the pump came on.

Here you can see in my 011D distroplate I actually measure my coolant drop from the port to the first screw hole just below it. When it drops to the first screw hole I top it off again with some spare fluid or fresh distilled water depending on what I have on hand. You can also see condensation within the unfilled portion of the distroplate. After adding my 2nd 360 rad and refilling it took only about 3 weeks to drop to that level the first time (no leaks) despite my efforts to bleed the bubbles out. The 011D is heavy and it gets pretty tiring tilting and moving the thing around so I don't spend much time doing it anymore and just keep spare liquid around expecting to top it off until it stabilizes.

 

[Vya Domus]

You also have a lot of radiator. It will take some for any air trapped there to make it back to the reservoir.

My problem is the exact opposite, air from the reservoir makes it's way back into rads.

 

[A Computer Guy]

My problem is the exact opposite, air from the reservoir makes it's way back into rads.

If you flip the 2 rads from port side down to port side up it might be easier to get the air out of them back into your res. Some tactical quick disconnects would allow you to hookup a 2nd pump to help drive the air into your res and/or connect a 2nd res at the top of your loop to collect air until it's finished bleeding out.

 

[BoggledBeagle]

He would be fine just with those two cross flow 360s. His PC is simply overraded.

 

[Vya Domus]

If you flip the 2 rads from port side down to port side up it might be easier to get the air out of them back into your res. Some tactical quick disconnects would allow you to hookup a 2nd pump to help drive the air into your res and/or connect a 2nd res at the top of your loop to collect air until it's finished bleeding out.

There is no way to flip the radiator on the back because it interferers with the port from the top one, I thought about it but there is no easy way to do it. What I find strange is that the side rad never develops this pocket of air despite being in the same orientation.

He would be fine just with those two cross flow 360s.

Of course it would be fine but if there is a spot where I can fit a rad I will.

 

[BoggledBeagle]

What I find strange is that the side rad never develops this pocket of air despite being in the same orientation.

You have this rad as a first item after the pump and the flow in it is probably so lively (turbulent?) that it can carry away any air collecting on top. On the other hand, the problematic rad is almost the last item.

BTW there is a pressure drop along the water flow route and lower pressure means higher degassing from water. The higher pressure, the more air can dissolve in the water.

The smallest pressure is of course in the reservoir, but when you calculate how long is route which the water flows in the loop, the problematic rad is at 2/3 of the route and it has two major pressure drops in behind it (the blocks). If I calculated that correctly, your loop is almost 4 meters long.

You have 11 120mm fan rad surface, are you sure you could not survive with just 9 and remove the 240 rad?

 

[Vya Domus]

You have 11 120mm fan rad surface, are you sure you could not survive with just 9 and remove the 240 rad

This is a small annoyance, I am not going to remove it just for this.

 

[Noci]

You have this rad as a first item after the pump and the flow in it is probably so lively (turbulent?) that it can carry away any air collecting on top. On the other hand, the problematic rad is almost the last item.

This is actually the other way around, in the rad (no matter if it is Cross-flow or U-flow) the liquid slows down due to the relative large passthrough in the end chambers, thats where air bubbles accumulate untill they are large enough to get caught again by the flow.

BTW there is a pressure drop along the water flow route and lower pressure means higher degassing from water. The higher pressure, the more air can dissolve in the water.

Really? Do you realise we are talking about a delta-P of just some millibars in a system like this. Don't confuse degassing with evaporation, which seems to be the fact in your dialogue.

The smallest pressure is of course in the reservoir, but when you calculate how long is route which the water flows in the loop, the problematic rad is at 2/3 of the route and it has two major pressure drops in front of it (the blocks). If I calculated that correctly, your loop is almost 5 meters long.

I would love to see a detailed calculation of this . In front of the blocks there is actually a pressure increase, not a decrease or drop as you call it. Due to the (intended) design of the blocks with a jet-plate and narrow finstack canals there is a restriction, which results in a small pressure increase and also locally higher velocity of the liquid. This significantly improves the efficiency of these parts.

You have 11 120mm fan rad surface, are you sure you could not survive with just 9 and remove the 240 rad?

More heat exchanging surface is always better (up to the point of diminishing returns) to increase the efficiency of the cooling system as a whole. More rads does not mean more cooling it is just easier to dissipiate the heat with lower airflow through the rads, so lower rpm's on the fans = less noise.

The length and size of a loop don't matter on this small scale, as long as the pump is capable of circulating the cooling liquid with approximately 100 l/h for the best results.

So before posting BS and confusing peeps with false information, get your facts right.

 

[Steevo]

Water molecules will evaporate through plastic tubing. Oxygen dissolves in water, otherwise fish would die. Same principal allows an equal volume of C02 and soda or beer in a can or bottle.

 

[BoggledBeagle]

I would love to see a detailed calculation of this . In front of the blocks there is actually a pressure increase, not a decrease or drop as you call it. Due to the (intended) design of the blocks with a jet-plate and narrow finstack canals there is a restriction, which results in a small pressure increase and also locally higher velocity of the liquid. This significantly improves the efficiency of these parts.

When a part creates flow restriction, there is higher pressure of the input water and lower pressure of the output water. The pressure drops.

You can see in the reviews on this site, that radiators have up to 1 PSI and CPU water blocks up to 1,5 PSI pressure drop. If you have 4 rads and two blocks in serie, you have total of 7 PSI pressure drop along the loop, if you ignore all tubing and other restrictions. That is 50 kPa difference.

BTW in a soda bottle you have pressure cca 250 kPa. Note the amount of gas coming out of the beverage when you pour it in a glass.

So before posting BS and confusing peeps with false information, get your facts right.

Ok...

 

[redeye]

the pump should be in the lowest part of the loop. your distro plate is in the top, and most likely is not under water when off, because air rises and displaces the water. (i’m guessing)
move the distro block to the bottom of the case.
i have a water fountain feature in my case, and do not have gurgling sounds etc. i assume it is because the pump is the lowest part of the case,

 

[Vya Domus]

the pump should be in the lowest part of the loop. your distro plate is in the top, and most likely is not under water when off, because air rises and displaces the water. (i’m guessing)
move the distro block to the bottom of the case.

The water level is definitely above the pump inlet at all times even when off but I don't see how moving the distro pump will do anything seeing as the 240 rad is always above it's level anyway.

 

[Shrek]

My central heating system has a device has a float to leak off gas; is there anything simillar for these small systems?

Maybe a T with a transparent tube sticking up to catch gas might be a nice addition.

It is indeed one of the mysteries of water cooling. I don't buy the Oxygen theory as in heating/cooling of the liquid it is very unlikely, it just doesn't make sense in chemistry.

Heat some water and one can see the dissolved oxygen being released long before any boiling... which leads to the idea of boiling the coolant before use...