Nutrient Pulse Reactor (DIY DyMiCo filter)

[dartier]

I have set out to build a DIY version of the DyMiCo filter. This thread is for documenting the process. Since my filter will not be an actual DyMiCo, I have decided to call it a Nutrient Pulse Reactor, NPR for short.

For the people who have never heard of a DyMiCo filter, it is a type of filtration system from a company in Europe called Eco Deco. They make large industrial size units for public aquariums and are now bringing out hobbyist targeted models. At this point they are not available outside of Europe. What makes them unique is that they are intended to be the main filtration for a tank, with no protein skimmer and no calcium reactor. They also allow one to forego the need for water changes, since the filter processes the excess nutrients and creates a plankton rich environment.

At this point I should probably mention that I have never actually seen a DyMiCo filter, nor have any knowledge of the their control methodology. I only have what I have been able to determine by reading online articles, watching videos online, reading their patent, and studying the user guide from their hobbyist filter.

Ok, enough about them, lets talk about how I am going to approach this. The tank this is going to be going on will be a new 300 gallon deep dimension aquarium. I will be using a 75 gallon tank (48x18x20) as the filter. The filter will be divided into 3 chambers using glass baffles. The first chamber will be 6 inches in length and will contain the mixing pump, CO2 feed, ORP probe and PH probe. The next chamber will be 36 inches in length and will contain the 2 types of media (sand and coarse media) separated by a perforated PVC plate. The final chamber will be 6 inches in size and will contain the return pump. The sizes of the chambers may change, once I have all the equipment on hand. The glass baffles have been ordered and should be ready next week. Each baffle will have 5 holes drilled into it for 5 bulkheads. 2 of the bulkheads will connect to the mixing pump, and the other 3 will be the return from the anaerobic chamber (holding the coarse media) and also the feed for the return pump.

To act as a controller and to provide the control signals I will be using a Reef Angel. I will also be using a Raspberry Pi to acts a higher level controller for the Reef Angel. I expect to need to do a lot of data logging to fine tune the operation of the system, so the RPI interfaced to the RA will be perfect for this. This will also allow me to test control methodologies in Python before I create the Wiring version for the Reef Angel. Even after the Wiring version is ready, I will probably keep the RPI to act as the interface for the system, perhaps with a webpage for monitoring operation and the RA will be the failsafe for the RPI should the RPI go down. So the reactor will be under the control of the RPI and the RA will take over should the RPI fail to act.

For sensors, I will be using 1 ORP probe and 2 PH probes. The ORP probe will be for monitoring the anaerobic process and will be located in the mixing chamber. 1 PH probe will be in the mixing chamber as well and will be for controlling the PH of the anaerobic chamber for alkalinity and calcium replenishment. The other PH probe will be located in the sand bed (aerobic section). This probe will be for monitoring the nitrification process and hopefully will be useful in determining the nutrient load. I should note, the DyMiCo version does not use a PH probe in the sand bed. They control their filter with a single ORP probe. I want to explore the possibility of using the PH of the sand bed to detect the end of the nitrification of incoming ammonia.

For a carbon source, I will be using my version of the DIY NoPox from the thread on here. For a carbon pump, I will be using a peristaltic pump.

Once I have the baffles back from the glass shop, I will start to post pictures.

Dennis

 

[mr9iron]

Sounds like an interesting idea. I can't wait to see it in action.

 

[ravi197699]

Very Interesting.. Tagging along..

 

[dartier]

Hmm, lets see if I can post photos. It has been a challenge ever since they discontinued Picasa Web.

That seems to work. I noticed that the DyMiCo manual stresses the need to make sure that the return pump is set to flow 700 - 1000 L/H. I expect this is because their software needs to be able to know how much inflow is entering the filter. The water flowing in will be in direct proportion to the return pump pushing water back to the tank. By measuring and adjusting the flow of the return pump, and then operating the return pump for a fixed period of time, the volume of water being returned to the filter can be controlled. My expectation is that this will be important for the right amount of inflow water to move the water that was in the sandbed for nitrification, to the lower part of the filter to be processed for denitrification.

I searched around for a local supply of flowmeters. I wanted a simple one, much like they (EcoDeco) uses. Later I can add a digital one that the controller can monitor for changes in pump efficiency (signalling a cleaning is due). I came up with these ones from eBay, that look pretty much like the ones used in the DyMiCo version. Unfortunately that also means they are a metric sized fitting. After a lot of searching I finally realized that BRS sells adapters. So 20mm to 1/2" adapters are one the way.

Dennis

 

[ravi197699]

Very nice, I read some of their articles as well but how do you determine what size pulse reactor do you require. what is the size of of the tank that you are going to use this for and is this going to be installed on our existing system or this is experiment on new set up, This method makes sense but I really could not get clear picture of what are the required numbers of PH and Redox..

 

[dartier]

Very nice, I read some of their articles as well but how do you determine what size pulse reactor do you require. what is the size of of the tank that you are going to use this for and is this going to be installed on our existing system or this is experiment on new set up, This method makes sense but I really could not get clear picture of what are the required numbers of PH and Redox..

I based the size of the reactor to mirror their model 2000 filter. They provide the dimensions in their manual, and by working backwards from the stated media volume in the manual, I was able to determine that a 75G 48x18x20 should be just about right (with 6" at each end dedicated to the process and return pump).

This is going on a new 300DD tank that I am building. The model 2000 filter is spec'd for up to a 2000L tank, and my 300DD only works out to 1100L or so. I may run the tank empty (no livestock) for awhile both to cycle the filter and to develop the control algorithm. To stand in for livestock, I am thinking of dosing Ammonia. That should limit phosphate build up while I tweak the system.

The manual states the filter operates between +250mv and -250mv during the startup period, but online comments indicate that 0mv and -200mv are the normal range. It may be that they have fine tuned the operation of the production units (the comments were from when the units were in BETA).

The PH is not pertinent to the operation of the filter. It is used solely for the calcium reactor aspects of the filter. They mention setting the PH to be 7.5 for startup as a rule, and then lowering it further if supplementation is inadequate. They also state that a high load hard coral tank is usually fine with 6.7.

The big unknown at this point, is the operation of the carbon dosing pump. An online video mentions that the filter doses carbon once an hour, but I have also read comments that carbon is dosed on each cycle of the return and process pump, which makes sense.

The other thing involving the carbon dose is if the system modulates the amount of carbon based on nutrient loading. This would make the most sense. What I need to work out is how to determine how to pace the cycling of the filter during normal operation. I am leaning towards a PID type of controller that optimizes for x cycles per hour and moderates the carbon dose to achieve it. The manual mentions a on time for the pumps for 30 seconds per hour at minimum and 16 minutes at maximum. I might be able to get away with a simple control loop that adjusts based on the pacing of the cycle time. My thinking is that the system needs to be able to recognize the need for an increase in carbon dose and a decrease where the water column is providing nutrients that can be utilized rather than just the carbon being dosed. The added carbon is the energy for the heterotrophic bacteria that are doing the denitrification. I am wondering if intermittent carbon additions (not every cycle) may have some beneficial effects for lowering other water column nutrients (like dissolved organic compounds).

Dennis

 

[dartier]

Denitrification Cycle Control

Yesterday’s post on controlling the denitrification cycle made me start to think just how am I going to attack the carbon dosing pacing problem.

As I see it, the filter should be in 1 of 3 modes at any one time. A steady state where the carbon dose, and the bacteria population that it supports, matches the nitrate production of the sand bed. A ramp up mode where the carbon dose is insufficient to support a bacteria population capable of denitrifying the nitrate inflow from the sand bed. Lastly a ramp down mode where the carbon dose is too high supporting an over population of denitrifying bacteria that is able to denitrify the incoming nitrate too quickly and therefore will cause the cycle timing to be exceeded by too many cycles occurring in a target period. Why would this be a bad thing (cycling too quickly)? In of itself, I am not sure it is bad, but my instinct is that the carbon dose should be kept to a minimum where possible. It is the added carbon that powers the denitrification process and having unused carbon making its way back to the tank would be undesirable to say the least.

Target Cycle Pacing

My initial plan is to target a cycle pacing of 15 cycles per hour with each lasting 4 minutes. I have chosen this based on a bunch of assumptions from information that I have extracted from the DyMiCo manual. The minimum cycle will be 1 cycle per hour. So if the filter has not detected the completion of a cycle by the 1 hour mark, it will still flush the current water back to the tank and start a new cycle. During the startup period where the filter has not yet cycled, I expect this will be the ‘normal’ mode of operation. Once the filter is fully cycled, then the cycle time will start to be lower than 1 hour and my cycle time optimization targeting can take over.

Detecting Cycle Completion

As anyone who has used an ORP probe can attest to, ORP is not what I would call deterministic. It tends to bounce around and be pushed higher and lower by outside factors. For instance ORP tends to swing higher during the day when lights are on and fall during the night when lights are off. As well dosing of additives can affect ORP and cause it to read higher or lower for a period of time. This will make detecting the completion of cycle challenging to say the least. Just to be clear, the completion of the cycle is defined as all the incoming nitrate has been converted to nitrogen. So we need a way to detect when the cycle is complete and all nitrate is now in the form of nitrogen.

The simplest way would be to use the ORP level with a preset target (Eg. <-200mv ) and expect by that point all the oxygen molecules have been stripped from the nitrate (and nitrite) and utilized by the heterotrophic bacteria. This should work in the common case, but will be dependant on how ORP is being pushed or pulled higher or lower by other factors. Although we are talking about a large volume of water here (in the filter) so it is possible that the variations in the incoming water from the tank, are not able to move it too much to cause a simple level based approach to break down. However calibration of the ORP probe will be important, because that could cause a lot of divergence from true readings. The DyMiCo manual recommends a weekly cleaning and monthly calibration of the probes.

Another method that is used in waste water treatment, and that I experimented with using a sulfur denitrator, is to watch for what is called the nitrate knee. This is the point where all the nitrate has been converted to nitrogen and the ORP level starts to fall at a increased pace. On a graph the ORP level shows a gradual slope downward with a sudden plunge at the point where the knee occurs. The benefit to this method is that it does not have to be a particular level of ORP reading. It is really just watching the rate of change. My hope is that probe mis-calibration and outside factors would have less of an impact on the systems ability detect the nitrate knee occurring. That being said, it is also a lot harder to implement because often the knee is most obvious when seeing it on a graph, and after the fact.

Here is a graph showing nitrate knees during my earlier experiments with a sulfur reactor.

Later I will outline what I hope the extra PH probe I have planned for in the sand bed of my filter will bring to the table.

Dennis

 

[ravi197699]

I based the size of the reactor to mirror their model 2000 filter. They provide the dimensions in their manual, and by working backwards from the stated media volume in the manual, I was able to determine that a 75G 48x18x20 should be just about right (with 6" at each end dedicated to the process and return pump).

This is going on a new 300DD tank that I am building. The model 2000 filter is spec'd for up to a 2000L tank, and my 300DD only works out to 1100L or so. I may run the tank empty (no livestock) for awhile both to cycle the filter and to develop the control algorithm. To stand in for livestock, I am thinking of dosing Ammonia. That should limit phosphate build up while I tweak the system.

The manual states the filter operates between +250mv and -250mv during the startup period, but online comments indicate that 0mv and -200mv are the normal range. It may be that they have fine tuned the operation of the production units (the comments were from when the units were in BETA).

The PH is not pertinent to the operation of the filter. It is used solely for the calcium reactor aspects of the filter. They mention setting the PH to be 7.5 for startup as a rule, and then lowering it further if supplementation is inadequate. They also state that a high load hard coral tank is usually fine with 6.7.

The big unknown at this point, is the operation of the carbon dosing pump. An online video mentions that the filter doses carbon once an hour, but I have also read comments that carbon is dosed on each cycle of the return and process pump, which makes sense.

The other thing involving the carbon dose is if the system modulates the amount of carbon based on nutrient loading. This would make the most sense. What I need to work out is how to determine how to pace the cycling of the filter during normal operation. I am leaning towards a PID type of controller that optimizes for x cycles per hour and moderates the carbon dose to achieve it. The manual mentions a on time for the pumps for 30 seconds per hour at minimum and 16 minutes at maximum. I might be able to get away with a simple control loop that adjusts based on the pacing of the cycle time. My thinking is that the system needs to be able to recognize the need for an increase in carbon dose and a decrease where the water column is providing nutrients that can be utilized rather than just the carbon being dosed. The added carbon is the energy for the heterotrophic bacteria that are doing the denitrification. I am wondering if intermittent carbon additions (not every cycle) may have some beneficial effects for lowering other water column nutrients (like dissolved organic compounds).

Dennis

Thank you for your reply Dennis and this will be great build, I have watched videos and articles but video does not show the entire system just water draining over sand bed at slower rate and returning to the tank, Tanks look amazing and coral and polyps extension on corals are the the fullest, I have built all of my reactors, protein skimmers etc for my 700 gallon system my self and I am big fan of DIY and new approaches.. I would love to provide my input as help where I can..

 

[dartier]

Detection of the Nitrification Cycle

As I mentioned I am going to be inserting a PH probe into the sand bed of my filter. This is a probe that the DyMiCo folks do not employ. The PH probe included in their system is used for control of the Alk/Ca supplementation. What I hope to gain from the inclusion of this extra PH probe is the ability to observe and perhaps develop a useful signal from the change in PH that accompanies the completion of the incoming Ammonia being oxidized into nitrate. This effect is termed the Ammonia Valley. The PH falls as the nitrification progresses until the Ammonia has been completely oxidized and then the PH reverses direction and climbs.

You can see a obvious example in Figure 1 at the following site. http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3571785/

I am hopeful that this is going to provide some useful indication of the amount of waste loading in the incoming water. Decreases in the period of the nitrification cycle should be indicative of a decrease in ammonia entering the sandbed (or the growth of the sandbed microbes). An increase in the period should show the opposite effect. As I plan to cycle and operate the filter initially with ammonia, it should make for a good method of testing the hypothesis.

I should mention, I am not sure if this effect will even be measurable in a saline environment, all the papers I found on it only discuss freshwater applications in the waste water treatment area. That being said the same base processes are occurring within our tanks as they leverage in the treatment space (nitrification / denitrification), so perhaps the ammonia valley will also be present.

I plan on logging the PH of the sandbed during the cycles for the filter to look for correlations between the period of the nitrification cycle using the ammonia valley as the indicator and the period of the denitrification cycle using the nitrate knee. I suspect that because the effect is observable as a relative change, it should not be prone to probe calibration errors (similar to the nitrate knee, except relative change rather than rate of change).

That is my plan for the second PH probe and what I hope to gain from it. My next update will be my thoughts on how the Alk/Ca supplementation portion of the filter could be managed.

Dennis

 

[gregkn73]

Detection of the Nitrification Cycle

As I mentioned I am going to be inserting a PH probe into the sand bed of my filter. This is a probe that the DyMiCo folks do not employ. The PH probe included in their system is used for control of the Alk/Ca supplementation. What I hope to gain from the inclusion of this extra PH probe is the ability to observe and perhaps develop a useful signal from the change in PH that accompanies the completion of the incoming Ammonia being oxidized into nitrate. This effect is termed the Ammonia Valley. The PH falls as the nitrification progresses until the Ammonia has been completely oxidized and then the PH reverses direction and climbs.

You can see a obvious example in Figure 1 at the following site. http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3571785/

I am hopeful that this is going to provide some useful indication of the amount of waste loading in the incoming water. Decreases in the period of the nitrification cycle should be indicative of a decrease in ammonia entering the sandbed (or the growth of the sandbed microbes). An increase in the period should show the opposite effect. As I plan to cycle and operate the filter initially with ammonia, it should make for a good method of testing the hypothesis.

I should mention, I am not sure if this effect will even be measurable in a saline environment, all the papers I found on it only discuss freshwater applications in the waste water treatment area. That being said the same base processes are occurring within our tanks as they leverage in the treatment space (nitrification / denitrification), so perhaps the ammonia valley will also be present.

I plan on logging the PH of the sandbed during the cycles for the filter to look for correlations between the period of the nitrification cycle using the ammonia valley as the indicator and the period of the denitrification cycle using the nitrate knee. I suspect that because the effect is observable as a relative change, it should not be prone to probe calibration errors (similar to the nitrate knee, except relative change rather than rate of change).

That is my plan for the second PH probe and what I hope to gain from it. My next update will be my thoughts on how the Alk/Ca supplementation portion of the filter could be managed.

Dennis

Very interesting project. I wonder why you expect any significant nitrification occur , into the sandbed of your filter. Nitrification bacteria colonising rocks,sand bed and every other available surface at your DT, should make the majority of this process. Otherwise with such low flow to your filter, you might have problems with ammonia, at your DT.

 

[dartier]

Very interesting project. I wonder why you expect any significant nitrification occur , into the sandbed of your filter. Nitrification bacteria colonising rocks,sand bed and every other available surface at your DT, should make the majority of this process. Otherwise with such low flow to your filter, you might have problems with ammonia, at your DT.

Yes, the planned filter will be very low turnover compared to a normal setup, but even at the low turnover rate, it will filter the entire tank volume (300G) every single day (once it is fully cycled). The sand bed will still be instrumental in helping to process the solid waste that has not been broken down into ammonia by the time it enters the filter. That will generate its own ammonia as it is broken down. One of the desirable aspects of this type of filter is the possibility of high stocking levels, and the ability to feed freely. We will see if my DIY version is able to pull this off.

It will be interesting to see if I am able to detect nitrification occurring in the sand bed or not.

Dennis

 

[dartier]

Parts are trickling in

Parts are trickling in

Parts are slowly arriving for the build. I went with Sicce Syncra 3.0 pumps for the process and return pumps. As near as I can tell, these are pretty close to the Tunze ones that OEM filter uses. The head height for the return pump from the filter back to the tank should be around 5 feet. So the Syncra 3.0 that I am using should provide about 200% of the target flow (1000 L/H).

I went with the exact same model of pump for the process pump as this is what the OEM filter uses. Seems a bit excessive to me, the smaller model 700 filter uses a lower flow pump for the process pump, but for the model 2000, with the dual pipes into the gravel bed, they use the same capacity of pump for both process and return. We will see if it ends up causing too much surface movement of the water in the process chamber.

Dennis

 

[ravi197699]

looking good, What else you are waiting for and are you going to use apex to control system..

 

[dartier]

looking good, What else you are waiting for and are you going to use apex to control system..

Still waiting on the glass baffles. They were supposed to be done on Wednesday, but when I called, they said their "hole" person had not yet drilled the 10 holes required (5 in each piece). They adjusted the ETA to today or tomorrow, I just called a little while ago to see if "his holiness" had graced me with holes. Grrr, they were not sure so they are going to call me back.

I plan on using a Reef Angel rather than an Apex. I already have a RA, so I am somewhat familiar with coding for it. The other reason is that, if this works out well, and I want to build more filters for some of my other systems (frag tank, quarantine, etc.), I know that I can run the RA software on a plain Arduino and save on having to buy x controllers. Apex does not strike as being as hackable (as a Reef Angel).

One way or another, I should have some photos this weekend of the filter starting to come together.

Dennis

 

[dartier]

Ok, so I finally got the baffles. They were still not done when I went today, but I said I would wait will they completed the drilling. That way I was pretty much guaranteed that they would get done.

They came out OK, but I did have a few glitches. The alignment of the holes was pretty good, off by perhaps 1/16" for the matching hole in each piece. Though when I got home and went to insert the Lifeguard bulks heads in they did not fit.

The recommended hole size was 1" 1/4", but the glass shop could only do 1" 1/8" or 1" 1/2". I measured the bulkhead at the counter and decided that 1" 1/2" would be too big, and the 1" 1/8" should just fit. Today I got home and found none of the bulkheads would not push through the holes :facepalm:

I ended up taking a piece of emery cloth and grinding down the threads of each bulkhead ever so slightly to just take the points off. Now they fit.

Here they are with some unions pressed into them.

Once I had the bulkheads in I noticed that I should have planned better on the spacing. Had I used CAD to lay it out I would have noticed that the 2 outside holes in the bottom row should have been moved closer to the edge of the baffle to make the distance from the higher holes uniform (as compared to the bottom center spacing).

Here is how I tried to accommodate for the spacing snafu.

Tomorrow I will see about siliconing the baffles into the tank that will be the body of the filter. I will work on the plumbing for each pump before I actually affix them though just to make sure I have enough room.

... man the tag in this BBcode does not allow resizing?

Dennis

 

[glaukos]

Very interesting project. I wonder why you expect any significant nitrification occur , into the sandbed of your filter. Nitrification bacteria colonising rocks,sand bed and every other available surface at your DT, should make the majority of this process. Otherwise with such low flow to your filter, you might have problems with ammonia, at your DT.

Interesting project indeed,dennis!
It's easy for someone just sitting on his couch guessing/saying/wondering/criticising that it will or won't work but instead trying things and sharing them,makes this hobby evolve:thumbsup:
So thanks for trying and most importantly...sharing the whole thing:beer:

 

[ravi197699]

Hi Dennis, I was wondering the same about holes, They seem little too close to edge, you will have to play with this a little when you silicone these into the sump but other than this looks good,

 

[dartier]

Hi Dennis, I was wondering the same about holes, They seem little too close to edge, you will have to play with this a little when you silicone these into the sump but other than this looks good,

Do you mean vertically from the bottom edge? It just looks that way in the photos. There is about 1" of clearance between the edge of the bulkhead and the bottom edge of the glass.

However I am going to have to take the baffles back to be adjusted. I brought the tank with me to the glass shop when I ordered them so measurements could be taken on the spot. I learned from experience in the past (and lots of return trips) that it is easier to bring the tank with me. This time though, I measured it in the parking lot with the counter staff, decided on the correct dimension to allow a reasonable gap for them to be siliconed in. Things went awry though when the glass technician came out after and did the "actual" measurement. The final size is too big. They fit in the tank, but I am not able to get a 1/16" spacer between the glass and the baffle on one end, and only 1 spacer on 1 side at the other end of the tank. The end that is for the process pump will not be under pressure (same water level as filter), but the return end needs to hold back 18" depth of water (and gravel/sand).

I will go back and see if they can take an 1/8" of the side of each to give proper clearance for a bead of silicone between the baffle and the glass. I might see about getting some 1" wide stips of glass that can be siliconed on the tank to provide a lip for the baffles to butt up against as well.

This filter is turning into an engineering project :lolspin:

Dennis

 

[OllieNZ]

Do you mean vertically from the bottom edge? It just looks that way in the photos. There is about 1" of clearance between the edge of the bulkhead and the bottom edge of the glass.

However I am going to have to take the baffles back to be adjusted. I brought the tank with me to the glass shop when I ordered them so measurements could be taken on the spot. I learned from experience in the past (and lots of return trips) that it is easier to bring the tank with me. This time though, I measured it in the parking lot with the counter staff, decided on the correct dimension to allow a reasonable gap for them to be siliconed in. Things went awry though when the glass technician came out after and did the "actual" measurement. The final size is too big. They fit in the tank, but I am not able to get a 1/16" spacer between the glass and the baffle on one end, and only 1 spacer on 1 side at the other end of the tank. The end that is for the process pump will not be under pressure (same water level as filter), but the return end needs to hold back 18" depth of water (and gravel/sand).

I will go back and see if they can take an 1/8" of the side of each to give proper clearance for a bead of silicone between the baffle and the glass. I might see about getting some 1" wide stips of glass that can be siliconed on the tank to provide a lip for the baffles to butt up against as well.

This filter is turning into an engineering project [emoji38]spin:

Dennis

They should be able to easily if they have a polishing machine. One of my friends is a glazier and I've had him do a few bits for me and if there were and size issues we could run the glass through the polisher to take off a very precise amount.

 

[dartier]

They should be able to easily if they have a polishing machine. One of my friends is a glazier and I've had him do a few bits for me and if there were and size issues we could run the glass through the polisher to take off a very precise amount.

The edges of the baffles are nicely bevelled and polished. The last time I got baffles from there (about 3 years ago), they were only able to do a rounded edge, that still had grinding marks apparent. These new ones look a lot better with a flat edge and 2 bevels, polished completely smooth. I guess they got some upgrades since the last time.

Hopefully they do it without charging me more ...

Dennis