Yes - that latest pic Quickmcj is exactly what I'm trying to get at. It'll work even better if you can tilt the entire roof of the res like that to get air to gather at a single point up there. And IMHO you don't need a second air-trap at the top. You just need to make sure that when you initially fill the system that you get as much air as possible out of the rad and block by tilting them around a bunch before mounting. Very small air pockets should be pushed away by the pump and should gather in the res where the speed of the liquid is much slower and they can escape the flow again.
And then there are only 2 things you need to worry about
:
1. If the rad and block are both still higher than the top of that refill area of the rad then with the pump off gravity may pull the water down out of the high points of the system, and it may flow back up out the fill point. But now that I think about it there should be no way for air to get into the rad/block area up there - which means water can't flow out (nothing to take it's place except water in the res, which means that water-level in the res won't move), which means it should be safe. I'd recommend opening that valve with the system out of the computer and the rad lifted up good and high just to make sure though.
2. If you move the system around (eg. LAN) - you need to keep the res vertical so that air bubble can't find its way into one of the hoses/the pump. Easiest way is just to transport the case vertically (thats how I keep my air-bubble at the top).
And of course - I can't make a post without a rebuttal to the flames
Quote:
|
However, if the inlets extend in the water, and the res is open, it is still a closed loop (I think you got that!)
|
Maybe one of us needs to re-think our definition of a closed-loop system, because what I just quoted in my mind would not be closed-loop. The definition I've had in mind would basically state that if it holds pressure, it's closed-loop. If the res is open, it doesn't hold pressure. (Not counting during maintenence when you may open a valve to fill/drain the loop).
Quote:
|
The thing is though, the volume of coolant will fluctuate, so you'll have to make sure that this long tube is pretty filled, so that the coolant is sucked in by the rad, but not so full that it spills, if the coolant expands.
|
The volume of coolant in your system should definatly NOT fluctuate. If it does, you have a leak somewhere and are constantly re-filling to compensate. The water-level in your res may fluctuate if you have an air-bubble trapped in the system and the bubble is moving around. But volume of coolant is a constant in virtually every water-cooling system except for bong-coolers.
And as for blocks (or anything else) that traps air..., what you have is something like the situation in my newest crappy picture (I avoided jpeg this time though). Water moving fast around the corners will leave small air-pockets in the dead spaces just after the corner. By turning the pump off and back on, many of these bubbles will have moved out of the corner and may be picked up (at least partially) by the water-flow, and can then be trapped in the air-trap, wherever it is in the system. If you happen to be really unlucky one of these dead-spots in the flow may also be in the highest position in it's area, so that it doesn't move at all when flow is shut off. When this happens you need to move the block around (eg, lay on side) and turn the pump on and off a few times to try and dislodge the air-bubble. Though - if there is a dead-spot in the flow anyways, I'm not sure if it makes any difference in cooling ability of the block if there's non-moving air, or non-moving water in that spot.