I am making plans for a v2.0 of this filter for use in my (long running) tank build. After using the filter for the past year or so, I have come up with some tweaks and rules for operating that a new build can incorporate.
Some of the improvements over the original design will be
- Pleated filters on return to tank (10 micron and 5 micron already added to existing filter)
- Parallel glass bead filters
- Back flushing capability for all canisters
- Built in drains for all canisters
- Minimize tank water usage
The pleated filter addition really helped my confidence that the filter was removing the bound LaPO4. Their addition was also helpful to gauge when a back flush is needed. The glass beads will filter down to 3 microns, but as the filter becomes packed with precipitate, some precipitate will eventually get pushed past as the back pressure builds up. The pleated 10 micron filter, which is next in line, will start to show brown blotchy stains as the precipitate that got past the beads gets trapped. The 5 micron is just for insurance and shows only minor tinting and no obvious signs of particulate still being in the effluent at this point.
The canister containing the glass beads does the majority of the work with this system. It is where precipitated LaPO4 is trapped and then back flushed down the drain. The hard shell canister insert with the small grating from BRS that this is based on, only comes in the 10" x 2.5" size unfortunately. I would prefer a deeper one, like a 20" canister, but with only the 10" available, then I will need to design around that. For that reason, the new design will have 4 of the glass bead canisters operating in parallel. This should extend the time between flushes. With the current design, flushing every 6-8 hours is needed. Once you get to 12-14 hours without a flush, participate will get pushed past, and the flow will drop down to the point where alkalinity can start to bind instead of phosphate. I expect the parallel bead filters will auto adjust the flow as they become packed with precipitate, with the ones with least back pressure taking over for ones that have been blocked with precipitate.
Currently only the bead filter can be back flushed. The pleated filters have to be unscrewed and removed for cleaning in the sink. I plan on adding back flushing for each individual canister (4 glass bead, 2 micron filters). This will negate the need to remove the micron filters for cleaning. It is also handy when I get this thing fully automated. The back flushing in the current v1 system involves actuating a bunch of valves in order. I plan to use 3 way ball valves for the new design with will both help it to be more compact and to require fewer valves. Though each individual valve will be more expensive.
Starting up the v1 design requires tank water to be flushed to the drain to push out the (most likely stagnant) water from the reaction canister. The filter sits unused more than in operation, so water sitting in the reaction chamber in the v1 design, tends to get skunky with the potential for hydrogen sulfide to be present. To avoid putting this water into the tank when the filter resumes, the reaction canister's volume of water plus additional water gets flushed to the drain in the v1 design. For v2, I plan to add a 1/4" drain in the bottom of the reaction chamber hosing and possibly in the micron filter housings as well. I will use the same 3 way ball valves, they have an 'off' setting where both directions are closed. These will be perfect for draining the canisters. My plan for the v2 design is to have the filter drained and ready to be started up after each use. Currently there are manual maintenance in the v1 design, removing the and cleaning the micron filters, flushing the reaction canister, that have to be completed between runs or during the start up. With the addition of liberal drains, both to waste and back to the tank, I hope to minimize the wasted tank water and make it easier to store the filter in a ready state.
For the v2 design, I hope to add automation. This will involve making mounts for high torque solenoids to be mounted to the 3 way ball valves. With the addition of a flow meter, the need for a back flush can be monitored and determined. At that point, the bead filter canisters would be drained back to the tank, then each bead filter will get flushed in sequence until all have been flushed. Pre-draining the bead filters before the flush should help to minimize tank water loss. Several minutes before the flush is started, the LaCl dosing can be suspended to allow both sides of the bead filters to be drained back to the tank with no risk of LaCl being present.
My new build has a filter with a big bed of substrate that I don't want to become bound with concentrated PO4, so I need a plan ready to be able to deal with it in an ongoing way. This redesigned v2 filter will hopefully be the solution to that.
Dennis
