Oversized Chillers, Short-Cycling and Dirty Coils: The Cheapest Efficiency Wins

The most expensive four words in building services

"Size up to be safe." It sounds like prudence. If the building might one day be hotter, fuller, or busier than expected, surely a bigger chiller is the responsible choice? So the consultant adds a margin, the contractor adds another, and the machine that lands on the roof is comfortably larger than the building will ever need.

This is the single most expensive intuition in cooling. An oversized chiller does not sit there quietly holding spare capacity in reserve. It actively wastes energy, wears itself out faster, and controls comfort worse than a right-sized machine — every hour of every day. And it does so in a way that never shows up as a dramatic fault. Nothing breaks. The building just quietly costs more to cool than it should.

This part is about the cheap, physical wastes that hide in plain sight: an oversized chiller that switches on and off like a nervous light switch, and coils so coated in grime that the very surface meant to move heat has become an insulator. Many complaints of "the chiller is struggling" are solved not with a new machine, but with a hose, a filter change, and — at replacement time — some honesty about what size you actually need.

Short-cycling: cooling too fast is not a good thing

Let us start with a term. Short-cycling is when a chiller (or any cooling machine) runs for only a brief period, shuts off, then starts again soon after — repeating this stop-start pattern instead of running steadily. The "short" refers to the short run-times between shutdowns.

Here is why oversizing causes it. A chiller's job is to hold the chilled water at a target temperature — its setpoint. When the building's cooling demand is, say, 60% of what the chiller can produce, a right-sized machine settles into a comfortable steady run, matching output to demand. But an oversized machine can dump far more cooling than the building is asking for. It races the water down to setpoint in minutes, discovers there is nothing left to cool, and switches off. A few minutes later the water has warmed slightly, so it fires back up, blasts the temperature down again, and shuts off once more.

Think of driving in heavy traffic by flooring the accelerator, braking hard, flooring it again — instead of holding a gentle, steady speed. You use more fuel, you wear the brakes and engine, and the ride is jerky. That is short-cycling.

Why the start-up is the expensive part

The reason this matters so much is that starting a chiller is its least efficient moment. At start-up, the compressor draws a surge of current, oil and refrigerant are still redistributing, and the machine has not yet reached the smooth operating conditions where it sips energy most efficiently. A chiller that runs steadily for an hour spends most of that hour in its efficient zone. A chiller that starts six times an hour spends much of its life in that costly start-up phase and never fully settles.

So short-cycling wastes energy three ways at once:

- Repeated inefficient start-ups — you keep paying the most expensive part of the cycle, over and over.
- Extra mechanical wear — every start stresses the compressor, contactors and motors. More starts means shorter life and more maintenance. (We will not pretend a monitoring layer diagnoses that wear — it does not; but the operator who sees the machine cycling knows to look.)
- Poor comfort control — and this one surprises people.

Short-cycling makes the building clammy

In part 5 we saw that cooling does two jobs at once: it lowers temperature (sensible cooling) and it wrings moisture out of the air (latent cooling, or dehumidification). Removing moisture takes time — the air has to stay in contact with a cold surface long enough for water vapour to condense out.

A short-cycling chiller never gives it that time. It satisfies the temperature quickly and shuts off before it has properly dried the air. In a tropical climate, where the outdoor air is heavy with humidity, this is exactly the wrong failure. You end up with a space that hits its temperature target on the thermostat yet feels cold-and-clammy, encourages mould, and prompts occupants to drop the setpoint lower — chasing comfort that a steadier machine would have delivered at a higher, cheaper temperature. An oversized chiller can leave a building both colder on the meter and less comfortable to sit in.

Right-sizing: match capacity to the real load

The fix for oversizing is right-sizing — matching installed cooling capacity to the building's real connected load, the actual amount of cooling the spaces need, rather than to a stack of safety margins.

The catch is that most sizing decisions are made before the building exists, from calculated worst-case assumptions. Those calculations tend to be conservative, and then everyone in the chain adds a little more "just in case." The result is a plant sized for a peak that occurs rarely, if ever.

The honest way to size a replacement plant is to look at what the building actually did. True peak load, and the load profile — how demand rises and falls across the day, the week, the seasons — are only revealed by measuring over time. A building that was assumed to need 500 RT might, on the evidence of a year's data, peak at 380 RT and spend most of its hours below 250 RT. That is not a number anyone can guess from a nameplate; it is a number you read off a trend. Sizing the next plant from that evidence, rather than from a default margin, is one of the highest-value decisions an owner ever makes.

Staging beats one big machine

There is a second, related move: staging. Instead of one enormous chiller, install several smaller ones and switch them on as demand rises. This is how good chiller plants are built, and the reason is straight physics.

A chiller is usually most efficient when it is fairly heavily loaded — running near, but not at, its full output. A machine loafing along at 20% load is often working inefficiently (we dug into part-load behaviour in an earlier part). So imagine you need 300 RT of cooling right now:

- One 600 RT chiller runs at 50% load — and if demand dips, it slides toward short-cycling.
- Three 200 RT chillers, with two running at 75% each, keep both operating machines in their efficient zone, with the third idle in reserve.

Staging lets the plant follow the load: add a chiller as the building fills up and heats up, drop one as the evening empties out. Each running unit stays efficiently loaded, rather than several machines all running lightly and wastefully. It is the difference between one person carrying an unbalanced load and a team each carrying a sensible share.

Dirty coils: when the heat exchanger becomes an insulator

Now to the cheapest waste of all — and the one most likely to be sitting in your building right now.

A coil is a heat exchanger: a maze of finned tubes designed to move heat between air and the refrigerant or water inside. Its whole purpose is to present a large, clean metal surface so heat crosses easily. There are two you care about: the cooling coil (the cold coil that chills the air blown into the space) and the condenser coil (the hot coil that dumps the building's heat outside, on air-cooled systems).

Both live in moving air, and moving air carries dust, pollen, grease and grime. Over months, that debris settles on the fins and builds into a felt-like blanket. And here is the cruel irony: that blanket is an insulator. The exact surface engineered to transfer heat gets wrapped in a layer that resists heat transfer. You have, in effect, put a jacket on your radiator.

What fouling actually costs

The numbers are stark, and they are worth memorising:

- A dirty cooling coil can deliver as little as about 70% of its rated capacity. A machine you paid for as 300 RT is quietly handing you 210. The plant compensates by running harder and longer for the same comfort.
- A dirty condenser coil can raise energy consumption by roughly 40 to 50%. When the coil that is supposed to reject heat cannot breathe, the refrigerant condenses at a higher temperature and pressure, and the compressor must work dramatically harder to push heat "uphill" against that resistance. This is one of the fastest ways to wreck a chiller's efficiency, and it is caused by nothing more exotic than dust.

Fouling is a leading cause of low delta-T

In the previous part we introduced delta-T — the temperature difference between the water returning from the building and the water supplied to it — and why a healthy, large delta-T is the signature of a well-run chilled-water plant. Coil fouling is one of the leading causes of low delta-T.

The logic is direct. A fouled cooling coil cannot pull heat out of the air efficiently, so the chilled water passing through it barely warms up. It returns to the chiller almost as cold as it left. That small temperature rise is a low delta-T, and — as we saw — low delta-T forces the plant to pump far more water to move the same heat, running pumps harder and dragging chiller efficiency down across the whole system. One dusty coil in one air handler doesn't just under-cool one room; it taxes the entire plant.

Air-side fixes are the cheapest measures you own

Here is the good news, and the reason this part exists. The remedies for all of this are almost embarrassingly cheap:

- Clean the coils. Often a hose, a soft brush and a coil cleaner. Restoring a fouled condenser coil to clean condition can claw back a large slice of that 40–50% energy penalty for the price of an afternoon's labour.
- Change the filters on schedule. Filters exist precisely to stop dust reaching the coil. A clogged filter also chokes airflow, so the fan strains and the coil starves. New filters are cents-on-the-dollar insurance for the expensive coil behind them.
- Control fan speed. Running fans flat-out when the building needs only part of the airflow wastes energy continuously. Fitting a variable-speed drive — an electronic controller that slows the fan motor to match demand — is one of the highest-return measures in building services. (The electronics behind it are worth understanding: see Power Electronics: Rectifiers and Inverters in our Electricity Fundamentals course, and we covered fan-speed control in cooling in an earlier part.)

These are the lowest-hanging fruit in the entire building — cheaper than new chillers, cheaper than controls upgrades, cheaper than almost anything. So why are they so often neglected?

Because an air handler with a filthy coil looks exactly like an air handler with a clean one from the outside. The waste is invisible. A coil losing 30% of its capacity produces no alarm, no error code, no puddle on the floor. It just makes the whole plant work a little harder, forever — a cost that hides inside the total electricity bill where no single person is accountable for it. These fixes only get prioritised when the performance loss is actually measured — when someone can point to a coil and say, "that one is costing us."

The takeaway

The most expensive chiller decision is often "make it bigger to be safe" — because oversizing breeds short-cycling, which burns energy on endless start-ups, wears the machine out, and leaves the air clammy instead of comfortable. Right-sizing to the real measured load, and staging several smaller machines so each runs efficiently loaded, beats one big machine every time. And before you blame the chiller at all, check the cheapest things first: a dirty cooling coil can rob 30% of your capacity, a dirty condenser coil can add 40–50% to your energy bill, and both are among the leading causes of the low delta-T that drags the whole plant down. A hose and a filter change are the best-value tools in the building.

The one thread running through every fault here is invisibility: an oversized chiller, a fouled coil and a collapsed delta-T all look normal from the outside and only reveal themselves when capacity and efficiency are trended over time — the kind of picture a chilled-water and energy monitoring layer like CobiNeural is built to draw. See it, and the cheapest wins stop hiding.

Next, we turn from the machines to the mind that runs them: how building automation and BMS systems tie a mixed, multi-vendor plant together — and where their strengths end.