Why Your New Tank’s Ammonia Spike Won’t Drop After Two Weeks of Cycling

The bottle of Dr. Tim’s Ammonium Chloride sat on the shelf for six weeks before the owner of a sixty-gallon tank in a rented apartment in Bukit Batok finally understood what it was there to do. Not just dose the water—that part was easy—but teach patience through the worst kind of failure. The first batch of test strips had turned a deep, alarming green within three days. By day five, ammonia was at 4 ppm, nitrites barely visible, nitrates absent. Textbook, supposedly. By day fourteen, nothing had changed. Same green. Same frustration.

What happened in that tank happens in hundreds of others across Singapore every month. The ammonia climbs, then stays. The owner waits. The internet offers contradictory advice: add more bacteria, add less bacteria, do a water change, never do a water change, wait longer, it’s broken. None of it helps because the real problem isn’t on the forum threads. It’s physical, and it’s hiding in plain sight.

A foam block rated for thirty gallons

Most beginners buy the bottled bacteria and assume the job is done. A capful goes in. The bottle goes back on the shelf. The bacteria are supposed to multiply. But a capful of bottled bacteria contains a finite number of cells, and those cells need somewhere to live. If the tank’s filter media is sparse—a single foam block in a small HOB filter, or a canister packed mostly with ceramic rings that arrived dry and dusty—there isn’t enough surface area for a colony to establish fast enough to process 4 ppm of ammonia daily.

The owner of a sixty-gallon nano reef setup in a Jurong HDB once spent three weeks watching ammonia refuse to budge before pulling the filter apart and discovering the foam was rated for a thirty-gallon tank. Not a defect. Just the wrong part. The solution was doubling the media volume and adding a mat of coarse sponge cut to fit the intake. Ammonia dropped to zero within ten days after that.

For a freshwater tank, the same principle applies. Sponge, ceramic media, even a handful of inert lava rock in a mesh bag—whatever offers surface area. The bacteria don’t care what the media looks like. They care how much of it there is.

Twenty-five degrees, not twenty-eight

Air-conditioned apartments in Singapore run at about 24°C. A tank in the living room, next to the AC vent, might hold at 25°C. That’s comfortable for fish. It’s slow for bacteria. Nitrosomonas and Nitrobacter species, the workhorses of the nitrogen cycle, operate faster at 28–30°C. Not dramatically faster. But over two weeks, the difference between 25°C and 29°C can mean the difference between a spike that’s falling and one that’s stalled.

The obvious fix is a heater. Most beginners buy one anyway, for the fish. But many don’t turn it on during cycling because they assume the bacteria don’t need warmth the way livestock will. Set the heater to 28°C before adding ammonia. Let it stabilize for a day. The difference shows up in the test results within a week.

One aquarist in Tiong Bahru kept detailed logs across three identical ten-gallon tanks started on the same day. Tank A ran at 25°C. Tank B at 28°C. Tank C at 30°C. Tank B cycled in eleven days. Tank A took nineteen. The third tank, at 30°C, finished in nine days but developed a cyanobacteria bloom that required manual removal—a reminder that faster isn’t always cleaner.

A bag of crushed coral in the sump

The ammonia spike that won’t drop often isn’t an ammonia problem at all. It’s a pH problem. Adding ammonia to water, particularly in a tank with no buffering capacity, drives pH down over time. Soft water—common in Singapore’s tap supply, which comes from reservoirs rather than groundwater—drops faster than hard water. At pH below 6.5, the bacterial colony slows. At pH below 6.0, it nearly stops.

The test kit still reads high ammonia because total ammonia includes both toxic NH₃ and ionized NH₄⁺. At low pH, most of that reading is the harmless form. But the bacteria can’t use NH₄⁺ as efficiently, so the number stays high while the colony struggles to process what it can.

A reef keeper in Kembangan solved this by adding a small bag of crushed coral to the filter sump. The coral dissolved slowly, stabilizing pH around 7.8. Ammonia dropped within a week. The same trick works in freshwater tanks, though the target pH is lower—7.0 to 7.5 is fine. A teaspoon of baking soda dissolved in tank water and added slowly can also lift pH temporarily, but it’s a temporary fix, not a long-term strategy.

Three days with the lights off

A tank cycling in a brightly lit room, with the light on twelve hours a day, will develop algae. The algae consumes ammonia—this part is true—but it doesn’t consume it fast enough to eliminate the spike, and it doesn’t help the bacterial colony establish. Worse, the algae competes for available surface area and oxygen, particularly in the daytime when photosynthesis raises dissolved oxygen but also supports biofilm growth that can physically block filter media.

A common mistake: leaving the light on during cycling because the owner wants to watch the process. The owner of a planted tank in Pasir Ris ran lights on a ten-hour schedule through a six-week cycle before realizing the algae was the primary consumer of the ammonia, not the bacteria. Every time the ammonia level dropped, a water change would reset the cycle because the algae had outcompeted the bacteria. The fix was a three-day blackout—lights off, tank covered—followed by reduced lighting to six hours daily. The bacterial colony took over within two weeks after that.

Hardware-store ammonia with detergent in it

Not all ammonia sources are equal. The most common method is adding pure ammonium chloride, available from aquarium stores or online. But some hobbyists use household ammonia, which sometimes contains surfactants or fragrances that inhibit bacterial growth. A friend-of-a-friend story that circulates among the older keepers in the Singapore Aquarist Society involves a man who used a brand of clear ammonia from a neighborhood hardware shop, cycled for a month with no results, and finally tested the bottle—it had a small amount of detergent added for cleaning purposes. The tank had never stood a chance.

If the ammonia source doesn’t produce a clear, immediate spike when dosed, test the source itself. Add a drop to a sample of dechlorinated tap water, shake, and test. If the result doesn’t match the expected dose, the bottle is the problem.

Thirty percent on day eighteen

Conventional wisdom says not to do water changes during cycling because they remove the ammonia the bacteria need to eat. This is true for the first few days. But by week two, with ammonia still at 4 ppm, the tank is not stalled because of a lack of food. It’s stalled because the bacterial colony hasn’t reached critical mass. High ammonia can actually inhibit the growth of the nitrite-oxidizing bacteria that show up second in the cycle.

A partial water change—thirty percent, done carefully to avoid disturbing the filter media—can drop ammonia to a level the existing colony can handle, then the cycle can proceed. The owner of a forty-gallon tank in a Clementi terrace house did a thirty-percent change on day eighteen, saw ammonia drop from 4 ppm to 1.5 ppm, and had zero ammonia by day twenty-five. The colony had been overwhelmed, not starving.

The trick is testing immediately after the water change and again the next day. If ammonia drops and stays low, the colony was ready to take over. If it climbs back to 4 ppm within twelve hours, the colony wasn’t large enough yet, and the water change just reset the clock.

Months later, the owner of that stalled tank in Bukit Batok keeps a sticky note on the hood: “Check the filter. Check the temp. Check the pH.” The ammonia spike that wouldn’t drop taught more about how tanks actually work than any cycling guide ever did.