Automatic Biomass Scrubber


New member
Let me first say, this is a project journal. My goal is the sharing of concepts and results. Second, I admit to being a home-built tech junky. This build will be complex and time consuming. It's not my intent to guide anyone else down this path.

As with most projects, this is built on the foundations laid by others. I have always been intrigued by the concept of natural biological filtration systems and I have been attracted to vodka dosing concepts since reading about it some years ago. More recently, I got into a couple of threads on Reef Central: Donovan's Nitrate Destroyer and Nutrient Pulse Reactor DIY DyMiCo system. Just as a recap, these previously mentioned threads utilize organic carbon / vodka dosing into a specially designed reactor.

Most of the working theories available on organic carbon dosing go something like this:
Vodka, sugar, vinegar or some form of organic carbon is dosed into the system to cause bacteria populations to grow exponentially. As bacteria increase exponentially in deep sand beds or inside the surface of live rock, they deplete oxygen and start to consume nitrate to complete their biological processes. The threads above differ from standard vodka dosing by providing an alternate and specific area for this low oxygen condition to occur.

I have been running several variations of the Nutrient Pulse Reactor on a couple of tanks. The difficulty, long term, is keeping the bacteria mat (slime) in check. When I got too aggressive with the organic carbon dosing, the system would eventually clog up and stop functioning properly. Reading Donovan's post, I saw some of the same issues as others tried to implement his system.

Outside of the difficult balancing act, there can be no doubt that these systems work. They will outperform any algae turf scrubber or refugium when it comes to nitrate and phosphate reduction. However, in order to achieve the desirable low nutrient targets, requires either more reactor volume than I have available or more organic carbon than could be dosed, long term. The best balance I could achieve was a system that cycled stably but never reached low nutrient status. Nitrate would hover around 4.0 ppm and phosphate at around .016. When I tried to drive them lower by increasing the carbon dose, the system would clog within weeks.

Consistent / uniform dosing proved to be an issue as well. Every low volume peristaltic pump I tried, even homemade units, didn't perform consistently at anything less than 1 full revolution of the pump. There were always dead spots. This meant that I was diluting the organic carbon so that I could get at least a full revolution on the pump. The catch there is that when you dilute vodka too much, eventually the bacteria creeps into the dilution and consumes the carbon source. This changes the potency, again affecting consistent operation. I even tried vinegar since it is much less potent than vodka but still had to dilute the mix and eventually ended up with bacteria slime and reduced potency in the carbon source.

On one such occurrence, I noticed the system was floundering (not cycling normally). I found slime in the bottom of my carbon source. I cleaned the bottle and mixed up a new batch. Then, I turned on the pump to prime the system and got side tracked. When I finally noticed, I had dumped almost a week's worth of carbon into the system. It took about half a day to clear the reactor of the massive dose. The tank suffered no apparent ill effects, other than a little cloudiness but the nitrate went from a consistent 4.0 ppm to 0.25. Phosphate went all the way down to .01 to .02 ppm, from 0.16. All this happened in less than 12 hours. I turned off the carbon dosing and let the system self flush for a few days before starting things back up. My point is to show the potential for affecting these nutrient levels with this method but the following reading, two days later, was also interesting. With the system running a flush cycle every few minutes and no additional carbon going in; the next reading showed. nitrate at 0.5 ppm but phosphate had rebounded to 0.1. This leads me to believe two things.

1. Nitrate is being exported from the system, probably as the theory suggests, as nitrogen gas.
2. Phosphate is not being exported but is being bound up within the increased bacterial population.

If phosphate were actually exported it would build back more slowly when the system was inactive. Since the nitrate was exported from the water, it was building back slowly, as food was added and cycled within the system but phosphate was quickly released back into the water as the bacteria population changed suddenly when I stopped dosing carbon. In fact, I ran a crude test on some of the bacterial slime, adding about 1 to 2 ml to RO water and using the same Red Sea tests and found high levels of both nitrate and phosphate, ten to twenty times higher than the tank water. I also believe that this also may be tied to the theory that live rock can absorb or leach phosphate. Possibly this phenomenon is really related to the bacterial populations within the rock changing. One more point to add is that most vodka dosing regiments require a good skimmer for final nutrient export. The exception is the DiMiCo system, although they do recommend GFO.

So, I decided to try the third upgrade of my homemade sump to test the feasibility of periodically removing the bio-mat and actually export phosphate while driving the system to it's full potential. I'm calling it the Automatic Biomass Scrubber or ABS for short.

First a little historical perspective...

My first sump was a homemade dual refugium with a central protein skimmer. It was 43 X 23 X 19 (external dimensions) and was my first real attempt at sump construction. Each side had a 2" sand bed, planted with caulerpa prolifera. Both sides wound around a baffle to the skimmer chamber and then to the dual return pumps.

The main mistake I made with this system was to place filter sponges in such a way that there was no way for water to flow around or over them when they became clogged. This drove water levels up in the proceeding chambers and required constant attention.

About that time the Roller Mat came out on the US market. I was really ready to get rid of the need to constantly attend to sponges and filter pads and this seemed like the way to go. Of course, I didn't want their prefabricated version, because I thought, yeah, I can do that and I can incorporate it into the sump the way I want it.

Again, I had the outer "œbox" of the sump built by a local plastics company due to the size and concern over leaks. I always do this. I am just not set up, with the tools and jigs to handle the long seams without leak concerns. I had the new box built with ¼ in plexiglass, again keeping with relatively the same footprint (43 X 23). I went an inch shorter on the depth this time, to only 18 inches deep. I decided, with all the internal baffles and bracing, I did not need to have anything thicker than ¼. This sump has been up and running for a couple of years now without issue. The primary method of nutrient export was a sand bed in four stacked trays under an algae turf scrubber. I also have a Reef Octopus protein skimmer. The continuous felt changer was between the ATS and skimmer.

This system ran for about a year prior to my interest in the system described in the Nutrient Pulse Reactor post.

This new ABS system will be composed of 4 primary components:

Remote Deep Sand Bed
Bio Reactor with motorized agitation
Rotating Chaeto Reactor / Refugium
Continuous Felt Changer

The first thing to keep in mind is: this ABS side is not operated as a continuous-flow system. This means water does not flow through this section continuously. It is a batching system. Water is pumped into the system, carbon is dosed and then a wait period allows for the reaction to occur. ORP is used to determine when the process is complete and then another batch is processed. This was the primary reason for combining the chaeto refugium; to have a parallel section with continuous flow. The chaeto will also hopefully increase the pH as well.

I'm trying to reuse as many pre-existing components as possible. For example, I removed the baffles and structures from my original (ver 1) 43x23x19 sump and will build the new components into this box. Also, the Remote Deep Sand Bed will be constructed from a 24x24x6 box, from a previous project. This box is vertical so it is 24 inches deep and long but only 6 inches wide. It will fit above the sump, between the back of the tank and the wall.

Locating the DSB away from the Bio Reactor solves two problems. First, it provides for fluid separation between the sand bed and Bio Reactor and it's carbon source. I had issues with other versions of this system due to the close proximity of the two. Organic carbon would migrate into the sand bed and the resulting bio mat would quickly clog it. Secondly, locating the DSB away from the Bio Reactor gave me more reactor space and allows me to reuse these existing boxes. The sand bed will also need to be layered with a plenum space between the three sand layers. I added a bottom panel, resting on two 3/8 inch supports with ¼ inch holes and a baffle on one end. A shower-head cover will diffuse the force of each cycle, preventing the sand from being disturbed. I've used this concept before to diffuse the water into a compartment. It works well.

Each cycle or batch of water will then be directed down through each layer, exiting through the bottom panel and out the side, on it's way to the bio-reactor. The concept here is to provide a place for nitrification to take place prior to being pushed to the bio-reactor for de-nitrification. Then on a weekly basis, I will turn on the rotating arms in the bio reactor and flush out any loosely attached bacterial slime into the CFC for removal. The CFC will also capture and remove and loose pieces of cheato.


  • IMG_0091.jpg
    54.1 KB · Views: 0
  • Version 1.jpg
    Version 1.jpg
    30.2 KB · Views: 0
  • Version 2.jpg
    Version 2.jpg
    33.6 KB · Views: 0
  • Version 3 Concept.jpg
    Version 3 Concept.jpg
    41.3 KB · Views: 0
  • Deep Sand Bed (remote).png
    Deep Sand Bed (remote).png
    61.7 KB · Views: 0


New member
Interesting!. I am still using my Donovan's nitrate destroyer for full 2 years now without any problem. Yes, clogging is an issue when nitrate is abundant, but easily dealt with less vodka. I am no longer emphasizing pumice stones usage, that will solve clogging issue permanently (non issue to me as I purposely clogged my reactor once in a while to provoke bacteria population).

I am on a complete nutrients recycling for several months (no skimmer) now. I am using cryptic zones, live oysters and activated carbon in my sump. Other filter feeders and sponges is acting as filter socks.

I haven't test NO3/PO4 in months but judging from my tank I think it is within range.

Definitely following your build, there is a lot of things to learn. Can't wait!


Active member
Seems way too complicated. On my 300 I have 2 filter socks I clean weakly, a modified ASM G4 skimmer, and a Refugium with macro algae. When I started getting some GHA I added A simple alge turf scrubber with a little vinegar and iron dosing. It was all I needed to drop my nitrate and phosphate to almost zero. I stoped dosing vinegar dosing a few months back so nitrate would climb a little. I only dose 5 ml of iron once a week or so now. I also cut back on the lighting on my ATS to 8 hours a day from 24 originally. For my ATS I just put a piece of roughened mesh in the overflow and lit it with a Chinese red & blue led grow light. I have never had my nitrates and phosphates as low in any other tank I have currently.
I have 25 fish including 4 large fish( my sail fin tank is dinner plate sized). I feed heavily, and do not have any other reactors.


New member


I made good progress. I cleaned out the old sump body (cleaner than it was) and removed the remaining old baffles and compartments. I had taken most out earlier, so there were only a few left. Then, I started cutting and assembling the parts for the bioreactor area.

I am using a clam-shell design that I have used before. I will cut 8 ribs (4 on each half) that will support 1/16 inch polycarbonate shell with 3/8 inch holes. I've used this method before and it works well; providing a consistent round curve that is very sturdy after welding. Counting back, I cut and installed over 50 individual plastic pieces to build this section. It can be slow at times. Since I don't have detailed drawings, just a concept in my mind, I periodically I must stop and try to visualize how everything must be assembled. I'm trying not to make mistakes I can't recover from. This means a lot of dry-fitting and pre-assembly prior to welding.

This compartment, which consumes 16 inches of the right side of the box; contains an 18 X 14 cylinder, in two halves, which should hold about 12 gallons of bio-balls. I have installed a center distribution pipe that also has 3 arms (spaced at even intervals) to stir the bio-balls. A syncra-silent 1.5 pump will provide circulation within the reactor. This is necessary to ensure the carbon dose is distributed evenly. I made the compartment big enough for a 2.0 if the flow provided by the 1.5 isn't sufficient. I also had to make sure I could access and service the pump. The plan is to have a motor with a pulley drive the stirring system for about 15 minutes, once a week. This should break up the bacteria bio mat, which will then be collected and removed by the CFC (continuous felt changer). The big unknown here is how much force will be needed to stir, without getting stuck or breaking the agitator. I plan on using round bioballs that float. I'm hoping this will minimize the torque required. I'm using belts for two reasons. I have access to some and I needed a way to drive a system that will be underwater. Since I cannot get the motor wet, this seemed like the best solution. If I had the ability to make a large spur gear system, that would probably have been my preference but I don't have a 3D printer and I have never seen plastic gears, as large as I would need, for sale.

I also completed the DSB by patching up the old outlet area and adding the main baffle and bottom panel. The box was 6 inches (ID) constructed from 3/8 acrylic. I drilled ¼ holes in the bottom panel and installed a new 1 ¼ inch outlet on the side. I also built the showerhead that will diffuse the water going into the DSB. I also added a reinforcement brace in the middle to help with the pressure from the sand and water.

Pictures attached below:
Ribs, shaped by router.
Bottom clam-shell with top ribs dry fit.
Both clam-shells assembled with distribution pipe and arms dry assembled.
View of the bottom 1/16 shell dry fit.
DSB with plenum separators.

At this point, I will probably stop and complete this bio-reactor section later. I want to at least, dry-assemble all the components first. This allows me to confirm the entire layout and gives me the ability to make some modifications, if necessary. Seeing the rest of the components in position usually give me a better perspective. I will work on the paddle wheel for the CFC next.


  • IMG_0513.jpg
    52.9 KB · Views: 0
  • IMG_0516.jpg
    45.4 KB · Views: 0
  • IMG_0519.jpg
    49.5 KB · Views: 0
  • IMG_0531.jpg
    31.2 KB · Views: 0
  • IMG_0550.jpg
    48.5 KB · Views: 0


New member
Thanks for the input / update and your original post.

It appears one of the major differences between the system you build and the Nutrient Pulse, is the batching vs continuous low flow used by your system.

The biomass buildup definitely impacts the batching systems more. Although the flow through the biofilm decreases as the population increases, it doesn't seem to completely clog. I've read several articles about biofilm and the bacteria seem to work together to keep paths open for water flow. It's just not enough flow for these batching systems, I guess.

A couple of questions:

1 Are you still doing water changes? If so how often? I was wondering, with the skimmer off, how phosphate was being exported?
2 You mentioned that clogging is an issue when nitrate is abundant and you dealt with that by reducing the carbon source. Given that the amount of carbon seems to be linked to the reduction capability of these systems, how did you deal with the increased nitrate?

You are right; it is way too complicated. "¦glad your system works so well for you.


New member
As mentioned above, nutrient recycling and conversion to bio mass reduces testable no3/po4. I do have a 2' algae through made from cut out 4" diameter PVC pipe to grow some algae. I don't remove the algae, I fed my tangs (except dragon breath which they don't seems interested at). No water changes since I started 4 years ago.

My daily dosing is more than enough to keep nitrate below 5ppm. Two doses, between 2ml and 5ml depending on visible flow coming out from the reactor. I purposely doing overdosing once in while, everytime I did this the flow will slow down significantly and will strip nitrate to zero. I never experienced nitrate build up, except on several occassion where I added my chromis (7 of them) and anthiases (killed by my dragon wrasse) which I overfed the tank. I remember at one time nitrate shoots up to 20ppm, but quickly brought down below 5ppm with additional vodka dosing for few days.

My reactor is quite dynamic, once you know how it works and how well it react to changes you made, fine tuning it's effectiveness is very simple.


New member


I completed the paddle wheel for the CFC. This component is probably the most time consuming and tedious piece to construct. I made this one clear so I could see into it better. I had a snail get stuck in this section on the version two sump so this should allow me to see better if that happens again.

The two ¼ X 13 inch disks are connected by (41) 5.75 X ¾ inch pieces set one inch apart. This forms the 6.25 inch wide support for the Roller Mat felt. Two 1/8 X 13.25 disks are attached on the outside as guides and to seal things up. This should give me over 200 square inches of filter surface area at full contact. Running depth in this compartment will be about 13 inches with the wheel about a half-inch off the bottom.

I also removed a lot of material in the ¼ sides by drilling holes, to reduce the weight. This was a lot of extra work and I'm not sure I would do it that way again. In the end, it did reduce the weight by a pound, overall. This wheel is 13.25 inches in diameter and 6.5 inches wide. It weighs in at 3.55 lbs. It is constructed of almost 50 pieces of acrylic and polycarbonate. There were over 35 holes drilled as well.


  • IMG_0530.jpg
    46.6 KB · Views: 0


New member


After completing the wheel for the CFC, I decided to change the original design / configuration. This new layout (turned 90 deg) allows me to place all of the systems in a better position to be visually monitored. I had to recut one of the outer 13.25 disks. I originally cut it in an opaque white acrylic but after I changed the configuration, I made it entirely clear. Luckily, I hadn't welded the outer disks. I am also thinking of relocating the process pump into the upper half of the bio-reactor. This would allow for easier access to the process pump and would simplify the design by eliminating the access door but I've not fully committed to the concept yet.

I found a concerning crack near the bond line in the front right of the sump. I have made repairs but I'm still concerned. I will definitely do a prolonged water test before installing it. Maybe it wasn't the best idea to use the old sump box, given the critical nature of a failure.

The change in design cost me some volume on the cheato reactor (about an inch in width but consolidating the flow path in front, where it can be visually monitored, will be better in the long run. This also places all of the plumbing out of the way, in the back of the sump.

The chaeto reactor will have four 15 watt pucks. I purchased 90 mm aluminum disks / circuit boards on eBay. They work pretty well. I use eight 660 nm (red) with five 460 nm (blue) and two 10K white LEDs per puck. They are powered by two 650 mw constant current power supplies (one for two pucks in parallel). I have used this configuration before and it seems to mimic most of the LED grow bulbs on the market which have strong peaks in the blue and red spectrum.

The LEDs must be force-air cooled, which means a small fan. Without it they heat up to almost 150 deg F. Anything above 95 and the LEDs start to lose their intensity. The losses can become significant at the higher temps. It only requires a little air movement to keep them relatively cool.

I prefer to have the chaeto tumble constantly. I have used several methods to ensure this, from pumps to channeling the in-flow into a circular motion with vanes and baffles. The pumps work really well but requires some attention. This build will utilize a guide vane that will channel the water from the overflow down and under the chaeto. At the end of the refugium, water will be redirected up and across the surface by an air-stone and vane. Hopefully this will provide enough circular motion to keep the ball of chaeto tumbling and also provide some additional gas exchange to increase the pH.

This will be the last post until I'm through the holidays. Cheers!


  • CFC Rear View.jpg
    CFC Rear View.jpg
    46.8 KB · Views: 0


New member
I finally moved back to the drive system for the Bio-Reactor's agitation arm. I guess I've been avoiding spending the time on it because of the tedium but it is a very critical component and I need to see if it will work as designed. It's also very cold, which means I have less desire to spend time in the garage.

I am using ribbed Kevlar belts from the food processing industry. One longer belt (the wet belt) will be used to drive the main pulley, attached to the arm. This belt will be mostly under water all the time. I am using a second "œdry" belt to separate the moisture from the motor and also to extend the motor above the sump. This will help prevent the motor from getting salt water in it, regardless of how high the water level gets. Hopefully, the use of two belts will also prevent moisture from getting on the drive shaft, which is metal. I constructed the box for the motor, which is a 12 v 6 rpm geared setup. The shaft is the only thing that is exposed and most of it is in the first pully. Again, I got it off eBay. They claimed it was a "œhigh torque" 80KG / cm motor. It seems to live up to that billing. It uses about ½ an amp, no load but jumps up to 2 amps with me trying to stop it. It seems like it will do the job.

The main drive pulley is cut from two 3/8 plexi disks welded together. The wet belt is just over 5/8 inches wide. I cut two 3/8 inch disks with a 4" hole saw and welded them together. I then cut groves or slots about 1/8 inch deep into the welded disk with a table saw. These groves will line up with the ribs on the belts. This is very tedious: Turn the saw on, make a pass over the blade, turn the disk slightly and repeat. I then added 1/16" poly-carbonate disks on the outside as guides. This process was repeated for all three pulleys. I cut 16 circles out to make all of them and that was just counting the final product. I also had several mistakes and remakes. It took a long time.

The final reduction was only 3 to 1 instead of 6 to 1 so my arms turn at 2 RPM and not the 1 RPM I had planned for. I just didn't feel comfortable cutting the small pulleys any smaller and I didn't have enough belt to make the big ones larger and keep the motor out of the water. It's still pretty slow and the motor seems to have plenty of torque. I'm going to fill the reactor with dry bio-balls and test it out next. I am also thinking I need a tensioner on the larger wet belt.


  • motor&box.jpg
    48.6 KB · Views: 0
  • Belt& Pulley.jpg
    Belt& Pulley.jpg
    67.6 KB · Views: 0


New member


I ran my first dry test with 800 bioballs and the results were less than spectacular. Things moved OK at first and then began to tighten up after a few minutes. I opened things back up and found that the balls seemed to be "œbonding" to each other, like atoms forming a molecule. First they would pair up but when they locked the third ball into their "œmolecule" things really got tight. They became very dense and difficult to drive around the system. The belts started slipping and the motor labored. The bad thing is that the 800 bioballs barely filled about 1/3 of the cylinder. That means I will need about 2400 to completely fill it. I'm going to move on to the water test in the hope that the load will dramatically decrease once they start to float.


  • bioballmol.jpg
    27 KB · Views: 0


New member
Water Test

Water Test

For the water test, I started out with just enough water to float the balls. Everything moved without issue or excessive load. I then added 800 more balls and put in enough water to allow them to float slightly. Again, no excessive load was perceived. Although the balls still clumped together after some time, they didn't stick as tightly as they had when I ran the dry test. It appears that the more force applied to moving them translates into a tighter more compact molecule. I guess it's sort of a run-away reaction or action. I have a short video but I'm not sure how to link that in. The attachments seem to be limited to pictures, probably due to size.

I also don't have any more balls to test with. The other 800 are being used in one of the other test systems. When the time comes to install the new filter, I will removed and rinse those and add them to the other 1600 to complete the bioreactor. Hopefully this will give me a head start on growing the biofilm throughout the reactor.

I've thought about modifying the ball structure to prevent them from bonding to each other. I think I can do this by melting the open ends just a little bit by touching them to something hot but that would require a lot of effort. "“two sides times 2400 balls, ugh.

I also took the first water sample and sent it off to Triton. I'm fairly confident that I can remove the nitrate without water changes. My hope is to also remove the phosphate without the need for a GFO or Lanthanum Chloride reactor. Also, I hope to understand what trace elements are removed by running the ABS vs just the chaeto system over the next year or so. I'm also hoping that there will be increased biodiversity and plankton food sources will be present. This is one of the claimed benefits of the DyMiCo system.

I need to work on the bubble diffusers and some of the other plumbing and flow components. I also need to complete the drive system for the CFC. I have been going back and forth between a direct drive system, another belt driven system and a gear driven system that I have used in the past.

The gear driven system would be easiest. It's a "œknown animal" but I'm out of large nylon gears so I would need to reuse ones from one of my previous systems or find more gears. I'm really moving away from the direct drive system because I'm thinking that if something gets stuck, as it occasionally does, breakage will occur. Right now with the belt drive or the gears, there is enough play to allow them to slip rather than break so I'm focused in that direction.


New member


Well"¦ I'm back to working on the ABS again. I had been doing some travel for work and was having difficulty finding time (and motivation) to work on the project. I have finished the lights on the refugium as well as the plumbing, bubble diffusers and vanes for the tumbler.

I also started working on the CFC again. I just finished the take-up spool (pictured). I decided to go with the belt drive in the end. The spool spins on two 8M nylon bolts. The previous two had gear drives so I'm curious to see if this is any better. Again, I built the pulleys for the ribbed belts on the table saw. The feed spool will be located higher than the take-up spool. I didn't have enough room to make them level and I sort of wanted the feed spool above the potential high water line. If the felt roll gets wet on one side, they become really heavy on that side and cause feed / take-up issues.

I also got the results back from the Triton test. Everything was good except the iron was crazy high at 6.2 ug/l. I am not sure what caused this. I was dosing some iron, along with other trace elements (Red Sea). I of course stopped the iron dose when I saw the results. I am not convinced the Red Sea was the source. I am concerned that it might be related to the mud I have in the stacked trays so I am hoping to get the new system finished so I can pull out the old. By the way, the Red Sea iron test was showing zero. So much for that test.


  • IMG_0614.jpg
    56 KB · Views: 0
  • IMG_0615.jpg
    48.7 KB · Views: 0


New member
CFC is running

CFC is running

I finished and tested the belt drive system on the CFC. The take-up motor is a 12 v 2 RPM (reduced to 1 RPM by belts). It doesn't generate a lot of torque but it seems to be enough. Currently the felt moves at a snails pace (under .1 in / sec) but as the roll moves to the take-up spool the speed will increase to just under .5 in per sec as it nears the end of the roll. This is better (slower) than I was able to achieve with the gear drive but I wish I had gone with the gears. With that system, I had the motor mounted on the lid. When installed, it sat on the take-up spool gear and the weight of the lid kept things engaged. I did have to add some weight and near the end of the roll, there was some slippage as the speed increased. This produced a loud thumping sound as the lid went up and down, when the gears slipped. This did alert me that the roll needed to be changed but it always annoyed me. The good thing about the gears option was, it was much easier to change the rolls. I just took the lid off and lifted the spools from the slots. Now, with the belt drive, I will need to removed the motor or dis-engage the belt in some way, to get the felt off the spool.

I added three floats on the bioreactor input side, at different levels. I anticipate needing to trigger actions based on water levels, such as stirring the bio-balls. As water is pumped into this area, by the process pumps as well as the flush pumps these floats will provide feedback on how clogged the reactor has become.

I also added a snail guard near the take-up spool. I am hoping this will knock off the young trochus snails before they get pulled out of the water with the felt. With the version two sump, most of these would get wrapped up in the take-up side before i noticed and would die.


  • IMG_0619.jpg
    55.8 KB · Views: 0
  • IMG_0616.jpg
    63.9 KB · Views: 0
  • IMG_0617.jpg
    56 KB · Views: 0


New member
As built

As built

So this is pretty much the finished product. i still have some small adjustments and things to clean up but I will probably start to find a sturdy, level place to run the water test.


  • screenshot_101.jpg
    43.1 KB · Views: 0
  • IMG_0620.jpg
    53.6 KB · Views: 0


New member
Testing and fine-tuning

Testing and fine-tuning

I have been doing extensive water tests on the system over the past week, with salt water. I've been transferring water from each side at higher than normal running levels to stress test the seams. Everything looks good, including my repairs.

The lights on the refugium look good and I've tested the tumbling effect with some chaeto.

I also ran the bioreactor stir, extensively and without issues. I still don't have it completely filled (only 1600 balls out of the full volume of 2400) but everything looks good. The water causes the balls to float ,which reduces the necessary torque. I added a tensioner pulley for the wet belt. Also, there was no moisture transfer between the wet and dry belts during the tests so I am pleased that result as well.

I fine tuned the flow of the process and flush pumps and adjusted the float switch heights. I also removed one of the three floats. I had a ¾ inch fitting on the process pump that went straight into the inlet side of the reactor, so it was running at zero head. The water level was about 2 ½ inches higher than the reactor compartment so I went to a 5/8 fitting and that dropped it down to about 2 inches above. I then, went on down to ½ inch and it looks better at about 1 inch above the water level in the bioreactor. I set the first float switch height to trigger at 2" so when the biofilm builds up to the point where the water goes from 1 to 2 inches, it should trigger a stir and flush cycle. I'm guessing the Syncra Silent 1.5 process pump will be running around 200 to 250 gph with this flow reduction. That should turn the entire chamber over about 10 times per hour.

I also adjusted the flush cycle, reducing the Syncra 2.0 flush pump, from ¾ inches to 5/8 inches. This gives me about 3 inches difference between chambers and the top float, which will stop the pump if it reaches 5 inches. I am figuring this new flow rate will be in the 400 gph range. I will also trigger a reduction in speed of the Waveline variable speed return pump to compensate for the transfer of water into the bioreactor, during the flush cycle. This will prevent the water in the return chamber from dropping too low and causing the pumps to suck air.

I also added a cleanout port in the back of the bioreactor. This is my biggest regret, not considering this during construction. At least I did catch it before I installed it. The cleanout port allows me to run a round brush, on the end of a flexible shaft, into the distribution / stirring pipe for cleanout. I assume that biofilm will eventually build up inside the 1 inch pipe and will need to be cleaned. I can see into it from the front so that is one positive takeaway. Hopefully this doesn't happen too often but it was one of the maintenance steps called for in the DyMiCo system, which some of this is based on. If I had 20/20 hindsight, I would have put this port up front for easier access and actually incorporated it into the sump, rather than having to go through the outer 3/8 inch wall. The second regret was that I didn't use a larger size for the distribution / stirring pipe.

I also wish I had added a second pump at the bottom of the bioreactor that would have allowed me to pump it down completely. This would have been a, "œnice to have" for cleanout and maintenance but also from an experimental perspective, I would like to have been able to completely empty the chamber once the ORP reached it's target level, before adding the next batch of water, just to see what the results were.

I plan to pull out the old (version 2) system and install the new one this weekend. I will update once complete.


  • IMG_0637.jpg
    47.9 KB · Views: 0
  • IMG_0638.jpg
    37.6 KB · Views: 0


New member


Part of the install process requires me to remove the smaller pulse nutrient system I had running on this tank. I need the other bioballs that are in this system and the area I am installing the remote deep sand bed is in this location. So, after emptying that system and letting it drain overnight... this is the biofilm that is left. As you can see, it's pretty messy looking. Sort of an orangish brown color. I would describe it's consistency as that of pudding. It's not strongly attached to anything; not sticky, just thick. It doesn't have a strong odor. It smells very similar to live sand. It's thicker near the top, bottom and sides of the tank. I'd guess that there are several pounds of it. I have some better pictures but for some reason, they won't upload.


  • IMG_0644.jpg
    51.8 KB · Views: 0


New member
Up and Running

Up and Running

I got the sumps switched out this weekend as planned. I still have a few minor details to work out and had to work through a few glitches but so far everything seems to be working.

I started the Bio Reactor this morning after priming the system for a day. Attached is the first screen shot from the Apex.

The system has automatically run 15 batches of water through the remote DSB and then into the reactor, since 6:00 this morning when I started it. I am guessing a batch is about 4 to 5 gal of water.

A couple of notes on the Apex...

For some reason, they don't read ORP as a negative value. This means I have to tell the Apex that a pH probe is attached to it and then convert the pH reading to mV. For reference, one point on the pH scale is about 53 mV. A pH of 7,0 corresponds to 0 mV. So for the readings you see below, the a pH reading of 8.5 is roughly -80 mV. This is my trigger point for a new batch of water. This means the graph is also upside down. High is -mV and lower is positive mV or ORP.

Also, the Apex Fusion only records data points every 10 minutes and automatically smooths the curve to those snapshots, so it's not always an accurate representation of what is really happening.

At this time, the biomass is not impeding flow through the system so no stir events have been triggered. It may take a few weeks to build up enough biofilm to start having an effect on flow through the reactor.


  • screenshot_102.jpg
    38.6 KB · Views: 0


New member


So, 24 hours after I started the system yesterday and the unit is now batching better than every half hour. You can see 15 batches over the last 6 hours. There are a couple of short batches. These are due to the limitations of the Apex. In this case, I am using an oscillate command that is triggered by the ORP reading. If the ORP doesn't reach the trigger point until midway through the cycle, I get a short one. Again the graph doesn't really show what is actually happening. The Apex updates every few seconds and is quickly responding to the ORP values but since they only log a snapshot reading on their ASP every 10 minutes and then smooth the curve, based on that reading, the graph starts to look strange. In reality, the system is topping out at around 50 mV with each batch and then bottoming out at around -80 to -120 mV before another batch is processed.

More importantly, I tested nitrate and phosphate late Monday night, just prior to starting the system. I found Nitrate up around 4.0, maybe slightly higher as the Red Sea test tops out there. The Phosphate had crept up to 0.12. This morning, after just 24 hours, nitrate came in at 0.25 ppm and phosphate at 0.03. I am guessing at the 0.03 a little, as the color looked to be between the 0.02 and 0.04 readings on the Red Sea phosphate test. Assuming I can maintain this rate of reduction, I have massively overbuilt / over-sized the reactor for a 210 gallon tank.

The question now will be how long to wait for a stir. I originally planned to do a timed stir, every few days or weeks but then I added the floats and thought maybe I'd trigger off how quickly the biofilm was building up in (clogging) the reactor. I guess I will try both and see how things look.


  • screenshot_103.jpg
    63.6 KB · Views: 0


New member
Sunday update

Sunday update

On Thursday I noticed the water on the inlet side had increased about 1.5 inches above the level it was before I started the system up. This restriction came much sooner than I anticipated. As a result, I decreased the dose substantially to about 25% of what I had been dosing. I also triggered a manual flush & stir, which did little to reduce the water level. I then added a stir during each batch and cut out all carbon doses except for one that occurs after each batch has settled in the bioreactor. This reduced the level that the ORP was topping out after each batch by about 50 mV. I assume because I was dosing sooner and more consistently after the batch. The water level hasn't increased since Thursday. If it holds where it is, I should be OK.

I also set up an automatic flush and stir to occur each day at 2:00 in the afternoon. The reason I selected this time was that it typically has the highest pH of the day and the flush tends to drive pH down because the bioreactor pH runs around 7.5.

I had to remove about a gallon of bioballs. They were so tightly packed that entire group moved with the arms, as one mass, in the cylinder. There was very little mixing or agitation occurring. With some removed, the space allowed the majority of them to float and then the arms could move through them instead of with them. This also reduced the torque.

On Sunday, I monitored the ORP readings every 30 seconds, by hand; to get a better picture than the Apex Fusion is providing. The result seems to indicate that the de-nitrification is occurring at a much lower ORP then I was achieving so I lowered the target to around -130 mV. I will let the system cycle for a few days with these new parameters and let things stabilize before making any more changes. It takes about two hours after the flush cycle before the bioreactor recovers back to the trigger point of -130 mV. At this point, I am not sure if there is anything I can do to make it recycle faster. I have a few things I can try. First, I may use a larger dose of carbon at the end of the flush cycle in order to get the biofilm active and growing again.

The first batch after the flush reacts very differently than the others. The ORP doesn't rise above zero and quickly recovers to the trigger point in short order, triggering a second batch within about 10 to 15 minutes. Things eventually work their way back to about 30 minutes per batch / cycle.

During the flush cycles I can see bits of biomatt being dislodged; washed out by the flow and gathered up by the CFC. Also, the CFC is running about twice as much felt as the previous system; about an inch an hour. This will cut my roll life down to about 70 to 75 days. This is about half the roll life I was getting prior to switching out the Version 2 for the V3 sump.

One other noteworthy mentionable is that I was seeing a lot of foam generated by the bubble diffusers. The overflow / air escape was located toward the return pump section so there was no leakage but there was a layer of foam there and the overflow hole would burp out more every few seconds. This decreased when I reduced the carbon dosing. I supposed if I wanted, I could cut a new hole on the opposite side and collect / export the foam like a built-in skimmer.