water coolers

[littlerichie]

came back from a computer show!
finally found water cooler distributor.
SOB?.
no wonder your temps do not fall fast enough?
those water connectors are too damn big!
your pumping water through your blocks faster
than it has time to transfer heat from the block to the
water!
slow the pump down significantly then you should notice
a serious drop in temperatures
david.

 

[donnie]

I agree with you. I am useing a fuel pump from a 1987 chevy p/u (an in tank)it pumps 25 gph. It keeps the water in the radiator long enough for the fan to do it's job of cooling the water.

 

[stool]

There have been a lot of posts, discussions, and different studies posted for a long time on this topic. Several of these studies seem to indicate that highflow pumps actually remove more heat that lowflow. I believe you may be able to find a couple in the Watercooling section of Tips and Techniques.

 

[lightning]

the speed of the flow rate cannot be too slow!! it should be at mid range and depend how u arrange the system!! small flow rate can get more time for the water to absorb the heat, but if too slow, the water will transfer the heat back to the cpu!! which mean they tend to achieve equalibrium state!! almost same in temp!! this is what i learn from fluid mechanic!!

 

[0dan0]

This is not true. The reason this doesn't work is because water is everywhere in the tubing, therefore water is constantly absorbing giving up heat. When the water moves too slowly, the water passing thru the waterblock reaches a plateau in the amount of heat it can absorb, and the water in the radiator reaches the limit of how much heat it can give off. When there are faster flow rates, no 'heat pockets' can form in the water, so the radiator is always giving off maximum heat from the water, and the water in the waterblock is also absorbing maximum heat from the CPU. Your main point is that if water moves more slowly in the radiator it can give off the maximum heat, but if you think about it the heated water will *always* be in the radiator no matter what speed it is going, and the faster the flow rate the faster that cooled water leaves and the faster hotter water comes in.

-Dan

 

[William]

okay, this is an interesting debate. Slow water rate should yield higher water temps due to the slower rate in the jacket, but it should dissipate well cause it goes through the radiator really slow. If it is faster, you should get the same cpu temps, just a lower water temp, and even though it flies through the radiator, your water does not have as much heat to dissipate. I will have to ask my physics prof which is better.

 

[TT120]

I saw this on 'another' forum and the physics person said you can not have the water running too fast for best cooling but I have had real life experience with this in my camaro and here's what I found out. My camaro was running a bit hotter than I liked, so I thought, taking out the thermostat to let the water flow more freely would help. It had the opposite effect. The water was flowing faster and didnt stay in the radiator long enough to cool down. I took a hose clamp on the upper radiator hose and tightend it down so as to constrict the water flow and bam, the temps dropped back down. Their arguments to the contrary sound very good but I know from personal experience that you can have the water flowing too fast for good cooling.

 

[Murphy]

Physics

Ok, I've been in this subject some times before. I'm a chemistry student, and have followed courses of Industrial Innovation, about pipe flows/ heat transfers/ reaction rates and kinetics, you name it... Here's what I think:

First let's start with a given constant: Your heat transfer (Watts) is linearly dependant of your Temperature difference. This is far more important than waterflowrate, though you have to adjust your flow according to your temp differences.

To go short: my opinion is that the efficiency of your cooling setup (the right water flow) is dependant of the ratio between waterblock capacity and radiator capacity. 2 examples:

- radiator (inclusive the air flow) heat transfer more effective than block: Radiator benefit MORE from a average medium temp (accomplished by high flow), than from a partial high temp (in the beginning) and partial low temp (the end of the radiator. Especially when water has passed the first half, there is almost no more heat radiated because of the low temp difference. Actually I should have a graph right here to point this out, but I'm short on time, maybe next weekend I can do some more research if there is any interest. When water flows RELATIVELY cool though a waterblock (because of high radiator capacity), you want high temp difference, that means high flow, not that it's warmed up halfway.

-the opposite.. well it's about diner time, it isn't hard to figure out it's quite the opposite. Water get's hotter, so you have more heat out of it by your poor radiator, with a low residence time or a low surface area/ thermal conductivity.

I use a heater core, it's alu and is far superior for my athlon barking out about 60 watts, not peltiered (yet, I'm starting to get facinated and bored because of not much oc'ing action lately).

About the ability to pick up heat and flow rate, there is a certain point when the water flow goes from Laminar flow (=linear speed distribution for fluid from the middle off your pipe to the wall) to Turbulent flow (no linear relation, constant mixing of water near the walls and in the middle. It's quite clear that turbulent = better. You can accomplish this by higher flow, an irregular water path, but (where many go wrong..) a smooth surface!! Because if you don't, you have 'pockets' of fluid heating up and lowering the heattransfercoefficient. Best blocks are smooth, have complex tunnels and a large surface area to exchange heat. Lower inner diameter increases local water flux(don't dare to say flow), helps as long as your pump can handle it.

I think I've made my point for now, if there is any interest, I can setup a mathematical example. But many of the numbers are very difficult related to real life performance, this is where you enter the realm of true industrial innovators. As I'm just a student.. but I'm an overclocker!

p.s. I really like the article about evaporation water cooling, maybe I'm going to try it too soon!