Why is my aquarium filter not working?

If the motor hums and nothing moves, and refilling the housing did not fix it, the impeller is the next suspect. It is a small magnet on a shaft, it sits in a chamber that collects slime and grit, and it is the only moving part in the whole device.

Most of the time it does not need replacing — it needs pulling out, rinsing, and reseating, because a single grain of substrate or a build-up of slime is enough to stop it turning. Replace it when the fins are visibly chipped, the magnet is cracked, or it still will not spin after a clean.

This is the assembly for the AquaClear 30, which is the hang-on-back this site recommends most often.

The honest costs. It is a part, not a diagnosis: a filter that hums and moves no water is more often air-locked than broken, and swapping an impeller that was never the fault wastes the money. Impellers are model-specific — the 20, 50 and 70 take different assemblies, so match the number on your filter before ordering. And the shaft it rides on is a separate part again, and is the piece that more often snaps.

A filter that is running, moving water, and simply moving less of it than it used to is almost never broken. It is blocked, and the block is in the first thing the water meets.

This is also the failure that does the most quiet damage, because reduced flow means reduced biological capacity long before anything looks wrong in the tank.

The right move is to rinse the existing mechanical media in old tank water first and only replace it when it is physically falling apart. Keep a pad in the cupboard so the choice is never between a clogged filter and a trip to the shop.

The honest costs. It is a mechanical medium only — it polishes water and does nothing biological or chemical. You cut it to fit, which means a pair of scissors and a slightly annoying five minutes. Packed too tightly it becomes the clog rather than the cure. And it is a consumable, so this is a recurring cost in a way that rinse-and-reuse foam is not.

Every fault on this page is a fault of powered filters: losing prime, jamming an impeller, air-locking after a clean. A sponge filter has none of those parts and therefore none of those failures.

Running one permanently alongside your main filter means that the hour you spend with the hang-on-back in pieces is an hour the tank is still being filtered, and that the bacteria colony is not sitting entirely in a device you have just opened.

It is also the cheapest insurance against the failure that is not the filter's fault at all — a power cut, covered on our outage page.

The honest costs. No pump, airline or check valve in the box, so the real purchase is three or four items and the pump is the noisiest thing in the setup. It is visible inside the tank. No mechanical polish, so fine particles stay suspended. And it is a nano device: on a stocked 20 or 30 gallon tank it supplements a main filter rather than replacing one.

Worth owning for a second reason this page cares about: if the filter is going to be off for more than an hour or two, the thing the fish actually need is surface movement, not filtration.

A small pump with an air stone covers that for the length of any repair, and costs less than the part you are probably about to order.

The honest costs. No airline tubing and no check valve included, and the check valve is not optional — a stopped pump will siphon tank water down the airline onto the floor. No flow control. And the diaphragm is the wearing part, so expect to replace the unit rather than service it.

A motor unit that is seized, burnt out or has been run dry long enough to cook is not economically repairable on a filter at this price, and that is the point to stop troubleshooting.

The reason this is the default replacement rather than the cheapest one is the basket: you rinse and reuse the media instead of buying cartridges, which is the difference between a filter that costs nothing to own and one that costs a few dollars a month forever.

If you are replacing rather than repairing, move as much of the old media as you can into the new filter. That media is your bacteria colony, and transferring it is the difference between a seamless swap and a small cycle.

The honest costs. At full output on a 10 gallon tank it is more current than a planted nano or a betta wants, so the flow knob is not optional there. No published minimum flow figure, so you cannot check where that knob lands. The cascade return is a surface-churner by design. And rated flow is measured with clean media and no head height, so the real figure is lower — our method divides by 0.7 to allow for it.

Included because a filter failing is an unplanned expense, and the honest cheapest option that still works belongs on the page rather than only the one we would pick ourselves.

On a lightly stocked 20 gallon tank it does the job, and the base plate that braces it against the glass is a genuine contributor to how quietly it runs.

The honest costs. It is cartridge-based, so every replacement discards the bacteria living on that cartridge and the recurring cost never ends. No published minimum flow. And at 130 GPH on a fully stocked 20 it is short of the 170 our method asks for, so treat it as a filter for a lightly stocked tank.

The skip entry, and it is here because it is the single most common wrong response to this page's problem.

Flow drops, the owner looks at a dirty cartridge, and the instructions on the box say replace monthly. So they replace it — and discard, in one motion, a large share of the nitrifying bacteria that were keeping the tank safe. The flow comes back. A week later the tank has an ammonia reading nobody can explain.

The correct move with a clogged cartridge is to rinse it in **old tank water** and put it back, repeatedly, until it is physically disintegrating. A cartridge that looks terrible and still passes water is doing two jobs; a brand-new one is doing one.

If you must replace it, do it in stages — new alongside old for a few weeks — so the colony has somewhere to live in the meantime.

The honest costs of the skip. Cartridges are genuinely convenient, and a keeper who will not do media maintenance may get better results from a cartridge they actually change than a basket they ignore. The objection is to replacing on a calendar rather than on condition, and to doing it as a response to weak flow.