DIY Denitrator Idea?

[Wallace]

Howdy,

I am fairly new here, but I have been keeping fish for many years so at least I hope I know what I am doing. Anyway, I have been wanting a denitrator and have sorted through many of the more common “coil denitrator” plans. There are lots of success stories, but very commonly I’ve heard people say they tend to cog up in a year or so especially when thinner tubing are used, and fine-tuning the flow rate could be a pain. Since the coil denitrator is not without its shortcomings, I thought to myself maybe there is a better way to do it.

That being said, first of all let me make sure my understanding of denitrator is correct. According to the various coil denitrator plans, it is really a two-stage design where in stage one the aerobic bacteria will deplete all the oxygen in the water through the normal nitrogen cycle, and then in stage two anaerobic bacteria will convert nitrate to nitrogen gas. The trick is to allow enough time for the oxygen to deplete and also enough time for the anaerobic bacteria to work out all the nitrates before the water is released back into the main tank.

With the above approach, my concept is that instead of using a “constant flow” design (aka coils), why not use a store-and-flush design? If I build a completely sealed acrylic box with one water intake and one water exhaust, fill it with bio balls (or one of those “BioSlab”) before I seal it, and then hook up a pump to it that regularly “flush” that “filter” (say, once a week), would it achieve the same result because the box would be filled with stagnent water in between flushes? Conceptually, if both type of bacteria can co-exist and stay alive inside the “box”, whenever fresh water enters the box (“flushed”), aerobic bacteria will begin consuming the oxygen in the water, depleting it. From there on, anaerobic bacteria take over and convert all nitrates into nitrogen gas. Once we know all nitrate has been consumed, we run the pump for maybe 3 minutes to flush the “box” - release water back into the sump and fill it with "fresh" nitrate-infested water, and the cycle begins again.

I am sure lots of minor detail will surface, for example how long it will take for the bacteria to colonize the filter or how frequent it should be flushed etc. But I think the bigger question is whether the two type of bacteria will live and coexist: whenever its flushed, will the fresh oxygen kill the anaerobic bacteria? And in the same token, once oxygen is depleted (and stay depleted for maybe days), will the aerobic bacteria stave and die? The hardware part is really easy, I can build it for less then $50. But it’s the biological life cycle of these two types of bacteria that I do not fully understand.

Do you think its possible? Am I out of my mind? Thoughts?

Thank you!




Wallace

 

[dskidmore]

Well.... Interesing idea. There certianly should be a way to do this without the thin coils, but there are some challenges.

Going with that weekly period, you will see great fluxuations in the populations of bacteria. The areobic bacteria are going to run out of food before the anerobics have finished their work. The anerobics are going to be suffering from too much oxygen when the areobics are thriving. On top of that, you're having a weekly major fluxuation in the tank nitrates. (This would be on top of the PWC change. No denitrator can replace PWCs.)

I have a counter proposal. What if instead of using coils to increase surface area and travel distance, we used a fine sand? A dosing pump would be hooked up to the filter for slow flow, wate would come in to the top of the pile of sand where the areobic bacteria grow, and flow down through the sand, through an anerobic region? Might need something to trap the sand from going through the return back to the tank.

For cleaning a clogged filter, you'd just have to stir up the sand and run fresh water through it. Might even have an auto clean feature that occasionally runs water backwards through the filter and pushes the backflow through a floss filter. (Would not want to autoclean freqently, as this would introduce oxygen into the bottom of the filter.)

Many people use deep sand beds for the detritrifying effect. It should be possible to set this up with optimal flow for maximum effect.

 

[Wallace]

Hum... interesting idea. If we reverse your water flow, and have the dosing pump move water into the BOTTOM of the sand bed, and let it overflow from the top back into the tank, would that eliminate the need to chean the sand? We could use a tall PVC pipe to ensure the water flow through most of the sand before it exit? Are there cheap dosing pump available that would, say, run once a day and put maybe 30ml of water on each "turn"?

Now, going back to a weekly flush: as far as flutation in the nitrate level, since the "box" is relatively small (probably no more then a couple gallon), on a relatively large system the effect should be minimal. On a 50 gallon tank a 2 gallon flush would be 4% impact on nitrate and oxygen level... the other water chemistry should remain fairly similar. I think that can be tolerated?

Finally, just for the record - I have no problem with PWCs, I built my filtration system in the basement and its constructed in a way that a 25% water change only takes literally about 2 minutes to perform. But doesn't matter how often I perform water change, I can never bring the nitrate level down to near-zero because I am only changing 25% of the water. Thus the denitrator idea...

Thoughts?



Wallace

 

[dskidmore]

Reverse flow would reduce the chances of getting sand into the pump. I would want some sort of prefilter to remove solid debris, otherwise they'll get trapped on the bottom and produce more nitrates.

I'd be more interested in a distribution manifold along the bottom of the filter. The vertical pipe starts to get rid of the seperate anerobic/arobic zones, doesn't it?

I wouldn't be able to guess the appropriate timing/volumes of the dosing pump. I would hope that it could handle more than 30 mL per day. 30 mL every 30 minutes would still be a light trickle. Use a nitrate test kit to calibrate the timings/volumes. Ideally, the nitrate in the water coming out should be significantly lower than the nitrate going in. Might have to wait a week between tweaks, so the bacteria populations stabilize.

If the effect is that minimal, what is it's purpose? Denitrator will help matters, but you may also look at nitrogen inputs into the system.

 

[Wallace]

Pre-filter can easily be done on the intake side of the dosing pump, since the flow rate is so low the need to replace / clean that filter should be minimal.

Regarding seperation of the two bacteria colony, I actually think a vertical pipe will help clearly define it. Think about it this way: with a 2" PVC pipe that is 20" tall, you will have a fairly narrow passage for the water to flow "up". Thus, the aerobic bacteria will grow on the bottom section of it where fresh water enters, up until the point where oxygen is fully consumed. Thereafter, anaerobic bacteria will start to colonize. Assuming the dissolved oxygen level is stable, the filter should automatically establish that "virtual line" between the two colonies based on flow rate. Now, honestly I couldn't care less where that line finally ended up to be... as long as its not close to the top of the filter (which translates to inefficienty). What I am concerned is that we need to make sure water will flow evenly through all the sand on their way up - e.g. its not going to just flow through the center of the pipe, leaving the water around the pipe wall completely stagnent. Is that what you mean by decreased seperation and the need for "distribution manifold"?



W.

 

[dskidmore]

Yeah. the water will naturally flow through the sand from the injection point to the overflow point. Make an X or H or something in the bottom of the bucket and distribute the water around. Water flows up through the sand much slower than it would through the vertical pipe, because the diameter of the bucket is larger than the diameter of the pipe.

 

[Wallace]

Ok all sounds good... lets build it!

For the container, I am thinking about this:
http://www.petsolutions.com/Phosban+Reactor-I-48142292-I-C-1017-C-.aspx

"By pushing water from the bottom upward through a dispersion plate, the reactor gently fluidizes the media and forces an even distribution of water through it, which prevents channeling."

Should take care of the water-flow-distribution issue?

Now the hard part... dosig pump. So far this seems like the right choice, since I can program it to run on a fixed intervel, allowing easy control (and adjustment) of water flow rate:
http://www.marinedepot.com/md_viewItem.asp?idproduct=AQ4531

But its pricy... any suggestion of a cheaper dosing pump that is also "programable"?



W.

 

[dskidmore]

"By pushing water from the bottom upward through a dispersion plate, the reactor gently fluidizes the media and forces an even distribution of water through it, which prevents channeling."

Should take care of the water-flow-distribution issue?

Except you're aiming for a lower flow rate than what it takes to fluidize the bed. It looks like a good plan though.

... any suggestion of a cheaper dosing pump that is also "programable"?

Sorry I'm not very useful on that score.

 

[Wallace]

I think the only problem left is cost... unless there is a more economical dosing pump, the projected cost could be close to $300. That's not really in the tradition of DIY... lol

I wonder if I could use a water pump on the inlet to pressurize the unit, then use some Jaco fitting on the exhaust to restruct output flow rate, and go back to the "count your drip rate" method. Since its pressurized, it shouldn't clog except the Jaco valve itself, which could be opened briefly once in a blue moon (1/2 seconds?) to "flush" itself.

Now granted if I am building something pressurized that has the ability to be "flushed", I could've gone back to the tube denitrator...

Thoughts? Anyone?



W.

 

[dskidmore]

Would a small powerhead and a wavemaking timer be any cheaper?

Or a small powerhead and a float switch?

5 gallon bucket and some PVC would likely be cheaper than the reactor you picked out.

 

[Wallace]

The reactor being $33 isn't that bad of a deal... think about it, you would probably end up paying close to $20 for PVC fittings and glues etc. anyway, plus your time. Although you could most likely build a much bigger filter without spending much more though. Anhow, the point is, its the dosing pump at $220 that kills.

In regards to powerhead and wavemaker... is there even a wave maker that allow you to set it to turn on for 1 second then turn off for X hours? I think most waive maker timer are based on minutes. So even if you set it at the minimum, at 1 minute with a small powerhead we would've likely still flushed the whole damn thing (unless its a 5 gallon bucket we are building)... now, if the unit CAN be flushed, then I could've just go back to my very original design and use X-10 timers (which I already have).



W.

 

[dskidmore]

The greater volume would work for you here. You could get away for a higher flow rate without disrupting the anaerobic range.

I've realized that we can calculate the maxiumu flow range, if we pre-determine the boundary between the aerobic and anaerobic ranges. (May not be optimal, as I don't know what the optimal ratio of areobic to anaerobic volumes should be.) You never want to displace more liquid at one time than the volume of the aerobic range, because then you'll be pushing fresh water into the anerobic range.

The low end powerheads loose alot of power when pushing water uphill. Se if you can find some performance curves that get low enough at the hight distance from your aquarium to the top of the reactor.