Efficient Water Flow: Designing a PVC Tubing System Without Air

[KingNothing]

I want to design a tubing system using flexible PVC (1/2" diameter) tubing that carries water in a loop. Obviously there is a pump too, but that's not the focus here.

What I want is a way to get all the air out, so that the loop contains only water. There is the idea of a "fill port", where there is an extra tube stemming from the loop which you use to fill it with water, and allow air to escape.

However, some manufactured products contain liquid loops that have no air at all, and don't have "fill port". How can I accomplish this?

 

[FredGarvin]

If it's a closed loop, it's not too terribly difficult. Things to watch out for are, obviously, high spots in your tubing in relation to your reservoir and pump. Also avoid sharp corners. In low pressure fuel systems I use here, these can be buggers because they tend to trap air bubbles. At the highest spot it is a good idea to tee off a small ball valve to use as a bleed. Once it is bled and you don't open the system up to atmosphere, you should be good to go.

There are a few finer points to consider, but I'll hold off until you give some more details of what your system is like and it's desired result. You could be complicating this without knowing. I'd be most interested to know what kind of pressure you want to run this at and if there will be appreciable heat transfer.

 

[KingNothing]

What I am trying to do is streamline a water-cooling system for a computer I plan to build. But a forewarning to others: Please, don't discuss computers here. This is about tubing.

It really only involves a pump, radiator, and 'block'. I'm not sure what the pressure is, but the manufacturer's spec for the pump is 140GPH.

One idea I had was to use a 'temporary reservoir' type of system just for the bleeding process. What I was thinking is to make a small hole in the tubing, about 3mm in diameter, and letting it rest within the 'temporary reservoir' to bleed air from and take water in. Imagine a straw on top of a glass, with the middle pushed into the water. Then, hopefully when it's completely bled, I could seal it off with some underwater epoxy?

 

[DaveC426913]

1/2" diameter is pretty big for a cooling system. You might consider smaller tubing (such as aquarium air hoses - they come in various diameters). Smaller tubing provides better heat transfer than larger tubing.

 

[KingNothing]

No. The liquid is what carries the heat in this application. It doesn't matter at all if the tubing itself is a good insulator or not. What matters is the amount of water flwoing through.

 

[Artman]

Removing all of the system's air does not allow for compression when the heat is added. You need to provide for expansion of the water or at least try and determine the effects expansion/contraction will have on your system. In a small enough system of minimal temperature rise, it may not matter, the system may be able to withstand the increase in pressure, otherwise you may at least have leaks, or at worst, a burst in your computer.

 

[Danger]

Perhaps an auxilliary pressure relief would be appropriate, similar to the coolant return tank attached to a car's radiator cap?

 

[Artman]

Perhaps an auxilliary pressure relief would be appropriate, similar to the coolant return tank attached to a car's radiator cap?

Probably just a high spot in the system such as a small tank or piece of larger diameter pipe, where air could collect out of the flow stream and act as a sort of compression tank. It may be overkill, but if the system water expands more than the system and the pressure exceeds the system capability, it will give out, somewhere.

 

[KingNothing]

There should not be any air at all. The tubing itself is flexible enough to handle the expansion. Air ruins water pumps.

 

[FredGarvin]

Perhaps an auxilliary pressure relief would be appropriate, similar to the coolant return tank attached to a car's radiator cap?

In any system, ideally there should always be a relief for safety purposes and to save the pump.

Is this going to a water jacket or is the tubing wrapped around something?

I would simply put a maximum high spot in the loop somewhere and a cheap-o small ball valve from the hardware store as a bleed port. That would be the easiest way to do it.

There are still places where you can run into trouble here. You haven't said if you plan on it being a sealed system or not. Also, do you have any idea of the delta T for the water will be? If you come close to vaporization, you can expect some other issues as well. You can also run into problems of air evolving out at low pressures. If you have any restrictions, etc... that will lower the static pressure, you may face some issues there as well.

 

[Q_Goest]

Hi KN, The way this is typically done is to have your pump suction drawing water from the bottom of a resevoir or tank of some kind and have the return stream come in near the top. Keep the tank at atmospheric pressure by allowing it to vent to atmosphere. By doing that, you've created a phase seperator which would effectively remove any air from the water before it gets back to the pump. It also gives your pump a very low suction pressure which will minimize seal leakage problems. And it also ensures your system pressure remains stable with atmospheric pressure at the pump suction and maximum pressure is then dependant on your pump.

To start up the system, turn on the pump (it only gets water in). Water gets pushed through the system and only some high spots, dead ends, (or corners such as FredG mentions) would retain any air. By cycling the pump a few times much of that could be removed. Dead ends and high spots need a bleed valve on them, or if its ok to leave the air in those sections (many times it is) you need not bleed the air at all. I've even seen automatic bleed valves you don't have to do anything with. I've not used the automatic ones so I can't recommend a supplier, but I'd bet there are some in your local hardware store or online at McMaster Carr http://www.mcmaster.com/ (search on "vent valve")

 

[DaveC426913]

No. The liquid is what carries the heat in this application. It doesn't matter at all if the tubing itself is a good insulator or not. What matters is the amount of water flwoing through.

You misunderstand. It's not about the thickness of the tube, or its insulating properties - it's not about the tube at all. For the purpose of calculation, the tube is assumed to be 0 thickness and a perfect conductor.

It's all about the amount of coolant at the heat exchange interface. You want to maximize the surface area (where heat exchange takes place) while minimizing the coolant volume (or most of your coolant will be dead weight).

A 1" section of 1/2" pipe holds pi/16 cubic inches of coolant and has pi/2 square inches of surface area.
Wheras FOUR sections of 1/4" pipe - holding the same amount of coolant (4pi/64) - has TWICE the surface area (4Xpi/4 = pi) for heat exchange.

That's why the tubing in cooling units is as narrow diameter as possible (you never see 1/2" pipe in cooling units unless they're very large and very heavy duty).

Moral: Before beginning construction: redesign it with the narrowest diameter tubing you can get away with. This is an important factor in your success - don't ignore it.

 

[KingNothing]

You misunderstand. It's not about the thickness of the tube, or its insulating properties - it's not about the tube at all. For the purpose of calculation, the tube is assumed to be 0 thickness and a perfect conductor.[/B]

Don't make such bold assumptions. Do you know how a watercooling system works? The whole idea is to concentrate and direct the heat transfer. Tubing is meant to carry the thermal interface (the liquid/water). One component in the system is the radiator, which is designed to radiate heat from the water. Moral: The tubing is not supposed to radiate heat.

Please, stick to the topic. I am looking for ways to completely seal off a tubing system after the bleeding process. How are the vent valves at McMaster used? From what I gather, they simply allow the coolant to flow through while bleeding the air out. This would be a nice feature for future minor bleeding, but how would I fill the system with water to begin with?

 

[DaveC426913]

OK, it was my misinterpretation of your description. I thought the tubing you were referring to was the coiled tube in the heat exchanger (wherein volume vs. surface area as well as conductivity is important).
It seems you are talking about only the transport portion of the coolant loop. I stand corrected.

 

[Artman]

There should not be any air at all. The tubing itself is flexible enough to handle the expansion. Air ruins water pumps.

Contain the air in a bladder in your reservoir. Then you can completely fill the system around the bladder.

 

[deckart]

You will not going to damage your water pump if you keep your reservoir above the pump creating a positive head pressure to the pump inlet. It's really not an issue. 140 GPH is 140 gallons per hour/2.33 gallons a minute. Overkill for heat dissapation in my opinion.

 

[KingNothing]

OK, it was my misinterpretation of your description. I thought the tubing you were referring to was the coiled tube in the heat exchanger (wherein volume vs. surface area as well as conductivity is important).
It seems you are talking about only the transport portion of the coolant loop. I stand corrected.

I'm sorry if I sounded angry or anything, your explanation of tubing for radiation will be a good read for anyone who happens to read this thread who is trying to design that.

Ideally I do no want a reservoir at all. I would like a completely closed loop, with no air and only enough water to fill the loop.

 

[Danger]

Maybe you could set it up as a reservoir system initially, then isolate and remove the reservoir once it's up, running, and bled?

 

[FredGarvin]

You should have a reservoir for a lot of reasons. The main one is that you will be constantly adding heat in two forms: direct heat transfer from the computer and work energy due to pumping/compression. You'll have to dissipate that heat somewhere.

 

[KingNothing]

Please, I beg PF to stick to the topic. I am simply trying to create a closed, fully bled loop of water without a reservoir. I have plenty of reasons for not wanting a reservoir. I should not have to fully explain how a computer water cooling system works to get help.

Maybe you could set it up as a reservoir system initially, then isolate and remove the reservoir once it's up, running, and bled?

This is essentially what I'd like to do. But I don't know how. Such is the question of this thread. I was thinking about possibly running the system for a few days in a box, then moving the box to a large freezer and letting the coolant freeze. Hopefully I could then chip the ice away, cut, and re-join and seal the tubing to remove the reservoir. Not sure how "solid" of a plan this is. Especialyl considering I may not use water, but rather something with a lower freezing point.

 

[Q_Goest]

KN, air conditioning and refrigeration systems are completely closed loop. Suggest you look into how that's done, but here's what I would suggest as a starting point:

- Keep the temporary reservoir as Danger suggests, basically set it up as mentioned before in order to bleed the system. Try to eliminate all high points in the design. If you have points in the system where air can collect and not be eliminated, ask yourself if it is ok for the air to remain there. If not, perhaps the entire assembly can be rotated so the high points become low points and all air can be flushed out by rotating in 3 dimensions. That's not very desirable for a production system obviously. You might consider having flexible lines in a few points so parts of the system can be rotated without moving other parts of the system. If so I'd recommend metal bellows hoses as opposed to plastic since plastic is very slightly permeable to air.

- To seal the system, you will need a bypass around the reservoir and back to the pump suction which is valved off when purging. It would look like this: imagine there not being a reservoir, the flow comes from the radiator which is rejecting the heat, goes through a T in the line (we'll come back to this) through a valve (valve A), another T in the line, and into the pump. The first T goes off through another valve (B) and into the reservoir. The water in the reservoir exits at the bottom, goes through another valve (C) and into the T at the pump suction. Run/cycle the system with A closed, B & C open till all air is out, then shut off the pump, close B & C, and open A. Remove the temporary reservoir by removing at the outlet of B and inlet of C (B & C are part of the cooling system). Cap or plug the lines. Another way lines are plugged is to have a copper line, crimp it closed, cut it and solder the end closed.

- Because this is a completely sealed system, you can't afford mechanical seals such as O-rings. The valve can be bellows or diaphragm sealed. Use metal bellows/diaphragms, not elastomer. The pump similarly shouldn't have external seals, use a canned pump such as the scroll pumps used in A/C systems.

- I would still be a bit concerned about thermal expansion of the water. Perhaps you'd want to test the apparatus without any feature to compensate, and if it becomes a problem, install a fully sealed accumulator such as a metal bellows sealed one.

Hope that helps.

 

[FredGarvin]

Please, I beg PF to stick to the topic. I am simply trying to create a closed, fully bled loop of water without a reservoir. I have plenty of reasons for not wanting a reservoir. I should not have to fully explain how a computer water cooling system works to get help.

Well then. Have fun. Don't let the door hit you on the ass on the way out.

 

[DaveC426913]

Well then. Have fun. Don't let the door hit you on the ass on the way out.

Um, FG speaks for himself.

 

[Artman]

Q_goest good post.

KingNothing, you were talking about a closed loop, without air, and adding heat. This is a recipe for disaster without expansion provision. You did not explain how the heat is to be removed from your system, we know ways to remove the heat, but how are you doing it? This effects the possible trapping of air.

An open system has air in it and the pump pulls from an open reservior. A closed system has to absorb expansion or release it. Air helps to absorb it in typical HVAC applications, but you don't want any air, so you have to count on your piping to absorb the expansion, this is possible, but be aware that seals will be stressed so if some water escapes and your system is closed, when the water cools it will try and make up that contraction (you don't have provision for that other than your piping) so in comes air through the seals or at fittings. This air has no escape route (you have left no provision for such) so it goes through your pump (oh by the way I hear that air damages pumps).

It's a system which means it all has to work together.

 

[Danger]

I'd certainly toss out the idea of freezing it. You do know that water expands with tremendous force when it turns into ice, right? Even if your lines are stretchy enough to handle it, you'd for sure blow your pump to pieces.