COP, EER and kW per Ton: How to Read Chiller and Aircon Efficiency Numbers

Same room, twice the bill

Picture two identical offices, side by side, both held at a comfortable 24°C on a hot afternoon. From the inside they feel exactly the same. But one is fed by a thick electrical cable that runs warm to the touch, and the other by a thin cable that stays cool. Same cooling delivered, wildly different electricity burned to deliver it.

That difference is efficiency, and it is invisible from inside the room. You cannot feel it, and the thermostat will not tell you. The only way to compare two cooling machines fairly is with a number that divides cooling delivered by electricity consumed. This part is about those numbers — and there are, annoyingly, several of them for the same idea. By the end you will be able to read any of them and grade your own aircon or chiller plant.

Why the number can be bigger than 1

Here is the fact that surprises most people first. A good chiller has an efficiency of five or six — it delivers five or six times more cooling energy than the electrical energy you feed it. If you have ever seen an electric heater rated "100% efficient," how can a cooler beat it several times over?

Because a chiller does not make cold. Cold is not a substance. What the machine actually does is move heat — it picks up heat from the water or air inside your building and dumps it outside. Moving something is far cheaper than manufacturing it. Think of a bucket brigade carrying water out of a flooded room: the workers spend a little energy carrying, but they shift far more water than the effort would suggest. The electricity you pay for runs the "pump and carry" machinery; the heat itself rides along for a comparatively small fee. (We unpack how this heat-moving cycle actually works — evaporator, compressor, condenser — in an earlier part on the refrigeration loop.)

So an efficiency of "6" is not magic and it does not break any law of physics. It simply reflects that we are counting moved heat against the electricity spent moving it.

Scoreboard 1: COP — higher is better

The cleanest measure is the Coefficient of Performance (COP).

$$\text{COP} = \frac{\text{cooling output}}{\text{electrical input}}$$

Both quantities are in the same units — watts of cooling over watts of electricity — so COP is a plain ratio with no units attached. That is its great virtue: it is unambiguous anywhere in the world.

- A COP of 3 means three units of heat moved for every unit of electricity paid.
- Higher is better.
- A small split aircon typically sits around COP 2.3 to 3.5.
- A large, well-run water-cooled chiller reaches COP 5 to 6 or better.

Chillers beat small aircon units because they use water to carry heat away to a cooling tower, which is a far more effective heat-dumping route than blowing hot air off a small outdoor box. More on that trade-off in the part on air-cooled versus water-cooled machines.

Scoreboard 2: EER — a snapshot in mixed units

The Energy Efficiency Ratio (EER) measures the same thing — cooling out over electricity in — but in mixed units. It divides cooling in BTU per hour by electrical input in watts.

Because it mixes an imperial cooling unit with a metric electrical unit, EER lands on unfamiliar-looking numbers like 10 or 12, and you cannot read it as a plain multiple the way you can COP. The two are directly related:

$$\text{COP} \approx \frac{\text{EER}}{3.412}$$

So an EER of 12 is about COP 3.5. Same physics, different dress. EER is a snapshot taken at one specific test condition (a fixed outdoor temperature and load). Real weather is not one fixed condition, which leads to the next scoreboard.

Scoreboard 3: the seasonal number — and what Malaysia actually prints

A single snapshot flatters a machine, because efficiency changes with how hot it is outside and how hard the unit is working. A seasonal rating averages efficiency across a realistic spread of operating conditions, so it better reflects a full cooling season.

Internationally you will see SEER (Seasonal Energy Efficiency Ratio) used for this. But if you are standing in a shop in Malaysia looking at the energy label on a split unit, you will not see SEER. Malaysia's Suruhanjaya Tenaga (Energy Commission) runs a Minimum Energy Performance Standards (MEPS) scheme, and the label you actually read carries:

- a star rating (up to 5 stars — more stars, more efficient), and
- a CSPF figure — Cooling Seasonal Performance Factor, the tropical seasonal efficiency measure designed for our climate.

CSPF is the honest local equivalent of a seasonal rating: cooling delivered across a season divided by the electricity used, tuned to a cooling-only, warm climate. When you compare two split units for an office or a home, the 5-star MEPS label and its CSPF are the numbers to trust — not an imported SEER figure from a US catalogue. Keep EER and COP for understanding the physics; use stars and CSPF for buying the actual box.

Scoreboard 4: kW/ton — the one your chiller engineer quotes

Walk into a plant room and ask the engineer how good the chillers are, and nobody says "COP." They say kW/ton (kilowatts per refrigeration ton), sometimes written kW/RT.

This metric deliberately flips the ratio upside down:

$$\text{kW/ton} = \frac{\text{electrical kilowatts drawn}}{\text{tons of cooling delivered}}$$

It puts electricity paid on top and cooling delivered on the bottom — the opposite arrangement to COP. The consequence is simple but catches people out:

- For kW/ton, LOWER is better. Fewer kilowatts spent per ton of cooling is the goal.
- It reads like a fuel-consumption figure for a car (litres per 100 km): the smaller the number, the more you get for what you burn.

From an earlier part we know 1 refrigeration ton (RT) = 3.517 kW of cooling. That single fact ties kW/ton and COP together exactly:

$$\text{COP} = \frac{3.517}{\text{kW per ton}}$$

So the two scoreboards are the same information wearing opposite clothes:

- 0.59 kW/ton → COP 6.0 (excellent)
- 0.80 kW/ton → COP 4.4 (average)
- 1.00 kW/ton → COP 3.5 (act soon)

A higher COP and a lower kW/ton both mean the same thing: good. If you ever see a table that calls 0.5 kW/ton "excellent" and then in the next breath praises "COP 4.4" as top-tier, someone has muddled the two — 0.5 kW/ton is actually COP 7, while COP 4.4 is a middling 0.80 kW/ton. Always convert with the line above and the contradiction disappears.

The electricity in the ratio is real power

One subtlety worth flagging: the "kW" in kW/ton is real electrical power — the kilowatts that actually do work, the same kW your meter bills. Motors driving compressors and pumps draw current that is not perfectly in step with the voltage, and losses inside those motors are real. If you want to understand why the electrical side of this ratio behaves the way it does, the Electricity Fundamentals course covers it in power vs energy: kW and kWh and, for why motor loads inflate the current, the kW, kVA, kVAR power triangle. The cleaner the electrical side, the fewer wasted kilowatts land in your efficiency ratio.

Grading your own plant

Here is a practical scale for a water-cooled chiller plant in our climate. Numbers are whole-plant kW/ton unless noted:

- Excellent: ~0.5 to 0.6 kW/ton (COP ~5.9 to 7.0) — modern, well-maintained, well-controlled.
- Average: ~0.9 to 1.0 kW/ton (COP ~3.5 to 3.9) — typical of older or loosely tuned plant.
- Act now: above ~1.1 kW/ton (COP below ~3.2) — money is leaking; investigate.

For context, a modern chiller on its own can run near 0.5 kW/ton, whereas 1970s-era equipment sat closer to 0.9 kW/ton — roughly a doubling of efficiency across a few decades of engineering. And as a regional target worth remembering, Singapore's building code requires new commercial developments to hit a water-side plant efficiency no worse than 0.63 kW/ton. That is a well-chosen line for "a genuinely good plant," and it is a fair yardstick to hold your own building against.

Air-cooled machines and small split units live higher up the scale — a good split unit around COP 3 to 3.5 is perfectly respectable for its class, because it lacks the water-and-cooling-tower advantage of a big central plant.

Chiller-only versus whole-plant — don't grade the wrong thing

The single most common mistake in reading these numbers is confusing the chiller with the plant.

A chiller's nameplate might proudly say 0.55 kW/ton. But the chiller is not the only thing drawing power to cool your building. A complete water-side plant also runs:

- chilled-water pumps (pushing cold water to the air handlers),
- condenser-water pumps (pushing warm water to the cooling towers),
- cooling-tower fans, and
- assorted auxiliaries.

Add all of those, and a plant whose chiller alone reads 0.55 kW/ton might deliver 0.75 or 0.85 kW/ton as a whole system. When you benchmark, be clear which figure you are quoting. The number that matters for your electricity bill is the whole-plant kW/ton, because you pay for every one of those motors, not just the compressor.

Nameplate is a best case, not your reality

One more warning. Every efficiency figure stamped on a machine — COP, EER, CSPF stars, kW/ton — is measured in a laboratory under ideal test conditions: the design load, clean coils, perfect water temperatures, brand-new components. Your plant on a humid afternoon, with fouled tubes, part-loaded chillers and pumps running flat-out regardless of demand, will do worse. Nameplate is the score the machine got on its best day, not the score it gets every day.

The takeaway

There is only one underlying idea here — cooling delivered per unit of electricity — dressed up in four outfits. COP is the clean ratio (higher is better). EER is the same thing in mixed units, a single snapshot. CSPF and the Suruhanjaya Tenaga star rating are what a Malaysian buyer actually reads on a split-unit label. And kW/ton is what chiller engineers quote (lower is better), tied to COP exactly by COP = 3.517 / (kW per ton). Learn to convert between them and no efficiency table can fool you.

But remember the last warning: nameplate is a best case. The only way to know your plant's true kW/ton is to measure the real electrical power going in and the real cooling coming out — continuously, and trend it over months rather than trusting a laboratory sticker. That is precisely what continuous energy monitoring with CobiNeural is for.

Next, we put this metric under load — the part ahead looks at why a chiller's kW/ton changes as it runs at part-load through the day, and why the annual average is the number that really sets your bill.