Key takeaways
- A heater mounted in still water makes its own warm pocket, so the thermostat shuts off early and the far end of the tank runs cold.
- Beside the filter intake strainer or directly in the outflow are both correct; a hidden back corner is the worst position available.
- Leave at least 3cm between the tube and the substrate, and 2cm between the tube and the back glass.
- Tanks longer than about 120cm need two heating positions rather than one, whatever the total wattage.
- Under 0.5C measured across four points in the tank means the placement is right; over 1C means the heater is sitting in dead water.
A 120cm community tank, one 200W heater, and a 1.9C spread from one end of the glass to the other. The end nearest the heater held 25.8C (78F) and the far corner behind the driftwood sat at 23.9C (75F).
Nothing on that tank was faulty. Correct wattage, accurate thermostat, a probe thermometer that had been checked against an ice bath. The heater was hanging in the wrong twenty centimetres of glass.
Position is the cheapest thing you can change on a heating setup and the last thing anyone checks, because a heater looks equally busy wherever you stick it.
The element needs water moving past it
A heater warms the thin skin of water touching its tube. In still water that skin lifts away as a narrow plume and the rest of the tank waits on slow convection to even things out, which takes hours in a long tank.
The thermostat sits inside that plume. It reads the hottest water in the box, concludes the job is finished, and opens the contacts while the far end is still short of target.
Hang the same unit in a current and both faults disappear together. Heat leaves the glass as fast as the element makes it, the sensor reads something close to the true tank average, and the heater settles into shorter, more frequent cycles.
Beside the outflow or beside the intake
Either works. Both put the element in the fastest water available, which is the only thing that genuinely matters, and choosing between them comes down to small tradeoffs.
Mounting in the outflow gives the quickest mixing, since warmed water is thrown down the length of the tank the moment it leaves the tube. Mounting beside the intake strainer routes that water through the filter first, which warms the media a little and returns it at the opposite end.
On a canister setup I lean toward the intake side. It keeps both pieces of hardware in one corner and leaves the outflow free to drive surface movement. With a hang-on-back filter, put the heater directly beneath the return lip.
| Position | Mixing | What the thermostat reads | Verdict |
|---|---|---|---|
| In the filter outflow stream | Fastest | Close to tank average | Best on long tanks |
| Beside the intake strainer | Fast | Close to tank average | Best with canisters |
| Under a powerhead or air stone | Good | Slightly above average | Workable |
| Mid panel with no flow nearby | Poor | Its own warm plume | Expect roughly 1C of error |
| Rear corner behind hardscape | Almost none | A hot pocket it created | The worst option in the tank |
Horizontal, vertical or angled
Horizontal and low is the strongest choice for even temperature. Heat then rises through the full water column instead of leaving from one point near the surface, so convection works for you rather than against you.
Vertical is perfectly fine and is what most suction cup brackets encourage. It costs a little mixing and gains a lot of convenience, which is a fair trade on any tank under about 80cm.
Angled at roughly 45 degrees with the cap uppermost is the compromise pictured on most boxes. Read the label before committing to a flat mount: a few designs place the thermostat capsule immediately below the cap and specify a maximum tilt, and lying those down will make them read wrong. Manufacturer mounting diagrams, including the ones published by Fluval, are worth two minutes before you fit anything.
Whatever the angle, the whole element must stay submerged, including after a fortnight of evaporation. Every heater carries a minimum immersion mark on the tube. A tube running part dry does not fail gently: the glass cracks, and it usually cracks at the moment cool water reaches it again.
How far off the substrate
Leave at least 3cm between the bottom of the tube and the gravel, and 5cm is better. Substrate insulates, so an element resting in it holds its own heat against the glass instead of shedding it into the water.
Sand is worse than gravel because it packs and it drifts. Two of the cracked heaters on my bench last year had a scorched band exactly where 20mm of sand had piled up around them over a few months.
There is a livestock reason as well. Corydoras, kuhli loaches and sleeping plecos will rest against anything warm, and a bare element leaves a mark. That is a large part of the argument for fitting a guard on any tank with bottom dwellers.
And how far off the glass
Two centimetres of clearance behind the tube lets water circulate all the way round it. Cranked flat against the rear pane with tight suction cups, one entire side of the element is heating glass instead of water.
On thin panels that is a small but real stress point. More often it simply produces the permanent white scale stripe on the inside of the back glass that no scraper ever shifts.
Why the far corner is the worst spot in the tank
People put heaters in the back corner because it hides them. On almost every tank that corner is also the deadest patch of water in the box.
Return flow runs down the length of the tank and turns before it reaches the space behind the intake. Hardscape and a season of plant growth make it worse. The thermostat then sits in a pocket of its own warmth, so the unit short cycles while the rest of the tank runs cool.
On tanks over 100cm long I measure spreads of 1.5C to 2C from that single mistake. It is also the corner you cannot see properly, so a fogged tube or a lifting seal goes unnoticed for weeks.
Long tanks need two positions, not one
Past roughly 120cm, no single position serves both ends. A 150cm tank with everything at the left will run a gradient no matter how good the filter is.
Divide the wattage and put one unit at each end, both in flow, both set to the same figure. Spread usually drops below 0.5C, and you pick up the failure protection that makes running a pair rather than one large unit worth the second socket.
If a gradient survives that change, the problem is circulation rather than heating. Check whether the turnover suits the tank volume before you buy any more heaters, and check that the return is not aimed straight into a wall of stem plants.
Sumps, canisters and getting it out of the display
With a sump, the heater belongs in the return chamber, in the flow, ahead of the return pump. Never in the drain chamber: the level there moves with the siphon, and an element exposed by a falling level is the classic sump failure.
An inline heater on the canister return hose removes the question altogether. The unit lives in the cabinet, every drop of circulated water passes through it, and the display glass stays empty. The differences between glass, titanium and inline designs are mostly about where you want the sensing to happen.
Proving the position was right
One thermometer moved around beats four thermometers you have never checked against each other. Leave the tank alone for six hours after any change, then read four points with the same probe, five minutes at each.
- Mid water, about 10cm from the heater
- Mid water at the opposite end of the tank
- Just above the substrate at the far end
- Just below the surface at the far end
Under 0.5C across all four is a well placed heater. Between 0.5C and 1C is worth improving. Over 1C and the unit is either sitting in dead water or is too small for the room, which sends you back to the wattage the tank actually needs.
A 90cm tank fixed by a 20cm move
A 160 litre (42 gallon) planted tank, 90cm long, 150W heater standing vertically in the right rear corner behind a bogwood root. The owner reported that the corydoras spent every day piled at the right hand end and the crypts on the left had stopped growing.
Measured spread was 1.4C. We shifted the heater about 20cm so it stood beside the canister intake strainer, still hidden, still 6cm clear of the sand. Same heater, same dial position, same filter, nothing else touched.
Six hours later the spread read 0.3C, and the tank average had risen 0.4C without anyone adjusting a setting. The fish redistributed within two days. The repair cost nothing at all.
Mistakes worth not making
- Heater tucked in a corner with no flow past it
- Tube resting on, or slowly buried by, the substrate
- Suction cups pulling the element flat against the rear pane
- Part of the element above the waterline after a fortnight of evaporation
- Thermometer mounted in the same corner as the heater, so a gradient can never show up
- Heater left in place during a water change that drops the level past the tube
That last one accounts for more cracked tubes than anything else. Switch off before the water change, wait fifteen minutes, and lift the heater out if the level is going below it. A dial reading true on a properly calibrated unit counts for nothing once the tube is sitting in air.
Questions readers ask about this
Both are good. The outflow mixes fastest because warmed water is thrown straight down the tank. The intake side routes it through the filter first and keeps your hardware in one corner, which suits canisters. What matters is that the element sits in moving water rather than a still pocket, so pick whichever is easier to service.
Horizontal low in the tank gives the most even heating, but keep at least 3cm of clearance above the substrate and check the manual first. A few heaters place the thermostat capsule just below the cap and state a maximum tilt, and laying those flat makes them read the wrong temperature by a degree or more.
Two, positioned at opposite ends, even if a single unit would supply enough watts. Anything past about 120cm cannot be served from one point, and a 150cm tank heated from one end typically shows a 1.5C to 2C gradient. Split the total wattage between two matched units and both should be sitting in flow.
Very much. Put the thermometer at the opposite end from the heater and out of direct flow, mid water. A thermometer sitting beside the heater reads the warm plume, agrees with the dial, and hides any gradient in the tank. If you only own one, that far position is the one that tells you something useful.
Only if the water level stays above the immersion mark on the tube. Switch it off at the wall first and wait about fifteen minutes so the glass cools. An element left powered in air, or a hot tube meeting refill water, is the most common way people crack a heater that had nothing else wrong with it.
