You are at The Krib ->Plants ->Plant People ->Darn Plant Tank (Olson) [E-mail]

That Darn Plant Tank
Part 4: Plumbing and Heating

text and photographs by Erik Olson
October 1997

In pervious installments of this series, I discussed the hardware involved in building a tank stand and lighting hood. Today I'll delve into Plumbing and Heating, not necessarily in that order.


In my previous tank, I used a combination of substrate heating cables and a normal heater. There has been much said about the benefits of in-gravel heating cables to the plants: they allow gentle circulation (much slower than that in an undergravel filter, for instance) of water through the substrate along the temperature gradients, bringing in new nutrients. It's not clear how beneficial substrate heating is in the short term, though it may provide a long term nutrient replenishment through the currents.

The biggest problem with heating cables has been their cost, somewhere from $100-150 for the cable alone, and if one buys a full-blown Dupla setup with controller and transformer, it can run upwards of $500. Ouch! To get around this, I used a lower-wattage cable that could be left ``on'' all the time without overheating the tank, thus removing the need for the controller. I purchased the transformer at a surplus electronics store for $10. This left only the cable, which I did eventually buy from Dupla. Yes, I could have made the cable myself, but at the time I just didn't feel it was right to chance electrocution or other disasters.

With the new tank, I didn't spend a lot of time worrying about the heater. For one thing, I knew when I set it up that I would only be running it for about a year before I moved, so the need for good in-gravel cable heating was not so great. To add to that, there was a temporary shortage of the 100 watt cables from Dupla when I was setting up the tank, so I opted to just use an old 50 watt cable I had in the old 45 gallon tank. The problem with this is that there just isn't enough length with the 50-watt cable to cover the footprint of the tank. I ended up just covering the back half of the tank rather than try to thin out the spacing over the whole bottom.

As I mentioned before, fifty watts is not enough to heat the whole tank; most of the warmth (and all the control) is provided by some traditional Ebo Jäger tube heaters in the sump. The water pump also generates some heat as well. The only time this arrangement poses a problem is during the hottest of summer months when the temperature can rise above 80 degrees F.


[Photo of Sump]

The sump is the nerve center of the whole operation, holding twelve gallons of water, the CO2 reactor and the heaters. The water returns are powered by a Rainbow Lifeguard Quiet One pump.

[Close-up photo of Sump]

I have to say that I'm very proud of my sump; it's made of heat-bent smoke-brown and clear acrylic, is totally overdesigned and looks like it should be a tank by itself. I chose to build it because of problems drilling holes. You see, the Quiet One pump is not submersible; it cannot be dropped into the sump like a powerhead; it needs to be externally plumbed to the sump. Rather than deal with the whiny local glass shop who didn't want to cut a hole in the side of a 10-gallon glass tank (which I also managed to smash during a short stop in my car on the trip home), I thought it cheaper to buy scrap acrylic and a hole saw.

At the time, I wasn't sure I wanted to invest in bulkhead fittings for the plumbing from the sump. I constructed my own crude versions from PVC pieces and rubber washers. I think that this is actually more trouble than it's worth, and will invest the six dollars in a real bulkhead fitting from now on.

The pump goes into a weird PVC candelabra-like manifold. Four separate returns are regulated by ball valves: Two run back into the tank via vinyl tubing and over-the-side PVC contraptions. A third was added as a spare, but now runs a ``satellite'' 10-gallon tank. The fourth return is a drip irrigation manifold, hooked to the CO2 reactor and other low-flow ``applications'' (which I'll talk about later). The ball valves are an absolute necessity, because the flow would be far too great otherwise. Something I didn't know about pumps such as The Quiet One is that there is nothing wrong with reducing their flow via ball valves; the literature even says they use less energy that way. The important thing is that the flow must always be restricted on the returns, never the intake!

[Manifold Diagram] [Manifold Photo]


The purpose of injecting carbon dioxide into the aquarium is to supplement the CO2 found in the atmosphere, allowing (together with high light and adequate fertilization) better plant growth. CO2 also regulates pH along with carbonates in the water (CO2 lowers it, carbonates raise it).

There is a very wide range of technology available in CO2 delivery systems. The source can be as simple as a 2-liter coke bottle refilled with yeast and sugar every two weeks, or as elaborate as a compressed gas cylinder with output regulated by tank pH. After five years I still take middle ground: a completely manual CO2 injection system from a compressed cylinder, actually a 15-pound refurbished fire extinguisher. The pressure is reduced by a 2-gauge regulator (measuring both input and output pressure, so I know when to refill) and an inexpensive needle valve (the ARO Model NO-1). These three pieces cost me $170. Smaller tanks will reduce the cost a bit.

[Diagram of Reactor]

From the needle valve, the slow flow (a few bubbles per second) is fed into a homemade reactor in the sump, where it dissolves in a trickle of water (which returns to the tank). I've gone through a few different design revisions before settling on something that is extremely efficient, and not surprisingly looks similar to the Dupla reactor (it even uses Dupla Minikaskade bioballs, one frivolous expense of my project). The reactor is a 7 inch high, 2.5 inch diameter acrylic cylinder cemented to a flat acrylic base, held to the bottom of the sump with suction cups. CO2 and water enters the top through a drip irrigation manifold that has four low-flow taps mounted in a 1/2'' threaded plug (why four? that's what they had at the hardware store! I leave two of them capped off), which in turn is screwed into a series of reducing bushings to match the 2.5'' mouth of the acrylic. The top is epoxied to the body (because acrylic solvent didn't seem to bond well to PVC). CO2-laden water exits through a hole near the bottom. On one of the prototypes, I epoxied a single tap to the exit hole so the unit could be operated outside of the sump.

CO2 is injected 24 hours a day, as compared to some systems, which shut off the CO2 at night via a solenoid valve. The pH does drop slightly at night (for instance, from 6.3 to 6.1), which I find is better than having it rise to it's equilibrium pH, over 7.0, which would be the case if the CO2 was shut off. As I mentioned, I could opt to buy or build a controller for another couple hundred dollars, which would more accurately regulate the pH, but heck, I haven't killed any fish in this tank yet.

The Overflow Skimmer:
Problems, Problems, Problems

The two hurdles to jump in keeping a sump are (as you might expect), getting the water down to the sump, and returning the water back to the tank. I've already discussed the return; the far more problematic part is getting water to the sump. This can be done by either drilling a hole in the bottom or side of the tank (the ``standpipe'' method), or by installing an overflow skimmer box that siphons the water over the top of the tank. Both have their ups and downs (ho ho ho): standpipes require modifying the tank (that whiny glass shop again), but do not suffer from as many reliability problems...

[Photo of Overflow]

I almost got this tank drilled for a standpipe, but decided to stick with my traditional homemade overflow siphon box used in my old tank. I've built three of these boxes in an effort to ``get it right'', and I just used the Mark III version from the old tank. This required some adjustments. First, the flow rate was much different with the new pump, causing air bubbles to quickly build up in the siphon... and after a week the siphon started to break. I solved this by drilling a hole for a tiny ``side-siphon'' in the top of the skimmer, which constantly sucks bubbles out of the main overflow. Even this, though, has its own drawbacks. Snails and small pieces of debris get caught into the little hole, rendering it useless until I clean it out, and whenever the system turns off (for instance in the recent power outages), the side siphon keeps sucking water until it breaks the main siphon!

A second problem with my overflow is that the diameter of the drain was too small for the larger flow rate. This meant loud sucking sounds 24 hours a day until I switched to 1-1/4'' PVC and sauna hose.

After getting the first two problems solved, a third emerged. The higher water flow through the small hole was pushing the sponge against the opening, where it quickly clogged. My solution was to ``hollow out'' a portion of the sponge near the drain, so that there was no way for the sponge to get sucked into the opening. This has worked flawlessly, and since that adjustment, I am able to run the overflow for many weeks with no adjustments.

One last problem that has always been around is fish jumping into the overflow box. A disadvantage of my design is that the overflow can cause debris and stray fish to be ``beached'' on the prefilter. Someday I might help reduce the number of fish sucked in by routing little ``teeth'' in the skimmer. Guess that will be the Mark V Model.

Next, I'll talk about setting it all up, planting, fish... you know, the good part.
Up to Darn Plant Tank (Olson) <- Plant People <- Plants <- The Krib This page was last updated 29 October 1998