Top 10 Hang-on-Back Filters for 20-Gallon Aquariums in 2027

The tank that keeps clouding up three days after every water change

Picture a standard 20-gallon high — 24 inches long, 12 wide, 16 tall — stocked with eight harlequin rasboras, six corydoras, and a small bristlenose pleco. The tank has been running four months on the filter that came bundled in the starter kit: a cartridge-style unit rated "up to 30 gallons," one slot, one disposable pad of floss wrapped around carbon granules. Every four weeks the instructions say to pull the cartridge and drop in a fresh one. The owner does exactly that, and every time, roughly 48 to 72 hours later, the water goes hazy and a test strip shows ammonia at 0.25 ppm.

This is not a mystery, and it is not a sick tank. It is the single most common failure mode of cheap hang-on-back filtration, and it is structural rather than accidental. In a 20-gallon community tank, the overwhelming majority of nitrifying bacteria — *Nitrosomonas* converting ammonia to nitrite, *Nitrospira* and relatives converting nitrite to nitrate — live on filter media, not in the water column and not primarily in the gravel. The filter is the tank's biological engine. When the entire media volume is a single disposable pad and that pad gets replaced wholesale, the colony goes in the trash with it. What follows is a mini-cycle: a bacterial bloom (the haze is heterotrophic bacteria feeding on dissolved organics that the nitrifiers are no longer outcompeting) plus measurable ammonia until the colony rebuilds over one to three weeks.

The same tank with a basket-style filter behaves completely differently. Coarse sponge stays in the basket for two to three years, rinsed in dechlorinated tank water every four to six weeks. Ceramic or sintered-glass biomedia sits underneath and gets rinsed maybe twice a year and replaced almost never. Carbon, if used at all, is a small pouch swapped monthly — and swapping carbon removes essentially no bacteria because carbon is the last thing water passes through and holds a trivial fraction of the total colony. Nothing that matters ever leaves the tank.

The scenario is worth spelling out because it reframes the whole buying decision. Hang-on-back filters look interchangeable on a shelf: a plastic box, an intake tube, a waterfall lip. The specification that actually determines whether your tank is stable is *whether the design forces you to discard biological media on a schedule*. Everything else — noise, price, brand, flow rating — is secondary to that one architectural fork. A $25 basket filter will hold a 20-gallon tank more stably than a $45 cartridge filter, because one of them is designed around consumable revenue and the other is designed around keeping media in the tank.

How a hang-on-back filter actually moves and cleans water

The mechanism is simple enough to reason about precisely, which is useful because it tells you which specs matter.

An impeller — a small magnet-driven rotor in a wet chamber at the bottom of the unit — spins and creates suction. Water is drawn up an intake tube ("the lift tube") from near the tank floor, pushed into the media chamber, forced through whatever media sits in the flow path, and then spills back over a lip into the tank. That spill is the aeration step: the waterfall breaks the surface, drives off dissolved CO₂, and pulls oxygen in. In a 20-gallon tank, that surface agitation is usually the *only* gas exchange the tank gets, which is why a hang-on-back filter doubles as your aeration system and why an air stone is generally redundant alongside one.

Water passes through three functional stages, though they're physically stacked in one box:

Mechanical. Sponge or floss traps particulates — uneaten food, plant fragments, fish waste. This is what makes water look clear. It is also the stage that clogs, and clogging is what causes the flow-rate decline you notice around week four. Coarse sponge in front, fine floss behind, is the correct order: coarse catches the big stuff and protects the fine layer from blinding over in a week.

Chemical. Activated carbon adsorbs dissolved organics, tannins, medication residue, and some odors. Carbon is genuinely optional on an established tank. It exhausts in roughly three to six weeks and after that it is just an expensive sponge with a large surface area — which is why leaving spent carbon in place is harmless, it simply becomes biomedia. The one time you want fresh carbon is after dosing medication, to strip the remainder out.

Biological. Porous ceramic rings, sintered glass, or the sponge itself hosts the nitrifying colony. Surface area and steady oxygenated flow are the only two things this stage needs. It does not need to be replaced, cleaned aggressively, or "recharged."

The pump is the failure point over a multi-year horizon. Impeller shafts wear, impeller wells accumulate calcium and biofilm, and the rubber bushings at either end of the shaft degrade. A hang-on-back filter that suddenly rattles or loses flow almost always needs the impeller pulled, the well swabbed with a cotton bud, and possibly a $6 replacement impeller — not a new filter. Any unit whose impeller is not user-removable should be disqualified on that basis alone.

Two consequences fall directly out of this diagram. First, media order is not cosmetic — putting carbon or biomedia in front of the sponge means they silt up with debris and lose function within weeks, which is why a filter with a *customizable* basket beats one with a fixed cartridge slot. Second, priming matters. A hang-on-back unit sitting above the waterline has to fill its chamber before the impeller can move water. Self-priming designs hold enough water in the pump housing to restart on their own after a power cut; non-self-priming designs run dry, whine, and can burn out a motor if the power comes back while you're at work. For anyone running a tank in a region with frequent brief outages, self-priming is the difference between an inconvenience and a dead pump.

Sizing, flow, and the numbers that actually apply to 20 gallons

The rule of thumb worth internalizing: target roughly 4× to 6× tank volume per hour in turnover for a normally stocked community 20-gallon, and 6× to 10× for a heavily stocked or messy-fish setup. That translates to a practical target of 80 to 120 GPH for a planted or peaceful community tank, and 120 to 200 GPH for goldfish, cichlids, or a tank with a pleco producing serious waste.

Notice what that implies about box ratings. A filter labeled "up to 20 gallons" typically pushes 80–100 GPH, which is the *floor* of the useful range with zero headroom. A filter labeled "20–50 gallons" typically pushes 150–200 GPH, which after real-world flow loss lands right in the sweet spot. Buy one size up and throttle down is the correct move, for three reasons:

1. Manufacturer GPH is measured with clean media, no head height penalty, and no intake sponge. Real installed flow is commonly 25–40% below the box number once media is loaded and partially clogged.
2. A larger unit has a larger media chamber. Media volume is the real currency — a chamber holding 2–3 cups of media supports a far more resilient colony than one holding 0.5–1 cup.
3. Running a larger pump throttled to half is quieter and puts less wear on the impeller than running a small pump at full tilt continuously.

Media capacity by class, for calibration: compact nano units in the "up to 15 gallon" bracket generally hold well under a cup of media. Mid-size units rated 20–30 gallons hold roughly 1 to 1.5 cups. The larger 20–50 gallon class holds around 2 to 3.5 cups. On a 20-gallon tank, anything under about 1.5 cups of usable media volume should be treated as a supplemental filter rather than the primary one.

Noise. Quality hang-on-back units run in the high-20s to low-30s dB range at a meter's distance — quieter than a refrigerator hum. In practice, perceived noise on a 20-gallon has less to do with the model and more to do with three installation variables: the water level in the tank relative to the outflow lip (a lower tank level means a longer, louder waterfall drop), whether the filter body is touching the rim glass and transmitting vibration, and whether the impeller well is clean. Raising the tank water level to within a half-inch of the lip and slipping a strip of foam or a rubber band between the filter housing and the rim eliminates most complaints without spending anything.

Cost math over three years, which is where the basket-versus-cartridge decision pays off. A cartridge filter using proprietary pads at roughly $3–5 each, replaced monthly, runs about $36–60 per year, or $108–180 over three years — on top of the unit price. A basket filter running reusable coarse sponge (replaced maybe once every two to three years) plus a bag of ceramic media (effectively permanent) plus optional bulk carbon costs on the order of $10–25 per year. Over three years the basket unit is typically $80 to $140 cheaper in total ownership even if it costs $20 more up front. That gap is the entire business model behind cartridge systems, and it is why proprietary-consumable designs consistently rank below basket designs in any list built around long-term keeping rather than shelf price.

Head height and placement. Hang-on-back units lose flow as the vertical lift between water surface and outflow increases. On a standard 20-gallon high with the water topped off, that lift is minimal. On a 20-gallon *long* (30 inches long, 12 tall) the lift is smaller still. But if you keep the water line two inches low — common when people fear splashing — you add lift, lose flow, and increase noise simultaneously. Keep it topped up.

What you give up with each design, and when a different filter type wins

No single hang-on-back unit is correct for every 20-gallon tank, and the honest framing is a set of trade-offs rather than a ranking.

Basket-with-modular-media versus cartridge. Baskets win on cost, biological stability, and customization. Cartridges win on convenience — genuinely, for someone who will not do maintenance, a cartridge system that gets changed on schedule beats a basket system that gets neglected into a clogged brick. If you know you are the "set it and forget it" type, a cartridge filter with a *supplementary* sponge or bio-media insert added behind the cartridge is a reasonable compromise: the cartridge handles mechanical duty and gets tossed, the added biomedia stays put and preserves the colony.

Bio-wheel designs. A rotating wheel driven by the outflow exposes bacteria to air and water alternately, which is genuinely effective nitrification per unit volume. The trade-offs are real though: wheels can seize when the axle grooves wear or when debris binds them, a seized wheel is a dead wheel biologically, and the ratcheting sound of a wheel with a worn axle is the most common noise complaint in this category. Bio-wheels also handle plant debris poorly in a heavily planted 20-gallon.

Surface skimming. A skimmer intake pulls from the top film rather than only mid-water. In a planted 20-gallon with modest surface agitation, protein film builds up within days and blocks gas exchange. A built-in adjustable skimmer solves this; a bolt-on skimmer attachment solves it for a small additional cost on units that don't have one. If your tank is planted or you run a low outflow for shrimp, treat skimming as close to mandatory.

Sponge filters as a companion, not a rival. An air-driven sponge filter costs very little, has enormous biological surface area, produces almost no current, and is completely safe for fry and shrimp. It does poor mechanical polishing — water stays slightly hazy — and it needs an air pump, which is its own noise source. The strongest 20-gallon setup for a shrimp or fry tank is often a modest hang-on-back for polishing plus a sponge filter for biology and safety.

Canister filters. A small canister on a 20-gallon gives you far more media volume, near-silent operation, and hidden plumbing. It costs substantially more, takes 20–40 minutes to service versus 5 for a hang-on-back, and introduces leak risk from hose connections. On 20 gallons it's an aesthetic and enthusiast choice, not a necessity.

Internal filters. Fully submerged units are quiet and immune to priming failures, but they occupy tank volume you can't spare in a 20-gallon and are more awkward to service.

Where 20-gallon setups go wrong, and the fixes

Replacing all media at once. The cardinal error, and the cause of the clouding scenario above. Rule: never change more than one media type in a single session, and never rinse sponge and swap carbon on the same day. If a sponge is genuinely disintegrating, cut the new one in half, run half-new/half-old for four weeks, then remove the remainder.

Rinsing media under the tap. Chlorine and chloramine in municipal water kill nitrifiers on contact. Always rinse in a bucket of water siphoned from the tank during a water change, or in dechlorinated water. Squeeze the sponge three or four times until the water runs murky-but-not-black — you are removing sludge, not sterilizing.

Over-cleaning. A filter that looks dirty is a filter that is working. Rinse mechanical media on a four-to-six week cadence, or when flow has visibly dropped, whichever comes first. Scrubbing biomedia until it looks new destroys the colony you spent two months growing.

No intake pre-filter with small livestock. A bare intake tube on a 20-gallon will take in shrimplets, fry, and the occasional small fish, and will jam with plant leaves. A coarse sponge pre-filter slipped over the intake fixes all three problems, adds meaningful biological surface area, and costs a few dollars. Rinse it every water change — it will load up far faster than internal media.

Flow too high for the livestock. Bettas, some gouramis, and shrimp genuinely struggle in strong current, and a 200 GPH outflow in a 20-gallon high creates a washing-machine effect. Fixes in order of preference: use the flow adjustment lever; redirect the outflow along the back glass rather than into open water; place a sponge or a piece of filter floss against the outflow lip to diffuse the fall; or attach a spray bar if the model supports one. Do *not* restrict the intake to reduce flow — starving an impeller of water increases wear and heat.

Not topping off the water line. Evaporation on a 20-gallon uncovered tank runs meaningful over a week. As the level drops, the waterfall gets louder, splash increases, and flow drops. Top off weekly with dechlorinated water — evaporation removes water, not minerals, so top-offs should be plain water, not a partial water change.

Assuming the filter cycles a new tank on its own. A brand new filter on a brand new tank has no bacteria in it. It needs an ammonia source and four to six weeks, or a seeded sponge from an established tank to shortcut the process. The fastest legitimate accelerator is transplanting a handful of established ceramic media or a used sponge into the new basket.

Buying on flow rating alone. Two units advertising the same GPH can differ by 3× in media volume. Compare chamber capacity in cups before comparing GPH, and treat both alongside whether the impeller is user-serviceable and whether replacement parts are still sold. A discontinued model with no available impeller is a two-year filter regardless of how good it is today.

Running carbon permanently in a planted tank. Carbon adsorbs some trace nutrients and, more importantly, occupies basket space that would host far more biomedia. In a planted 20-gallon, pulling carbon and filling that compartment with additional sponge or ceramic media is a straight upgrade.

Related questions

What GPH do I actually need for a 20-gallon tank?

Aim for 80–120 GPH for a normal community tank and 120–200 GPH for messy or heavily stocked setups. Since real-world flow runs 25–40% under the box rating once media loads up, buying a unit rated for 30–50 gallons and throttling it down is the reliable approach.

Can I run two hang-on-back filters on one 20-gallon?

Yes, and it's a good redundancy strategy. Split them across opposite ends of the tank to eliminate dead spots, stagger their maintenance by two weeks so you never disturb both colonies at once, and throttle both down so combined flow stays in your target range.

Should I keep carbon in the filter all the time?

Not necessarily. Carbon exhausts in roughly three to six weeks and mainly earns its place after medicating a tank or removing tannins. On an established, planted 20-gallon, replacing the carbon compartment with extra sponge or ceramic biomedia gives you more useful capacity.

How do I make a hang-on-back filter quieter?

Top the tank water up to within half an inch of the outflow lip to shorten the waterfall drop, put a foam strip or rubber band between the housing and the rim glass to kill vibration transfer, and pull the impeller to clean the well and check for a worn shaft.

Is a sponge filter enough on its own for 20 gallons?

Biologically, often yes — sponge filters have excellent surface area and are safe for shrimp and fry. What they don't do is polish water mechanically, so it stays slightly hazy. Pairing a sponge filter with a modest hang-on-back gives you both stability and clarity.

FAQ

How often should I clean a hang-on-back filter on a 20-gallon aquarium?

Rinse mechanical sponge or floss every four to six weeks, or sooner if flow visibly drops. Swap carbon monthly if you use it at all. Rinse ceramic biomedia lightly once or twice a year and replace it essentially never. Always rinse in tank water or dechlorinated water, never under the tap, and never service mechanical and biological media in the same session.

Can I use a filter rated for 50 gallons on a 20-gallon tank?

Yes, and it's usually the better choice — provided the unit has working flow adjustment. Oversized units give you a larger media chamber, which is the main driver of biological stability. Throttle the flow down to roughly 100 GPH for a community tank. The only setup where an oversized unit is a problem is one without adjustable flow paired with low-current livestock like bettas or shrimp.

Why does my tank get cloudy after I change the filter cartridge?

Because the cartridge held most of your nitrifying bacteria, and replacing it wholesale triggered a mini-cycle. The haze is a heterotrophic bacterial bloom, and you'll usually see a small ammonia spike alongside it. The permanent fix is switching to a basket-style filter with reusable media, or adding a sponge insert behind the cartridge that stays in place when the cartridge is swapped.

Do I still need an air stone if I run a hang-on-back filter?

Generally no. The waterfall outflow breaking the surface provides the gas exchange a 20-gallon needs. Add supplemental aeration in specific circumstances: sustained high water temperature above roughly 82°F, heavy medication dosing that reduces oxygen availability, or a very heavily stocked tank. If you throttle the outflow way down for shrimp, add a sponge filter to restore surface movement.

What's the difference between a cartridge filter and a basket filter, and does it matter?

A cartridge filter has one fixed slot for a proprietary disposable pad. A basket filter has an open chamber you pack yourself with sponge, carbon, and ceramic media in whatever order you choose. It matters enormously: the basket design lets biological media stay in the tank permanently, and it typically saves $80–140 in consumables over three years compared to buying proprietary cartridges monthly.

My filter stopped pulling water after a power outage — is it broken?

Usually not. Non-self-priming units lose the water in their chamber when the pump stops and can't restart dry. Pour a cup of tank water into the filter body to reprime it and it should catch immediately. If it whines but doesn't move water after priming, pull the impeller, clean the well and shaft with a cotton bud, and reseat it. A replacement impeller is an inexpensive part on any model still supported by its maker.

Sources

• EPA — Aquatic Life Criteria for Ammonia
• University of Florida IFAS Extension — Ammonia in Aquatic Systems
• University of Florida IFAS Extension — Biological Filtration in Recirculating Aquaculture
• USGS Water Science School — Dissolved Oxygen and Water Quality
• CDC — Drinking Water Disinfection with Chlorine and Chloramine
• NOAA Fisheries — Recirculating Aquaculture Systems
• Michigan State University Extension — Nitrogen Cycle in Aquaculture Systems
• Fluval — AquaClear Power Filter Product Line
• Seachem — Tidal Power Filter

Related on PULSE

• [Top 10 Hang-on-Back Filters for 20-Gallon Freshwater Aquariums](/knowledge/aq0765)
• [Top 10 Hang-on-Back Filters for 20-Gallon Long Tanks](/knowledge/aq0499)
• [Top 10 20-Gallon Aquariums 2027](/knowledge/aq0023)
• [What size aquarium heater is safe for a 20-gallon tall tank?](/knowledge/aq0487)
• [Top 10 Hang-On-Back Filters for Nano Reef Aquariums](/knowledge/aq0932)
• [Top 10 Hang-on-Back Filters for Betta Tanks Under 10 Gallons](/knowledge/aq0896)