AIR CONDITIONER

The Opening

42°C outside. Walk into a building: 22°C. Your sweat evaporates, heart rate drops. In 10 seconds you go from survival mode to comfort. The machine that did this moves heat OUT of the building against the temperature gradient — the same thermodynamic trick as a refrigerator, but at building scale: moving megawatts of thermal energy.

You need a machine that:
├── Cools 200 m² from 42°C to 22°C
├── Removes humidity (from 70% to 50% relative)
├── Filters air to remove dust, pollen, PM2.5
├── Does it quietly — under 40 dB indoors
├── Uses minimal electricity — COP above 3.0
└── Runs 3,000 hours per year without breaking

That COP number means: for every 1 kW of electricity you put in, you move 3 kW of heat. The machine doesn't CREATE cold — it MOVES heat from where you don't want it to where you don't care about it.

Let's build one.

───

PHASE 1: Move Heat Backward

Heat flows from hot to cold. Always. The second law of thermodynamics is absolute. So how does an air conditioner move heat from a 22°C room into 42°C outdoor air? You can't break the law — but you can PAY to move heat the wrong way. The currency is work.

The trick: use a fluid that boils at very low temperature. At low pressure, the refrigerant R-410A boils at -51°C. It WANTS to evaporate at room temperature. When it evaporates, it absorbs heat from the room. Then you compress it — raising both its pressure and temperature above outdoor temperature. Now it WANTS to condense, releasing that absorbed heat to the outdoors.

The Vapor Compression Cycle — Four Steps

INDOOR UNIT                    OUTDOOR UNIT
     ┌─────────────────┐           ┌─────────────────┐
     │   EVAPORATOR     │           │   CONDENSER      │
     │                  │           │                  │
     │  cold liquid     │           │  hot gas         │
     │  → boils at 5°C  │  ──────→ │  → condenses     │
     │  absorbs heat    │  high P  │    at 50°C       │
     │  from room air   │  hot gas │  rejects heat    │
     │                  │           │  to outdoor air  │
     └────────┬─────────┘           └────────┬─────────┘
              │                              │
              │ low P                        │ high P
              │ cold gas                     │ warm liquid
              │                              │
     ┌────────┴─────────┐           ┌────────┴─────────┐
     │  EXPANSION VALVE  │  ←────── │   COMPRESSOR      │
     │  drops pressure   │  warm    │  squeezes gas     │
     │  liquid cools     │  liquid  │  from 5°C → 80°C  │
     │  dramatically     │           │  uses electricity │
     └──────────────────┘           └──────────────────┘

    Room air (22°C) → passes over evaporator (5°C) → loses heat
    Outdoor air (42°C) → passes over condenser (50°C) → gains heatThe refrigerant is the shuttle bus. It picks up heat at low temperature (evaporator), gets compressed to high temperature, then drops off heat (condenser). The compressor is the engine — it's where all the electricity goes.

The Carnot COP — Theoretical Maximum

Carnot gave us the theoretical maximum efficiency for moving heat between two temperatures:

COP_Carnot = T_cold / (T_hot - T_cold)

Where temperatures are in Kelvin. For our system:
├── T_cold = 22°C = 295 K (room temperature)
├── T_hot = 42°C = 315 K (outdoor temperature)

COP_Carnot = 295 / (315 - 295)
COP_Carnot = 295 / 20
COP_Carnot = 14.75

That's the theoretical maximum. For every 1 watt of electricity, you COULD move 14.75 watts of heat. In practice, real air conditioners achieve COP of 3 to 5 — about 20-34% of the Carnot limit. Where does the rest go?

├── Compressor friction and motor losses: ~25%
├── Heat exchange inefficiency (finite ΔT needed): ~30%
├── Pressure drops in piping and valves: ~10%
├── Fan power consumption: ~5%
└── Superheat/subcooling irreversibilities: ~10%

Why COP Is Not Efficiency

A COP of 4 doesn't mean 400% efficiency. It means: for every 1 joule of work input, 4 joules of heat move from inside to outside. The first law of thermodynamics still holds — the condenser rejects Q_cold + W:

Q_hot = Q_cold + W
COP = Q_cold / W

So if your AC draws 3 kW electrical and has COP = 4:
├── Heat removed from room: 12 kW
├── Heat rejected outdoors: 15 kW (12 + 3)
└── The outdoor unit dumps MORE heat than the indoor unit removes

This is why cities with millions of AC units create urban heat islands — every unit pumps its waste heat plus its electrical consumption directly into the outdoor air.

DESIGN SPEC UPDATED:
├── Cycle: vapor compression (evaporate → compress → condense → expand)
├── Refrigerant: R-410A, boils at -51°C at atmospheric pressure
├── Carnot COP (22→42°C): 14.75 theoretical maximum
├── Real COP: 3-5 (20-34% of Carnot limit)
└── Q_hot = Q_cold + W — outdoor unit rejects MORE heat than indoor unit removes

───

PHASE 2: Size the Load

Before you buy a compressor, you need to know: how much heat must you remove? Too small and the room never reaches 22°C. Too big and the system short-cycles — turning on and off every few minutes, wasting energy and never dehumidifying.

The cooling load has three sources: heat leaking IN through the building envelope, heat from the sun, and heat from everything inside (people, lights, computers).

Envelope Heat Gain — Q = UA ΔT

The fundamental equation for heat transfer through a wall:

Q = U × A × ΔT

Where:
├── U = overall heat transfer coefficient (W/m²·K) — how leaky the wall is
├── A = surface area (m²)
├── ΔT = temperature difference (K)

For a 200 m² apartment with typical construction:

Component          Area (m²)    U (W/m²·K)    ΔT (K)    Q (W)
─────────────────────────────────────────────────────────────────
Exterior walls     80           0.45          20        720
Windows (double)   25           2.80          20        1,400
Roof               0            —             —         0
Floor (interior)   0            —             —         0
Infiltration       —            —             —         600
─────────────────────────────────────────────────────────────────
Envelope total                                          2,720 WWindows are 6× leakier than insulated walls per unit area. They're 31% of the wall area but 51% of the envelope heat gain. This is why window selection dominates building energy performance.

Solar Gain — The Invisible Furnace

The sun delivers up to 1,000 W/m² to a surface perpendicular to its rays. Windows transmit a fraction of this directly into the room:

Q_solar = SHGC × A_window × I_solar

Where:
├── SHGC = Solar Heat Gain Coefficient (fraction of solar energy transmitted)
├── I_solar = solar irradiance on the window (W/m²)

For 25 m² of east/west-facing windows at peak sun (SHGC = 0.40, I = 500 W/m²):

Q_solar = 0.40 × 25 × 500
Q_solar = 5,000 W

That's nearly DOUBLE the envelope gain. On a hot sunny day, the sun through your windows contributes more heat than all your walls combined.

Internal Gains — People, Lights, Machines

Everything inside the building generates heat:

├── Human body at rest: 75 W sensible + 55 W latent = 130 W total
├── Human body active: 130 W sensible + 130 W latent = 260 W total
├── Desktop computer: 150 W
├── Laptop: 50 W
├── LED lighting: 8 W/m² × 200 m² = 1,600 W
├── Cooking (kitchen): 1,000-3,000 W

For an office with 10 people, 10 computers, and LED lighting:

People:     10 × 130 = 1,300 W
Computers:  10 × 150 = 1,500 W
Lighting:             = 1,600 W
────────────────────────────────
Internal total:         4,400 W

Total Cooling Load:

Envelope:   2,720 W
Solar:      5,000 W
Internal:   4,400 W
Safety (10%): 1,212 W
────────────────────────
Total:      13,332 W ≈ 13.3 kW

In BTU (the unit AC manufacturers love): 13,300 × 3.412 = 45,400 BTU/hr. That's a 4-ton system (1 ton of cooling = 12,000 BTU/hr = 3.517 kW).

DESIGN SPEC UPDATED:
├── Envelope load: Q = UAΔT → 2,720 W (windows dominate)
├── Solar gain: SHGC × A × I → 5,000 W (biggest single source)
├── Internal gains: people + computers + lights → 4,400 W
├── Total cooling load: ~13.3 kW = 45,400 BTU/hr = ~4 tons
└── With COP 4, electrical draw = 13.3/4 = 3.3 kW

───

PHASE 3: Dehumidify or Die

Cool air to 22°C but leave it at 80% humidity. You've created a cave. Mold blooms on walls within a week. Your lungs feel heavy. Sweat won't evaporate — your body can't thermoregulate. A room at 22°C and 80% humidity feels worse than 28°C and 40% humidity. Temperature is only half the comfort equation.

Humidity is measured as relative humidity (RH) — the percentage of maximum water vapor the air can hold at that temperature. The key insight: cold air holds less water.

Dew Point — Where Water Falls Out of Air

Air at 42°C and 50% RH contains about 26 g of water per kg of dry air. At 22°C, air can hold a maximum of 16.5 g/kg at 100% RH. So if you cool outdoor air to 22°C, it MUST release:

26 - 16.5 = 9.5 g of water per kg of air

But you want 50% RH at 22°C — that's only 8.3 g/kg. You need to remove even more:

26 - 8.3 = 17.7 g of water per kg of air

To do this, the evaporator coil must be colder than the room target. Much colder.

Air condition              Moisture    Dew Point
──────────────────────────────────────────────────
Outdoor: 42°C, 50% RH     26 g/kg     27.5°C
Target: 22°C, 50% RH      8.3 g/kg    11.1°C
Coil must be BELOW         —           ~7-10°C

The coil runs at 7-10°C. Air passing over it:
1. Cools below dew point (27.5°C → 10°C)
2. Water condenses on the cold fins
3. Air leaves the coil at ~12°C, nearly saturated
4. Mixes with room air to reach 22°C, 50% RHThis is why AC units drip water. That water was dissolved in your air. A typical residential unit in humid climates removes 2-5 liters per hour — enough to fill a bucket every few hours.

Latent vs Sensible Heat

Cooling has two components:
├── Sensible heat: temperature change (you feel it)
├── Latent heat: moisture removal (you feel it indirectly)

Removing water vapor takes enormous energy because of the latent heat of vaporization:

h_fg = 2,450 kJ/kg at room temperature

For our 200 m² space, removing 17.7 g/kg from an airflow of 0.5 kg/s:

Latent load = 0.5 × 0.0177 × 2,450,000
Latent load = 21,682 W ≈ 21.7 kW

Wait — that's LARGER than the sensible cooling load of 13.3 kW. In hot, humid climates, dehumidification can be 50-60% of total cooling energy. This is why oversized AC systems are bad — they cool the air too fast, shut off, and never run long enough to dehumidify.

The Sensible Heat Ratio (SHR)

SHR = Q_sensible / Q_total

├── Desert climate (Phoenix): SHR ≈ 0.85 — mostly temperature, little humidity
├── Humid climate (Miami): SHR ≈ 0.55 — almost half the load is moisture
├── Moderate climate (London): SHR ≈ 0.75

The AC system must match the SHR of the space. If it doesn't, you either:
├── Cool too fast, never dehumidify → clammy 22°C
└── Dehumidify too much → dry 22°C, uncomfortable skin and throat

DESIGN SPEC UPDATED:
├── Coil temperature: 7-10°C (below dew point of incoming air)
├── Moisture removal: up to 17.7 g/kg, ~2-5 liters/hour condensate
├── Latent heat of water: 2,450 kJ/kg — dehumidification is energy-expensive
├── Latent load can exceed sensible load in humid climates (SHR < 0.6)
└── Oversized systems cool fast but fail to dehumidify — worse comfort

───

PHASE 4: Move the Air

The evaporator coil is 7°C. The room is 22°C. But the coil is a 0.5 m² block of aluminum fins tucked inside a wall unit. How does its cold reach every corner of a 200 m² space? You need to MOVE air — thousands of cubic meters per hour, silently.

A fan doesn't cool air. It moves air across the cold coil, and then pushes that chilled air into the room. The amount of air you move determines everything: cooling capacity, noise, and comfort.

The Fan Laws — Physics of Moving Air

Three laws govern fans. All relate to RPM:

Q ∝ RPM           (airflow is proportional to speed)
ΔP ∝ RPM²         (pressure rise goes as speed squared)
P ∝ RPM³          (power goes as speed CUBED)

That third law is devastating. Double the fan speed:
├── Airflow: 2×
├── Pressure: 4×
├── Power: 8×

This is why running a fan at 80% speed instead of 100% saves nearly half the power:

Fan Speed (%)    Airflow (%)    Power (%)    Noise
───────────────────────────────────────────────────
            100            100          loud
             80             51          moderate
             60             22          quiet
             40              6          whisper
             20              1          silentAt 60% speed you get 60% of the airflow but use only 22% of the power. This is why variable-speed fans dominate modern HVAC — they cruise at low speed most of the time.

Duct Sizing — Speed vs Noise

Air moving through a duct makes noise. The noise comes from turbulence, and turbulence increases with velocity. The rule:

├── Below 3 m/s: silent — can't hear the system
├── 3-5 m/s: acceptable for offices and bedrooms
├── 5-8 m/s: noticeable — fine for retail, kitchens
├── Above 8 m/s: loud — only for industrial spaces

For our 200 m² space, we need about 2,000 m³/hr of airflow (roughly 10 air changes per hour of the conditioned volume). Through a main duct at 4 m/s:

A = Q / v
A = (2,000 / 3,600) / 4
A = 0.556 / 4
A = 0.139 m²

That's a duct roughly 400 mm × 350 mm — bigger than most people expect. This is why ductwork dominates the ceiling void in commercial buildings. Want to halve the duct size? Double the velocity. But then the noise quadruples and the power octuples.

Air Throw and Distribution

Cold air is denser than warm air. Blow it horizontally from a ceiling diffuser and it drops. The distance air travels before losing momentum is called throw.

For comfortable distribution:
├── Supply air temperature: 12-14°C (10°C below room temp)
├── Throw distance: 4-6 meters per diffuser
├── Terminal velocity: 0.25 m/s at the edge of the occupied zone
├── No draft on necks — keep supply velocity below 0.2 m/s at head height

Ceiling slot diffusers use the Coanda effect: air clings to the ceiling surface. This extends the throw and prevents cold dumps. The air mixes with room air as it travels, reaching the far wall at near-room temperature.

DESIGN SPEC UPDATED:
├── Fan laws: Q∝RPM, ΔP∝RPM², P∝RPM³ — cube law dominates energy use
├── Running at 60% speed uses only 22% of full power
├── Duct velocity: 3-5 m/s for quiet operation
├── Main duct: ~400×350 mm for 2,000 m³/hr
└── Supply air at 12-14°C, Coanda effect keeps cold air on ceiling

───

PHASE 5: The Outdoor Unit

Walk past the back of any building. Hear that roar? See those big metal boxes with spinning fans? That's where all the heat from inside goes. The condenser has the hardest job in the system: reject heat into air that's ALREADY hot.

The condenser must dump Q_cold + W into outdoor air. For our system:
├── Q_cold = 13.3 kW (heat removed from room)
├── W = 3.3 kW (compressor electrical input)
├── Q_hot = 16.6 kW (total heat to reject outdoors)

The condenser coil temperature must be ABOVE outdoor air temperature — typically 50-55°C when outdoor air is 42°C. That 8-13°C difference drives the heat transfer.

Fin-and-Tube Heat Exchanger — Why Units Are So Big

outdoor air flow →→→→→→→→→→→

    ┌─┐ ┌─┐ ┌─┐ ┌─┐ ┌─┐ ┌─┐    ← aluminum fins
    │○│ │○│ │○│ │○│ │○│ │○│      (0.1 mm thick)
    ├─┤ ├─┤ ├─┤ ├─┤ ├─┤ ├─┤      (2 mm spacing)
    │○│ │○│ │○│ │○│ │○│ │○│    ← copper tubes
    ├─┤ ├─┤ ├─┤ ├─┤ ├─┤ ├─┤      (9.5 mm diameter)
    │○│ │○│ │○│ │○│ │○│ │○│
    ├─┤ ├─┤ ├─┤ ├─┤ ├─┤ ├─┤    hot refrigerant gas
    │○│ │○│ │○│ │○│ │○│ │○│    flows through tubes,
    └─┘ └─┘ └─┘ └─┘ └─┘ └─┘    condenses to liquid,
                                  releases heat to fins,
    ○ = copper tube (refrigerant)  fins transfer to air
    │ = aluminum fin (air side)The fins multiply the air-side surface area by 15-20×. Without fins, you'd need a condenser the size of a room. With fins: a box that fits on a balcony. The tradeoff: fins clog with dust, reducing performance by 10-20% per year without cleaning.

Heat Rejection Calculation

The condenser follows the same Q = UAΔT, but the air-side heat transfer coefficient is poor — air is a terrible heat transfer fluid compared to water:

Q = h × A_fin × ΔT_lm

Where:
├── h = air-side heat transfer coefficient ≈ 50-80 W/m²·K
├── A_fin = total fin surface area (m²)
├── ΔT_lm = log-mean temperature difference ≈ 10°C

For Q_hot = 16.6 kW, h = 60 W/m²·K, ΔT_lm = 10°C:

A_fin = 16,600 / (60 × 10)
A_fin = 27.7 m²

27.7 m² of fin surface. Packed at 500 fins per meter, that's a coil face area of about 0.6 m² — roughly 800 mm × 750 mm. Add the fan, compressor, and casing: a unit about 900 × 350 × 800 mm.

The Airflow Requirement

The condenser fan must push enough air to carry away 16.6 kW. Air has a specific heat of ~1,005 J/kg·K and density ~1.1 kg/m³:

Q = ṁ × c_p × ΔT_air

For a 10°C air temperature rise across the coil:

ṁ = 16,600 / (1,005 × 10)
ṁ = 1.65 kg/s

Volume flow = 1.65 / 1.1 = 1.5 m³/s = 5,400 m³/hr

That's a LOT of air. The fan is typically 500-600 mm diameter, spinning at 800-1,200 RPM. This is why outdoor units are loud — the condenser fan moves more air than the indoor unit by a factor of 3.

DESIGN SPEC UPDATED:
├── Heat rejection: Q_hot = 16.6 kW (Q_cold + W)
├── Condenser coil: 50-55°C, fin-and-tube, ~28 m² fin area
├── Condenser airflow: 5,400 m³/hr through 500-600 mm axial fan
├── Air-side h ≈ 60 W/m²·K (poor — this is why coils must be large)
└── Dirty fins reduce capacity 10-20% — cleaning is the #1 maintenance task

───

PHASE 6: Variable Speed

It's 3 AM. The building has been cool for hours. Solar gain is zero. Internal gains dropped to a few hundred watts from standby electronics. The cooling load is now 2 kW — about 15% of peak design. Your 13 kW system turns on, cools the room to 21°C in 4 minutes, turns off. Waits. Temperature drifts to 23°C. Turns on again. Repeat, all night. This is called short-cycling, and it wastes 30% of your electricity.

Fixed-speed compressors have one mode: ON or OFF. The solution is an inverter compressor — a variable-speed motor that adjusts its RPM to match the load.

Part-Load Reality

A building spends about 80% of its operating hours below 50% load. The design condition (42°C outdoor, peak sun, full occupancy) happens maybe 200 hours per year. The other 2,800 hours are partial load.

Load (%)     Hours/yr    Fixed COP    Inverter COP    Energy (fixed)    Energy (inv)
────────────────────────────────────────────────────────────────────────────────────
100          200         3.5          3.5             760 kWh           760 kWh
75           400         3.5          4.5             1,143 kWh         889 kWh
50           800         3.5          5.5             1,524 kWh         970 kWh
25           1,000       3.5          6.5             952 kWh           513 kWh
10           600         3.5          7.0             229 kWh           114 kWh
────────────────────────────────────────────────────────────────────────────────────
TOTAL        3,000                                     4,608 kWh         3,246 kWhThe inverter saves 30% of annual energy — not by being more efficient at full load, but by being much more efficient at PARTIAL load, where the system operates 80% of the time. SEER (Seasonal Energy Efficiency Ratio) captures this reality better than COP.

How the Inverter Works

A standard compressor motor runs on 50 Hz AC power — fixed at 2,900 RPM (for a 2-pole motor). An inverter converts:

AC (50 Hz) → DC → AC (variable Hz)

By varying the output frequency from 15 Hz to 120 Hz, the compressor speed ranges from ~900 to ~7,200 RPM. At 15 Hz, the compressor barely turns — delivering just 10% capacity, sipping power.

The efficiency improvement comes from thermodynamics:
├── Lower speed → lower compression ratio needed
├── Lower compression ratio → less work per kg of refrigerant
├── Lower speed → less friction and mechanical losses
├── Lower condenser load → smaller ΔT → COP approaches Carnot

At 25% load, the inverter COP can reach 6.5 — nearly double the full-load value.

SEER vs COP — The Real Efficiency Number

COP is measured at one condition. SEER (Seasonal Energy Efficiency Ratio) averages across all conditions:

SEER = total cooling (BTU) / total energy (Wh)

Conversion: SEER ≈ COP × 3.412

├── Minimum legal SEER (US, 2023): 14 (COP ~4.1)
├── Good inverter unit: 20-22 (COP ~6.0)
├── Best available (2024): 33 (COP ~9.7)
└── The gap between worst and best: 2.4× energy cost difference

DESIGN SPEC UPDATED:
├── Inverter compressor: 15-120 Hz, speed matches load
├── 80% of operating hours are below 50% load
├── Part-load COP: up to 6.5 at 25% load (vs 3.5 fixed)
├── Annual energy savings: ~30% vs fixed-speed
└── SEER captures seasonal reality — best units reach SEER 33

───

PHASE 7: Split or Package

You've designed the cycle, sized the load, picked an inverter compressor. Now: where does the hardware physically go? You have two options — put everything in one box, or split it into two pieces connected by copper pipes.

Split System — The Dominant Design

A split system separates the quiet part (evaporator + fan) from the loud part (compressor + condenser + fan):

INDOOR UNIT                        OUTDOOR UNIT
            (wall/ceiling mount)                  (balcony/ground)
          ┌──────────────────┐              ┌──────────────────┐
          │  evaporator coil │              │  condenser coil  │
          │  blower fan      │  ←─copper──→ │  compressor      │
          │  air filter      │    pipes     │  condenser fan   │
          │  drain pan       │  (2 tubes)   │  expansion valve │
          │  control board   │              │  electrical      │
          └──────────────────┘              └──────────────────┘
           ~900×300×200 mm                   ~900×350×800 mm
           25-35 dB                          48-55 dB
           ~12 kg                            ~45 kg

          Connected by:
          ├── Liquid line: 6.35 mm (1/4") copper — high-pressure liquid
          ├── Suction line: 12.7 mm (1/2") copper — low-pressure gas
          ├── Drain hose: 16 mm PVC — condensate water
          └── Control cable: 3-4 conductorThe split system puts the noisy compressor outside and keeps only the whisper-quiet fan indoors. This is why it dominates residential and small commercial — noise at the occupant is 25-35 dB, quieter than a library.

Refrigerant Line Length — The Hidden Limit

The copper pipes between indoor and outdoor units carry refrigerant. Longer pipes mean:

├── More pressure drop → lower COP (compressor works harder)
├── More refrigerant charge needed → higher cost, more leak risk
├── More oil management issues (lubricant travels with refrigerant)

Distance (m)    COP Loss    Charge Added    Practical?
──────────────────────────────────────────────────────────
0-5             ~0%         nominal         ideal
5-15            2-5%        +5-15%          standard
15-30           5-10%       +15-30%         acceptable
30-50           10-20%      +30-50%         requires care
50-75           20-35%      +50-80%         VRF systems only
75+             >35%        custom          not recommendedMost residential splits are designed for 5-15 m. VRF (Variable Refrigerant Flow) systems can reach 75 m by using larger pipes and more sophisticated oil management.

Packaged vs Split — When to Use Which

Packaged unit: everything in one box. Goes on the roof or beside the building.
├── Pros: single installation point, no refrigerant piping on-site
├── Cons: ductwork needed to distribute air, noise enters building
├── Used for: commercial rooftops, schools, retail

Split system: two boxes connected by pipes.
├── Pros: quiet indoors, no ductwork needed, flexible placement
├── Cons: refrigerant piping on-site, max distance limited
├── Used for: residential, offices, apartments

Multi-split: one outdoor unit, 2-8 indoor units.
├── Pros: one outdoor unit serves multiple rooms, individual control
├── Cons: complex piping, one compressor failure affects all rooms
├── Used for: apartments, small offices, restaurants

VRF (Variable Refrigerant Flow): one outdoor unit, up to 64 indoor units.
├── Pros: massive buildings with one system, heat recovery possible
├── Cons: expensive, complex commissioning, refrigerant leak risk
├── Used for: hotels, large offices, hospitals

DESIGN SPEC UPDATED:
├── Split system: indoor (evaporator) + outdoor (compressor+condenser)
├── Connected by 2 copper pipes (6.35 mm liquid, 12.7 mm suction)
├── Max practical line length: 15 m residential, 75 m VRF
├── Indoor noise: 25-35 dB (split), outdoor: 48-55 dB
└── For 200 m²: multi-split with 3-4 indoor units or ducted system

───

PHASE 8: Filter the Air

You're breathing 15,000 liters of air per day. In that air: dust mites, pollen, mold spores, bacteria, cooking particulates, tire dust from the road, and — in cities — PM2.5 from combustion. Your lungs are filters, but they weren't designed for industrial civilization. The AC system is your first line of defense.

The Particle Size Spectrum

Not all particles are equal. Size determines everything — how they behave, how deep they penetrate your lungs, and how hard they are to filter.

Size (μm)     What                            Lung Penetration    Filter
─────────────────────────────────────────────────────────────────────────
         Sand, visible dust              Nose catches it     any
          Pollen, mold spores             Upper airways       MERV 8
           Dust mite fragments             Bronchi             MERV 11
5           PM2.5 — combustion, smoke      Deep lung (alveoli) MERV 13
0           Bacteria                        Deep lung           HEPA
3           MPPS — most penetrating size   Deep lung           HEPA
1           Viruses, ultrafine particles    Deep lung + blood   HEPA
01          Nanoparticles                   Crosses into blood  ULPA0.3 μm is the Most Penetrating Particle Size (MPPS) — too big for diffusion capture, too small for inertial capture. This is where filters are weakest. HEPA is defined as 99.97% efficient at exactly this size.

How Filters Actually Work — Three Mechanisms

Filters don't work like sieves (holes smaller than particles). They use three physical mechanisms that depend on particle size:

1. Inertial impaction (particles > 1 μm):
Large particles can't follow the air as it curves around a filter fiber. Their inertia carries them straight into the fiber. Like a truck that can't make a sharp turn.

2. Interception (particles 0.3-1 μm):
Medium particles follow the airflow but pass close enough to a fiber to touch it and stick. Van der Waals forces hold them.

3. Diffusion (particles < 0.1 μm):
Tiny particles are buffeted randomly by air molecules (Brownian motion). They wander into fibers by random chance. Slower air = more time to wander = better capture.

At 0.3 μm, none of these mechanisms works well. Impaction is too weak (not enough inertia). Diffusion is too weak (not small enough for strong Brownian motion). This is the MPPS — the valley between two capture mechanisms.

HEPA — The Gold Standard

HEPA: High Efficiency Particulate Air
├── Definition: 99.97% capture at 0.3 μm (MPPS)
├── Construction: glass microfiber, fiber diameter 0.5-2 μm
├── Depth: 50-80 mm of pleated media
├── Pressure drop: 250-500 Pa (significant — needs strong fan)

The catch: HEPA filters need high fan pressure. A typical split AC fan produces 50-150 Pa. HEPA needs 250-500 Pa. So most residential AC units use MERV 8-13 filters instead:

├── MERV 8: captures 70% of 3-10 μm, 20% of 1-3 μm — basic dust
├── MERV 11: captures 85% of 3-10 μm, 65% of 1-3 μm — good residential
├── MERV 13: captures 90% of 3-10 μm, 85% of 1-3 μm — hospital standard
├── HEPA: captures 99.97% at 0.3 μm — clean rooms, operating theaters

Every step up in filtration requires more fan power. A MERV 13 filter adds ~80 Pa of pressure drop. The fan must work harder, consuming more electricity. There's always a tradeoff between air quality and energy.

DESIGN SPEC UPDATED:
├── MPPS: 0.3 μm — hardest particle size to capture
├── Three capture mechanisms: impaction (>1μm), interception (0.3-1μm), diffusion (<0.1μm)
├── HEPA: 99.97% at 0.3 μm, needs 250-500 Pa (too much for most AC fans)
├── Practical choice: MERV 11-13 for residential (80-150 Pa pressure drop)
└── Filter pressure drop directly increases fan energy consumption

───

PHASE 9: Control the Comfort

Set the thermostat to 22°C. The room reaches 22°C. The compressor shuts off. Three minutes later: 23.5°C. Compressor turns on. Five minutes later: 21°C. Overshoots. Turns off again. The temperature oscillates ±1.5°C around your setpoint. You never actually experience 22°C — you experience a sawtooth wave.

Simple on/off control creates oscillation because of thermal mass — the building's ability to store heat. Furniture, walls, floors, and the air itself all have thermal capacity. When the compressor stops, the room doesn't instantly warm — the stored cold slowly dissipates.

PID Control — The Three-Term Solution

Modern thermostats use PID control — Proportional, Integral, Derivative:

Output = Kp × e(t) + Ki × ∫e(t)dt + Kd × de(t)/dt

Where e(t) = setpoint - actual temperature.

Term          What It Does                      Without It
──────────────────────────────────────────────────────────────────
P (Proportional)  Output ∝ current error          No response at all
                   Far from setpoint → full power
                   Close → low power

I (Integral)      Output ∝ accumulated error      Permanent offset
                   Fixes the "droop" where P alone   (22.5°C instead
                   settles slightly above/below      of 22.0°C)

D (Derivative)    Output ∝ rate of change         Overshoot
                   Slows down as you approach       (cools to 20°C
                   setpoint — "anticipation"        before stopping)With PID + inverter compressor, the system continuously adjusts compressor speed. Room temperature holds within ±0.3°C of setpoint — you can't feel the difference. On/off systems oscillate ±1.5°C.

Thermal Mass — The Room's Memory

Every material stores heat. The time it takes a room to warm up after the AC shuts off depends on:

τ = (m × c_p) / (UA)

Where:
├── m × c_p = thermal capacity (J/K) — how much heat the room can store
├── UA = heat loss rate (W/K) — how fast heat leaks in

For a typical 200 m² apartment:

Thermal capacity:
├── Air: 200 × 2.5m × 1.2 kg/m³ × 1,005 = 602 kJ/K
├── Furniture: ~3,000 kg × 1,500 J/kg·K = 4,500 kJ/K
├── Interior walls: ~8,000 kg × 840 J/kg·K = 6,720 kJ/K
├── Floor slab: ~12,000 kg × 880 J/kg·K = 10,560 kJ/K
────────────────────
Total: ~22,382 kJ/K

UA ≈ 200 W/K (from our Phase 2 calculation)

τ = 22,382,000 / 200 = 111,910 seconds ≈ 31 hours

The thermal time constant is 31 hours. That means if you turn off the AC at night, the room takes many hours to warm up significantly. This is why pre-cooling works — cool the building's mass during cheap nighttime electricity, and it coasts through the morning peak.

The Dead Band Problem

A thermostat has a dead band — a temperature range where it does nothing. Set to 22°C with a ±0.5°C dead band:
├── Below 21.5°C: heating or compressor OFF
├── 21.5-22.5°C: no action
├── Above 22.5°C: cooling ON

Narrow dead band (±0.2°C) → more precise but compressor cycles more frequently
Wide dead band (±1.0°C) → fewer cycles but temperature swings noticeably

With an inverter compressor and PID, the dead band shrinks to ±0.1°C because the compressor modulates speed rather than cycling on/off.

DESIGN SPEC UPDATED:
├── PID control: P (proportional) + I (integral offset correction) + D (anticipation)
├── Inverter + PID holds ±0.3°C vs ±1.5°C for on/off
├── Thermal time constant: ~31 hours for typical apartment
├── Pre-cooling strategy: use thermal mass to shift peak load
└── Dead band: ±0.1°C with inverter, ±0.5-1.0°C with on/off

───

PHASE 10: The Global Problem

There are 2 billion air conditioners on Earth. By 2050, there will be 5.6 billion. Air conditioning consumes 10% of all global electricity — more than the entire continent of Africa uses for everything. And the refrigerants inside those units, if leaked, are greenhouse gases thousands of times more potent than CO₂.

The Cooling Demand Explosion

Most of the world's population lives in hot climates. As incomes rise, AC is the first major appliance people buy. The growth is staggering:

Region               2020         2050 (projected)    Growth
────────────────────────────────────────────────────────────────
China                570M        1,050M              1.8×
India                120M        1,140M              9.5×
SE Asia              90M         470M               5.2×
Africa               30M         380M               12.7×
Middle East          70M         170M               2.4×
USA                  380M        450M               1.2×
Europe               110M        275M               2.5×
Rest of World        130M        350M               2.7×
────────────────────────────────────────────────────────────────
TOTAL                ~2.0B       ~5.6B              2.8×India alone will add more than 1 billion AC units. Africa will grow 12.7×. The question isn't whether this happens — rising temperatures and rising incomes guarantee it. The question is whether the grid and the atmosphere can handle it.

The Refrigerant Problem — GWP

The refrigerant inside your AC is a potent greenhouse gas. GWP (Global Warming Potential) measures how much worse it is than CO₂ over 100 years:

├── CO₂:        GWP = 1 (reference)
├── R-410A:     GWP = 2,088 (current standard for residential AC)
├── R-32:       GWP = 675 (newer alternative, gaining market share)
├── R-290:      GWP = 3 (propane — nearly zero, but flammable)
├── R-1234yf:   GWP = 4 (automotive AC replacement for R-134a)
├── R-744:      GWP = 1 (CO₂ itself — used in heat pumps)

A typical split AC contains 1-2 kg of R-410A. If it leaks:

Equivalent CO₂ = 1.5 kg × 2,088 = 3,132 kg CO₂-equivalent

That's the same as driving a car 12,500 km. A single AC leak equals a year of driving. With 2 billion units, even a 5% annual leak rate means:

Annual emissions = 2B × 1.5 kg × 0.05 × 2,088 = 313 million tonnes CO₂-eq

The Kigali Amendment (2016) mandates an 80-85% phase-down of high-GWP refrigerants by 2047. The industry is shifting to R-32, R-290, and CO₂.

The Electricity Problem

AC creates a vicious feedback loop:

Hot day
       │
       ▼
    Everyone turns on AC ──→ electricity demand spikes
       │                              │
       ▼                              ▼
    Grid burns more     ◄────  power plants ramp up
    fossil fuels                      │
       │                              ▼
       ▼                      more CO₂ emissions
    Climate warms                     │
       │                              ▼
       └──────── next summer is hotter ◄──────┘

    Peak cooling day in Delhi: 75% of grid capacity is AC
    Peak cooling day in Dubai: 70% of grid capacity is AC
    Peak cooling day in Houston: 50% of grid capacity is ACAC causes the very problem it solves. The only way out: higher efficiency (COP 3→6 halves electricity), clean grids (solar peaks when cooling peaks — natural alignment), and better buildings (reduce the load so you need less AC).

The silver lining: solar generation peaks at almost exactly the same time as cooling demand. A building with rooftop solar and a high-efficiency AC can be nearly grid-neutral on the hottest days.

DESIGN SPEC COMPLETE:
├── Global AC: 2B units today, 5.6B by 2050
├── 10% of global electricity → AC, up to 75% of peak grid in hot cities
├── Refrigerant GWP: R-410A = 2,088 → shifting to R-32 (675) and R-290 (3)
├── Kigali Amendment: 80-85% phase-down of high-GWP refrigerants by 2047
└── Solar + high-COP AC + better envelopes = the path forward

───

FULL MAP

Air Conditioner
├── Phase 1: Move Heat Backward
│   ├── Cycle: vapor compression (evaporate → compress → condense → expand)}
│   ├── Refrigerant: R-410A, boils at -51°C at atmospheric pressure}
│   ├── Carnot COP (22→42°C): 14.75 theoretical maximum}
│   ├── Real COP: 3-5 (20-34% of Carnot limit)}
│   └── Q_hot = Q_cold + W — outdoor unit rejects MORE heat than indoor unit removes}
│
├── Phase 2: Size the Load
│   ├── Envelope load: Q = UAΔT → 2,720 W (windows dominate)}
│   ├── Solar gain: SHGC × A × I → 5,000 W (biggest single source)}
│   ├── Internal gains: people + computers + lights → 4,400 W}
│   ├── Total cooling load: ~13.3 kW = 45,400 BTU/hr = ~4 tons}
│   └── With COP 4, electrical draw = 13.3/4 = 3.3 kW}
│
├── Phase 3: Dehumidify or Die
│   ├── Coil temperature: 7-10°C (below dew point of incoming air)}
│   ├── Moisture removal: up to 17.7 g/kg, ~2-5 liters/hour condensate}
│   ├── Latent heat of water: 2,450 kJ/kg — dehumidification is energy-expensive}
│   ├── Latent load can exceed sensible load in humid climates (SHR < 0.6)}
│   └── Oversized systems cool fast but fail to dehumidify — worse comfort}
│
├── Phase 4: Move the Air
│   ├── Fan laws: Q∝RPM, ΔP∝RPM², P∝RPM³ — cube law dominates energy use}
│   ├── Running at 60% speed uses only 22% of full power}
│   ├── Duct velocity: 3-5 m/s for quiet operation}
│   ├── Main duct: ~400×350 mm for 2,000 m³/hr}
│   └── Supply air at 12-14°C, Coanda effect keeps cold air on ceiling}
│
├── Phase 5: The Outdoor Unit
│   ├── Heat rejection: Q_hot = 16.6 kW (Q_cold + W)}
│   ├── Condenser coil: 50-55°C, fin-and-tube, ~28 m² fin area}
│   ├── Condenser airflow: 5,400 m³/hr through 500-600 mm axial fan}
│   ├── Air-side h ≈ 60 W/m²·K (poor — this is why coils must be large)}
│   └── Dirty fins reduce capacity 10-20% — cleaning is the #1 maintenance task}
│
├── Phase 6: Variable Speed
│   ├── Inverter compressor: 15-120 Hz, speed matches load}
│   ├── 80% of operating hours are below 50% load}
│   ├── Part-load COP: up to 6.5 at 25% load (vs 3.5 fixed)}
│   ├── Annual energy savings: ~30% vs fixed-speed}
│   └── SEER captures seasonal reality — best units reach SEER 33}
│
├── Phase 7: Split or Package
│   ├── Split system: indoor (evaporator) + outdoor (compressor+condenser)}
│   ├── Connected by 2 copper pipes (6.35 mm liquid, 12.7 mm suction)}
│   ├── Max practical line length: 15 m residential, 75 m VRF}
│   ├── Indoor noise: 25-35 dB (split), outdoor: 48-55 dB}
│   └── For 200 m²: multi-split with 3-4 indoor units or ducted system}
│
├── Phase 8: Filter the Air
│   ├── MPPS: 0.3 μm — hardest particle size to capture}
│   ├── Three capture mechanisms: impaction (>1μm), interception (0.3-1μm), diffusion (<0.1μm)}
│   ├── HEPA: 99.97% at 0.3 μm, needs 250-500 Pa (too much for most AC fans)}
│   ├── Practical choice: MERV 11-13 for residential (80-150 Pa pressure drop)}
│   └── Filter pressure drop directly increases fan energy consumption}
│
├── Phase 9: Control the Comfort
│   ├── PID control: P (proportional) + I (integral offset correction) + D (anticipation)}
│   ├── Inverter + PID holds ±0.3°C vs ±1.5°C for on/off}
│   ├── Thermal time constant: ~31 hours for typical apartment}
│   ├── Pre-cooling strategy: use thermal mass to shift peak load}
│   └── Dead band: ±0.1°C with inverter, ±0.5-1.0°C with on/off}
│
└── Phase 10: The Global Problem
    ├── Global AC: 2B units today, 5.6B by 2050}
    ├── 10% of global electricity → AC, up to 75% of peak grid in hot cities}
    ├── Refrigerant GWP: R-410A = 2,088 → shifting to R-32 (675) and R-290 (3)}
    ├── Kigali Amendment: 80-85% phase-down of high-GWP refrigerants by 2047}
    └── Solar + high-COP AC + better envelopes = the path forward}

───

← Washing Machine     Oil Refinery →