Robert C. Michelson's Aquarium

This is the show tank in my family room. When I designed my home, I left a place in a wall separating the family room from the breakfast area to have a large marine “reef tank.” It is an acrylic tank measuring four feet long, by two feet wide, by 30 inches tall (though the water level is maintained at the spillway height of 28 inches). The choice of an acrylic tank over glass was a key decision prior to construction.

This aquarium has fish, but is designed to sustain invertebrates and especially corals. To do this I had to design a lighting system that would simulate the solar spectrum-- at least in the range of plant photosynthesis. To this end, I have incorporated two 400W metal halide bulbs (initially 5,500 degree Kelvin color temperature, but I use 6,500 K bulbs now), and six 30W actinic flourescent tubes to provide the high frequency blues to stimulate photosynthetic activity. In addition, I have two red night-observation lights.

These lights are controlled by a computer system that I designed and built to cycle on incrementally in order to simulate sun rise and sun set with the actinics coming on first one after another, and finally, the metal halides turning on one at a time until “high noon” is achieved. After a period of exposure to the full flux, the bulbs are cycled off in reverse order. Right before the last of the primary light sources has been extinguished, the incandescent red lights cycle on and decrease in intensity until darkeness is achieved. All ballast transformers have been located remotely in the basement to keep the heating of the show tank to a minimum.

The reef was constructed from Fiji and Australian coral rubble that was drilled and tie-wrapped to a PVC scaffolding (shown here under construction) when the tank was first filled. This allows a much larger interstitial space for fish to hide inside the reef while giving the appearance of a much larger amount of “live rock” than actually necessary. The overall weight of the system is similarly less than had the reef been constructed from a solid pile of rock.

Extreme water purity starts with filtered deionized well water and is maintained using the “Berlin” method in which coral rubble from Fiji and the Australian Great Barrier Reef provide surface area for Nitrosomonas and Nitorbacter bacteria to thrive. These bacteria reduce the primary waste products of respiration and consumption: ammonia. This ammonia is reduced to nitrites, which are in turn reduced to nitrates. Bacteria in a biological anoxic filter located directly below the show tank in the basement “Aquarium Command Center” further reduce the nitrates to free nitrogen gas.

Calcium is introduced into the system in the form of saturated kalkwasser from an automatic Nilsen Calcium Reactor which dispenses the calcium hydroxide directly into the sump (without atmospheric contact) in order to maintain calcium levels, alkalinity, and a high pH. The entire system is gravity-fed and doses the calcium hydroxide solution as replenishment for pure fresh water evaporating from the system.

As the water flows over the spillway from the show tank above, it falls approximately 10 feet into the basement by way of flexible PVC pipes which allow the water to pass through a mechanical filter consisting of baffles to trap detritus. Water can be drained from sumps between these baffles to eliminate the trapped detritus by means of a ball valve connected to a manifold that draws from each baffle area.

The water then drops into the biological filter where it passes over an aragonite bed. A pleneum under the aragonite (calcium sand) fosters the growth of the nitrate-consuming bacteria. From the biological filter, the water passes into a sump where it is circulated through a dual-tube down-draft protein skimmer that I designed. This adds a tremendous amount of oxygen to the water and also eliminates many hydrophobic/hydrophillic molecules, heavy metals, and detritus from the water by mechanical means. The water is then pumped back into the show tank through a sinusoidally varying valve arrangement that causes a left-to-right, and then a right-to-left surge across the tank above.

All of these systems I designed and built by myself over a period of about 9 months from new and surplus materials. The total system cost in 1995 was approximately $3,500 (not including live rock and livestock) as shown in the table below.

Since the basic system was completed in December of 1995, I have built many additions to improve its level of automation and reliability. Included in these are such items as the automated aquarium refill system, show tank surge devices, and an emergency life support system 8 KW power backup with two phase zoned switching and load monitors. Some upgrades have been done as well. For example, an additional 75 gallons of system volume was included to bring the overall amount of water to 325 gallons.

TOTAL START UP COST FOR BASIC SYSTEM

Acrylic custom-made show tank (2 x 4 x 2.5 ft) ........................ $850
   (with upper surface milled for twin 18 x 12- inch acrylic covers)
Water distribution pipes/fittings (both flexible and rigid PVC) ........ 200
Main circulation pump .................................................. 115
Skimmer circulation pump ............................................... 115
"Michelson" twin down-draft protein skimmer ............................. 75
Acrylic skimmer sump .................................................... 30
Jaubert sump 2-inch hole-in-glass cutting charge ........................ 25
(surplus 40 gallon long glass aquarium - free)
Aragonite for Jaubert sump and show tank ................................ 45 
Detritus filter glue (materials to be glued were surplus - free) ........ 10
Miscellaneous glass lids to limit evaporation in spillway ............... 35
   and Jaubert sump
Floating polyethylene balls (1000) to limit evaporation ................. 30
   in skimmer sump
Twin 400W 5500 degree Kelvin metal halide bulbs and mogul sockets ....... 51
   (Have since gone to twin 6500 degree bulbs at 4 times the price)
Ballast transformers for metal halide bulbs ............................ 240
Heat sink for metal halide ballasts (surplus - free) ..................... 0
Stainless steel hood fabrication (material surplus- free) ............... 50
Six 30W fluorescent fixtures (for actinic bulbs) ........................ 90
Six 30W actinic bulbs .................................................. 150
Twin red night observation lights ........................................ 3
Four muffin fans for cooling of ballasts and air movement ............... 40
   across metal halide bulbs 
Miscellaneous wiring and junction boxes ................................ 300
Isolated 15A ground-fault breaker ....................................... 30
Computer controller for individual light and pump control .............. 240
Sinusoidal water diverter for oscillating surge in show tank ............ 50
Miscellaneous fasteners ................................................. 75
   (tie-wraps for rock-to-scaffolding interface, scaffolding, etc.)
200 gallons of sea salt ................................................. 60
Heater in Jaubert sump .................................................. 25
Show tank finish carpentry and sump support tables ..................... 550
Two twenty five gallon Nalgene make-up water drums (surplus) ............ 16
                                          	TOTAL START-UP COST = $3,500

As a final note, if I’ve seen one yellow tang picture at various aquarium web sites, I’ve seen them all. This site is really about the engineering of a research aquarium, in particular how to automate the system and do it affordably. Nevertheless, I get lots of E-MAIL from people who want to see the inhabitants of the aquarium. For those who require proof that things can actually grow in this robotic life support system, please refer to the pictures found here.