Photo on the right: View of the UGJ system before all of the substrate is added to the tank.

In my opinion under gravel jet systems (UGJ) are one of the most significant new developments in aquarium technology in the last decade or so. This is an excellent article outlining the concepts and giving instructions how to built one by Cichlid-Forum's Marc Elieson (VatoElvis), who to the best of my knowledge can be credited with having had the original idea for UGJ systems. In a nutshell, an UGJ system is kind of the opposite of an under gravel filter (UGF). While an UGF sucks all debris into the substrate, where it is either decomposed by bacteria or removed by gravel vacuuming (mostly the latter), an UGJ prevents debris from settling on the substrate, and keeps it moving around the tank, so that eventually it has no choice but to land in the filter.

Under Gravel Jet is actually a bit of a misnomer, since UGJ systems work best with sand as substrate, or maybe very fine gravel. With coarse gravel, debris will be caught in the nooks and crannies between the particles, and the UGJ system can't do its job. The only catch I have experienced is that debris also has a tendency to get stuck in plants - especially rangy plants like Java moss. However, even with this slight drawback, UGJ systems greatly reduce the need for gravel vacuuming, and they can create a gentle current that most fish like to play in. I believe any distraction in their otherwise fairly dreary tank existence is more than welcome. The reason you don't hear more about under gravel jet systems is probably that there is no way for anybody to make money from them - or at least nobody has figured out a way so far - and that's why nobody puts money into marketing the concept. In other words UGJ systems are little known, super cheap, and highly useful - kind of the opposite of canister filters with easy-prime-gadgetry, which are ubiquitous in any aquaristic catalog, get more expensive by the day, and are completely useless - since anybody who can't prime an Eheim classic series canister filter should not be allowed to keep fish in the first place!

Another view of the UGJ system before substrate is added to the tank.

Anyway, I am digressing. All you need to set up a successful UGJ system is an understanding how they work, and plumbing parts for a few dollars from your hardware store. As mentioned before, while Home Depot shines in the lighting department - bad pun intended - Lowes seems to stock a far better selection of plumbing parts, or at least that's true in North East Ohio, where I live. An UGJ system is connected to the outlet of a pump in form of a powerhead, the return pump from your sump, or even the pump in your canister filter. It consists of a series of pipes ending in jets (see Marc Elieson's article about how to make these), which serve to keep any debris from settling on the ground. Again, the idea is that if the debris is kept moving around your tank, sooner or later it will have to end up in your filtration system.

I trialled an UGJ system in a 29G South American setup before installing one in my 240G long. For the 29G, I used a six jet setup powered by an Eheim 2217 canister filter. The current created by this setup is slight, and will not disturb even the most non-current loving fish. Yet it seems very effective in preventing debris from settling in any area. With the outlet of the filter completely submerged, and creating little or no surface agitation, an alternative means to increase aeration had to be found. I installed an Aquaclear 301 powerhead, which seems ideal for this purpose. During the trial period the tank was lightly stocked with 4 adult Bolivian Rams (Mikrogeophagus altispinosus), some Amano shrimp (Cardinia japonica), some Otocinclus sp, some tetras (Characidae) or pygmy sword tails (Xiphophorus pygmaeus). The exact stocking list varied somewhat over time. I fed once daily with Tetra Spirulina flakes and rarely gave other food as a treat (frozen red blood worms, life black mosquito larvae, algae wafers, etc). Waterchanges were 50% weekly. Over the course of 2 years I gravel vacuumed the tank only once or twice, but even on these occasions there was hardly any debris there to be removed from the substrate. Since the filter is generously dimensioned and aeration is plentiful, it seems very effective in breaking down waste material. It is cleaned about once every six month, and maintenance intervals could be even longer without causing any problem. As an aside, the tank won a tank of the month award on Cichlid-forum in March 2005, the Rams bred and I got BAP points from the OCA and GAAS, and the dominant male Rams won 3rd place in the South American dwarf cichlid class at the 2006 OCA Extravaganza. I guess you could say this has been quite a successful tank for me.

Dyeing an UGJ system to camouflage it

The UGJ system in the 29G was build using 1/2" CPVC pipe labeled "CPVC 4120 100PSI @ 180F hot-cold water tubing". This pipe seemed to have just the diameter I was looking for, and instead of being stark white like regular PVC pipe, it is sand colored, which I thought might make it a little less obtrusive. However, since I used a black substrate, that didn't work out quite as well as I expected. I coated the top of the pipe with silicone, and sprinkled substrate on it to hide the pipes, but the jets still look out and are screaming for some technique to camouflage them better. That's why I decided to dye the pipes for the UGJ system of the 240G, where I wanted to use black substrate as well. I had read on Cichlid-forum that PVC pipe can by dyed using RIT dye, a clothing dye that is available for a few dollars in any drugstore.

So, off I went and built a huge 21 jet UGJ system using CPVC pipe - mostly ¾", but reduced to ½" for the jets. One of my other hobbies is cooking. Well, since I am now a stay-at-home-Dad it's more like a job, but I am digressing again. I own a nice big stockpot of about 3-4G, and in this beast I attempted to dye the fittings. I added 1 packet of RIT dye, about half a cup of salt, and about 2 tablespoons of washing powder (Cheer), and filled the pot with water. All this is according to the instructions on the dye packet. I then brought the whole contraption almost to a boil, switched the hotplate off, and let everything stand over night. By morning the fittings had hardly changed color. After consulting the Cichlid-forum thread again, I decided that more vigorous boiling might be necessary, and I placed the pot outside on the hotplate attached to my gas BBQ so that I would not have to worry about it boiling over. After about 1h the fittings were certainly changing color in the blue/green direction, but they were a long way away from looking black. I doubled the amount of RIT dye, washing powder and salt, and continued boiling the mixture vigorously for another 3h (!), occasionally adding some more water (preheated to boiling point), to replace the water that had evaporated, and ensure that the fittings remained covered by the solution. During that time the fittings might have become somewhat darker blue/green, but not anywhere near black.

It occurred to me then that there might be difference between PVC and CPVC for dyeing purposes. After all, PVC is not meant to ever come in contact with hot water, while CPVC is made for hot water applications, so there might be a big difference in how they take up dye in boiling water. I happened to have few spare parts of PVC lying around the house, and put one of them in the dye mixture to see what would happen. Within seconds that part was jet black in a nice matte color - exactly what I had envisaged. At the time I was short of cash and in a hurry to get the tank set up, so that I decided to keep on boiling my CPVC parts. After about a week of boiling the 3-4 batches of pipes I had, they turned a dark blue/purple kind of color but never black.

I used these parts as they were, and they actually don't look too bad in my tank - better than stark white anyhow. However, if I had to do things over again, I would stay well clear of CPVC and use regular PVC pipe, dye that black with RIT dye, and in addition apply a coat of black silicone to the top of the pipes and sprinkle substrate over that. That should camouflage the UGJ system as well as possible, even when fish keep digging part of it out, which most cichlids tend to do.

A jet of the UGJ system (right) looking out of the substrate with a snail shell (escargot) for size comparison.

Addendum January 2008:

After running UGJ systems for about 3 years now, I thought it worth mentioning that there is a bit of a difference between using an UGJ system with a canister filter, and using one with a sump system. With a canister, the intake of the filter is usually submerged, whereas most sump systems draw water exclusively from the surface of the tank. Any UGJ system will have a much easier time driving debris into a submerged filter intake, while transporting debris over the rim of an overflow box into a sump can be a real challenge. Despite this problem, in my experience UGJ system are at their best when used with a sump, because the drainpipe will automatically provide excellent aeration of the water. This means that there is no disadvantage in having a submerged pump outlet in form of an UGJ system. A canister with submerged outlet, on the other hand contributes virtually nothing to aeration, and if you choose to use your canister to drive an UGJ system, you will have to provide for sufficient aeration in another way, for example by use of a power head with Venturi feature or an air pump with a bubble stone.

My 125G Tropheus tank I chose to filter with an Eheim 2260 canister filter, and wanted neither additional power heads nor air pumps in that tank. This caused me to forgo the use of an UGJ system, and rely solely on the strong current created my the powerful filter to drive debris into the filter intake. This also allowed me to use one of the excellent Eheim diffusers to provide for aeration.