I think the point that wseaton is making is that, with most HOB
skimmers, the pump intake is so close to the skimmer output that a percentage of already skimmed water is being pulled back through the skimmer. I think this depends a lot on the currents in the tank though. Also, a surface skimmer makes an unbelievable difference in the efficiency of any HOB
skimmer. You can also locate the pump/powerhead away from the skimmer. The reason a sump skimmer does a better job is that it receives a fresh supply of water that has been skimmed from the surface of the tank. It also, in most cases, dumps the skimmed water into an area of the sump where it will not be recirculated through the skimmer.
On the subject of skimmer efficiency, the dwell time of the bubbles is what makes the skimmer efficient. The V-8 Chevy powered skimmer was a joke of course. If you are forcing water through the skimmer at such a high rate that the DOC
coated water molecules are not contacting the rising bubbles, then your efficiency will drop. What you want is a stream of microbubbles rising slowly through a downward flow of water. This lets the water molecules collide with the bubbles and shed the attached DOC
's. This is why counter-current skimmers work so well...bubbles always going up....water always going down. Co-current skimmers have the water and bubbles going in the same direction thus reducing the contact time considerably. Verturi skimmers are like super charged co-current skimmers in the way they function. Most of the good ones have a double or triple pass system that gives you a degree of counter-current effect as well as the initial co-current effect. I hope this makes at least some sense. If not, I can suggest reading (I think...can't find the darn thing right now) "Aquatic Systems Engineering" by P. R. Escobal. After you read it, you'll really be confused