QUOTE(jmill @ Mar 8 2009, 01:11 PM)
...All your doing is pushing the coolant through twice the surface area twice as fast. You get the same cooling with a single pass because you pass the coolant through half the surface area half as fast. In the end it's a wash...
not necessarily true*. without getting too deep into it, if fluid is flowing at a turbulent Reynolds value (directly related to velocity - higher velo = more turbulent) the convection is much higher at the boundary layer between the coolant and the aluminum tube than if it is in laminar flow or near turbulent flow.
For pure water, at purely laminar flow, the convection coeff can be around 300 W/square meter-K whereas at highly turbulent flow, that can be well over 10,000! And temperature drop across a cooling device is linearly dependent upon that coeff, so if you can increase it to 30 times its base value, then do so. Having high velocity flow can be very important.
*If you find you are highly turbulent through the single pass, then the dual won't help you.
Here in the laser world I have helped the R&D guys with a few cooling block designs - they tend to drill a single, big hole thru the cooling block to pass liquid coolant through, and they had temps running away on some of the current designs. For the same volumetric flow rate, I decreased the diameter and made 4 reversing runs in the block, getting both a higher conv coeff and a bit more surface area and temps decreased dramatically. It is mostly due to the higher velocity through the cooling block.