Originally Posted by joesaiditstrue
somebody did some tests, I can't remember if it were xtremesystems or martins, or some other site but if i recall, they noted higher temps the more tubing that was added
interesting that you find it to be the opposite, i'll try to find the article...
Actually you are both right. The more tubing you have, the more water is in your loop just like adding small increase volume to reservoir, so it will take another 15 seconds or so to heat the water to final equilibrium. But final equilibrium will be ~ 0.05C higher, since the flow will decrease about 0.05gpm for extra foot of tubing. So temps will slightly rise by ~0.05C from decreased flow to rad and block, but just like adding more water to a res, the time to loop equilibrium will increase a few seconds. Neither effect would you be able to measure by increasing 1 foot of tubing, have to add 10 feet, measure, then extrapolate back to be able to measure the effect assuming you have multiple dallas 1 wires in water and measuring each air intake. (swiftech white paper
showing increasing flow to rad does increase cooling, in case anyone want to incorrectly argue the opposite)
When you increase the diameter of tubing, then you are increasing flow per unit time, and then you would decrease temps slightly since increasing flow per unit time to rad and cpu block slightly. However going from 3/8 to 1/2 inch ID tubing the increase in flow is only about 0.15 gpm and effect on temps only about 0.15C, Cathars testing here
Loop order doesnt matter since it takes ~400W to heat water by 1C at flow rate of 1.5gpm (typical flow in loops). For example, when I put multiple probes (capable of 0.1C accuracy) in my water loop at various points, and each air intake... With 1 x 360 rad, 1 cpu loaded 130 watts when I first start prime my water temp is 25.5C, ambient is 25C. With 130 watt loaded cpu, the temp of water increases by 0.3C going through my cpu block with each pass. The rad cannot cool water per pass by 0.3C since rad is not efficient at cooling 25.8C using ambient of 25C, needs higher temp differential between ambient and water temp to cool more efficiently. So each pass for 7-8 minutes the water temps slowly increases by 5C to 30C, now with 1x360 rad at steady state the water temp is 30C, ambient 25C (here rad now efficient enough to cool cpu by same it heats water 0.3C per pass, and cpu temps for example are 75C at load. Delta air to water is 5C at steady state, but on each pass, the water temps is 0.3C hotter coming out of block, and 0.3C cooler coming out of the rad. So if I had a gpu at load in there, the 200W load means a pass difference of 0.5C, so I could get 0.5C lower temps by going to cpu first then gpu, with 2 gpus, get 1C better temps. But running prime gpus are idle, so really only get 60W with 2 gpus idle, so only about .15C better temps if making sure cpu temps get water first. If I add a second rad, I decrease that delta air to water by half, ie to 2.5C since doubled rad surface area (assuming same fan speed), and now temps 72.5C.Edited by opt33 - 8/14/12 at 5:41am