HOW TO DETERMINE THE SIZE OF A FILTER?

This is THE question most frequently asked by those who want to be certain that the water in the pool is very clear and that “everything is going well”.

A pond with clear water is easy to obtain: but whether “everything goes well” is another matter. To understand this, you will have to approach our page dedicated to mechanical and biological filtration before continuing this page.

Mechanical filtration must eliminate or retain all suspended particles, while the bio filter must – by means of bacteria – eliminate toxic substances contained in the water (ammonium and nitrites), and the two systems require different approaches.

One thing is clear: you cannot filter a pond with something the size of a shoebox! In fact, the real question to ask is first to know what maximum quantity of food will be distributed daily, and in what volume of water: this is really what makes it possible to determine the size of the filter.

On the Internet you will find various methods on the subject: the problem is that, depending on the process used, the difference in the results obtained can go from simple to double! Such a difference is due to the fact that the basic data and, above all, the way of calculating these decisive parameters (recycling time, proportion of ammonia from food, etc.) varies widely. This inevitably leads to different sized solutions.

METHOD BY AN EXAMPLE

It is possible to try a realistic approach; However, this remains an empirical method and therefore has no scientific value. Otherwise, one would not only have to take into account the type of food, but also the pH, temperature, alkalinity, oxygen concentration, recycling time, type of media, porosity, water velocity , the type of circulation in the filter, the volume of the pool, the dead insects or plants in the pool, etc. In short, quite a puzzle!

The method proposed for more than 15 years has proven to be effective and realistic. It is specific to Koi ponds and not to aquaculture which works differently.

Today in 2024 we have adapted it to new modern materials but the calculation method has been the same for 15 years. It consists of several stages.

- determine the stocking density, therefore the quantity of food to process per day.
- determine the volume of mechanical and biological filtration.
- check if the filter meets a few basic benchmarks.

BASIC DATA

To evaluate the size of a biological filter, you must know the quantity of food poured into the pond each day. To do this, we take into account the number of fish and their size. These Koi will normally all eventually reach respectable sizes, so it would be absurd to underestimate the filter sizing when considering 15cm fish. But for your information, we will indicate below the average ratio between size and weight of Koi.

Good practice indicates that the quantity of food to provide, per day, depends on the temperature and in summer, when the Koi's needs are high, it is considered that a quantity of food whose weight is equivalent to between:

1 to 1.5% of the body mass of the fish contained in the pond (*)

3 to 5% for the youngest (less than 15 cm)

(*) Some give even more, but with the aim of rapid growth during the Koi Show for example.

This quantity varies with the temperature of the water but in all cases, the food should be distributed each day in several installments:

Once a day or every 3 days depending on appetite from 7° to 10°C

1 to 2 times a day at 10 to 15°C

3 to 4 times a day between 15°C and 27°C

1 to 2 times a day above 27 – 28°C (the quantity is reduced due to the drop in the oxygen concentration in warmer water, which reduces the assimilation of food).

Food to be poured into the pool which should normally be consumed between 5 and 10 minutes (large Koi eat more slowly) otherwise to be corrected.

Finally, the percentage of protein contained in the diet is variable (around 28 to 45%): we also adapt this percentage according to the water temperature.

Size - cm	10	15	20	25	30	35	40	45	50	55	60	65	70	75	80	85	90
Weight - grams	15	65	150	290	480	700	1100	1550	2200	2900	3700	4600	5900	7000	8500	10300	12000

The pellet feed distributed to the fish produces approximately 37 g of ammonia per 1000 g of food (we accept food with 40% protein).

Knowing the speed of nitrification according to experimentally established kinetics (depending on pH, temperature, oxygen concentration - see nitrogen cycle) and, knowing various (constant) characteristics concerning the metabolism of bacteria, we retain that A surface area of 4 to 5 m² of bacterial support is likely to receive a bacterial biofilm capable of transforming 1 g of ammonia at a temperature of 22°C in 24 hours.

This reference surface is sufficient even for smooth bacterial supports (the least favorable), at a minimum oxygen concentration in the filter of 4 mg/l as well as a KH of 6° minimum.

When designing a pond, it is normal to first focus on aesthetic or practical considerations; However, we must not lose sight of the fact that for a specific volume of water, a population density must be respected.

By introducing too many fish into your pond, you inevitably open yourself up to problems: a simple rule of thumb says to maintain one Koi per cubic meter of water maximum. The ideal being a lower density of 1 Koi per 2 or 3 m3 of water. We often read here and there, a density of 1 Koi per m3: that’s too much!

Algae problem, health problems, biological filter often too small….

If you still decide not to follow this simple common sense rule, you will have to oversize your installation and regularly change a larger part of the water.

DETERMINE THE AMOUNT OF AMMONIUM TO TREAT

Depending on the composition of the diet, only a fraction of the food is assimilated by the fish. A part is therefore eliminated in the basin, in different forms, we retain the following distribution of nitrogen:

+/- 30% withholding

+/- 10% eliminated by feces,

+/- 30% excreted in ammoniacal nitrogen

+/- 30% lost as dissolved organic nitrogen.

That is approximately 70% of residues to be treated: from this percentage we estimate the quantity of nitrogen N that the biological filter must transform. This for a usual food containing 40% protein.

A specific food containing a very high percentage of proteins (specific food for fry > 50%) will therefore require an adaptation of the filtration size.

The method used is that described by Kjeldahl (1883), and by convention, it is accepted that proteins contain 16% nitrogen (the nitrogen content of proteins varies from 14 to 18%).

According to the above, let us remember that 1000 g of food generates 37 g of ammonium / ammonia

To help you, here is a practical example that may be useful to you: an 18 m³ pond that can accommodate a maximum of 10 60 cm Koi. In the most unfavorable conditions to be sure that the biological filter is not a little tight.

Or for 555 g of food > 20.6 g to process for our example.

DETERMINE THE SURFACE OF BIOLOGICAL SUPPORTS

The bacterial support surface necessary to ensure this operation: 20.6 x 5 = 103 m².

Then, simply determine the volume of bacterial support that it is useful to install to obtain this surface area. The data of the supports most used in fixed or mobile beds allows this operation.

Matala black: 190 m² / m³

Matala blue: 380 m² / m³

Matala green: 295 m² / m³

Matala gray: 470 m2 / m3

Japanese mat: 275 to 300 m²/m³

Hel-X 17: 595 m2/m3

Hel-X 13: 956 m2/m3

Biocerapond 1400 m2 / m3

Hel-X Flakes 5000 m2: m3

DETERMINE FILTER VOLUME

Depending on the performance of the supports, we will see that a filter can see its volume decrease or increase.

Please note, the gross volume of a room can only contain 50 to 60% of the carpet volume (since we intersperse 1 carpet with a water space)... in this case, it is therefore necessary to adapt the volume of the carpet. filtered.

In the case of Hel-X you will only be able to obtain good results by pouring 40% of media into a raw volume (eg: 400 liters of Hel-X maximum in 1000 liters of water). We can push for a filling of 50% by forcing a little on the ventilation.

Using the classic Japanese rug: to obtain the 103 m² of surface area in our example: (1000 / 275) x 103 = 375 liters of Japanese rug.

Other characteristics should still attract your attention: the pumping rate, the speed of the water in the filter and the contact time between the bacterial support and the water to be treated.

A contact time of between 5 and 10 minutes is a good reference when using media from the list above. By using supports with a poor colonization surface (pieces of non-porous plastic pipes) this contact time can increase to 20  minutes! 

By using efficient media this time can be reduced to 2 or 3 minutes. Let's not forget that the biological filter does not remove all the pollution in a single pass of the water through the filter. Modern Combi filters (drum with integrated bio chamber) are very effective despite a contact time sometimes approaching a minute!

This contact time of 5 to 10 minutes applies especially to multi-chambers where we attempt to sediment the particles whereas with a drum, the smallest particles are retained and evacuated regularly. that's a big difference.

An average water velocity in the biological filter of 0.02 meters/second maximum is also OK.

Pumping flow: the volume of the pool must pass through the filter every two hours minimum (three hours maximum for very large volumes > 100 m3).

By filtering a pond in 1 hour, suspended particles are reduced, especially if the mechanical filtration is efficient (drum or endless sieve filter).

The above demonstrates once again that a small filter cannot be suitable: even by filling it with a very efficient media in terms of surface area, it becomes impossible to establish a correct relationship between the contact time, the turnover of the pool and the speed of the water in the filter.

HARMONIZE VOLUME AND CONTACT TIME

Let's return to our practical example: if we use for our 18 m3 pool, a pumping rate of 9 m3/hour or 3600 to obtain 0.0025 m3/second.

At a speed of 0.02 meters/second, the minimum section of the filter is (0.0025/0.02) or 0.125 m2. (Case of a single chamber filter with horizontal water movement)

To obtain this section, we can for example opt for width x height = 35 cm x 36 cm (= 0.126 m2). This is an example, always opt for the largest dimensions possible, the water flow will be slower...

As the volume of the carpet is 375 liters or 0.375 m3, the length of the filter is given by (0.375 / 0.126) or 3.19 meters or 297 cm. Of course we can widen or increase the height of the filter in order to reduce its length... As the speed of the water is slowed, this poses no problem.

Do not forget to include in this theoretical useful height, the height of water under the mats (from the bottom of the filter to the retaining grid, count around ten cm) as well as the height between the top of the mats and the edge upper part of the filter (5 to 10 cm on average). All this is theoretical: the example proposed can be used as is for media such as ceramics: but as indicated above, mats require a space between two elements, it is therefore necessary to further increase the volume of the filter according to these spaces.

Last data which allows us to evaluate a good filter, the contact time. This time, between 5 and 10 minutes, allows good exchanges between the water and the supports.

In our example: let's assume that the water circulates horizontally in this long filter which measures 0.35 m x 0.36 m x 3.19 m, i.e. a useful volume of 401 liters.

Let's add 2 x 10 cm of space under the grids and above the filter masses, this now gives: 0.35 x 0.56 x 2.97 = 0.582 m3 or 582 liters. This for a flow rate of 9 m3/hour or 150 liters/min.

Contact time is 582 / 150 = 3.88 minutes. This time is not very low... Nothing serious because we still have to widen the filter since the mats must be spaced from each other.

If we move to 0.45m wide x 0.6m high x 2.97m long, which is 0.801 m3 or 801 liters, that now gives 748 / 150 = 5.34 min.

ASSESS WATER FLOW SPEED

Water speed is not to be confused with flow!

Here's how to estimate the average water velocity in a filter:

- measure the section of the filter (or the surface of a chamber if the water circulates vertically) for example 45 cm by 60 cm, this gives a water passage surface of 0.27 m2

- you normally know the flow rate of your installation, for example 9 m3/hour or 3600 = 0.0025 m3/second

The average speed is 0.0025 / 0.27 m2 = 0.009 m / second.

In the hydraulicity section, you will find information concerning the pipes to use for the construction of the basin.

OXYGEN CONCENTRATION

Without going into detail, a biological filter MUST be aerated, even a fixed bed filter!

Bacteria consume 4.33 g of O2 to transform 1 gram of ammonia.

The above method applies to multi-chamber filters with fixed beds or normally ventilated moving beds. Regardless of the mechanical filtration that must precede: brushes, pre-filter or drum, always aerate the biological media: this helps to provide oxygen, to reduce possible preferential paths in the filter and to develop the biofilm.

FINAL REMARKS!

We calculated the volume of a filter for “almost extreme” conditions: food dosage of 1.5%, high protein percentage, etc. but for an estimated “average” nitrification yield. So with average water parameters and a normally very clean pond containing a normal density of Koi.

If the pH (or KH hardness, etc.) of the water is no longer optimal, if the pool contains a lot of organic matter, the diagram above becomes correct.

It is now up to you to decide whether or not it is useful to slightly oversize this filter to overcome these negative points which can always arise. In practice, we too often see overcrowded pools with a bottom filled with silt or waste; the physical and chemical parameters of the water are not always monitored very well, so taking the precaution of oversizing is not useless. For my part, my filter is well oversized by 30%, fortunately, I still benefit from quality local water and I have had no problems for more than 12 years.

Finally, the current trend is the exclusive use of Hel-X, you will see on our sales site that most drum Combi include these supports. It works very well and the last 10 years have proven it to us. However, a mix of Japanese mats and Hel-X (50% - 50%) is considered by many to be ideal in terms of biological filtration.