System Almost Complete - Plus Initial Test Numbers

It's been a long time coming.  With other projects taking priority, the AP system took a back seat.  Finally, the time has arrived to complete it.

First, installation of the duraskrim liner:

After trying to follow along with a video on YouTube that showed how one farmer installs his liner, mine (after folding the ends) came out like this:

It's not easy to see from the picture, but the folds actually proceed between the liner and the wall of the DWC tank, resulting in a smooth side wall.  On the far side, I had to redo half the fold because the fold itself fell across where I had placed a hole for the water pump's bulkhead fitting.  Not wanting to cut through multiple layers of liner, I double-folded that part of the liner and was able to move it sufficiently out of the way.

The final setup - minus many other pieces of plumbing and the nailer boards (which will go over the liner ends) - looks like this:

This actually came out exactly as designed (see one of my earlier posts), right down to the fact that the outlet on the fish tank is actually several inches too low!  In the above picture, this situation had already been resolved by building a makeshift stand for the IBC container:

I ran some basic tests to check for leaks and to try out the double-head bell-siphon setup.  Here are some notes on that:

• The standpipes are 2" diameter, and pass through 2" bulkhead fittings.
• Below, the far standpipe makes a 90 degree while the near pipe falls into a tee.  Thus, both are joined to a common drain pipe.
• All the pipe below the standpipes is plumbed with 1 1/4" pipe and fittings.
• If the near pipe (the one with the tee) overflows first, the far pipe (the one with the elbow) won't siphon.  It appears that the trapped air in the remainder of the drain pipe prevents the siphon from starting on that end.
• If the far pipe starts first, the near pipe will be drawn into the siphon action.
• Siphoning appears to be working reliably, but isn't pulling as fast as I was hoping it would.  I think this may be caused by the 1 1/4" pipe.  It should be noted that I believe this reduction in diameter is helping to start the siphon (thanks to the bottleneck effect), however, after the tee, I'm considering widening the pipe back to 2" to accommodate the combined flow of both siphons.

I did a quick timing test on the siphon.  Right now I have no media in the grow bed and am only letting it fill to about 4.5 inches before the siphons kick in.  At that depth, it is holding around 42 gallons.  It took around 96 seconds to drain until siphon cut-off.  That comes out to about 1600 gallons per hour of flow rate (0.44 gallons per second).

The fill pump is rated to a maximum flow rate of 1000 GPH.  It has to make about a 6 foot rise, which should drop the rate (according to the chart on the box) to between 600 and 800 GPH.  Expected fill rate for 4" of water is 127 seconds at 800 GPH, 170 seconds at 600 GPH.

I tested fill and drain with the pump running, and the grow bed successfully drained completely (the siphons cut off as expected).  I have not yet timed what a filling drain rate is.

Here's the fish tank, with the top cut out.  I tried to make it a hinged top, but the flexible container walls make it too flimsy and pointless to keep.  Instead I'll put some sort of separate cover or mesh to keep the fish in and the children out.  The return pipe (top) is not final yet.  Neither is the grow bed feed pipe (bottom), which will run down into the bottom of the tank to help pull some of the waste material up.

Here's a picture of the temporary standpipe (elbow-side) with the bell removed.  The other bell actually has a piece of Lexan super-glued to it, so that we could watch and ensure the siphon was actually working.

Aside from the nailer boards, the remaining work to be done on the system is as follows:

• The media guards need to be finished.  In the picture immediately above, one of them (teal) can be seen in the foreground).  I had drilled it with small holes at regular intervals along its surface, but I want to add more to ensure that it does not impede water flow out of the grow bed.
• The tank-to-grow-bed drain bulkhead fitting seal failed, and needs to be replaced.  I had glued it in place with 3M 5200 Marine Sealant, but the sealant did not bond well between the PE tank sides and the fitting.  Consequently, when the waterline in the tank went above the fitting, it started leaking.  I ended up dropping the pump into the tank and transferring about 80 gallons into the grow bed and DWC tank to stop the leak and repair the fitting.
• The actual feeder pipes need to be cut and installed.  These will go down into the grow bed and will be set up to ensure that fish water doesn't splash any of the plants.
• As mentioned above, an attempt will be made to improve drain rates a little more - it might be superfluous, but I think it's worth a shot, and will only require the purchase of a single fitting.
• The air lines need to be secured (they are now in place, but not pictured).
• The lighting needs to be built and installed.

After all that, we should be good to go!

Lighting Calculations - Notes

Lighting Calculations, Round 2

Per the Wikipedia article...

1120 W/m2 of direct and indirect sunlight hit the ground.
1050 W/m2 of direct light.
43% is visble (400 to 700 nm)

luminous efficacy is about 93 lumens per watt of radiant flux.

Direct light = 97,650 l/m2 of illuminance.
Visible light = 41,989 l/m2

1 m2 = 10.7639 sqft

DWC: 20 sqft = 1.85806 m2
GB:  16 sqft = 1.48645 m2

Total required illuminance
DWC: 78,018 lumens
GB:  62,414 lumens

lumens per sqft:  3900

This means that we need around 8000 lumens per linear foot of grow space, since the spaces are 24 and 30 inches deep (respectively).  Four 2000 lumen elements, or eight 1000 lumen elements, would provide a sufficient quantity of light, plus we can balance out the wavelengths to provide a nice, full spectrum.

This would need to be repeated each linear foot, or 8 times, yielding around 32 or 64 (respectively) LED elements per grow space, to meet the lighting target.

A search on Digikey provides some ideas on wattage.  Efficiency will be between 80 and 140 lm/W. 80 lm/W buys us 1005 lumen elements, and 64 of those will provide 64320 lumens.  0.0125 (1/80) W/lm times 64320 lm = 804 watts for the smaller array.  Going with 140 lm/W in the search provided options for both 1005 lm and 1770 lm elements, so maybe we'll go with those instead.

Current electrical rate (assume the worst): $0.12047 / kWh

Assume 30 days per month, 12 hours per day of run-time.

Here are the array figures:

DWC:

• 78018 target lumens
• LED quantities:
• 78x 1005 lm LEDs, or
• 44x 1770 lm LEDs
• Cost per LED (best price, rough-estimate):
• $4.00 per 1005 lm element ($312.00)
• $16.51 per 1770 lm element ($726.44)
• Power Consumption:
• 78 * 0.8 A * 9 V = 561 W for 78390 lm (139.7 lm/W)
• 44 * 1.05 A * 12 V = 554.4 W for 77880 lm (140 lm/W)
• Energy Cost (with above assumptions)
• 78 element array = $24.33 / month
• 44 element array = $24.04 / month

GB:

• 62414 target lumens
• LED quantities:
• 62x 1005 lm LEDs, or
• 35x 1770 lm LEDs
• Cost per LED - same prices as above
• 62 elements = $248
• 35 elements = $577
• Power Consumption:
• 62 * 0.8 A * 9 V = 446.4 W for 62310 lm
• 35 * 1.05 A * 12 V =  403.2 W for 61950 lm
• Energy Cost (with above assumptions)
• 62 element array = $19.36 / month
• 35 element array = $17.49 / month

Commentary
The above figures are rough estimates.  Specifically, the energy costs are LOW-END estimates, meaning the final costs will be higher once losses in the AC to DC conversion are taken into account.  Also, the arrays are not necessarily geometrically optimal, meaning that 44 elements will not fit neatly into 8 linear feet of space (5.5 elements per linear foot).  A staggered design will be beneficial nonetheless, so this isn't necessarily a problem.

Output values for the lighting was taken from flux at 85 degrees C.  Lower temperatures should yield better light output (more efficiency).  Also, the test current was used, but max current ratings can be upwards of double or worse.  Higher current usage means less efficiency and higher power bills.

The costs above also do not account for: the LED drivers, wiring, heat-sinks, and LED mounts.

Return on investment will be realized when we save $50 / month on our groceries (really $43.69, but I'm throwing in a wild-ass-guess for the pump and aerator energy costs).


Further reading

• Sunlight
• https://en.wikipedia.org/wiki/Sunlight
• Tables for direct solar radiation at northern latitudes:
• http://www.fs.fed.us/pnw/pubs/pnw_rp142.pdf

Project Update: Electrical Install

Before I lose access to parts of the porch, I have started running the electrical.  A snapped concrete drill bit today stopped my work before I could complete the first run, unfortunately.  Still, I managed to get quite a bit done.

The nearest receptacle is quite a bit away from the grow beds.  I have opted to run schedule 40 PVC electrical conduit from just above the existing receptacle, overhead of the porch door, and down to the right of the door (in the space between the door and the porch screen wall).  The conduit enters a 2-gang weatherproof box, where power will be delivered to a GFCI outlet.  From that outlet, power to the other outlets and the timer (for the lighting) will be drawn.  This means that everything downstream from the GFCI outlet will be ground-fault protected.

I had initially planned to use an existing single-gang electronic timer to drive the lights.  Upon further consideration and research, I decided that it would be better to use a standalone hardwire timer, one rated to significantly more wattage than the single-gang would have provided me.  I chose this one:  http://www.intermatic.com/en/products/timeswitches/electroniccontrols/24hour_7day/gm40ave.  I haven't had a chance to test it, since it's only installed and not wired.  The way I have it mounted, I can either run the return power back to the 2-gang box and manage an outlet there, or run a separate conduit down from the timer to a dedicated LED driver box.  I will most likely be pursuing the latter.

While I am trying to do everything to the NEC code (and I have the large, blue 2008 version sitting in my living room), I elected to leave the bottom plugs of the 2-gang box unsealed.  This was a wise decision, since I realized before this evening's run to our local home improvement store that the pump plug wire was not going to reach to the 2-gang outlet box.  I'll be running another conduit and placing an outlet right behind the IBC tank.

Please note that everything is being weatherproofed, so any stray sprays of water, splashes, or even a good soaking, should not cause any damage or death.  And, of course, the GFCI is there to help with that.  Everything, from the conduit to the wiring to the enclosures, is rated for wet locations, as defined in the Code.

As for the lighting, I have started doing more research on that again.  We recently purchased some LED flashlights.  The beams are adjustable, and with the focuser pushed all the way in you can see the actual LED element.  It's a single-lens array, square in shape, and the output is 350 lumens.  It is quite bright, and has caused me to reconsider just how much light my system actually needs.  I am thinking that it may be time to do some basic testing.  I have begun pricing out some components, and may make some purchases soon so as to get some experiments done.  What I would really like to do is set up a few makeshift grow chambers out of opaque plastic boxes (think the giant tubs for storing your junk in).  With these, I could set up three or four experimental lights and test a small tray of seedlings in each.  My control would need to be a similar tray positioned outdoors, for direct sun.  Unfortunately, I'm not sure I have either the time or the money to do this.  But then again, considering the thought of wasting money on the wrong lighting...

Blueboard, Pump Installed

Slow going, but going forward.  I have installed the XPS Styrofoam board into both boxes.  See below:

The installation was extremely straightforward, and the nails went in quite smoothly.  I have cut, painted and installed a piece of 3/4" plywood for the pump-side of the DWC.  This is to support the bulkhead fitting that will supply the pump, though I am beginning to wonder if the XPS would have performed sufficiently well.  Nailed to both the lower and upper horizontal members, it is surprisingly strong between them.

The pump shelf is now also finished and installed.  I affixed two triangular braces to the underside of the shelf, which stabilized it nicely.  The shelf is cut to fit only the pump, and achieves that.  I positioned it with hopefully enough room between the pump inlet and the DWC bulkhead fitting to allow for adjustment of the pipe, since it won't be an exact straight line.  I may yet need to get somewhat inventive with the plumbing in order to get the pipe position I need; this may be a series of elbows (ugh - no), or some heat-bending (yes!).

I have drilled the hole for the DWC bulkhead fitting, and test fit the fitting in the hole.  All is well, except that with the extra rubber o-ring, there are barely two full threads to screw the fitting nut on.  I may have to leave the o-ring off, which may indeed be recommended - need to go back and check the Green Acres instructions.

I will soon be placing the two bulkhead fittings for the GB.  I haven't yet decided where.  I must also determine the final plumbing configuration for the GB drain; one option is to run one drain to the other, and both down a pipe on the head-end of the DWC (farthest from the pump).  Another option is to drain directly in the middle of the GB, thus ensuring equally good flow from either siphon.  Yet another option is to put the main drop in or near the middle, but then pipe it around to the head-end, for better nutrient dispersion.

I'll be cutting the PVC for the actual siphons soon.  I must also rip and cut to length the boards that will top of the box edges, where the liner will be affixed.  It too must be painted, hopefully the last thing I need to paint.

Dry-fit Complete, Final Prep Beginning

I have finished as much of the painting as I'm going to do on the stands, the lower ramps, and the upper ramps.  The uppers and the stands each received two coats, but the lower ramps received one.  To help mitigate water intrusion from rain, I have decided to place 3 mil plastic underneath the lower ramps, so that it can be pulled up along the sides of the DWC tank.

To the right is a picture of the ramps, on the plastic, near final position.

If nothing else, the plastic will keep the ramps from adhering to the concrete floor.  I am not sure it will completely prevent water intrusion, but I also have no plans to completely or hermetically seal the plastic to the DWC, so there should at least be some free air movement.

Having placed the DWC on top of the ramps, and adjusting the whole configuration to sit roughly where it will probably be staying, I placed the stands into position.  Everything appears to be fitting together correctly.

After placing the upper ramps on top (not necessarily in their final resting positions), I noticed one of the ramps had twisted rather badly.  I may have to reduce the count to four ramps, which wouldn't be the end of the world.

I am generally pleased with the way the system is coming together.  I do remain slightly concerned about the headroom between where the DWC rafts will be, and the assembly that will support the GB.  All in all, there should be enough room to pull rafts in and out, complete with plants (assuming the plants are not very tall), or to pull out plants directly and harvest them one at a time.

Speaking of rafts, I finally received my new adhesive, but haven't had time to try it out.  That will happen soon.

Moving forward, I have put the GB onto the sawhorses for painting.  I'll need to procure more paint, but I'm tossed up between getting a gallon or a quart, considering that a quart is half the price of a gallon, and I may need two quarts to finish the job.

I'll soon be fitting the XPS Styrofoam to the DWC.  I need to make a decision, and soon: I must either double-up on the XPS for the end that will receive the pump plumbing, or I must cut and paint a piece of plywood for that section, OR I must put the pump inside the DWC.

My study of electronics is proceeding, slowly.  I now have enough knowledge that I can start to make some more intelligent explorations into available components.  I highly doubt I'll have the expertise in time to assemble my own current-source, let alone outfit it with all the safety features one finds on such devices as to be available in Digikey's catalog.  This is partly because I am rapidly approaching an area of the topic where differential equations will be required.  I didn't fare well in that course, and since diff-eq requires a solid grounding in the calculus, I am engaged in reading a calculus book.  So, while I will continue both the calculus and the electronics studies, I will need to concede that my know-how will be limited to simple resistive circuits, powered by robust and commercially available supplies.  And that's fine!

As far as the lighting choices are concerned, I must also concede a sort of defeat there as well.  But not really...  Instead of building out full arrays from the get-go, I plan to build some test assemblies and try some plants under them.  This isn't really ideal, but unfortunately it's the best I can do under the circumstances.  I just really can't afford to buy ten massive LED components at nearly $20 a piece, and find out I was wrong about them.  This isn't really a bad thing, though.  This is another opportunity to explore and experiment, and obtain a good working result that I can expand upon.  What I have to do now is resist the temptation to purchase (at around $500) a full spectrum quantum sensor from Apogee Instruments.  Theirs will work with LEDs and provide the necessary data to quantify the actual light output.  Of course, maybe that investment wouldn't be entirely unwise...

Foamboard and Paint

I have finished first-coating the stands, and the lower ramps.  The upper ramps are almost completely second-coated.  See below:

 The paint is going on well enough, but I fear one gallon will not be sufficient once I get to the upper growbed.  The stands are extremely annoying with all their nooks and crannies.  But once painted, it will be done.  I am a bit concerned about the feet of the stands, and the lower slants.  For the former, the notion of water intrusion (as the porch gets wet during storms), and for the latter, I fear they'll "glue" themselves down to the floor after being there for 6 months.

I returned the two undamaged 1/2" foam panels to the home improvement store today, and procured two "blue board" foam sheets.  The key difference is: this foam board is extruded polystyrene (XPS), rather than the white expanded stuff I had purchased previously.  The foam is much denser, doesn't flake, and is reportedly waterproof.

I tried to get the store to order the 2" thick foam I needed, but to no avail.  It seems no one carries it.  I even tried looking for DWC suppliers, but again came up dry.  I'm really at a loss to know what people use, if they don't use completely inert Styrofoam.

Not all is lost, though.  I purchased enough XPS to experiment, and ordered a specific adhesive from the local box store's website.  The adhesive is called Eco-Bond, and reportedly (for the "Farm-Safe" variety) is completely non-toxic, hypo-allergenic, etc etc etc, no evil chemicals, no distillates, waterproof, etc etc etc.  Certifications abound.  Seems too good to be true, but we're gonna give it a try.  Worst-case, I'm out $13.  The goal will be to laminate two or three (or maybe even four) sheets of XPS together, to form a decent DWC raft.  I'm still undecided whether or not I should pre-drill or pre-mark the hole locations, and then glue around them, or else just clue the whole board with lots of beads of glue, and hope I got good enough coverage to avoid splitting during use.
 

Paintwork and Rafts

Painting

I have begun painting the various assemblies for the project.  I managed to find a good exterior-grade latex paint with low-VOC and relatively waterproof.  I'm putting two coats on for the serious wear-areas (any place that will be subject to the occasional water splash or mechanical interaction.

So far, I have only mostly completed the DWC tank.  It's one coat on the top-facing surfaces away from completion, and I will probably only double-coat the sections that won't otherwise be covered by or in contact with other materials.  Due to limited space on the porch, I'm playing musical-components in an attempt to avoid running out of room while painting and assembling.  Thus, the DWC is now sitting roughly where it will stay.

The color white was chosen arbitrarily....and partly to save time in the paint department.  To date, I've gone through nearly half of my one and only gallon.  The second coat does require less paint, luckily, and the GB will be less area to paint as well.  I must also paint the stands, the stabilizer boards, and the ramps for the DWC.  I have one coat on the GB ramps already (which I call "slants" in other posts).

Rafts

I was going to also cut apart some of my rafts today.  Then I made an unfortunate discovery.  The FAO manual mentions making sure that the polystyrene one uses for the rafts be of a food-grade (or food-safe) material.  OK, that makes sense, especially for the fishes.  I looked at the 2" thick sheet that I had already hacked up into a non-returnable state: contains flame retardant.  Fuck.

According to 365 Aquaponics, blue board insulation, or extruded polystyrene (XPS), is the best.  I had purchased expanded polystyrene, which is not recommended (it evidently disintegrates, among its other attributes).  I hope the blue board will also be sans chems, as others have reported that any chemicals within the polystyrene will leach out additives and these additives can be harmful to the fish (to say nothing about us humans).

I'm now going to engage in a brief search to locate possible sources of food-grade XPS.  I'm hoping that home-improvement stores will NOT be the only option.  If they are, I'll have to either find 2" stock or figure a way to adhere two or more thinner pieces together...while keeping it food-safe.  Fully cured polyurethane is one possible route, as (reportedly) once it's fully reacted it thereafter becomes inert and safe to handle.  It should also be waterproof.