Work in progress, all open to challenge, plus additions!
Watercooling Myths:
Flowrate makes a difference to temperatures - True
Increasing the flowrate will improve cooling via increased turbulence, both in the CPU block and also the radiator.
As a minimum you want at least 0.5 gpm (or ~120lph), increasing this to 1 gpm (~240lph) will reduce temperatures by about 2C, after this diminishing returns mean that increasing flow to 3gpm (~720lph) only reduces temperatures by a further 1degC in a typical loop.
A pump like a DDC 18W or D5 will meet the minimum requirement in all and exceed it in most loop layouts. Use Martin's flow calc to give you an idea of what to expect as a guide. As such, though I've said true, with careful planning it's impact is in reality insignificant.
Flowrate is highest at the pump exit - Not True
Flowrate is the same at everypoint in the loop, it would break the Law of Conservation of Mass if it didn't. So in reference to the previous point the location and order of loop components makes no difference to the cooling performance of the loop, with the possible exception of:-
The waterblock should be immediately after the radiator / the water is coldest after the radiator - Insignificant
The water leaving a radiator will be the coolest point in a loop - but the difference between the hottest and coldest will be small. For example, in a loop with low flow rates (0.5gpm) 250W of heat will only raise the water temperature through a block by 1degC. So even with a high end CPU, plus SLI/CF fully overclocked and at full load (say 750W in total) the total increase in water temperature in the loop is only 3degC. Running a higher flow rate will decrease this by 1-2degC (refer to the flowrate myth) so the difference is even smaller.
It's not worth routing tubing any route other than the shortest and neatest route.
The waterblock should be immediately after the pump / CPU waterblocks perform better at higher pressure - Not True
How well a waterblock cools is completely independent of water pressure, so placing it as the first component in a loop after the pump will have no benefit compared to having it the last component in the loop. Increasing pump pressure by using a bigger pump will increase flow rates and therefore provide better cooling though.
The Reservoir should be immediately before the pump - True
I've put this as true, it doesn't need to be but having it elsewhere in the loop where it doesn't naturally drain down into the pump inlet just makes priming and then bleeding the loop a complete nightmare. If it doesn't let you fill the pump inlet then it's not worth the effort having the reservoir in the loop.
Tubing size makes a diference to flow rates - Insignificant
Any tubing from 8mm ID and over will have minimal effect on cooling performance (less than 1degC).
You can't mix aluminium with anything else - Not True
There is absolutely no problem in mixing metals in a watercooling loop, what I would say though is to avoid having aluminium in direct contact with copper or nickel and to use a corrosion inihibitor in other cases. If you can ensure that the aluminium is electrically isolated from other metals in the loop even then in theory you shouldn't need to use corrosion inihibitor as there is no path for electrons to flow to/from anyother metal to cause galvanic corrosion.
You must use distilled water as deionised water rusts metal - Not True
There are many claims that deionised water will "re-ionise" by sucking ions out of the metal in a loop causing it to rust so use distilled water instead. The very process of distillation will also remove any dissolved ions from the water also resulting in water that has been de-ionised. Neither causes any meaningful corrosion.
Note: EK Waterblocks recommend that if you use distilled/de-ionised water or silver or copper sulphate based anti-algae that you also use anti-corrosion additives to prevent corrosion issues with their nickel plated copper blocks
Using watercooling will make your room cooler as your CPU temperatures are reduced - Not True
This is not true, in fact with watercooling you will likely increase your overclock, generating even more waste heat, in turn making your room hotter!
Watercooling Myths:
Flowrate makes a difference to temperatures - True
Increasing the flowrate will improve cooling via increased turbulence, both in the CPU block and also the radiator.
As a minimum you want at least 0.5 gpm (or ~120lph), increasing this to 1 gpm (~240lph) will reduce temperatures by about 2C, after this diminishing returns mean that increasing flow to 3gpm (~720lph) only reduces temperatures by a further 1degC in a typical loop.
A pump like a DDC 18W or D5 will meet the minimum requirement in all and exceed it in most loop layouts. Use Martin's flow calc to give you an idea of what to expect as a guide. As such, though I've said true, with careful planning it's impact is in reality insignificant.
Flowrate is highest at the pump exit - Not True
Flowrate is the same at everypoint in the loop, it would break the Law of Conservation of Mass if it didn't. So in reference to the previous point the location and order of loop components makes no difference to the cooling performance of the loop, with the possible exception of:-
The waterblock should be immediately after the radiator / the water is coldest after the radiator - Insignificant
The water leaving a radiator will be the coolest point in a loop - but the difference between the hottest and coldest will be small. For example, in a loop with low flow rates (0.5gpm) 250W of heat will only raise the water temperature through a block by 1degC. So even with a high end CPU, plus SLI/CF fully overclocked and at full load (say 750W in total) the total increase in water temperature in the loop is only 3degC. Running a higher flow rate will decrease this by 1-2degC (refer to the flowrate myth) so the difference is even smaller.
It's not worth routing tubing any route other than the shortest and neatest route.
The waterblock should be immediately after the pump / CPU waterblocks perform better at higher pressure - Not True
How well a waterblock cools is completely independent of water pressure, so placing it as the first component in a loop after the pump will have no benefit compared to having it the last component in the loop. Increasing pump pressure by using a bigger pump will increase flow rates and therefore provide better cooling though.
The Reservoir should be immediately before the pump - True
I've put this as true, it doesn't need to be but having it elsewhere in the loop where it doesn't naturally drain down into the pump inlet just makes priming and then bleeding the loop a complete nightmare. If it doesn't let you fill the pump inlet then it's not worth the effort having the reservoir in the loop.
Tubing size makes a diference to flow rates - Insignificant
Any tubing from 8mm ID and over will have minimal effect on cooling performance (less than 1degC).
You can't mix aluminium with anything else - Not True
There is absolutely no problem in mixing metals in a watercooling loop, what I would say though is to avoid having aluminium in direct contact with copper or nickel and to use a corrosion inihibitor in other cases. If you can ensure that the aluminium is electrically isolated from other metals in the loop even then in theory you shouldn't need to use corrosion inihibitor as there is no path for electrons to flow to/from anyother metal to cause galvanic corrosion.
You must use distilled water as deionised water rusts metal - Not True
There are many claims that deionised water will "re-ionise" by sucking ions out of the metal in a loop causing it to rust so use distilled water instead. The very process of distillation will also remove any dissolved ions from the water also resulting in water that has been de-ionised. Neither causes any meaningful corrosion.
Note: EK Waterblocks recommend that if you use distilled/de-ionised water or silver or copper sulphate based anti-algae that you also use anti-corrosion additives to prevent corrosion issues with their nickel plated copper blocks
Using watercooling will make your room cooler as your CPU temperatures are reduced - Not True
This is not true, in fact with watercooling you will likely increase your overclock, generating even more waste heat, in turn making your room hotter!
