Why a Chiller's First Job Is Your Light and Pump

This page contains affiliate links. If you buy through one we may earn a commission, at no cost to you.

Not the room, at least not entirely. At the shallow end of a chiller's own published range, about two fifths of what it removes is heat that the tank's own light and pump put in, and that fraction is the only part of the problem you chose.

There are exactly three terms in the budget. Heat leaking in through the glass and the water surface, which depends on how far below the room you are holding the water. Heat from the equipment, which does not depend on ambient at all. And, if the box is standing in the same room, heat the chiller itself rejects, which raises the room and comes back around. Only the first one gets sized for.

The three terms, and the one that does not move

A 1/3 HP aquarium water chiller for tanks up to 79 gallons publishes a usefully specific claim: it cools fresh or salt water 5 to 15 degrees Fahrenheit below ambient, it ships with a pump and hose so the loop is complete, it runs at a stated 30 to 40 decibels, and it has an LED touch screen. Its own drawbacks are honest about the gap: "The listing omits exact dimensions, port sizes, and power draw."

That missing power draw matters, so hold onto it. First the tank.

Our page on why a 5 gallon and a 15 gallon tank take the same 50 watt heater establishes the two coefficients this whole category runs on, and this page uses them deliberately so the two agree: glass to still room air at about 7 watts per square metre per kelvin, and the open water surface at about 100 watts per square metre at a few kelvin of lift, which works out near 16.9 watts per square metre per kelvin. Four walls, as that page counts them.

Take a standard 75 gallon tank at 48 by 18 by 21 inches, which is the class this chiller is rated for. Its four walls are 2,772 square inches, or 1.7884 square metres, giving 12.52 watts per kelvin. Its water surface is 864 square inches, or 0.5574 square metres, giving 9.42. Total 21.94 watts per kelvin.

An Empire USA Starfire Glass 150 gallon aquarium is the one tank in this catalogue that publishes real dimensions, 60 by 24 by 24 inches, along with 12 millimetre low iron glass, a Eurobraced top and a built in overflow box for a sump. Run the same two coefficients and it comes to 18.21 plus 15.70, which is 33.91 watts per kelvin. Note that the volume doubled against the 75 and the loss rose by 55 percent, which is the same area against volume point that page makes.

Now the term that nobody counts.

Your pump and your light are a space heater you installed on purpose

An AQQA 581GPH canister filter with 3-stage filtration publishes a 35 watt motor, up to 581 gallons an hour, quartz balls and filter cotton, a flow control valve, a four buckle seal with silicone rings and a copper motor. Its record is correct and careful that "real turnover depends on head height and plumbing, so treat 581 GPH as a maximum."

Work out where the 35 watts goes. At 581 gallons an hour, which is 2,199 litres an hour, lifting against a representative 1.2 metre head, the hydraulic power is density times gravity times flow times head, which is 7.19 watts. Against a 35 watt input that is 20.5 percent wire to water, which is entirely normal for a small canister. The other 27.8 watts is heat in the motor and the bearings.

And then the 7.19 watts of hydraulic power does not escape either. It goes into moving water around a closed tank, where it is dissipated by friction and turbulence and ends up as heat in exactly the same water. There is nowhere else for it to go. Taking a representative 80 percent of the electrical input as reaching the water through the wet path, the filter is a 28 watt heater.

A Lominie 80W COB LED planted aquarium light with a dimmer publishes an unusually complete set: 80 watts, up to 4325 lumens, a 90 degree secondary lens, a body 4.4 inches across and 3.86 long, suitability for tanks up to 24 inches, an aluminium fin heatsink with a fan below 28 decibels, a 0 to 100 percent dimmer, and a stated operating temperature of 30 degrees Celsius.

Two of those numbers together give you the third. 4325 lumens from 80 watts is 54.1 lumens per watt at the wall. White light has a luminous efficacy of radiation near 300 lumens per watt, so the radiant efficiency is 54.1 divided by 300, which is 18.0 percent. In watts: 14.42 leaves the fixture as light and 65.6 watts is heat before any of it has been near the water. That is what the aluminium heatsink and the fan are for, and it is why the record publishes a 30 degree operating temperature.

Of the 14.42 watts that does leave as light, most is absorbed by water, substrate and plants rather than escaping through the glass. Take a representative 85 percent and the light contributes 12.3 watts to the water.

So a tank running both is receiving 40.3 watts of equipment heat, around the clock for the pump and for as long as the light is on, and that figure does not care what the weather is doing.

What 40.3 watts is worth, which depends entirely on the setpoint

Here is the result that reframes the whole question. The chiller claims 5 to 15 Fahrenheit below ambient. Those two ends of its own range are not the same job.

At 15 Fahrenheit, which is 8.33 kelvin, the glass and surface leak 21.94 times 8.33, which is 182.8 watts inward. Add the equipment and the load is 223.1 watts, of which the equipment is 18.1 percent.

At 5 Fahrenheit, which is 2.78 kelvin, the leak falls to 61.0 watts. The equipment term does not fall at all. The load is 101.3 watts and the equipment is 39.8 percent of it.

Two fifths. Set a chiller to hold a tank a modest 5 degrees under the room and you have bought a refrigeration machine whose main job is undoing the pump and the light that are sitting in the same tank. Dim the light to 50 percent, which this one can do, and roughly 6 watts of the load goes away, which is 6 percent of the total at that setpoint and 2.7 percent at the deep one.

And the capacity check, because it is good news. One third of a horsepower is 248.7 watts of input power. A small water chiller working at this kind of lift runs at a coefficient of performance somewhere near 3, so about 746 watts of cooling. Against 223.1 watts that is a 29.9 percent duty cycle at the deep setpoint and 13.6 percent at the shallow one. The rating and the published claim are consistent, with real margin. That is worth saying as plainly as the criticisms, because a claim that survives arithmetic is rarer than one that does not.

What is missing is the number a buyer actually lives through. Taking 79 gallons, which is 299.1 kilograms of water, down 8.33 kelvin takes 299.1 times 4,186 times 8.33, which is 10.43 megajoules. During the ramp the net cooling averages the 746 watts minus the equipment and minus the growing leak, around 614 watts, so the pull down is 4.72 hours. Nothing publishes it, and it is the difference between a machine you switch on before a heatwave and one you switch on during it.

Where the removed heat goes, and why the room matters

A chiller does not destroy heat. It moves it, and it adds its own input power to the pile on the way. Removing 223.1 watts while drawing 248.7 means rejecting 471.8 watts into the air around the unit.

If that air is the same room the tank is in, the loop closes. Take a representative room losing 150 watts per kelvin to the rest of the house and outdoors: 471.8 watts raises it 3.15 kelvin. Which raises the tank's leak by 21.94 times 3.15, or 69.1 watts. Which is 31.0 percent more load than the chiller started with, created by nothing except where the box is standing.

Our guide to choosing an aquarium chiller for your tank gets close to this and stops one step short. It says to "site the unit with airflow around its heat vents so it cools efficiently and stays quiet", which is right about the unit and silent about the room. The stronger version is that the 471.8 watts should leave the room, through a doorway, a vent or a different cupboard, and that a sealed cabinet under the tank is the worst available location for precisely the reason the sentence hints at.

The same guide opens with "Match capacity to tank volume. Chillers are rated by gallons and horsepower. Pick a unit that comfortably covers your tank size." Volume is the wrong variable, for the exact reason our heater page lays out: the leak is an area and the gallons only enter because bigger tanks happen to carry more glass. A 75 gallon tank and a 90 gallon tank sharing the same 48 by 18 inch footprint differ by 20 percent in volume and about 8 percent in loss. And the sentence has no term at all for the equipment, which at a 5 degree setpoint is two fifths of the answer.

Its advice to "Confirm a pump and hose are included or on hand" is correct and carries the same unstated bill. This chiller does include a pump, which is genuinely convenient, and that pump's electrical power is added to the heat the chiller must then remove. Its wattage is not published, so the figure cannot be put in. If it draws 20 watts, it is 8.8 percent of the load at the deep setpoint and 19.7 percent at the shallow one, all of it self inflicted.

The fan is a different machine with a hard ceiling

A SEAOURA aquarium chiller fan with a timer and temperature control publishes a drop of 3 to 7 Fahrenheit, a probe with thresholds settable anywhere from 60 to 104 Fahrenheit, ten speeds, one to three fan heads, a 5.4 inch gooseneck that rotates 720 degrees, a bracket for rimless glass from 0.2 to 1 inch, and a 24 hour timer. Its drawbacks are the right ones: "Cooling drop is limited to 3 to 7F" and "Evaporation also lowers water level, meaning you top up more often during heavy use."

Evaporative cooling has a floor it cannot pass, which is the wet bulb temperature of the room air. Our page on why an evaporative cooler runs out of degrees in humid air works that argument through for a room, and the same ceiling applies here with the water surface as the wet surface.

At 26 degrees Celsius and 60 percent relative humidity the wet bulb sits at 20.38, a depression of 5.62 kelvin, which is 10.1 Fahrenheit. So the fan's published 7 degree figure is 69 percent of the thermodynamic ceiling, which is an honest and achievable claim.

At the same 26 degrees and 80 percent relative humidity, which is an ordinary summer afternoon in a lot of places, the wet bulb rises to 23.34 and the depression collapses to 2.66 kelvin, which is 4.79 Fahrenheit. The fan's own upper figure is now impossible, and its lower figure of 3 degrees is 63 percent of what remains. The 3 to 7 range is not a range of fan speeds. It is a range of weather, and the listing does not say which end belongs to which.

Then there is the price in water, which is where the number gets surprising. Holding the 75 gallon tank 3.89 kelvin below ambient means removing 21.94 times 3.89 of leak plus 40.3 of equipment, which is 125.6 watts, entirely by evaporation. At 2.44 megajoules per kilogram that is 4.45 litres a day.

On 299 litres that is 1.49 percent a day, so 10.4 percent of the tank leaves as vapour every week and every dissolved solid in it stays behind. On a marine tank held at 35 parts per thousand, a week of that without topping off arrives at 39.1. On fresh water it is a steady climb in hardness and in everything else that does not evaporate. The record's note that you top up more often is correct, and 4.45 litres a day is what more often means.

And the comparison that makes the two machines genuinely different rather than two grades of the same thing. The compressor removes the identical 125.6 watts for 125.6 divided by 3, which is 41.9 watts of electricity, and no water at all. Same heat, 41.9 watts against 4.45 litres a day, and one of them has a ceiling at the wet bulb while the other does not.

The cleaning tool moves more degrees than either of them

An electric aquarium vacuum 5-in-1 gravel cleaner publishes a 28 watt pump, up to 1,700 litres an hour, a handle extending from 12.4 to 33.6 inches, a 2 metre hose, a reusable sponge filter cup, suitability for roughly 10 to 100 gallon tanks, and the claim that it can drain a 180 gallon tank in about 30 minutes.

That last pair is checkable. 180 US gallons is 681.4 litres, and in 30 minutes that is 1,362.8 litres an hour against a rated 1,700, which is 80.2 percent of the maximum. The 19.8 percent shortfall is exactly the head loss the canister filter's record warns about, applied honestly to its own figure. Another claim that survives.

What no listing mentions is that a water change is a thermal event larger than anything the temperature gear does. Replace 25 percent of 299 litres with tap water 6 kelvin cooler and the mixed temperature falls by a quarter of 6, which is 1.5 kelvin, or 2.7 Fahrenheit. That is 54 percent of the chiller's entire 5 degree lower claim, delivered in half an hour by a cleaning tool, and it goes the other way if the tap runs warm.

Which is a useful thing to know in both directions. If you are fighting heat, time the water change for the warm part of the day and let it do the work. If you are holding a temperature, that 2.7 degree step is the disturbance your heater or chiller then has to recover from, and on the 75 gallon tank recovering 1.5 kelvin takes 299.1 times 4,186 times 1.5, which is 1.88 megajoules, so about 51 minutes at the chiller's net rate.

What this category refuses to publish

No equipment in this aisle publishes a heat output, and heat output is the one property every single item has. The filter publishes watts, which is nearly the same thing and is the most useful figure here. The light publishes watts and lumens, which between them give the split. The vacuum publishes watts. The fan publishes neither a wattage nor an airflow, which are the two numbers that would turn 3 to 7 Fahrenheit into a prediction instead of a range. And the chiller, the one machine whose entire function is thermal, publishes no power draw at all and says so in its own drawbacks.

One third of a horsepower is not a cooling capacity and it is not an input power. It is a convention from compressor sizing, and the two figures it stands in for, watts in and watts of cooling at a stated temperature, are both absent. The 5 to 15 Fahrenheit claim is the closest thing to a capacity on the page, and it only becomes one once you supply a tank area, which the tank listings mostly do not publish either. The 150 gallon display is the exception and it is the one tank nobody would put a 1/3 HP chiller on.

The honest limit, and the three numbers to carry

What is exact here is every balance. A steady state offset is heat in divided by loss coefficient. The equipment term is published wattage times a path fraction. The radiant efficiency of the light follows from its own two published numbers and one reference constant. The wet bulb temperatures come from the Magnus relation and the standard psychrometric equation. The evaporative water loss is a power divided by a latent heat. The pull down is a mass times a specific heat times a temperature divided by a net rate. None of those is an estimate.

What is representative is the surrounding furniture of the calculation, and it is worth listing because several of these are load bearing. The coefficient of performance near 3 sets the duty cycle and the 41.9 watt figure for the fan comparison; at 2.5 both rise by 20 percent and at 3.5 they fall by 14. The 1.2 metre pump head sets the 20.5 percent wire to water figure but not the 28 watts, which comes from the published input. The 300 lumens per watt luminous efficacy of radiation sets the 18.0 percent split; a warmer white spectrum runs nearer 260 and would make the light's radiant share 20.8 percent instead. The 80 and 85 percent path fractions, the standard tank footprints, the 26 degree room, the 150 watts per kelvin room and the 6 kelvin tap offset are all in the same category. Every percentage that compares the equipment term with the leak term is far more robust than any absolute watt figure, because both sides scale together.

What this page deliberately does not compute is three things. It does not give a temperature any species should be held at, because that is an animal question rather than a heat transfer one and no listing here carries a defensible figure. It does not compute the chiller's capacity in watts, because that needs a compressor curve at a stated evaporating and condensing temperature and the listing publishes no power draw at all; the 746 watts is the published horsepower times a representative coefficient of performance and should be read as an order rather than a specification. And it does not compute how fast the fan drops the water when it switches on, because that needs the fan's airflow and the wetted area under it, and the listing publishes neither; all that can be said is where the floor is.

Three numbers to carry. Equipment heat does not depend on your setpoint, so the shallower the offset you want, the larger the share of a chiller's work that is undoing your own gear, reaching two fifths at the 5 degree end of this one's range. A chiller rejects its removed heat plus its own input into the room, which on the figures here is 471.8 watts and about 31 percent extra load if the room is closed. And evaporative cooling stops at the wet bulb and is paid for in water, at 4.45 litres a day for 3.89 kelvin on a 75 gallon tank, which is 10 percent of the tank a week with the salts left in.