Here's an example of a simple 29-gallon chaetomorpha refugium with a flow of about 960gph @ 4' head (Iwaki MD30-RLXT):
The upward water current comes under the first baffle to suspend the chaetomorpha under the light. The current holds a mesh tank divider against the second baffle. This mesh screen keeps the algae from the return compartment. The return compartment is large to accomodate evaporation and water changes. My 100-gallon system (75g main tank+29g refugium) loses about a gallon a day to evaporation in dry Colorado.
Chaetomorpha is almost neutrally buoyant so a much lower flow is sufficient to suspend it. The 960gph flow does not tumble the chaetomorpha but holds it rigidly at the surface. When I replaced the chaetomorpha with negatively buoyant gracilaria parvispora, the tumbling action is plainly seen. The G. parvispora sinks to the bottom on the left side of the compartment and is raised by the flow on the right side.
The advantages of chaetomorpha are that it serves as a mechanical filter to trap debris and it clumps together so you can remove it easily before cleaning the refugium. G. parvispora does not clump together, takes longer to remove, does not trap debris and requires stronger flow to suspend and tumble.
Since I have a carbon reactor, I cannot tell if chaetomorpha or G. parvispora exudes more yellowing compounds. I suspect that G. parvispora does due to the color it added to my quarantine tank. However, it may have exuded the color compounds as it was dying back in the sterile quarantine tank environment.
My DSB is in the main tank. Many aquarists employ a DSB in the refugium believing that they can avoid lighting. My experience is that a light is important for maintenance and cleaning so refugium lighting is necessary whether you have algae or not.
My advice regarding flow is to use a slow quiet return pump and supplement the main tank flow with powerheads. I can't wait to replace my noisy Iwaki MD30 with a quieter MD20. I need to purchase powerheads before doing so.
