Originally Posted by dctravis
Hey ive got another post that noone has replied to, but I figured this thread was a good place to ask. I was thinking about having parallel paths for the cpu, each gpu, and one for the vrm ram and lsi plx chips on my mobo. My thought is that with the coolant temperature being the same going into each main component that my performance would be more balanced and allow for all of the components to be overclocked more evenly or is this thought process wrong? I realize that overall the same amount of heat is being drawn into the coolant and therefore extra heat would not be expelled from the radiator. So overall its the same performance but with the same temperature at the inlet of each block. Im simply worried about the overclocking capability of each component if it were run in series.
Also would anyone mind confirming or correcting my assumptions for tubing ans nozzle diameter changes thriughout a loop? If I have a single 1" discharge from my iwaki pump with a discharge pressure of 36psi going to a several different sized lines 1/4 lines and then recombining prior to the radiator to three 1/4 inlets through the rad then a single 1/4 inlet to the res and then back to 1" for the pump suction. Please correct if needed but with the changes in diameter it should increase or decrease pressure and velocity of the fluid allowing the fluid to flow faster and therefore the same amount of mass will still flow. Although this resistance to flow will determine the volumetric flow rate I believe even with this pressure a iwaki or possibly two should be able to handle it correct? And with different flow paths of different restriction, is it possible that the higher restriction path would recieve relatively no flow without a pump directly behind it?
No, parallel for water temperature is not of value. With a good flow rate, the temperature differential for something like 200W of heat which is more than even a highly overclocked hexacore is less than .5C. Most loops don't see more than one degree differential across the loop. It does depend on flow rate, but here is one of my radiator tests where I measured water in and out.
What could be of value is to do series for some of the luxury items. The trouble with parallel is, a huge lack of data (pressure drop) on many or most parts. Without that information, it is hard to know what will happen. If restriction is significantly unbalanced the high restriction loop may only get a very small portion of the flow. That might be ok, but it may not be what you want. I would always keep the CPU block in series, parallel is ok on like blocks such as GPUs, but hard to guess on other stuff. It certainly can be a valuable tool to emphasize flow rate on the CPU, but without pressure drop on everything, I wouldn't recommend it.
Don't worry to much about tubing size. it has some restriction but very low in the 3/8"+ sizes. No problem to mix and match different sizes in the same loop. Use whatever size you like the looks of best.