abcha0s' 300G Ultimate Reef

[abcha0s]

I’ve seen some really nice and very creative sumps. However, most of these are meant to maximize a very limited space under the display tank. With such a large stand for this build, space isn’t really the primary consideration and a lot of the creative solutions don’t really apply.

I came up with a fairly basic design.



The compartments in order from left to right are:

• In – This is where water will enter the sump. The baffle will reduce turbulence before it enters the skimmer chamber.
• Skimmer Chamber – I designed it to be large enough to hold two skimmers. I will likely start with just one, but may add another.
• Frags or refugium or both
• Return (evaporation occurs here) – Live rock rubble.

Bulk Heads

• The two bulkheads in the first chamber are intended for a closed loop between the sump and my water room. More on that later.
• The two bulkheads in the last chamber are intended for dual return pumps.

Water Volume – Pumps Running

Note: I’m not concerned about precision here – numbers are rounded for glass thickness and to error on the side of safety.

• (7 x 10 x 24 / 231) + (18 x 9 x 24 / 231) + (27 x 8.5 x 24 / 231) + (12 x 7.5 x 24 /231) =
• 7.273 + 16.83 + 23.84 + 9.35 = ~57 gallons

Water Volume – Total Capacity and safety considerations

Total volume is: 65 x 24 x 16 / 231 = ~108 gallons

Free space = 108 – 57 = 51 gallons

1” of water in the display tank is: 36 x 72 x 1 / 231 = 11.22 gallons

I should be able to accommodate for 4 inches of water back siphoning from the main tank.

(50 / 12)

Water Volume – Return Chamber

This is important at it effects the time before disaster in the event that all of the ATO systems in place fail. If for example the ro/di reservoir runs dry, water will evaporate from this section of the sump until the pumps run dry.

• 1 inch in this chamber holds 1.25 gallons of water: (12 x 24 x 1 / 231)

The pumps will start sucking air when the water level reaches about 3.5 inches. Therefore, I safely have about 4 inches to work with or 5 gallons of water. The rate of evaporation will be dependent on the amount of evaporative cooling required. I would estimate this at somewhere just short of 24 hours.

I considered a down turned elbow to bring the pump intake to about 1” from the bottom. The benefit being increased water volume at the cost of increased flow resistance. I probably won’t do this.

Guiding Principal

The main consideration with the sump was flexibility. I find it very hard to plan everything without the spatial benefits of seeing the space available. The sump should be usable in ways that I hadn’t originally considered. Suggestions are welcome?

Last Minute Modifications – The builder offered to drill a second hole in the first chamber (IN). I’m not sure it was necessary, but I figured it couldn’t hurt. If I don’t use it, I can always plug it with a bulkhead and some plumbing. Also, the two holes drilled in the last chamber (return) were too close together, so I had a third hole drilled. I will cap the middle hole or possibly use it as a drain.

[abcha0s]

The tank should fit together something like this:



There is a ¾” sheet of maple plywood that sits directly on top of the stand. There is ¾” Styrofoam (Check manufacturer part #) that sits between the tank and the plywood. The idea is that any imperfections in the plywood surface are minimized by the Styrofoam. The design also builds in an extra degree of safety.



As I understand it, there is a little bit of a debate over whether glass tanks can or should be rested on Styrofoam. It seems to be a requirement for acrylic tanks, but there are two schools when it comes to glass. Many store bought tanks have a one-piece injection molded frame that goes around the edges of the tank. These tanks are intended to have a free floating bottom and Styrofoam is not recommended by the manufactures. I would add that some people still use Styrofoam anyway (including me on my previous tank), but others have had real problems including cracked bottom panels. Even though this tank is glass, the particular style of construction requires a Styrofoam layer between the glass and plywood.



Here you can start to get a feel for how the overflow plumbing will be brought into the lower section of the stand. It did become apparent after the fact that there may be insufficient depth in the stand for the overflow plumbing to be rigid. Flex PVC overcomes this obstacle and is probably a better solution anyway.

[abcha0s]

This was the fun part. If you've never watched an aquarium being built, it might be really interesting. For me, it was inspiring.

I have to say again how impressed I was with the tank builder. As you go through the following slides, I think you'll see that this was not their first tank.

<< Images are clickable – View the images in hi-res >>

I'll spare you the commentary. A picture is worth a thousand words.

.

.

.

.

.

.

.

.

.

The tank was assembled as shown in these pictures in one evening. The cleats and Eurobracing were attached during a second visit. The overflow was attached during the third visit and cleanup was completed on the forth and final visit.

It not's hard to understand why a site build costs more than having the tank built at the shop. If you consider the cost of delivering such a large tank, I was happy to pay a little extra to have it built on-site.

Unfortunately I don't have any pictures of the Eurobracing being assembled or the overflow. However, the final tank pictures show these features well.

[abcha0s]

Not sure what to say? - Here it is...

<< Images are clickable – View the images in hi-res >>

.

.

.

.

.

[abcha0s]

<< Images are clickable – View the images in hi-res >>

When I plumbed my 90G, it took me 6 months. Plumbing the 300 took me about 6 hours. I enjoy the creative aspect of plumbing.

A couple of tricks that I’ve learned over the year.
• Always use primer on hard PVC. Never use primer on flex PVC.
• Always dry fit everything first.
• Wipe off excess PVC cement. Both inside and out. Disassemble couplers when gluing so as to allow access to wipe of excess cement.
• Plumbing at the pump intake should be short and direct. If this is not possible, a larger diameter pipe should be used and the diameter should be reduced right before the pump. Minimize resistance on the intake.
• Elbows before the intake of a pump can be a source of microbubbles.
• Align the couplers uniformly throughout all runs. This allows any section to be replaced with a single coupler (one piece on each end of the section).
• Use two 45s rather than a 90 wherever possible.
• Use flex PVC wherever possible. Always attach pumps to a section of flex PVC.
• Don’t allow plumbing to stress the tank glass by acting as a lever.
• Use a silicone lubricant on all O rings.
• PVC is easy to ‘cut to size’ with a dremel. If a piece is just a little too big, cut it down. Make your own fittings.
• Make sure you have a way to isolate any pumps for maintenance or in case of a failure.

These aren’t really guidelines so much as rules. Short cuts here are just not worth it.

Return Pump Selection

For my return I selected the Eheim 1262. For redundancy, I will use two of them. This was not an easy decision and I researched many other great pumps, but in the end, the only other pump that came close was a Red Dragon. At 4 to 5 times the cost per pump and no spare parts, the Red Dragon pumps are really hard to justify.

The biggest limitation of the Eheim pump is that it does not have high flow like many of the other return pumps on the market. However, I’m not sure that high flow through the sump is really necessary. I only want enough turn-over through the sump to keep my skimmer happy. This is just short of 700gph or slightly more than twice the display capacity.

The Eheim specs that were important to me in making this selection are:

• External
• Low heat transfer
• Quiet operation
• Max output: 900gph (estimate at 600 after 8 feet of head loss)
- Two pumps should give apq 1200gph
• Power consumption: 80W per pump.
- Two pumps would be 160W

Another very important thing is the ability to get spare parts or a replacement pump if necessary. Just about everyone stocks these pumps and parts are often available.

Finally, Eheim has a great reputation and the average user ratings are almost always good. I have used them for other purposes and agree with the general consensus.

Modding the Eheim for PVC plumbing

Eheim pumps come standard with barbed fittings for flex tube. This is pretty standard stuff with fresh water tanks, but didn’t really work with my plans for plumbing. I really needed a way to connect a PVC fitting.

The pumps are fitted with ¾” FPT (Female Pipe Thread), but I’m not sure how standard the threading is. Typically this thread would be tapered causing the male and female threads to jam together forming a solid connection. However, the Eheim thread does not appear to be tapered. The obvious concern is an inadequate seal resulting in leaks. Teflon tape might address this, but I suspect a lot would be needed and it would need to be reapplied whenever the pump was disconnected for cleaning. The original part is shown here.



Note the ‘O’ ring just before the hex head. This is how Eheim maintains a water tight seal. I figured 'why reinevent the wheel'. If the 'O' ring works - keep it.

Using a dremel, I cut the barb adapter off.

. .

The top view shows a relatively straight inner core. I rounded it out a little in hopes of minimizing turbulence. The concern here is micro bubbles.

For the intake I used a 1” to ¾” PVC reducer. This worked really well because of the hex shape matching the Eheim hex to within millimeters. For the outtake, I used a ¾” to ½” PVC reducer. This also worked well although the hex heads were slightly different in configuration. Only the intake is shown here. The first image shows the IPEX markings on the PVC fitting. In the second image they have been ground off.

.

I used Mr. Sticky’s underwater glue to bind the two pieces together.

.

The above process was repeated for the intake and outtake of both pumps.

Plumbing the return pumps

With the modded Eheim fittings installed back on the pump, I continued to plumb the return plumbing.

. .

The first step was to attach the union fittings. For the intake it is a 1” union and for the outtake it is a ¾” union. The ball valve completes the pump plumbing.

On the tank side, I used a section of spaflex (Flexible PVC). I used spaflex so the I didn’t need to be concerned with an exact alignment with the pump and to minimize any transfer of vibrations from the pump to the sump.

.

The Styrofoam riser brings the pumps to the correct height and should also help to further minimize vibrations.

The astute reader would note that there are no valves for isolating the pump from the sump. The primary consideration here was space and a valve just wouldn’t fit. However, I do have a solution. The bulkheads are threaded on the inside. I can attach a threaded cap on the inside (wet side) to temporarily hold back the water. With a small Tupperware tub, I can catch any water that leaks while disconnecting the pump. Once the pump is off, I can attach a capped union such as is already installed on the center bulkhead. The net result is I do have a strategy for removing the pumps.

The first picture below shows the 5 tank fitted bulkheads and also shows the loc-line fittings. The bulk heads were installed with nothing more than a hand tight seal (no silicone or glue). The other 3 pictures show the vertical section from the return bulkhead into the stand.

. . .

I used hard PVC to plumb from the return bulkhead into the stand for a couple of reasons. This section of plumbing is visible and I wanted it to be straight. I had originally thought to use spaflex as far as the vertical section, but the flexible PVC doesn’t have a small enough bend radius to go from the roof of the stand to the vertical run without looking unfinished.

Note the check valves installed just below the top elbow. These serve an important role in the overall tank operational strategy. If one of the two return pumps fails, the check valve will prevent the failed return loop from back siphoning (or at least limit the rate of back siphon). It also allows the return pumps to be stopped without the water level dropping in the main display. I may want to stop these pumps for feeding or other maintenance such as sand siphoning. What these check valves do not do is protect against a flood. The sump has full capacity to hold all of the water that could potentially back siphon in a full power failure. In my opinion, check valves should never be trusted to be water tight.

If the check valve on the return line needs to be serviced (for cleaning, or in the event of a failure), the water level in the display can be dropped below the return lines. This is really only a couple of inches and should be a reasonable strategy for emergency repairs.

One final observation on this section is in regards to the ¾” bulkheads. I found these to be extremely poorly made. Specifically, the PVC pipe does not fit securely into the slip side of the bulkhead. This has me a little bit nervous, but I did use extra PVC cement in an attempt to get the best weld possible. The zip ties attaching the plumbing to the stand are intended to minimize any stress on the bulkhead weld due to twisting or turning of the plumbing. If need be, I will replace these bulkheads. Maintenance would involve lowering the water level and cutting them out with a dremel. This really would be ‘no big deal’, but I first have to find a better bulkhead.

. .

The spaflex is zip tied to the frame to hold it in place. This is a temporary measure. I figure it’s sufficient for the purpose of this thread, but I wouldn’t leave it like this for anything more than testing.

Plumbing the overflow

As noted in the planning section, I am using a modified beananimal overflow. In contract to the ¾” bulkheads (no name) used for the return plumbing, the 1.5” bulkheads I am using for the overflow are really very well made (Lifeguard Aquatics). This first picture shows the unions glued to the bulkhead.



One problem that I knew I was going to have to deal with was the overflow not aligning properly with the stand. This came about for a variety of reasons, but it was easily solved. I had considered using 1.5” spaflex here, but I really prefer working with hard PVC and the benefits of spaflex don’t really apply to overflow plumbing.

. .

Using two 45 degree elbows together creates a step. However, the step was way more than I needed, so I cut one side of the connecter in half. The result was a smaller step that was still too big, although manageable. You can see in the picture with the section attached that it enters the tank stand on a slight angle. I can live with this.

Another challenge that needed to be overcome is the orientation of the sump in relation to the overflow. The sump runs the full length of the tank from left to right and is designed for water to flow the full length before being returned to the main tank. The overflow is dead center.

The principals of the beananimal overflow make the problem somewhat easy to overcome. It is really only necessary for the ‘full siphon’ standpipe to enter the first chamber of the sump as this standpipe carries 90 percent of the total flow. The ‘open channel’ standpipe carries the remaining flow while the ‘emergency’ standpipe is dry under normal operation. Having the ‘open channel’ standpipe bypassing the skimmer section should really be inconsequential.

To relocate the ‘full siphon’ standpipe into the first chamber I used two sweep elbows.

.

The gate valve allows for fine tuning the siphon with an amazing degree of accuracy and ease. This task is much more difficult with a ball valve as you really need the accuracy. The weight of the plumbing is supported by the stand to ensure there is no additional stress on the bulkheads or glass.

.

The standpipes are different heights depending on their function. They are friction fit and can be easily removed. There are a couple of outstanding tasks such as fitting a strainer to the top of each standpipe and drilling holes in the top of the open channel standpipe. I’ll update the thread with a picture once I’ve done this. For now, the height of the standpipes is a best guess. I will only really be able to finalize the configuration once water goes into tank.

[abcha0s]

<< Images are clickable – View the images in hi-res >>

Mostly, I just wanted to see water in the tank.

Aside from the primary motivation, there are several things that I really did want to test.

• Validate that the tank, sump and overflow hold water.
• Validate that sump baffles are water tight.
• Validate that all of the bulkheads are properly sealed.
• Validate that there are no leaks in the plumbing.
• Validate that the overflow standpipes are water tight when using a friction fit.

I filled the overflow and basically just watched for drips into the sump. The seals are good.



Starting from the first sump chamber, I watched as water filled one chamber and then overflowed into the next. There were no leaks. All of the baffles are water tight. All of the bulkheads have a good seal.

. .

.

The height of the baffles decreases as you move from through the sump. The intention was to agitate the surface. This appears to work, although I won’t know whether it is sufficient until the tank is in full operation.
I allowed the sump to fill almost to the top and then started testing the return pumps. First the left only, then the right only, and finally both pumps at the same time. Everything worked as expected.

. .

Measuring the flow through the sump

The Eheim 1262 are rated at 900gph before head loss. To measure the actual flow through the sump, I timed how long it took for the two return pumps to fill the tank a total of 5 inches.

.

• 5 inches of tank water is equal to: 71 x 36 * 5 / 231 = 55.325 gallons
• Time taken was: 2m46s or 166s
• Return rate per second is 55.325 / 166 = 0.333

Return rate per hour is: 0.333 * 60 * 60 = 1199.8 or 1200GPH

Filling the tank

It took a total of 1h and 14m to fill the sump and the tank. That’s a lot of water inside a glass box.

.

Tuning the overflow siphon

With the tank full and the pumps running, I adjusted the gate valve on the full siphon until water was trickling into the ‘open channel’ standpipe. This silenced the siphon.

The first image shows the water level in the overflow after tuning the siphon. The second picture shows the turbulence in the first chamber of the sump before tuning the siphon and the last picture shows the lack of turbulence after tuning the siphon.

. .

I did learn a couple of things.

• The siphon should probably be an inch shorter to avoid sucking air from the surface.
• The open channel standpipe gurgles just a little bit. I will have to do some reading and see if there is a way to completely silence this. I suspect it might need to be turned down?

Bare Bottom

There is an interesting effect when viewing the tank bottom through the sides. I’d never really seen this mirror effect as I have always had a bottom substrate.

[abcha0s]

Details are important. Details cause delays. I was pretty much ready to go with the tank, but had to take a big step back to work with some of the details.

Black

I love the depth that a solid black background creates. Almost an infinity or an abyss. However, I really want to preserve the viewing from all 4 sides including the back. Granted that rear viewing is somewhat awkward given the tanks proximity to the wall, but for a motivated viewer it’s actually not too bad. The glass of the tank is roughly 16” from the wall and it’s not too hard to squeeze behind the tank or just peak around a corner.

I also wanted to hide the overflow plumbing, but with a viewable back, the white PVC really stands out.

Rather than painting the rear glass panel black, I painted the entire wall behind the tank black.



This really worked well. When viewing the tank from the front, the gap between the wall and the tank adds to the overall depth. When viewing the tank from the couch (10 feet away), the black wall makes the tank appear to float in space.

At some point I will replace the white electrical plates with black ones.

The overflow plumbing is also painted black. It should blend into the wall and while it will be visible, it should not be distracting. I used Krylon spray paint (for plastic) for the pieces that don’t touch water. For the in tank plumbing (inside the overflow) I used schedule 80 PVC which is naturally a dark colour.



Floor

Next was the floor. Our basement was previously carpeted and honestly, no matter how hard you try, it’s impossible not to spill when working on a tank. For the tank to be a long term success, the carpet had to go. We explored a number of options including cork, hardwood, laminate and tile. In the end, we settled on a vinyl tile (Armstrong Alterna). This floor is extremely durable, looks like natural slate even up close, but will not crack and is much warmer under foot than natural slate would have been. We are very happy with the results.



The biggest challenge was moving the tank around while the floor was being done. You can imagine this was a big project. The carpet came out, the concrete was resurfaced and then the tiles were laid.

To move the tank around, I used a car jack to lift it and build a platform with wheels underneath. This was a little bit tricky to accomplish but the end result was I could easily role the tank around. The carpet guys just worked around it. Getting it down off the platform was also not an easy task, but I managed.

HRV

Humidity was a real problem in our house. With my 90G tank, on a cold day we had puddles on our windowsills. Ultimately, I am sure this would have ruined the house. More to the point, my wife was not very impressed with this aspect of my hobby and the prospect of a much larger tank was daunting.

An HRV (Heat Recovery Ventilation) seemed the obvious choice.

The whole thing is a bit of a monster and the installation was challenging. You really need a flexible space to install one of these.



The HRV has 4 hookups for airflow.

1. Draws fresh air from outside to inside.
2. Vents fresh air to house – Split between tank room (66%) and house central heating.

3. Draws stale air in – Split between tank room (66%) and cold air return (33%)
4. Vents stale air from inside to outside.



It’s all about the heat exchange coil.



This next picture gives you an idea of how the HRV will service the tank.



The Stale Air Intakes are centered directly above the tank. The humidistat is positioned slightly to the left of the tank, but in close proximity. The Fresh Air Exhaust locations were chosen for their practicality in terms of ducting, but are within close proximity of the tank.

The result of installing the HRV has been nothing short of spectacular. The air quality in our house has never been better and our humidity problem is completely gone. There is a cost in terms of reduced heating efficiency during the cold months, but the tank adds some of that heat back in. I have no question that this was the right solution for me.

Screen Top

One of the principals that I am striving for is a “nothing over the top glass” look. This creates a nice platform for a screen top. I bought the clear mesh from BRS and used a screen frame from Home Depot to make this top.



For anyone who has ever attempted this, you will appreciate how difficult it is to complete this project perfectly. I really took my time and laid everything out, but there are flaws. The frame is slightly off square and the mesh is not perfectly straight across all lines. However, it is not immediately evident unless you really inspect it. For the most part I am satisfied with the outcome, but may attempt a version 2 at some point in the future.

[abcha0s]

The only disappointment with our new floor was that it isn’t level or planar. This was pretty obvious when we pulled the carpet out, so we contracted the flooring company to level it for us. Well, my advice is to never have a “flooring” company do anything other than installing a new floor. They just aren’t qualified. They did try to level the floor, but it’s hard to tell if there was any improvement or if it’s actually worse. I negotiated a reduced rate for this aspect of the installation, but really I shouldn’t have paid anything at all. The floor just isn’t flat.

In leveling the tank stand, it was important that I do so in a way that wouldn’t damage the new floor. That rules out any kind of leveling feet as they would surely damage the tiles. The obvious solution is simply to shim the stand, but I also didn’t want the stand sitting directly on the tile.
The solution I came up with is somewhat elaborate, but worked extremely well. It basically amounts to layers.

Layer 1: Underlayment - Dense Rubber Mat cut into strips. This is 5/8” workout mat.
Layer 2: Supporting Frame - 1/2” standard grade plywood. Single piece with center removed.
Layer 3: Shim layer
Layer 4: Platform - 3/4" Oak – Full sheet.

The premise is that the shim layer is between two layers of wood. The rubber layer protects the floor and will compress a little to even out very small imperfections. It’s a little more complicated than that, but that’s the general theory. I wish I had more pictures, but sometimes documenting is not my primary focus.



You can see in the above picture that the width of the rubber strips matches the plywood and that it is roughly 3.5” – This is 2” for the metal frame with ¾” on either side. For support, it really isn’t necessary for this to be any larger as the plywood will flex and there will be no weight support beyond about 3/4”.

The rubber was cut to length and then glued to the wood frame.

The middle of the frame is cut out for two reasons:

1. There is a slight rise (or hump) in the floor. By cutting out the middle, this hump will not affect the overall level.
2. There is a sheet of 3/4” MDF cut to the inside dimensions of 1/2" plywood that is glued to the underside of the platform. While it isn’t shown in the pictures, this is an important concept in the overall design. It provides necessary rigidity to the platform. To visualize this, think of it as a plug that completely fills the available space.

The resulting platform layer (Layer 4) is actually 2” thick and itself is made up of 3 layers.

Layer 1: Lower Plug – 3/4” MDF
Layer 2: Platform – 3/4” Oak Plywood.
Layer 3: Upper Plug – 1/2” Standard Grade Plywood.

The three layers of the platform are glued together with wood glue and is extremely strong. The size of the platform is such that 3/4" extends out on all sides. This will permit the panels to rest on the platform rather than directly on the floor (which isn’t perfectly flat).



In the above picture you can see how the shims are inserted. At some point I will clean this up and glue a finishing trim all around the base of the platform. The final result will be such that it appears as a solid piece of oak.

[abcha0s]

This post is part 1 of 2 dealing with water changes. Part 1 focuses on ro/di storage, salt mixing and the general layout of my water room. Part 2 will discuss the Continuous Water Change System.

The Water Room

I don’t have space for a fish room, so I’m calling what I do have a “water room”. This room is in fact my furnace room. It is physically about 10 feet from the aquarium, but there is a door and a hallway in between the two. The furnace room is of average size and is big enough for several reservoirs and all of the ro/di equipment. It also has plumbing for both a water source and a drain.

I do have access to the ceiling from several vantage points. Also, the wall behind the aquarium leads to a closet under the stairs. This means that I can run plumbing or any other system hookups between the two rooms with relative ease.



You can see the water heater and furnace on the right. The wall on the left was just vapor barrier over insulation, so I mounted 3/4" plywood. I sealed the floor with “Kitchen and Bath” caulking. We also did the floor with an epoxy so there is no exposed concrete anywhere. This is the starting point.



The picture above is the room as it is now with everything functional. The rest of the pictures in this post just attempt to show what’s going on.

There are a couple of details worth pointing out in this picture - The first is that the sink is supported by brackets (Lee Valley) rather than legs. These laundry tubs come with really cheap legs that aren’t much good for anything. The bracket also frees up space under the sink for things like salt and buckets. The second detail is the florescent lighting. The room is inherently dark and I think it's important that you can see the water that you are working with.



I’m not a plumber and learning to weld isn’t in the cards. Fortunately, the plumbing in my house is flexible. I used “Shark Bite” fittings for everything which makes plumbing really easy. This picture shows where I tee’d off one of the main lines. I installed a shutoff right after the tee so that I can isolate my plumbing from the rest of the house.



The RO/DI is a key part of the system. The RO unit on the left is an old Kent Marine Maxxima unit that feeds my ATO reservoir. The RO unit on the right is a Vertex Puratek 100GPD unit that is currently only used to fill my RO/DI reservoir for salt mixing. This gives my capacity to change roughly 25% of the total water volume in my tank every day.



The system is pressurized with the use of Lifegard Aquatics Quite One 3000 pumps. I’m not a fan of these pumps but for occasional use where noise (or random failure) isn’t a big deal, they fit the bill. The pumps and reservoirs are isolated with ball valves and knife valves, so replacing a pump is relatively straight forward.

The ro/di reservoir is the one on the left (closest to the wall). The reservoir next to this is the salt mixing tank. There is a third reservoir capable of holding 200 gallons of prepared saltwater that is used for salt water storage.



The picture above shows two lines coming into the saltwater reservoir. The line that is currently connected is a loop from the pump attached at the bottom of the saltwater mixing tank. This pump runs for 10 minutes every 4 hours just to ensure that the water doesn’t sit in the pipes. It can also be turned on as needed when I am moving water out of this tank.

The line that isn’t hooked up to anything is for RO/DI water from the other reservoir. This comes into play only when the lid is off as shown in the next pictures.



Inside the saltwater mixing tank there are two Korallia Evolution 1400GPH pumps. These run continuously to ensure the water is constantly agitated, but the primary function is to facilitate rapid mixing of salt. Consider that this is 2800GPH of flow in a 50 gallon tank. That’s over 50X turnover. The pumps are directed at the bottom of the reservoir to keep the salt suspended and avoid settling.

Note: Since taking this picture, I have found that by angling both powerheads inwards towards the wall of the barrel, I can create a circular water current similar to a whirlpool. This allows the flow to be cumulative instead of canceling. I've found that this is more effective at keeping unmixed salt in suspension.

I also used a paint mixer attached to a drill when adding salt to the mixing tank. It takes me about 10 minutes to make up a batch of saltwater. I would always try to wait 24 hours before using the new saltwater, but in a pinch it would probably be suitable for use after about an hour.



To allow the lid to close, the power cords had to be routed through a hole I drilled in the reservoir.

.

In these pictures, you can see the clear tubing on the outside of the reservoirs. These enable the water level to be observed without opening the reservoir. They are attached with threaded bulkheads and an L shaped MPT to barbed hose adapter. There is some Velcro that easily slides along the tubing that I use to mark the water levels. I’ve also taking the time to calculate and mark increments of 5 gallons on the tubes so that I can keep track of how much water I’m using.

Another detail in the above pictures is the platforms and leveling feet. Each reservoir has it’s own platform and is leveled independently. These are the heavy duty feet from Lee Valley that someone else pointed out in another thread (thank you). One of the requirements of installing all of this equipment is that it be removable if/when we want to replace the furnace or hot water tank. I won’t be fun to take it all apart, but it also won’t be difficult.



As I’m not a plumber, the only way I could hook up the sink was to the drain in the furnace room. This works well enough. At some point I may have a plumber come and hook it up to the main drain, but I have no plans for that yet.

Note: It was pointed out that the way my sink drains is actually against building code.



These taps (left) allow me to pump out either RO/DI or saltwater into the sink. More to the point, I can fill a bucket from here. Their positioning is adjustable by loosening the union - they are pushed out of the way when not in use.

[abcha0s]

This post is part 2 of 2 dealing with water changes. This post may be somewhat long, but for anyone setting up a similar system, hopefully it is interesting.

Continuous Water Change System

“Continuous water changes, despite their name, are not necessarily performed every minute of every day. The distinguishing feature of these changes is that water is added at the same time that it is removed. The actual rate of addition can be high or low. Reef aquarists (myself included) most often perform these types of water changes with two matched pumps, one that removes the old water and one that adds the new water.” - Randy Holmes Fraley

Water Changes in Large Tanks

I’m sure that almost everyone does water changes on their tanks, or at least wishes they could find the motivation. The number that I’ve heard most often is 30% monthly. The diligent hobbyist tends to achieve this by performing weekly changes of 7-10% of the total system volume. There is lots of literature that promotes this and most people generally accept it as a requirement.

Water changes in large tanks seem to be thought of slightly differently. The bigger the tank, the harder it is to keep up with 30% monthly water changes. It may be that the overall benefit to the system also changes. Many large public and commercial tanks have such efficient filtration that water changes are rarely required. Calcium and other trace elements can be supplemented and there are other more effective means of nutrient export.

A good example of this is Inland Aquatics who claim to have 40,000 gallons of water and change only 5% annually.

* Over time I may experiment with reducing the volume of water that is exchanged every day. With a new tank, I personally feel that water changes will be an important part of maintaining overall stability. Once the tank has matured, the cost may start to outweigh the benefits.

System Goals

Basically, I never want to do another water change again. There’s nothing stopping me from siphoning detritus or doing emergency water changes, but the week to week changes that are typically a never ending part of this hobby just don’t work for me.

Total System Water Volume = 300 Gallons (1,136 Liters)
Percent Water Change = 33% Monthly
Total Monthly Change = 100 Gallons (379 Liters)
Total Daily Change = 3.33 Gallons (12.62 Liters)

The system can run unattended for 60 days. Realistically, I will top up the reservoirs and ensure everything is calibrated once a month, but it’s nice to have extra time as needed.

• Simple is best
 
Water Storage

There has been some debate as to whether it is reasonable to store saltwater for extended periods of time. I am confident that if you have a clean vessel suitable for storing potable water and no contaminates are introduced that it can effectively be stored indefinitely.



The dimensions of this loaf tank are 58"Lx29"Wx37"H. It was made by paddleplastics - www.paddleplastics.com - I picked it up in Crossfield from Promould: 403-946-9920

The tank comes with a mainway and a bulkhead that can be installed at the time of purchase. I had Promould put the mainway on, but I put the bulkheads on myself. I installed the bulkheads on the bottom of the tank to allow maximum drainage.

Note: It’s somewhat important to mix the saltwater in a separate tank from the one used for ongoing storage. From observation, it can be seen that the mixing tank gets dirty from particles settling out of the newly mixed saltwater. By allowing this to happen in the mixing tank before transferring to the storage tank, very little contamination is transferred. My storage tank stays very clean whereas my mixing tank needs to be cleaned somewhat regularly.

Peristaltic Pumps and System Pressure

Even high quality peristaltic pumps are incredibly sensitive to changes in pressure at both the intake and the outtake. A reservoir slowly draining will cause a pressure variance at the intake of the pump as a factor of the water volume remaining in the reservoir. As the reservoir drains, pressure will decrease and the rate of flow through the pump will decrease.

To overcome this characteristic of peristaltic pumps, pressure within the system must be a constant. After considerable research into high tech solutions, I came up with a low tech solution that works perfectly. It's basically a constant pressure regulator that helps ensure the pumps stay balanced.



The water cooler can be purchased at Canadian Tire. Basically, the water enters at the top and fills the cooler to the point where the float valve shuts off the flow of water. As water is pumped out, new water from the main reservoir will replace it. There is a John Guest fiiting at the bottom of the water cooler that allows the peristaltic pump to connect at a low water point.

.

I know that you can get John Guest bulkheads, but they aren't really suitable here. The trouble with these bulkheads is they are difficult to tighten without access to both sides. The trick to installing the threaded adapter is make the hole using a drill bit slightly smaller than the size of the fitting such that it threads into the plastic. I used glue on the joint to ensure a permanent seal.

SpectraPure LiterMeter III Paristaltic Pump

I've tested a number of peristaltic pumps and the SpectraPure LiterMeter III system is, in my opion, by far the best. A good quality peristaltic pump is really important when setting up a continuous water change system or there will be a high probability of drift and/or premature system failure.

From the manufacturer’s description - "The new aquarium dosing pump is crafted from precision-machined aluminum fabricated to exacting aerospace tolerances. For example, such critical tolerances as the rollers are machined to +/-0.0005". The outer surface is polished to a mirror finish. Internal surfaces are held to a 63 micro inch finish and hard anodized to provide a durable and low-friction raceway. The pump's planetary direct drive is at an 11:1 ratio, thus providing enough torque to generate over 40 pounds of pressure. This incredible lift enables the LiterMeter III™ to pump over 60 feet above itself at a flow rate of over 250 ml/min. It can also draw up from 25 feet below. The motor in the LiterMeter III™ is made by a manufacturer of precision high reliability motors made to our specifications as a peristaltic drive motor. The pump is so reliable we now offer a five year limited warranty."

Some additional features that were important to me include:

• Calibrated by volume not time
• Remote pump support
• Programming will survive a power outage
• Ability to adjust dosing volumes on each pump independently by as little as 10ml/day without recalibrating
• Automatically tracks pump run time and will beep after 300 hours
• The desired daily volume is dosed in 150 equal parts throughout the course of one day

The System

The whole system really isn't complicated at all.



Black Tubing = Water from tank heading for the drain

On the tank side, I hooked up to the first chamber in my sump. This chamber has a constant water level and as such, the pressure will always remain the same.



The water is drawn up by Pump A (built into the controller) through about 10 feet of 1/4 inch polyethylene tubing. The total vertical rise is only about 3 feet. This is then pumped through another 30 feet of tubing through the ceiling and eventually to the drain.

Red Tubing = Water from reservoir heading for the tank

The reservoir has two bulkheads installed on the bottom. One is connected to a large Panworld pump (picture not shown) which is used to agitate the water once every 4 hours. It can also be used to pump water into the main tank for larger water changes.



The second bulkhead (shown above) has a John Guest fitting which connects to the pressure regulator water cooler. The floor of the main resevoir is elevated to approximately the same hight as the float valve installed in the water cooler thus allowing the main resevoir to completely drain.

This bulkhead also has an attachment for a drain line as can be seen by the cap at the end. The plumbing for the drain is stored detached to keep it out of the way, but can be easily reconnected as needed. The drain can be used in an emergency or for periodically cleaning the loaf tank.

. .

The remote LiterMeter pump pulls water from the pressure regulator water cooler and pushes it back to the tank through approximately 40 feet of tubing ran through the ceiling.

One interesting observation is that the run from the reservoir to the tank seems to have considerably more resistance. The calibrated flow rate through the LiterMeter pump B is about 80% of the calibrated flow rate through Pump A.

Heating the Reservoir

Heating the reservoir is not necessary. I keep a couple of spare heaters that could be used to heat the water in the reservoir if an emergency situation arose, but because of the very small volumes of water that are added to the main display tank, there is no measurable affect on temperature.

Tuning

The LiterMeter pumps are calibrated by volume. To accomplish this, the system asks you to fill a 500ml vessel and to stop the pump when complete. Once set, the system adjusts the calculated flow rate through each pump accordingly.

I found this difficult as the output of my tubing is nowhere near the controller. To overcome this challenge, I set the pumps to run continuously and timed how long it took to fill the 500ml. I repeated the test a number of times until I had an average that was within a reasonable margin of error. When I ran the calibration routine, I simply started a stop watch at the start of the test and stopped the pump at the appropriate time.

500ml - Pump A = 2.07s
500ml - Pump B = 2.47s

Once calibrated, you can then set the run time for each pump indendantly. Everything is done in metric. The initial configuration to achieve my desired rate of water exchange is:

Pump A: 12.62 liters per day.
Pump B: 12.62 liters per day.

The only reason we really care about precise calibration is to avoid any drift in salinity.

The SG of the saltwater in the storage reservoir is 1.025
The SG of the saltwater in the tank is 1.025

If the pumps are perfectly calibrated, then the tank will stay stable at 1.025.

• If there is a calibration error causing a slow increase in salinity, then the daily volume for pump A can be increased (or pump B decreased).
• If there is a calibration error causing a slow decrease in salinity, then the daily volume for pump B can be increased (or pump A decreased).

It's a little like guessing a number between 1 and 10 where the person who knows the number answers with "higher" or "lower".

The only other consideration is evaporation of tank water and the ATO. Ideally, this would be unaffected by the water change system as the volume of water in and out of the tank is balanced. However, if the margin of error is too high, the rate of evaporation could exceed the variance or vice versa. In any event, this is unlikely and the solution would be to recalibrate the pumps.

Safety

There are a couple of scenarios that warrant additional consideration.

Pump Failure or full blockage

These pumps are quiet. Without periodically checking, a failed pump could go unnoticed for weeks. The result would be a a proportionately rapid shift in salinity.

This scenario would not be immediately evident as the ATO would maintain a constant water level.

To mitigate this requires some diligence. Operation of the pumps should be checked at a minimum of every couple of days.

* I am considering installing flow sensors on the lines and connecting them to my Apex controller. I have one already that I am not using, but also haven't tested it yet. If I can trigger an alarm based on a period of inactivity, it may solve this problem.

Calibration drift or partial blockage

This would only be detectable through measuring salinity. Any unexpected change in salinity could likely be attributed to either calibration drift or a partial blockage of one of the lines.

• When the pumps are flagged for maintainence (every 300 hours), all of the lines should be cleaned to prevent salt buildup.
• Every second servicing, all of the lines should be replaced.
• Once a month, the time it takes to fill 500ml should be recorded. If there is a variance greater than 2 or 3 seconds then the lines should be cleaned and the pumps recalibrated.

Dry Reservoir

This one is pretty easy to deal with. The reservoir is very large so simply spot checking should be sufficient. However, I also plan to install a level sensor into the pressure regulating water cooler. If the water level drops for any reason, the Apex controller can alert me.

The system is expected to run for extended periods of time without requiring any adjustments. However, it does require some monitoring to ensure that all systems are opperating as expected.