You are at The Krib ->Plants ->Plant People ->George Booth [E-mail]

Filtration

Filtration for the SST is handled completely by a trickle filter. The filter is an Amiracle 150, rated for a 100 gallon saltwater tank. This rating seems a little optimistic since the media chamber only holds 4 gallons of media. However, based on our experience with the AOA, this is sufficient for a 80 gallon (true capacity), moderately loaded freshwater tank. The only requirement that seems relevant is to provide enough surface area for nitrifying bacteria, and since we measure 0 ppm ammonia in the AOA, we seem to have enough area. The AOA uses an unknown brand of blue, spiked bio-balls for media, with an unknown surface area per gallon.

The Amiracle filter has no provision for an external pump, so we had to drill a 1 3/4" hole in the end for a 3/4" PVC bulkhead fitting. The bulkhead fitting is connected to a 3/4" union and then goes to the pump via a 6" length of flexible vinyl tubing. Vinyl tubing was used to isolate any pump noise from the filter and to allow for any slight misalignment between the pump and filter.

A Rainbow Quiet One pump was used to return filtered water to the tank. This pump is absolutely quiet in this application and is far superior to any submersible pump we have tried, both in terms of flow volume and noise. Depending on the plumbing and head it pumps against, it can pump 400 to 1000 gallons per hour (GPH).

We initially used all 3/4" PVC to plumb the pump, providing for two returns to the tank. This allowed a maximum of 800 GPH and needed to be adjusted to the 400 GPH capacity of the trickle filter. We discovered that the 3/4" ball valves needed to be almost closed to get the right flow and made it very hard to perform fine adjustments. The increased back pressure also caused a few leaks in the threaded joints of the plumbing. After being unable to seal the leaks, the plumbing was rebuilt with 1/2" PVC. 3/4" flexible vinyl tubing was still used for tank returns. The 1/2" PVC reduced the maximum flow to about 500 GPH which is easily modulated with 1/2" ball valves. Besides the 2 tank returns, a third 1/2" ball valve supplies water to the CO2 reactor in the trickle filter sump.

The two tank returns use 3/4" Lee's siphon tubes to get the 180 degree bend required to get over the rim of the tank. Two more siphon tubes were cut to provide 90 degree elbows for directional control. One of the returns is directed across the bottom, aimed left to right and slightly towards the front of the tank. The other return is directed 2" below the top, from right to left and slightly toward the front. The intent was to provide good water movement in the tank in a circular pattern. This was verified before plants were added by mixing up a tablespoon of calcium carbonate in a cup of water and pouring it into the trickle filter sump. The calcium carbonate dissolves slowly and provides a cloudy solution that makes the water movement visible.

The trickle filter has the usual skimmer box and prefilter. A cylindrical piece of foam wrapped in a piece of drip plate filter pad replaces the DLS in the prefilter. We found that the DLS clogs easily and is very difficult to clean. The filter pad prevents the foam from clogging and is simple to replace. This acts as the primary mechanical filter in the tank.

The filter has a drip plate instead of rotating spray bars, providing a mechanically simple and trouble-free water distribution system. The filter comes with a thin pad over the drip plate to keep debris from the media chamber, but we found that the pad soon became clogged, causing most of the water to bypass the drip plate. The reason for this appears to be that just the small part of the pad over each hole passes any water. These little 3/16" diameter areas clog in about a coiuple of days, even with the prefilter.

Our solution to this problem was to cut a piece of DLS separator (the black pastic mesh) to fit the drip tray and to hold the pad 1/4" above the drip plate. A smaller piece of mesh was placed on top of the pad to prevent it from contacting the water inlet. With this in place, the entire pad area is available to filter the water and there are no "dry" areas on the drip plate. When the filter is running, we can see an even flow across the entire plate.

The media selected for the filter was the new Dupla BioKascade bio-balls. The intial reason was simply to have the latest style media, but it does seem to perform better than our old, generic balls. The BioKascades have an internal vane structure that appears uni- directional. The balls were individually placed so that the vanes are roughly horizontal. Our idea was to provide a horizontal cascading action rather than a vertical drop (bio "cascades"?). This appears to cause less outgassing of CO2, compared to the trickle filter with the standard bioballs. Of course, there wasn't any biological usage of CO2 when this was tested, so this is conjecture at this point.

The balls are 1 5/8" in diameter and happened to pack fairly well into the media chamber. We were able to get 238 balls into the 3.99 gallon space, using 59.6 balls per gallon. Dupla suggests that there are 54 balls per gallon, which would be true if they packed as perfect cubes.

We are not pumping any air into the media chamber as most manufacturers suggest. We feel that the photosynthesis in the tank will provide all the oxygen needed by the bacterial colonies. We also feel that air injection causes excessive CO2 loss in a system like this. We originally had air injection in the AOA filter but noticed no difference in any parameters, especially ORP, after turning it off. The only difference we noted was the lack of annoying air pump hum.

CO2 Injection

We feel that CO2 injection is a key element in a sucessful and stable plant tank. The SST, like the AOA, uses a fully automated CO2 system to provide the proper levels of CO2 for optimum plant growth and stable pH.

The heart of the system is a Sandpoint II pH and ORP controller. The ORP (Oxidation Reduction Potential) is not controlled, but is monitored to ensure that the water is high quality. Typically, ORP will begin to drop if we wait too long to change one of the filter pads. The pH is controlled by injecting CO2 when the pH is above the setpoint (6.8) by 0.05 pH units. The CO2 is turned off when the pH is 0.05 units below the set point, providing for a stable pH of 6.8 +/- 0.1.

The CO2 system is fed from a 20 pound CO2 bottle with a commercial two stage regulator, solenoid and fine control valve from local sources. The Air Products regulator has high and low pressure guages and a control to set the low pressure. We set the low pressure to be between 2 and 4 psi to allow fine adjustments of the actual flow. The a solenoid is capable of controlling up to 100 psi and plugs directly into the Sandpoint controller.

A Nupro needle valve provides the final flow adjustment into the reactor. With the low pressures we use, the valve can easily be set to a flow of 1 bubble every 10 seconds. Initially, we have it set for 1 bubble every 1.5 seconds.

A Dupla Bubble Counter provides a visual indication of the CO2 flow and provides a check valve to prevent water from backing up into the solenoid and regulator (something to be avoided!). Three millimeter silicon tubing is used to connect the bubble counter to the valve and reactor. The bubble counter is mounted on the front of the trickle filter so it is easy to monitor.

The Dupla Reactor "S" sits in the sump of the trickle filter and receives water from one of the ball valves after the Quiet One pump. The outlet of the reactor is in the sump and has a short piece of vinyl tube on it to direct the CO2 injected water towards the bulkhead fitting.

For initial setup, a Dupla CO2 Indicator was used to determine the flow needed to acheive the proper CO2 levels. For the KH and pH we desire, about 20 mg/l of dissolved CO2 seems to work best.

[Continue to part 4 or up to contents.]
Up to George Booth <- Plant People <- Plants <- The Krib
This page was last updated 21 February 1999