You are at The Krib ->Plumbing and Filtration [E-mail]

An Air-driven Reverse-flow Undergravel Filter

by Bob Dixon (IDMiamiBob at
October 1997

Okay it's like this:
I built it for a thirty gallon tank. I started with a drop ceiling "egg crate" from the home center. I glued marbles on the bottom for legs. I covered it with "plastic canvas" from a craft store to keep the gravel from falling through. In each of the two back corners I cut a one inch circle. Into this I glued a length of 1" aquarium filter stiff tubing from the LFS.

NOTE: I precut all tube lengths before gluing At the top of this tube I installed a 3/4" PVC street elbow from the home center. these street elbows are not carried by everyone. I finally found them at the Depot. The large end of 3/4 inch elbows and tees are slightly larger than the 1 inch tubing, and will hold on fine with silicone. To the other side of the elbow I installed a vertical 3/4" tee. I installed a 1 inch tube to the bottom of the tee, down to the bottom of the tank. Then another street elbow, and to the street elbow, a regular elbow pointing back up. I installed one more tube back up toward the surface and capped it one of those guards that come with power filters to protect against sucking up fish. I originally used a fan shaped guard, but found that the straight ones would allow me to better skim the protein and fatty acid slime off the surface down to the filter, where something seems to be digesting it. Stick an airline down through the tee and into the middle tube, all the way down. Stick an airstone on the end of it if you are so inclined, I get the flow rate I want without it.

You could do it with only two tubes, if you don't mind drawing water from the bottom. I kinda like the idea of sucking from the top.

You can build it cheaper by using 3/4 inch PVC, but you can measure flow rate when you can see through the tubes, and they are a bit less visible. Meeasure the flow rate by dropping a single drop of Methylene blue or even food coloring onto the surface next to the intake. If you know the length of the tube, and it's diameter, a little algebra and the time it takes for the dye to travel the length of one tube, and you can figure it out. If you get water up through the top of the tee before you achieve the flow rate desired, add some PVC tubing. I am getting 35 gph on each side without it. I have one and a half to three and a half inches of gravel for a substrate, more than enough for adequate bacterial filtration.

Erik wrote:

One last question, how do you deal with the mechanical filtration with this setup? It sounds like you would skim any surface muck down into the undergravel, and you'll just settle all your waste just on top of the gravel, right?

Not quite. The whole point of reverse flow is to get the stuff below the gravel. Everything colid settles on the bottom of the tank, below the filter plate. As far as I can tell, the fatty and proteinous stuff that skims off the surface floats back up through the gravel, and is consumed by plant roots, and bacteria in the gravel bed. Most people say you shouldn't put things like rocks or whatnot onto the surface of the substrate because it will create dead spots. [ But if you don't overstock, the dead spots should turn anaerobic. As far as I can tell, the stuff off the surface, which some people remove in salt-water setups by foam fractionation ( protein skimmers) , is the same stuff they reintroduce later into their anaerobic filter chambers as "bacteria food". ] The couple sentences I have placed between brackets is purely personal conjecture based on what I can glean from magazines and personal experience. If I'm right, this should help keep the nitrate levels down between weekly water changes. What I know for sure is that when I changed intake guards from the fan style to the straight to improvesurface skimming, the surface cleared up in a couple days.

Just a note on the solid waste build-up. It tends to form a paper-thick hard film. I have broken down tanks after six months and two years. The "cake" doesn't seem to get any thicker, or at least not very quickly. I conjecture that there is some kind of bacterial action that breaks it all down the same way a septic tank does. Plants seem to like it (the grass is always greener over the septic tank), but the folks on the live plants mailing list pretty much don't agree. Someone a week or so ago wrote that he had two ten gallon tanks with a substrate of gravel and fish mulm. The plants were doing great. He cleansed the gravel in one thoroughly in one tank, and the plants began to fail. When he restored the "dirt" in the gravel, they recovered. I have found root systems ensconsed in the builup under the plates, and I agree with his conclusions, but he seemed shocked that fish-pooh would actually be good as plant food.

So, I guess, in the end, we are all going to find what we like best and go with that. I think, for display tanks, and other systems where you want to keep the tank and gravel in presentable condition, the RUGF is an option that should be more popular. With the right manufacturing capabilities, all three tubes could be set up concentrically ( one inside the other) and improve the aesthetics of the thing.

I have my gravel thickness sloped up toward the back, with gravel "terraces" outlined by polished rocks around the bottom elbows of my filter tubes to hide the pvc. It has a very pleasing appearance, and valisneria planted in front of it can reduce its visibility even further.

Well, not much else I can think to add. I think it works great, and I am thinking of modifying some of my factory filters using 1/2 inch tubing so it doesn't overcrowd the tank.

Some of this material originally posted on the Apistogramma mailing list. HTML translation, editing and GIF by Erik Olson

A note from Bob, received December 15th, 1998

As you may recall, I have had my air-powered RUGF running for about a year and a half. I broke the tank down a couple days ago to add a more suitable substrate for plants, and found something unique. The tank is painted on the back, so I haven't paid much attention to what's been going on behind it. Well, on the wall, about 6 inches behind it, strating level with the top of the tank and going down the wall about a foot (where it is pretty much out of sight because of the light strip), and in line with the two uplift tubes (where the air escapes) are what look like coffee stains, little watery streaks and all.

My best guess is that some small amount of fractionation was occuring, and as the bubbles popped at the top of the tube, the DOCs were "atomized" and sent aloft. I can only wonder about why it all settled on the wall, and not the glass tank lid or the TV stand nest to the tank. Perhaps there was some sort of low-level static charge on the wall that attracted the airborne particles, like electro-static spray-paint techniques.

I'm going to reconsider the design, and see if I can come up with some sort of collector to catch the "skimmate" as it rises. Perhaps I can extend the tube up a bit, stuff it with cotton, and add a small side-path with a cup. Yes, I know skimming isn't very effective in freshwater tanks, but that is primarily because the bubbles are too large to get a good surface area/ volume ratio. But who knows? Maybe this set-up can somehow be coerced into changing that. I had no airstones, and got some amount of skimming action just with the huge bubbles that come out of the end of the airline. Maybe one of those glass Amano CO2 diffusers can change that.

Meanwhile, I have to get this scuzz off my wall. Any ideas?
Up to Plumbing and Filtration <- The Krib This page was last updated 20 December 1998