Canister filters have been in the hobby for a long time now, more than the lifetime of many of our current hobbyists. They most commonly are not directly attached to the tank, but connected by flexible hoses. Thus, they may be concealed in the tank stand or behind the tank. They should offer advantages in much greater filtration media volume and flexibility in selection of media, as well as in water pick-up and return. Canisters are conventionally packed to perform all desired filtration functions: mechanical, chemical, and biological. Stepped or graded levels of porosity (mechanical particle retention by sponges or fiber) are first in the flow path. Chemical media would follow the mechanical, if needed or desired. Biological filtration comes last in the flow, but sometimes it is topped with a fine floss or sponge for final water “polish” before return to the display. Additionally, canisters should be quieter in operation than air-driven in-tank filters or than HOT (hang on tank) or HOB (hang on back) power filters. They offer quite a bit of flexibility in the filtered water return to the tank. This allows the tank-keeper to customize the current and oxygenation to suit the needs of the fish (and plants) to be kept. Splashes, spills and overflows should not be an issue as canisters are closed systems.

Canisters are not without their drawbacks. They are significantly more expensive than HOT/HOB filters as a rule. The larger media volumes used may carry larger initial price tags as well. Historically they have been less than user-friendly. There seems to be a real learning curve in setting and using the older designs of canister filters. Once mastered, the opening, cleaning, and reassembling of the canister is easy enough. The process, however, is not ever as obvious, fast, or as easy as pulling a cartridge from a hang-on power filter. More modern units than the Classic Eheims that I use most are much more user-friendly, but still complex compared to HOBs. Media costs over the life of the unit should actually be much less than disposable cartridge replacement, as most media for mechanical and biological filtration in canisters can be selected in long-term reusable formats. Quality, well-engineered canisters should serve their function for decades, so the life cycle cost may be below to far below that of the seemingly “inexpensive” filters using disposable media or media needing routine replacement, and with possibly higher power use. Don’t ignore the filter’s power requirement. Filters operate 24/7, while lights are only half time or less, and heaters are intermittent operation depending on needs. The significant up-front expense may be intimidating or out of reach for many hobbyists, however. If you know that you are likely to continue in the hobby long-term, canisters are among the most economical filtration techniques by life-cycle costs.

Some canisters are designed to attach to the tank (HOT Magnums for example) in the manner of hang-on power filters. Several canister types are available that are internal filters (Fluval and Eheim both offer internal models, other brands may as well). These internals fit inside the tank, rather than outside having connection to the tank by hoses. These do tend to offer less flexibility than external types. They do have smaller media volumes along with limited media selection. These internal canisters may still be useful in some applications, one of which will be discussed later in this article. They are not in any way particularly aesthetic additions to tank décor.

My primary filtration technique has been the use of external canisters for more years than I care to admit. I will admit that even though I am married to this filter type, long ago I found opening the canisters every three months (a standard back then) for maintenance to be a chore. As I became ever more neurotic about water quality, I was offended that entrapped particulates from the tank, although removed from the water column quite well by the canisters, were held in the canisters for such a long time that heterotrophic and saphrophytic bacteria or fungi could digest them. This meant that these particulates (bits of fish poop, fine uneaten food particles, bits of plant debris, etc.) were continuing to contribute to the bioload (using oxygen), increasing nitrate and DOC (dissolved organic compounds), and other pollutants of the water during their digestion or mineralization (Endnote 1). I had already observed that nitrate levels in my RFUG-filtered tanks (Reference 1) were lower than those in my canister-filtered tanks with comparable bioloads and water changes. I attributed this to the fact that the sponges pre-filtering the RFUG powerheads and the internal mechanical-only canisters were rinsed frequently. This weekly rinsing must have been removing significant waste before it was fully digested and thus it was out of the system rather than increasing nitrate, DOC, or general pollution. In FO (fish only) saltwater tanks using sump wet/dry filters, the pre-filters for the wet/dry were rinsed even more frequently (daily) in efforts to hold down pollution, so this idea made sense in FW application as well. I decided that I could not and would not go entirely either RFUG or W/D (wet/dry). Nor was I willing to discard the significant investment I had in Eheims. So I needed to find modifications that would allow me to use the canisters, but still result in low-nitrate (and hopefully comparably low-DOC and other pollutants) tanks.


The obvious solution was to pre-filter the canister intakes, comparable to the W/D or RFUG techniques I use. Fine, Eheim offered a pre-filter intake for the canisters. So I bought several of those. They do work, but in my experience are not very user-friendly. Their loading or carrying capacity is not very great, nor is their particulate capture very fine. Okay, Tetra at that time made several sponges for their Billi or Brilliant air or powered sponge filter line. These should and did work well as pre-filters for the canisters. I either modified the canister intake hard tubing by drilling holes in the intake tubing area to be covered by the sponges, or shortened the intakes and spliced on the perforated tubes made by Tetra for these sponges, using flexible tubing for the connection. Both DIY techniques work well, but mid-range and larger canisters might need double prefilters. Those units are no longer manufactured by Tetra, but are currently available under the Hagen brand. There are several other substitutes available. The ones that I currently use most are FilterMaxI/FilterMaxII or FilterMaxIII units (Endnote 2). The former is a very fine sponge and is best serviced more than once per week in this use, the latter is coarser and is practical with weekly service/rinsing in my tanks (Endnote 3).

Having solved the original problem of keeping gross particulates out of the canister, I was left with another issue. A significant part of the canister was packed with Eheim particulate (mechanical) filter material, in several stepped grades, to trap and hold particulates. I no longer had significant particulates going into the unit. Why was I canister filtering for such? Okay, next solution, convert the canister packing to pure biofiltration. The first material in my conventionally packed canisters was Ehfimech, noodle-like cylinders of porous ceramic, designed as a non-clogging coarse mechanical filter material. This material appeared to be suitable for attachment sites for nitrification bacteria, so I tried it for such use. Ehfisubstrat was not yet on the market. I also tried some large rounded gravel and Dupla’s Minikaskade balls (a small form of the Biokaskade ball I had used in my SW and later FW W/Ds, the smaller type is suitable for small W/Ds or for submersed use in other filters such as canisters, Reference 2). Later I did try Ehfisubstrat (a porous foamed glass material), but gave it up due the suggestion that it be replaced regularly. It does not self-clean of old biofilms, the internal pores clog, and the use of a fine pad is required downstream of this medium to avoid the possibility of fine glass particles doing impeller damage in the drive unit or being pumped back into the tank. To me, an ideal aerobic biological attachment substrate should not just provide good attachment sites for nitrifying bacteria; it should as well shed the over-grown films and re-colonize with bacteria from adjacent/nearby areas. Anyone who has used W/Ds knows that under the “dry” (it is not dry, just not submerged) filter bed or “column” a layer of debris builds very slowly, no matter how well the intake is pre-filtered. This layer has been shown to be mostly dead bacterial films, shed from the media and washed down to the sump base. The sand particles or small synthetic beads used in fluidized bed filters shed old film perfectly. From prior experience, I knew that either the Dupla spheres or Ehfimech shed quite well to satisfactorily also. I was a bit uncertain that the Ehfimech could support sufficient bacteria to service the biofiltration requirements of my tanks, as it was manufactured and marketed solely as a mechanical medium. All three test materials (Dupla spheres, large river gravel, and Ehfimech) served for more than sufficient biofiltration beds in practice. I soon abandoned the gravel, as it seemed to build up significant retained debris inside the canister, quite contrary to what I wanted. My best guess is that water flow was not as turbulent with the gravel as with either of the other materials, and that the gravel bed could channelize more readily. Unfortunately, the Dupla material may or may not be presently available in the US. That availability has varied over the years. Both of the manufactured materials serve nicely for “pure” biofiltration. Neither builds up any significant debris inside the canister over periods of six months to a year. The Dupla spheres are far the better of the two in self-cleaning. No doubt there are other formats of bio-substrates which will also serve. Using these materials in my canisters, I can readily maintain low-nitrate (< 10ppm) aaaaatanks just as I had previously done with RFUG or W/D, following my normal water change techniques (References 345). Note that shed films inside these canisters will be pumped back to the tank for capture by mechanical filters, as it is from fluidized beds. This is a very tiny fraction of the total particulate debris in an operating aquarium. It has never been noticeable.

To achieve the low nitrate levels I wanted, and having given up the mechanical filtration portion of the canisters other than the pre-filter sponge, I did need to add more but separate mechanical filtration. Otherwise the prefilter sponge would need more frequent rinsing or to be much larger. I already used small internal filters (both earlier Fluval models and several different Eheim models) in several of my tanks for current (especially the RFUG filtered tanks), so I adopted these in my canister-filtered tanks as well. I use these small units as sponge-only “pure” mechanical filters. No efforts are made to preserve any microbes that develop in these sponges. For me they are present only to remove particulates before they can be digested. I actually prefer that these sponges have no viable bacteria to digest captured particulates. These are not attractive tank additions I agree, but some degree of concealment is not difficult. I tend to use two small internals, one at each rear corner. One of these will have the output aimed parallel to the back wall of the tank, the other diagonally forward toward the opposite front center to the far front corner to establish a roughly circular current in the tank. Ordinarily the filters are set just enough below the surface to ripple it slightly without actually breaking the surface. This promotes gas exchange without wasting too much CO2 from planted tanks. On tanks with canister biofiltration, the canister spray-bar return may be aimed vertically downward by simple release below the water surface; similarly mounted below the water surface but aimed diagonally at the opposite face of the tank; by vertically mounted spray bar in one front corner aimed parallel to the front wall of the tank (maximum circulation); or by spray bar mounted horizontally a couple to a few inches above the substrate along the rear of the tank, aimed forward, straight upward, or sometimes back against the rear wall (that last for minimum current). My biofiltration in general is done by canister (Endnote 4), RFUG, or W/D, depending on the tank, or even by two of the three filter types listed in large or heavily stocked tanks. Most of my tanks, currently all of my tanks above 10 gallons (other than shrimp-only tanks), include small internal canisters. Most of these are for current and mechanical filtration only (Endnote 5).

Currently the biofiltration-only canisters are back-siphoned (Endnote 6) periodically to check for any particulates present. They are only opened at six to twelve month intervals to clean the impeller and housing, and to clean the flexible tubing if needed.

Should chemical filtration be needed, it can be incorporated in the canisters, or into the internals, or into the sump of W/Ds. In practice, if I need chemical filtration, I tend to use a separate canister unit for such. That practice means the least possible disruption of my normal mechanical and biofiltration processes. That is really only practical when you have a surplus of canisters already.

Therefore, if your aim is the best possible water conditions, dividing filtration processes into separate components may be a beneficial part of the technique. Regular partial water changes and filter maintenance on a schedule optimum for each filter type can assist you in your efforts. Why risk disturbance of your biofilter when it is not needed? Doing only those things that need doing is just as beneficial a technique as is adding only those things that need to be added to your water. This technique may not be practical in very small tanks (15 gallons or less), and for me is most useful from 20-30 US gallons and up. When used, it does assist significantly in my efforts to maintain high water quality.


(1) Water Change Math – General or,11523.0.html

(2) Water Treatment Systems by Dupla

(3) Water Changes, Why, How Much, and How Often? or,10624.0.html

(4) Water Change Math – General part 1 or,10630.0.html

(5) Water Change Math – General part 2 or,10631.0.html



Endnote 1: Mineralization is the conversion of organic materials to inorganic forms. Frequently, some to much of the resulting mineral content is dissolved in the tank water. Some remains undissolved as fine particulates, known in the hobby as mulm. The brownish material rinsed from filter pads is mulm. If the brown, cloudy rinse water is held undisturbed, it will settle out to being clear again, showing that the material is particulate, not dissolved. Mulm can be a significant reservoir of minerals (including nitrate) in somewhat neglected filters and neglected unplanted tank substrates. In planted tanks the material can be plant food, but can still build up to excess.

Endnote 2: Made by American Aquarium Products. The FilterMaxI is a small fine-pore sponge. FilterMaxII is the same sponge material, twice the size of the model I. FilterMaxIII is the large size again, but with coarser sponge. The largest units are available from the Hydro Sponge line.

Endnote 3: To simplify my life, I have spares of such prefilter sponges. It is short and easy work to pull the sponges off the intake and replace with a pre-cleaned sponge. The used sponge is rinsed later and then set aside for future use.

Endnote 4: When I used canisters as all-in-one units (for mechanical, chemical, and biological filtration, or full capacity use), I tended to scale units toward the lower end on the range suggested for that model. In other words, I chose a unit where the smallest tank volume suggested was closest to my tank. That is the most conservative choice. For biofiltration only, with no internal mechanical filtration at all, I tend to pick a canister model toward the maximum tank volume suggested, or even a bit beyond. That is done because flow-through in these biofiltration-only canisters is relative unimpeded. Having no mechanical filtration internally does remove much of the resistance to flow. Thus I select what in earlier, full capacity use of the unit, would have been the least conservative choice, but in this use is fully satisfactory.

Endnote 5: Current alone could be provided by the use of small powerheads only. That might well be less obtrusive in appearance. But if I am using power in any case, I prefer to let it do as much work as is practical. Mechanical filtration, frequently rinsed, comes at no significant operating cost difference, so is essentially a cheap to free bonus.

Endnote 6: Back-siphoning is done by unplugging the canister and disconnecting the canister intake from its prefilter and coupling the Python hose to the canister tubing, then opening the canister and Python flow valves to pull water back through the unit via the return or spray bar, through the canister, and out via the Python to waste. A bit of moderate shaking and tilting of the external canister during back flushing will likely free any particulates in the unit. Generally, this is performed with the first water change in the month.



Robert T. Ricketts, a.k.a. RTR


Please enter your comment!
Please enter your name here