The Under-$100 Tank Automation Stack — and the Two Things You Must Never Put on a Timer
September 12, 2026
You can automate a freshwater tank for under a hundred dollars US. Genuinely — four things off a shelf, one afternoon with the stand pulled out from the wall.
The money was never the hard part.
The hard part is that about half the automation setups I see posted online have a device on a switch that should never, ever be on one. People spend real money and end up with a tank that’s less safe than the one they started with. So before the shopping list, the two rules.
Rule one: the heater never goes on a timer
A timer works on time. A heater needs to work on temperature. Those are not the same job and no schedule can bridge the gap, because your house gets cold on nights nobody put on a calendar.
Put the heater on a timer and you’ve built a tank whose temperature is decided by the clock instead of the water. That’s a swing, and swings are what stress fish into illness.
There’s a worse version. The most common way an aquarium heater dies is the internal thermostat sticking in the on position — the contacts weld shut and the thing just keeps heating. A smart plug running a schedule has no idea that’s happening. It doesn’t measure anything. At 6pm it turns the outlet on exactly as instructed and helps.
The only switch that belongs between your wall and your heater is one with a probe in the water — a proper temperature controller that cuts power when the tank hits your number, regardless of what the heater thinks. That’s a different device with a different job, and I’ve written up how to wire one so both parts have to fail before anything cooks.
Heater into a controller, or heater straight into the wall. Never into a schedule.
Rule two: the filter never goes on a timer either
This one surprises people, because turning the filter off at night sounds tidy and quiet and slightly green.
Your biological filtration isn’t the machine. It’s a colony of living bacteria sitting in the media inside it, and those bacteria need oxygenated water moving past them every hour of every day. Stop the flow and the water trapped inside a sealed canister goes oxygen-starved in about an hour or two. Do that nightly and you are slowly killing the thing that processes your fish waste, then wondering why your ammonia readings crept up.
Filters run around the clock. There’s no exception and no clever schedule.
If you want the full version of what happens when flow stops for hours — including the restart mistake that dumps a canister’s worth of foul water back into the tank — that’s in the power-outage piece.
Now the three things worth automating
In this order, because that’s the order they pay you back.
1. Lighting — the one job a dumb switch does perfectly
Lights are the ideal automation target: they genuinely should run on a clock, nothing dies if they’re late, and consistency is exactly what plants want. This is where your first plug goes.
Six to eight hours a day, not twelve. If you want more lit hours for viewing, split the block with a siesta in the middle rather than adding light at the ends — algae thrives on long, unbroken light and your plants won’t miss it. I’ve gone through the schedule and the app settings in detail elsewhere; for this build, all you need is a switch that can do two on-blocks a day.
Smart plug or the old mechanical dial? Honest comparison, because it isn’t obvious.
The mechanical timer wins on simplicity. No app, no account, no firmware, no WiFi. It fails in exactly one way and you can see it from across the room. Its weakness is the one I mentioned above: it keeps time off the wall frequency, so a blackout doesn’t pause it — it shifts it, permanently, by the length of the outage, until a human walks over and turns the dial. Come home from two weeks away after a four-hour storm outage and your lights have been running four hours late the entire time.
The smart plug wins on recovery. Its schedule sits in the plug’s own flash memory, so it survives being unplugged; when power comes back, it reconnects, pulls the real time off the network, and picks up where it should be with nobody touching it. That self-correction is the whole reason to pay the extra few dollars.
The smart plug’s weakness is that it has more ways to fail — a router that doesn’t come back, an app account, a firmware update that changes something at 3am. Which is why I own both. Lights on the smart plug, the mechanical dial in the drawer beside the tank as a five-minute swap if the smart one gets confused.
2. Feeding — and the feeder must run on batteries
I’ll keep this short because the feeder itself deserves its own write-up, but the one decision that belongs in this build is the power source.
Buy the battery-powered feeder. Not the one that plugs into the wall, and definitely not a plugged feeder on a smart plug.
The reasoning is the same as everything else here. A mains feeder is a device that stops working during precisely the event you’re not home for. Batteries make the feeder independent of your grid, your router, and your schedule. It’s a rare case where the cheaper, dumber option is also the more reliable one.
Load it with pellets, not flakes, run it for a week while you’re still home, and turn the portion down from what your gut says.
3. Monitoring — which is not a switch at all
The third slot in the stack doesn’t control anything. It watches.
A digital thermometer with a high and low alarm is the cheapest insurance in this hobby. Temperature is the fastest-moving thing that can go wrong in a stable tank, and a stuck heater can take a tank past liveable in well under a day. An alarm won’t stop it. It’ll tell you — which is enough if there’s someone in the house, and the reason to upgrade to something that can reach your phone once you start leaving for real.
Put the probe at the opposite end of the tank from the heater. Beside the heater you’re reading the heater.
The fourth slot everyone wastes money on
Top-off. And this is the one I want to be blunt about.
The idea is obvious: evaporation drops your water level over two weeks, so put a little pump in a jug and run it on a timer. Please don’t. A pump on a schedule has no idea whether it worked. It can’t tell a full tank from an empty reservoir. It runs for its ninety seconds either way. So it either pumps air after the jug empties, or it keeps adding water to a tank that didn’t need it, and you come home to a flood and a very expensive floor.
Automation needs a feedback loop. A timer isn’t one.
You’ve got two honest options and neither is a smart plug. Buy a real top-off unit with a float or optical sensor that fills until the level is right and then stops — I’ve covered sizing one for a two-week absence — or do what I’d suggest for a first build and remove the problem instead of automating it. A well-fitted glass lid cuts evaporation dramatically. No power, no pump, no failure mode. Design beats gadgets, every time.
Wiring it so it fails safe
Half an hour of care here is worth more than anything you bought.
Use a GFCI outlet. Water and mains belong nowhere near each other, and this is the device whose entire job is noticing when they’ve met. If the outlet behind your tank isn’t one, a plug-in GFCI adapter costs very little.
Drip loops on every single cord. Let each cable hang down below the outlet before it comes back up to plug in. Water that runs down a cord then follows gravity to the floor instead of into your wall socket. It takes ten seconds per cord and it is the single best habit in this hobby.
A check valve in the air line, if you run an air pump. If the pump sits lower than the tank’s water line, a power cut lets water siphon backwards down the tubing, over the rim, and onto your floor until the tank is empty. A check valve is a small one-way part that costs about as much as a coffee. There’s an arrow printed on it — the arrow points toward the tank. Get that backwards and you’ve just blocked your air.
Set the smart plug’s power-on state on purpose. Most let you choose what the outlet does when power returns: on, off, or back to whatever it was. Decide it rather than inherit it.
Label the always-on outlets. I use a strip of tape on the power bar. Filter and heater live on marked sockets that nothing schedulable ever gets plugged into. It sounds fussy right up until you’re reaching behind a stand in the dark.
The list, and the weekend
Four things: a pair of smart plugs, a mechanical timer as the backup, a battery feeder, an alarm thermometer. A check valve if you run air. That’s comfortably under a hundred dollars US and it’s most of the value in this hobby’s entire automation aisle.
Saturday morning: pull the stand out, fit the GFCI, dress the cords with drip loops, label the always-on sockets. Saturday afternoon: set the light schedule and let it run. Sunday: load the feeder, set the alarm band a degree and a half either side of your target, and then leave the whole thing alone for a week while you’re still home to watch it.
That last week is not optional. It’s the mock run, and it’s the only way you find out that your feeder drum clumps or your plug forgot its schedule while you’re still in a position to care.
What to add later, and in what order
Once the cheap stack has proven itself, spend in this order: a heater controller with a probe, because that’s the failure that kills whole tanks. Then a monitor that can message you instead of your empty house. Then a sensor-driven top-off if you’re regularly away longer than two weeks.
Notice what’s still not on the list — a big proprietary aquarium controller. They’re good units. But they’re one box, and one box is one thing to fail. Four cheap independent devices that each do a single job, with the two critical loads wired so they can’t be switched off at all, will outlast a fancier setup that fails all at once.
Boring, separate, and hard to break. That’s the whole trick.