Key takeaways
- Detritus settles at the slowest point of the circulation loop, so the fix is redirecting the loop rather than blasting the pile.
- Mount the pump about a third of the way down from the surface and angle it 10 to 15 degrees downward, never steeper than 20 on sand.
- Two powerheads facing each other cancel out and create a stagnant band, which is the single most common placement mistake.
- A pinch of flake food filmed on a phone and played back at half speed shows the return path along the substrate.
- Clearing dead spots sends more waste to the filter, so expect to clean mechanical media more often afterwards.
The instinct is to point the powerhead straight at the pile of detritus. That is almost always the wrong move, and it is why so many tanks end up with two pumps, a scoured patch of sand, and the same dead corner as before.
Detritus does not sit in a corner because nothing is pointing at it. It sits there because that is where the tank’s circulation loop slows down and drops what it was carrying. Blasting the pile scatters it for twenty minutes, then the same loop puts it back.
Read the tank before you move anything
Spend five minutes watching before you touch a mounting bracket. Look at where fine debris settles, which plant leaves stay clean, and which ones carry a dusty film.
Clean leaves mean moving water. Dusty leaves mean the flow there is below roughly 5 cm per second, which is about the point at which suspended organics start dropping out.
Note the pattern rather than the spots. In most tanks, the debris forms a line or an arc rather than a random scatter, and that shape is a map of the existing current.
The pinch of food test
Turn off the filter, wait thirty seconds, then drop a small pinch of flake food into the water near the surface at one end. Switch everything back on and watch where the flakes go.
Flakes are close enough to neutrally buoyant to trace the current honestly. Frozen brine shrimp works too, and so does a couple of drops of food colouring if you would rather not add food to the tank.
Watch three things: how far a flake travels before it stops, whether it ever reaches the far end, and where the flakes that fall out of the current end up. That last spot is your dead zone, and it is frequently not where the detritus visibly gathers.
Do the flake test with the lights on and a phone camera recording from the front. Playing it back at half speed shows the return path along the bottom, which is nearly impossible to follow in real time.
Why the debris ends up where it does
Water leaving a nozzle travels as a jet, entrains the water around it, slows, and eventually becomes a broad slow drift. Where that drift meets the glass it has to turn, and the turn costs it more speed.
Solids drop out at the slowest point of the circuit. In a standard rectangular tank with the filter output at one back corner, that point is usually the opposite front corner, or the strip of substrate immediately below the outlet where the jet has passed overhead without ever touching bottom.
The area directly under a powerhead is a classic dead spot for this reason. Flow shoots out and over, leaving a shadow beneath the pump that people rarely think to check.
Aiming patterns that work
Three arrangements cover the vast majority of freshwater and mixed tanks. Pick based on tank length, not on what looks tidy.
The closed loop
One pump at the far end from the filter intake, aimed low along the front glass, pushing water back toward that intake. The filter output travels one way along the back, the powerhead returns it along the front, and the tank runs as a single flat circuit.
This is the default for tanks between 90 and 150 cm. It clears both front corners and keeps solids moving toward the intake, where they belong.
The gyre
Two pumps on opposite end walls, both aimed along the same long wall in the same rotational direction, one high and one low. The whole body of water turns like a slow wheel.
Best on tanks over 120 cm and on anything deeper than 50 cm, since the vertical component is what stops stratification. It is also the pattern that controllable pumps on a freshwater tank are designed to produce and vary.
The wall bounce
One pump aimed at the opposite end glass rather than along the tank, letting the wall break the jet into a diffuse return. Useful for small tanks, shrimp tanks, and anywhere the direct current would be too strong.
It costs you reach. Do not expect a wall bounce to clear a corner more than 60 cm from the pump.
| Tank | Pattern | Pumps | Watch for |
|---|---|---|---|
| 60-90 cm standard | Closed loop | One, low output | Sand scouring near the nozzle |
| 90-150 cm standard | Closed loop | One or two | Front corners stalling |
| Over 150 cm | Gyre | Two, matched | Middle of the tank going slack |
| Deep or cube | Gyre, vertical | Two, offset in height | Warm layer at the surface |
| Shrimp or fry | Wall bounce | One, guarded intake | Livestock pinned to the glass |
Height and angle
Mount the pump about a third of the way down from the surface as a starting point, not at the top. High mounting ripples the surface and does very little for the substrate, which is where the waste is.
Angle it slightly downward, perhaps 10 to 15 degrees, so the jet reaches the lower half of the tank by the time it gets to the far end. Aim it dead level and it stays level, riding above the aquascape entirely.
Do not aim it into the substrate. Anything steeper than about 20 degrees on sand digs a crater and buries the plants downstream of it.
Common mistakes worth undoing
- Two pumps pointed at each other, which cancels both and creates a stagnant band in the middle
- A pump mounted in the same corner as the filter output, doubling flow along one path and starving the rest
- Flow aimed directly at the heater, which can fool a poorly placed thermostat
- A pump behind the hardscape, where 70 percent of its output hits rock 5 cm away
- Cables and suction cups blocking the intake grille, quietly halving output
The head to head arrangement is the most common of these and the most stubborn. It feels like more flow and produces less, because opposing jets convert their energy into turbulence rather than transport.
A 300 litre tank with a persistent corner
A 150 x 50 x 40 cm cichlid tank, external filter rated around 1200 litres per hour, output at the back right. The front left corner accumulated a visible grey drift within four days of every water change.
The owner had already added a powerhead aimed at the corner. It cleared the pile and dumped the same material along the front glass, where it settled in a longer, thinner line instead.
We turned that pump around. Mounted on the left end wall about 15 cm below the surface, angled 10 degrees down and aimed along the front glass toward the filter intake, it turned two half circuits into one full loop. The corner has stayed clean for eighteen months, and the filter now traps material that used to sit on the floor of the tank.
Clearing dead spots moves waste to the filter, which means the filter fills faster. Expect to clean mechanical media more often after a placement change, and take that as evidence the change worked rather than a new problem.
Retest, and give it time
Repeat the flake test 48 hours after any change, with the aquascape settled and the plants back in their normal position. A pattern that looks right on day one sometimes reveals a new shadow once leaves have leaned into the current.
Then leave it two weeks before judging. Detritus already on the substrate does not lift on its own, so vacuum the tank once and see whether it comes back.
If flow across the whole tank now looks weak rather than misdirected, the problem was never placement. Check total turnover against the target for your stocking level and, if the number is short, work through how pump capacity is matched to tank dimensions before buying anything.
When the pump itself is the problem
A powerhead that used to clear a corner and no longer does has usually lost output rather than aim. Impeller wear, a calcium crust on the magnet, and silt packed into the volute all cut flow gradually enough that you stop noticing.
Pull it and check. Stripping and cleaning the impeller assembly takes about fifteen minutes and often restores most of what was lost, and manufacturers such as Hydor publish exploded diagrams for their models.
Where a pump has stopped drawing water at all rather than simply weakening, that is a different fault with different causes, and a pump running dry or airlocked needs sorting before any placement work is worth doing.
Questions readers ask about this
On the end wall furthest from the filter intake, roughly a third of the way down from the surface, angled 10 to 15 degrees downward and aimed along the front glass. That closes the circulation loop and carries waste back toward the intake rather than stirring it up into the water column.
No. Opposing jets cancel one another, converting transport into turbulence and leaving a stagnant band where they meet. Point both along the same long wall in the same rotational direction, one mounted high and one low, so the whole body of water turns like a single slow wheel.
Drop a pinch of flake food in at one end with the filter briefly off, restart everything, and watch where it travels. Record it on a phone and play it back at half speed. Wherever the flakes fall out of the current is your real dead zone, which is often not where debris visibly gathers.
The jet leaves the nozzle travelling outward and passes overhead, leaving a shadow underneath the pump with almost no movement in it. Lowering the pump or angling it down 10 to 15 degrees usually clears the patch. Anything steeper than about 20 degrees will dig a crater in fine sand instead.
Retest with flake food after 48 hours, once the plants have leaned into the new current and the aquascape has settled. Then vacuum the substrate once and wait two weeks to see whether debris returns. Material already lying on the bottom will not lift by itself, so the first clean is part of the test.
