FishProfiles.com Message Forums

How do you determine how many gph your sump pump needs to output? I understand that I will filter only as much I am pumping, I just don't know much that needs to be. I'll give an example. If I had a 100 gallon tank, and a 30 gallon sump, how much water should I be moving?

Last edited by firecontrol at 28-Dec-2005 23:56

I don't know how to figure how much water you need to pump, but I do know that the amount of water you can pump is determined by the size of your overflow(s?)

Sorry, that's all I can help ya with there

I am aware of the importance of overflow size and tube/pipe diameter. I've been pouring over a bunch of DIY plans tonight, and I can't find any mention of the amount of flow recommended. Should I go with something comparable to my Fluval 404 (as far as rate of water movement)? More? Less?

Depends on how much you want to flow. Your sump will only see how much water your returning to the tank.

It also depends on the sump design itself. If you're cramming it with media, go with a lower powered pump so you're not bypassing all the media. If you've got a lot of coarse media, step it up to a beefier pump.

The general guideline for power filters is to snag a filter that will overturn the volume of the tank over 7 times per hour. On a 100 gallon tank, that'd be a filter (or combination of) that will pump a combined 700 gallons per hour. Since canister filters are more efficient, their flowrates are generally lower to achieve the same results. It's a quality over quantity issue with them.

I feel a sump is a good balance between quality of filtration due to the amount and difference in media you can put in them, and also you can pack them with some pretty high flowrated pumps. I'd shoot for a middle of the road pump, maybe something that will turn over the tank 4-5x per hour. When looking for pumps, remember that it's going to have to pump up, so look for something that will put out the volume at whatever height you'll be pumping up to.

Just remember there are a million variables, so experiment!!