For what it's worth, the proper flow rate according to Pedro Ramon Escobal, the man who literally wrote the book on filtration devices
http://www.amazon.co.uk/Aquatic-Syst.../dp/1888381051
... and wrote the book on rocket science.
http://www.amazon.co.uk/Methods-Astr...tt_at_ep_dpt_1
... then another book after the rocket scientists finally understood the first one.
http://www.amazon.co.uk/Methods-Orbi...tt_at_ep_dpt_2
According to Escobal, founder of Aquatronics (premier aquarium medication manufacturer) and Filtronics (innovative filtration device manufacturer), protein skimmers should have a 1.333333 throughput per hour for maximum efficiency. He uses the same flow through rate for sumps for obvious reasons. If you go on to read his recommendations for UV sterilizer throughput the math works out to 2 x per day, yes DAY turnover ratio.
Media filters such as fluidized beds operate with small powerheads so we're talking 300 GPH, at best. Refugia require slow flow for pod development and longer dwell time with algae.
As far as filter socks go, they work best with a slow flow so the detritus you collect isn't pulverized by water flow and converted from POC (particulate organic carbon) to a liquid form - DOC (dissolved organic carbon). keeping the detritus in the display tank suspended with dedicated water flow such as a closed loop or powerheads is the key to collecting more detritus. A flow rate of 1.33333 x the volume of the display tank is enough to collect the maximum amount of detritus without diminished returns.
Ozone is dosed into the protein skimmer so we can go with the magic skimmer number of 1.33333. Are you starting to see a pattern yet? All of these devices are operating concurrently, that is to say the water entering one, then goes into the next device until the water exits the sump.
There is absolutely no reason to deliver more water to filtration devices than they require to operate. It is a waste or resources and actually decreases performance. It skims a thicker water sample from the surface of the display which yields less surfactant proteins, pre-skimmed for the protein skimmer. The water travelling over the overflow box does not move faster (meaning more rapid surface skimming) when you increase the return pump output volume. The water simply climbs higher over the box drawing the excess water from below the surface "skin". This taxes your overflow drains causing more microbubbles, noise, turbulence, saltcreep and heat & energy from the pump.
The most efficient filtration system is a first in first out (FIFO) system. The water skimmed by the overflow box should be fed directly into the protein skimmer in the sump. That water should then be drained out of the skimmer and move on to the other devices you have selected in series/succession (one at a time). If your protein skimmer does not allow for direct feeding, you should locate the protein skimmer in a partitioned area at the beginning of the sump where water drains from the display tank at the same rate as the skimmer pump (likely 1.33333 x display tank volume per hr). The processed water that exits the protein skimmer should be directed over the partition so it enters the next stage of filtration, usually the refugium with media filters being last.
If you implement this simple system, you will process 100% of the water entering the sump, and do so only once. If you do not use a FIFO system your protein skimmer is filtering the same water over and over randomly, missing some display tank water entirely.
Now let's look at some scenarios if the return pump and skimmer pump are not matched exactly. If the skimmer pump is stronger than the return pump, more water will go through the skimmer than the amount entering the sump. This will result in the excess portion of water travelling "backwards" over the partition from the second stage back to the first to equal out the pumps. The disadvantage of this is that excess water will get processed a second time, but no water will go unskimmed. If the return pump is stronger than the skimmer pump, the excess flow will bypass the skimmer, but you will not process any water two times before it is returned to the display tank. If this is the case, it is a lot easier to restrict the return pump than mess with the skimmer pump. Some manufacturers allow you to adjust air flow which in turn lowers or increases water flow.
If we continue with this idea, the media filters could be housed in the second partition with a dedicated pump (powerhead). The effluent (exit/processed) water from the media filters would be directed over a second partition so it is processed only once. If you use a return pump bypass, the water will be reprocessed randomly.
In summary, the sump is not a filter, it's merely a vessel to old them. If we could live with the clutter hanging on the back of the display tank, we would not have to go though the expense, noise, flood hazard and bubbles that come with moving water to and from a sump. You don't have to be a rocket scientist to understand that less is more in the case of display tank flow through/throughput rates, but if you ask a rocket scientist he will agree that the flow rate is governed by the protein skimmer pump output which happens to be in the neighbourhood of 1.33333 x the volume of the tank.