DEV Community: Evgenii Konkin

Thermal Mass: Why a Concrete Building Peaks 5 Hours After the Sun Does

Evgenii Konkin — Fri, 06 Nov 2026 04:08:35 +0000

In a metal building, the cooling load peaks within an hour of peak sun. In a heavy concrete building, the peak shifts to late afternoon. The difference — thermal lag — changes equipment sizing.

Thermal lag by wall type

Construction                  Time Lag
──────────────────────────────────────────
Metal panel (lightweight)     0-1 hour
Wood frame + insulation       2-3 hours
4" concrete block             3-4 hours
8" concrete block             5-6 hours
12" poured concrete           7-8 hours
Brick veneer over block       6-8 hours

Heavy materials absorb solar heat, store it for hours, then release it to the interior.

Why the peak shifts

Lightweight metal building:
  Sun heats wall → heat enters interior in <1 hour
  Peak solar at noon → peak cooling load at 1 PM

Heavy concrete building:
  Sun heats wall → concrete absorbs and stores heat
  Heat enters interior 5-6 hours later
  Peak solar at noon → peak cooling load at 5-6 PM

Real case: concrete office in Dallas

Hour    Solar    Wall    Roof    Internal    TOTAL
────────────────────────────────────────────────────
Noon    peak     absorb  enter   80%         38 tons
3 PM    decline  release start   100%        44 tons
5 PM    low      release peak    100%        47 tons ← ACTUAL PEAK
7 PM    zero     still   still   dropping    35 tons

Engineer sized at noon: 42 tons.
Actual peak at 5 PM: 47 tons.
Result: 11% undersized. West side 3°F above setpoint every afternoon.

Multiple components peak at different hours

Component               Peak Hour (south/west exposure)
────────────────────────────────────────────────────────
East window solar       8-10 AM
South window solar      11 AM - 1 PM
West window solar       3-5 PM
Roof (metal deck)       12-2 PM
Roof (concrete)         4-6 PM
Wall (metal)            12-2 PM
Wall (concrete)         5-8 PM
People/lights/equip     9 AM - 5 PM

The total cooling load = sum of ALL components at EACH hour. The hour with the highest sum is the design hour — not the hour with the highest single component.

Lightweight vs heavyweight

Metal building:
  + Peak occurs near solar peak (predictable)
  + Load drops fast after sunset
  - Load spikes quickly with sun angle changes
  - Morning pre-cool needed less

Concrete building:
  + Morning stays cool longer (stored nighttime cool)
  + Load rises gradually
  - Peak delays to late afternoon
  - System must run later into evening
  - Nighttime "coast-down" still releases stored heat

The rule

For lightweight construction, the noon load calculation is close to the peak. For heavyweight construction, run the calculation at every hour from noon to 8 PM. The peak is almost always in late afternoon — and it can be 10-20% higher than the noon value.

Quick Tip: CFM Per Ton — The Fan Speed Setting That Changes Humidity by 13%

Evgenii Konkin — Thu, 05 Nov 2026 03:59:17 +0000

The standard residential airflow is 400 CFM per ton of cooling. But that number is a design compromise — and adjusting it up or down changes humidity control more than most people expect.

The tradeoff

Lower airflow (350 CFM/ton):
  Coil surface gets colder
  Air spends more time on the coil
  More moisture condenses and drains
  Sensible capacity drops ~8%
  → Best for humid climates

Standard airflow (400 CFM/ton):
  Balanced sensible + latent
  Equipment runs at rated capacity
  → Default for most applications

Higher airflow (450 CFM/ton):
  Coil surface stays warmer
  Air passes through faster
  Less moisture removed
  Sensible capacity increases 5-8%
  → Best for dry climates only

The numbers

3-ton system (36,000 BTU/hr rated):

350 CFM/ton = 1,050 CFM:
  Total capacity: 33,200 BTU/hr (-8%)
  SHR: 0.72 (28% of cooling = moisture removal)
  Room humidity: ~45%

400 CFM/ton = 1,200 CFM:
  Total capacity: 36,000 BTU/hr (rated)
  SHR: 0.78 (22% = moisture removal)
  Room humidity: ~50%

450 CFM/ton = 1,350 CFM:
  Total capacity: 38,800 BTU/hr (+8%)
  SHR: 0.85 (15% = moisture removal)
  Room humidity: ~58%

Same equipment. Same tonnage. 13 percentage points of humidity difference from fan speed alone.

Why lower airflow dehumidifies better

At 350 CFM/ton:
  Air velocity across coil: lower
  Contact time: longer
  Coil surface temp: colder (closer to refrigerant temp)
  More air reaches dew point on the coil surface
  More moisture condenses per CFM

At 450 CFM/ton:
  Air velocity: higher
  Contact time: shorter
  Coil surface temp: warmer (air warms it faster)
  Less air reaches dew point
  Less moisture per CFM — but more total air moved

When to adjust

Climate              Recommendation    Why
──────────────────────────────────────────────────────
Houston, Miami       350 CFM/ton       Humidity is the problem
Atlanta, Charlotte   375 CFM/ton       Moderate humidity
Most US cities       400 CFM/ton       Standard
Denver, Phoenix      425-450 CFM/ton   Maximize sensible capacity
                                       humidity is never an issue

How to change it

Residential systems:
  Most air handlers have multiple fan speed taps
  (low, med-low, med, med-high, high)
  Installer selects the tap during commissioning
  Each tap changes CFM by ~50-75 per step

Commercial systems:
  VFD on the supply fan
  BAS controls fan speed
  Can adjust dynamically by season or humidity

The warning

Below 300 CFM/ton:
  Coil surface drops below 32F
  → Ice forms on the coil
  → Airflow drops further
  → More ice → complete freeze-up

Above 500 CFM/ton:
  Coil cannot cool the air enough
  → Supply air temp too high
  → Delta-T drops below useful range
  → System runs but doesn't cool effectively

350-450 CFM/ton is the operating range. Below or above causes equipment damage or failure.

Calculate the CFM for your system with the CFM calculator on CalcEngineer — then decide where in the 350-450 range your climate needs to be.

Quick Tip: Superheat and Subcooling — Two Delta-T Measurements That Diagnose Refrigerant Charge

Evgenii Konkin — Wed, 04 Nov 2026 03:38:37 +0000

Superheat and subcooling are both delta-T measurements — just on the refrigerant side instead of the air side. Together they tell you the charge condition without connecting a single gauge.

What they measure

Superheat (SH):
  T_suction_line - T_saturation(at suction pressure)
  = how much the vapor heated above boiling point
  Normal (TXV): 10-15F
  Normal (fixed orifice): 15-20F

Subcooling (SC):
  T_saturation(at discharge pressure) - T_liquid_line
  = how much the liquid cooled below condensing point
  Normal (TXV): 10-15F
  Normal (fixed orifice): 10-18F

The diagnostic matrix

Superheat    Subcooling    Diagnosis
──────────────────────────────────────────────────
Normal       Normal        System properly charged ✓
HIGH         LOW           Undercharged (most common)
LOW          HIGH          Overcharged
HIGH         Normal        Restricted metering device
                           or low evaporator airflow
Normal       HIGH          Liquid line restriction
                           or slightly overcharged
LOW          LOW           TXV stuck open or
                           sensing bulb issue

Touch diagnostics (no thermometer needed)

Suction line (big pipe, vapor)

Cold and sweating (condensation on pipe)
  → Superheat is in range ✓

Warm and dry (no condensation)
  → Superheat too high
  → Likely undercharged or low airflow ✗

Ice-cold or frosted
  → Superheat too low
  → Likely overcharged or flooded evaporator ✗

Liquid line (small pipe, liquid)

Warm (outdoor temp minus 10-20F)
  → Subcooling is in range ✓

Very hot (close to outdoor temp)
  → Subcooling too low
  → Likely undercharged ✗

Cool (well below outdoor temp)
  → Subcooling too high
  → Likely overcharged ✗

Connection to air-side delta-T

Air delta-T across evaporator = symptom
Superheat/subcooling = cause

High superheat (undercharged):
  → Less effective evaporator area
  → Air delta-T drops below 14F
  → Low cooling capacity

Low superheat (overcharged):
  → Coil floods with liquid
  → Air delta-T may look normal
  → But compressor is at risk of liquid slugging

The numbers for a typical residential system

R-410A at 75F indoor, 95F outdoor:

Suction pressure: ~118 psi → T_sat = 41F
Suction line temp: 53F
Superheat = 53 - 41 = 12F ✓

Discharge pressure: ~350 psi → T_sat = 115F
Liquid line temp: 101F
Subcooling = 115 - 101 = 14F ✓

The rule

Air delta-T tells you the system is underperforming. Superheat and subcooling tell you why. Both are temperature differences — one on the air, one on the refrigerant. Together they give you the complete diagnostic picture.

Quick Tip: Wet-Bulb Temperature — Why 95F Houston Is More Dangerous Than 110F Phoenix

Evgenii Konkin — Tue, 03 Nov 2026 04:51:18 +0000

Dry-bulb temperature (what the thermometer reads) does not tell you how dangerous the heat is. Wet-bulb temperature does — because it captures both heat and humidity in one number.

Wet-bulb vs dry-bulb

Dry-bulb: air temperature measured by a standard thermometer.

Wet-bulb: temperature measured by a thermometer wrapped
          in a wet cloth with air flowing over it.

If the air is dry → evaporation cools the cloth → WB < DB
If the air is saturated → no evaporation → WB = DB

The gap between dry-bulb and wet-bulb tells you how much evaporative cooling is possible.

Why wet-bulb matters for humans

Your body cools itself by sweating. Sweat evaporates and removes heat. The rate of evaporation depends on the wet-bulb temperature — not the dry-bulb.

Phoenix: 110F dry-bulb, 15% RH
  Wet-bulb: 71F
  Sweat evaporates rapidly → body cools effectively
  → Uncomfortable but survivable

Houston: 95F dry-bulb, 75% RH
  Wet-bulb: 89F
  Sweat barely evaporates → body cannot cool
  → Heat stroke risk, even at "lower" temperature

The hotter city (Phoenix) is safer because the air is dry enough for sweat to work.

NIOSH/OSHA heat stress thresholds

Wet-Bulb     Risk Level          Action
──────────────────────────────────────────────────
< 77F        Safe                Normal work
77-82F       Caution             Rest breaks needed
82-86F       Danger              Limit heavy work
86-90F       Extreme danger      Only light work, frequent rest
> 95F        Fatal threshold     Body cannot cool — death without shelter

At wet-bulb 95F, even a healthy person in the shade with unlimited water will die from heat because evaporation stops completely.

Why wet-bulb matters for cooling towers

Cooling towers reject heat through evaporation — the same mechanism as sweat. The tower can cool water down to within 5-7F of the wet-bulb temperature (this gap is called "approach").

City          Dry-Bulb   Wet-Bulb   Tower Leaving Water
──────────────────────────────────────────────────────────
Phoenix       110F       71F        76-78F (excellent)
Denver         95F       62F        67-69F (excellent)
Atlanta        95F       78F        83-85F (adequate)
Houston        95F       82F        87-89F (marginal)
Miami          92F       80F        85-87F (marginal)
Singapore      90F       82F        87-89F (marginal)

Same tower. Same design. Completely different leaving water temperatures — driven entirely by wet-bulb.

Finding wet-bulb from dry-bulb and RH

Approximate formula (Stull, 2011):
WB ≈ DB × atan(0.151977 × (RH + 8.313659)^0.5)
     + atan(DB + RH) - atan(RH - 1.676331)
     + 0.00391838 × RH^1.5 × atan(0.023101 × RH)
     - 4.686035

Or: read it from a psychrometric chart.
Or: use a calculator.

The rule

When evaluating heat danger (for workers) or cooling tower performance (for systems), use wet-bulb — not dry-bulb. A 95F day at 30% RH and a 95F day at 80% RH are completely different conditions. Dry-bulb treats them the same. Wet-bulb does not.

Variable Speed vs Single Stage: Same 3 Tons, 34% Less Energy, Better Humidity

Evgenii Konkin — Mon, 02 Nov 2026 04:11:49 +0000

A 3-ton single-stage AC and a 3-ton inverter AC have the same nameplate tonnage. They deliver that tonnage in completely different ways — and the difference changes energy consumption, humidity control, and compressor life.

How single stage operates

85F day. House needs 1.5 tons.

Single stage (3 tons, one speed):
  ON:  3.0 tons for 8 minutes → house cools to 71F
  OFF: 0 tons for 15 minutes → house warms to 74F
  ON:  3.0 tons for 8 minutes → house cools to 71F

  Cycles: 3-4 per hour
  Runtime: ~35% of each hour
  Start-up surge: 5-7× running current, 4 times per hour

The compressor is either at 100% or 0%. It overshoots on cooling, then the house drifts warm until the thermostat calls again.

How variable speed operates

Same 85F day. Same 1.5-ton load.

Inverter (3 tons, variable speed):
  Runs continuously at 50% capacity = 1.5 tons
  House stays at 72F ±0.5F
  No cycling. No start-up surges.
  Compressor modulates between 25% and 100% as needed.

Why inverter uses less energy

1. No start-up penalty

Single stage: 4 starts/hour × 5× inrush × 3 seconds
  = 60 seconds of inrush current per hour

Inverter: 1 start per day
  = negligible inrush

2. Better part-load COP

Compressor efficiency by load:

Load        COP (single stage)    COP (inverter)
──────────────────────────────────────────────────
100%        3.5                   3.5
80%         N/A (can't do 80%)    4.2
60%         N/A                   4.8
40%         N/A                   5.0+

The inverter runs most hours at 40-60% load where COP is highest. The single stage can only run at 100%.

3. Lower duct losses

At 60% airflow:
  Velocity drops to 60%
  Friction drops to 36% (velocity²)
  Duct leakage drops (lower pressure)

The humidity difference

Single stage:
  Runs 8 min → coil gets cold and wet
  Off 15 min → moisture re-evaporates from coil
  Net dehumidification: limited
  SHR: 0.85 (15% of cooling = moisture removal)
  Room humidity: 60-65%

Inverter:
  Runs continuously → coil stays cold and wet
  Moisture condenses and drains steadily
  SHR at low speed: 0.70 (30% = moisture removal)
  Room humidity: 48-52%

In humid climates, this is the difference between comfortable and clammy — without a separate dehumidifier.

Sizing difference

Single stage:
  Must be 100-110% of design load
  Larger → short cycles → humidity problems
  Smaller → can't maintain setpoint on design days

Inverter:
  Can be 115-125% of design load
  Modulates down to 25-40% minimum
  Never short-cycles
  Handles mild spring AND extreme summer

Annual energy comparison

Same house, same 3-ton sizing:

Single stage (14 SEER):  3,200 kWh/year
Inverter (20 SEER):      2,100 kWh/year

Savings: 1,100 kWh × $0.13 = $143/year
Equipment premium: $2,400
Energy-only payback: 16.8 years

With avoided dehumidifier ($1,500 + $200/yr):
  Effective payback: 5-7 years

The rule

The tonnage number tells you the maximum capacity. It does not tell you how the system delivers that capacity. A 3-ton single stage and a 3-ton inverter are different machines with different energy profiles, humidity performance, and comfort characteristics.

Calculate the right tonnage for your space with the AC tonnage calculator on CalcEngineer — then decide how that tonnage should be delivered.

Transformer Derating: Why a 500 kVA Nameplate Delivers Only 445 kVA in a Hot Room

Evgenii Konkin — Fri, 30 Oct 2026 03:53:13 +0000

Every transformer nameplate assumes 40C (104F) ambient air. If the room is hotter, the transformer cannot deliver its rated kVA — and the kW you can actually pull drops with it.

Why temperature matters

A transformer generates heat from copper losses (windings) and core losses (iron). This heat must dissipate into the surrounding air. If the ambient air is hotter, the temperature difference between the winding and the air decreases — less heat dissipates — the winding reaches its thermal limit at a lower load.

The derating table

Ambient Temp    % of Rated kVA    500 kVA delivers
─────────────────────────────────────────────────────
30C (86F)       105%              525 kVA
35C (95F)       103%              515 kVA
40C (104F)      100%              500 kVA  ← nameplate
45C (113F)       95%              475 kVA
48C (118F)       89%              445 kVA
50C (122F)       87%              435 kVA
55C (131F)       80%              400 kVA
60C (140F)       72%              360 kVA

At 60C ambient, the transformer delivers only 72% of nameplate — a 140 kVA loss on a 500 kVA unit.

Real case: data center

Situation:
  500 kVA dry-type transformer (Class F, 155C)
  Server racks added next to electrical room
  Room temperature rose from 25C to 48C
  Nobody recalculated capacity

Load:
  410 kW at PF 0.92 = 446 kVA

Nameplate capacity: 500 kVA ← looks fine
Derated capacity at 48C: 445 kVA ← NOT fine

446 kVA > 445 kVA → thermal overload trip
at 2 PM on the hottest day of the year

The kW impact

The kVA derating directly reduces available kW:

kW available = Derated kVA × Power Factor

At 40C (full rating):
  500 kVA × 0.92 PF = 460 kW available

At 48C:
  445 kVA × 0.92 PF = 409 kW available

At 55C:
  400 kVA × 0.92 PF = 368 kW available

Lost capacity: 92 kW from a 15C temperature increase

Where hot rooms happen

- Electrical rooms next to server rooms
- Electrical rooms next to boiler rooms or kitchens
- Rooftop pad-mounted transformers in desert climates
- Underground vaults with poor ventilation
- Rooms where original ventilation was blocked or removed
- Any room where heat-producing equipment was added later

The fix options

1. Cool the room
   Add exhaust fan or supply air
   Target: below 40C during peak load
   Cost: $1,500-3,000
   Payback: immediate (prevents trips)

2. Reduce the load
   Move circuits to another transformer
   Shed non-critical loads during peak
   Cost: $0-5,000 depending on rewiring

3. Upgrade the transformer
   Replace with higher kVA rating
   Or replace Class F with Class H (180C limit)
   Cost: $15,000-40,000

4. Add forced cooling
   Fans on the transformer itself
   Some dry-type models allow fan-cooled rating (FA)
   Typical boost: 33% above self-cooled rating
   Cost: $2,000-5,000

The lesson

Data center case:
  Exhaust fan cost:     $1,800
  Downtime cost:       $23,000

  Room dropped from 48C to 35C
  Full 500 kVA restored (actually 515 kVA at 35C)
  Fan paid for itself before installation finished

Before loading a transformer, check the room temperature. The nameplate kVA is only valid at 40C. Above that, derate — or cool the room.

Quick Tip: Three-Phase Power Measurement — Why V I 3 Gives You 134% Too Much

Evgenii Konkin — Thu, 29 Oct 2026 03:50:47 +0000

A technician clamps one phase of a 480V three-phase compressor. Reads 62 amps. Calculates: 480 × 62 × 3 = 89.4 kW. Actual power: 38.2 kW. Where did 51 kW come from?

The two mistakes

Mistake 1: ×3 instead of ×√3
  Three-phase power uses √3 (1.732), not 3.
  The phases are 120° apart — they don't simply add.

Mistake 2: No power factor
  V × I = apparent power (kVA), not real power (kW).
  A motor at 74% PF draws 26% more current than its kW suggests.

The correct formula

Three-phase real power:
  P (kW) = √3 × V_line × I_line × PF / 1,000
         = 1.732 × 480 × 62 × 0.74 / 1,000
         = 38.2 kW

Compare to the wrong method:
  P = V × I × 3 / 1,000
  = 480 × 62 × 3 / 1,000
  = 89.4 kW  ← 134% too high

Three common errors ranked

Method                  Result    Error
──────────────────────────────────────────────
V × I × 3              89.4 kW   +134% ← worst
V × I × √3 (no PF)     51.6 kVA   +35% (kVA, not kW)
√3 × V × I × PF        38.2 kW     0% ← correct

Three correct measurement methods

1. True 3-phase power meter
   Clamp all three phases simultaneously.
   Meter calculates kW directly.
   Best accuracy. No PF needed.

2. Single clamp + PF
   P = √3 × V × I × PF
   Works if: phases balanced (±5%), PF known.
   Use clamp meter with PF function.

3. kWh meter reading
   Read utility meter. Wait 15-60 min. Read again.
   kW = ΔkWh / hours
   Most accurate field method — meter is calibrated.

Why it matters

Energy audit:
  Baseline at 89 kW → "VFD saves 30 kW" → approved
  Actual baseline 38 kW → VFD saves 11 kW → never pays back

Generator sizing:
  89 kW measurement → buy 100 kW generator
  Actual load 38 kW → 50 kW generator is plenty
  Wasted: $15,000+ on oversized equipment

Billing check:
  Your measurement: 89 kW
  Utility bill: 38 kW
  "The meter must be wrong" ← No, your formula is wrong

The rule

For three-phase equipment, always:

P = √3 × V × I × PF

Never V × I × 3. Never V × I × √3 without PF. The √3 accounts for the phase angle. The PF converts apparent power to real power. Skip either one and the number is fiction.

Fan Curves Lie When Air Density Changes. Here Is How to Read Them Correctly.

Evgenii Konkin — Wed, 28 Oct 2026 04:05:28 +0000

Every fan catalog assumes standard air: 70F, sea level, 0.075 lb/ft³. If your fan operates at a different temperature or altitude, the catalog numbers are wrong — and you need to know which ones change and which ones do not.

What changes and what stays the same

When air density decreases (hotter or higher altitude):

  CFM:            STAYS THE SAME
  Static Pressure: DECREASES proportionally
  Brake Horsepower: DECREASES proportionally

  SP_actual = SP_catalog × (ρ_actual / ρ_standard)
  BHP_actual = BHP_catalog × (ρ_actual / ρ_standard)

The fan moves the same volume (CFM) regardless of density. But lighter air produces less pressure and requires less power.

Real example: kitchen exhaust in El Paso

Location: El Paso, TX (3,740 ft elevation)
Air temperature: 130F (kitchen hood exhaust)
Actual density: 0.057 lb/ft³

Standard density: 0.075 lb/ft³
Density ratio: 0.057 / 0.075 = 0.76

Fan catalog:
  CFM: 8,000
  SP: 2.5 in. w.c.
  BHP: 5.2 HP

Actual performance:
  CFM: 8,000 (unchanged)
  SP: 2.5 × 0.76 = 1.9 in. w.c.
  BHP: 5.2 × 0.76 = 3.95 HP

The gauge reads 1.9 instead of 2.5. Nothing is broken — the air is 24% lighter.

Why the system still works

Key insight: duct friction ALSO scales with density.

Duct system designed for 2.5 in. w.c. at standard air:
  Actual friction at 0.76 density = 2.5 × 0.76 = 1.9 in. w.c.

Fan delivers 1.9 in. w.c.
Duct requires 1.9 in. w.c.

Operating point: same CFM, lower SP → system balances itself.

Both fan output and duct resistance drop by the same factor. The fan curve shifts down, the system curve shifts down, and they intersect at the same CFM.

The real danger: cold startup

Cold morning air is DENSER than standard:

  0F outdoor:  ρ = 0.086 lb/ft³
  -20F outdoor: ρ = 0.094 lb/ft³

  Density ratio at 0F: 0.086 / 0.075 = 1.15
  BHP at 0F = 5.2 × 1.15 = 5.98 HP

  If motor is 5 HP (sized for catalog BHP) → OVERLOAD

Low density = less power (safe for motor).
High density = more power (motor overload risk).

Density correction quick reference

Condition              Density     Ratio    SP Factor
──────────────────────────────────────────────────────
-20F, sea level        0.094       1.25     125%
0F, sea level          0.086       1.15     115%
70F, sea level         0.075       1.00     100% (standard)
100F, sea level        0.071       0.95      95%
70F, 5,000 ft          0.066       0.88      88%
130F, 3,740 ft         0.057       0.76      76%
200F, sea level        0.060       0.80      80%
200F, 5,000 ft         0.053       0.71      71%

The selection process

1. Calculate actual air density (temperature + altitude + humidity)
2. Calculate density ratio: DR = ρ_actual / 0.075
3. Select fan from catalog at STANDARD conditions
   (catalog CFM and catalog SP — do NOT pre-correct)
4. Verify actual BHP: BHP_catalog × DR
5. Size motor for COLD startup conditions
   (densest air = coldest morning = highest BHP)
6. Add 15-20% motor safety factor above cold-air BHP

The catalog selection is correct for CFM. The density only matters for verifying motor size at extreme conditions.

Get the exact air density for your fan application with the air density calculator on CalcEngineer — enter temperature, altitude, and humidity for the density ratio.

Quick Tip: Duct Velocity Limits — Why a Correctly Sized Duct Can Still Be Too Loud

Evgenii Konkin — Tue, 27 Oct 2026 03:51:16 +0000

Equal friction sizing gives you the right duct diameter for pressure drop. But it says nothing about noise. A duct that is correct for airflow can be far too loud for the space it serves.

The noise source

Air moving through a duct creates turbulence. Faster air = more turbulence = more noise. The relationship is roughly exponential — doubling velocity increases noise by 16-18 dB.

ASHRAE maximum velocities by space type

Space Type                    Max Velocity (FPM)
──────────────────────────────────────────────────
Residences (main trunk)        600
Residences (branches)          400
Conference rooms / libraries   600-800
Hotel rooms / patient rooms    800
Private offices                800-1,000
General offices / retail      1,000-1,200
Restaurants                   1,200-1,500
Industrial / warehouse        1,500-2,500
Mechanical rooms              2,000-3,000

Real example

Conference room duct: 2,800 CFM

Equal friction method (0.08 in. w.c./100 ft):
  Result: 18" diameter
  Velocity: 2,800 / (π × (18/24)² / 4) = 1,600 FPM
  → 2× the conference room limit

Velocity method (800 FPM max):
  Area = 2,800 / 800 = 3.50 ft²
  Diameter = √(4 × 3.50 / π) × 12 = 25"
  → Meets noise requirement

18" was correct for pressure. 25" is correct for noise. The duct needs to be 39% larger.

Quick sizing by velocity

Area (ft²) = CFM / Velocity (FPM)
Diameter (in) = √(4 × Area / π) × 12

Examples at 800 FPM (office/conference):
   500 CFM → 0.625 ft² → 11"
  1,000 CFM → 1.25 ft² → 15"
  2,000 CFM → 2.50 ft² → 21"
  3,000 CFM → 3.75 ft² → 26"
  5,000 CFM → 6.25 ft² → 34"

When space is too tight for the velocity-sized duct

Option 1: Lined duct (internal fiberglass)
  → absorbs 5-10 dB, allows ~20% higher velocity

Option 2: Sound attenuator (silencer)
  → 15-25 dB reduction, adds 3-5 ft length

Option 3: VAV terminal unit
  → noise break between high-velocity trunk and low-velocity branch

Option 4: Route trunk through non-sensitive space
  → long lined branch attenuates before reaching diffuser

The cost lesson

Correct sizing at design:     +$800 in larger sheet metal
Post-construction silencer:   $4,200 installed

Size for pressure AND velocity. The chart gives you airflow. The velocity table tells you if the space can live with the noise.

Roof Color Changes Cooling Load by 7 Tons on a 30,000 sq ft Building. Here Is the Math.

Evgenii Konkin — Mon, 26 Oct 2026 04:17:05 +0000

A dark roof absorbs 94% of solar energy. A white roof absorbs 22%. The difference on a 30,000 sq ft warehouse in Phoenix: 7 tons of cooling load from the roof alone.

Sol-air temperature

The roof does not see outdoor air temperature — it sees sol-air temperature, which includes solar radiation absorbed by the surface:

T_sol-air = T_outdoor + (α × I / h_o) - correction

Where:
  α = solar absorptance (1 - reflectance)
  I = solar irradiance (BTU/hr·ft²)
  h_o = outside surface coefficient (~3.0 BTU/hr·ft²·°F)
  correction ≈ 7°F for horizontal surfaces

Phoenix design day example

Outdoor: 110°F
Solar irradiance: 290 BTU/hr·ft²

Dark roof (reflectance 0.06, absorptance 0.94):
  T_sol-air = 110 + (0.94 × 290 / 3.0) - 7 = 194°F

White roof (reflectance 0.78, absorptance 0.22):
  T_sol-air = 110 + (0.22 × 290 / 3.0) - 7 = 124°F

Difference: 70°F before insulation enters the equation

The dark roof surface reaches 170°F. The white roof stays at 115°F.

Cooling load calculation

Q_roof = Area × U-value × (T_sol-air - T_indoor)

Roof: 30,000 ft², R-25 (U = 0.04), Indoor: 75°F

Dark roof:
  Q = 30,000 × 0.04 × (194 - 75)
    = 30,000 × 0.04 × 119
    = 142,800 BTU/hr = 11.9 tons

White roof:
  Q = 30,000 × 0.04 × (124 - 75)
    = 30,000 × 0.04 × 49
    = 58,800 BTU/hr = 4.9 tons

Difference: 7.0 tons

Roof color vs insulation — which matters more?

Combination              Roof Cooling Load
──────────────────────────────────────────
Dark roof, R-15          19.8 tons
Dark roof, R-25          11.9 tons
Dark roof, R-38           7.8 tons
White roof, R-15          8.2 tons  ← beats dark R-38!
White roof, R-25          4.9 tons
White roof, R-38          3.2 tons

A white roof at R-15 outperforms a dark roof at R-38. Reflectance is cheaper than insulation for reducing roof cooling load.

Cost comparison

White TPO vs dark membrane:
  Premium: +$0.30/ft² × 30,000 = $9,000
  Smaller HVAC: -$3,500 (7 tons × $500/ton)
  Annual energy: -$2,800 (7T × 1,200h × 1.2kW/T × $0.11)
  Net cost: $5,500
  Payback: 2.0 years

Where reflectance matters most

City           Solar (BTU/hr·ft²)   Reflectance Impact
────────────────────────────────────────────────────────
Phoenix           290               Very high
Las Vegas         280               Very high
Miami             260               High (+ humidity)
Dallas            250               High
Atlanta           230               Moderate
Chicago           210               Moderate
Seattle           180               Low

In high-solar climates, roof color is the single most cost-effective cooling load reduction.

The rule

When calculating cooling load, never use outdoor air temperature for the roof. Use sol-air temperature — it includes the solar energy the roof absorbs. And the easiest way to reduce sol-air temperature is to change the roof color.

Chilled Water Delta-T: How a 5-Degree Drop Cut Chiller Capacity in Half

Evgenii Konkin — Fri, 23 Oct 2026 03:54:40 +0000

In a chilled water system, delta-T is the temperature difference between return water (warm) and supply water (cold). When delta-T drops, the chiller delivers less capacity — even at full compressor speed.

The capacity formula

Q = 500 × GPM × Delta-T

Where:
  Q = capacity (BTU/hr)
  GPM = water flow rate (gallons per minute)
  500 = constant (8.33 lb/gal × 60 min × 1.0 BTU/lb·°F)
  Delta-T = return temp - supply temp (°F)

Design vs actual

300-ton chiller, 600 GPM:

Design:
  Supply: 44°F, Return: 56°F
  ΔT = 12°F
  Q = 500 × 600 × 12 = 3,600,000 BTU/hr = 300 tons ✓

Actual (low delta-T):
  Supply: 44°F, Return: 49°F
  ΔT = 5°F
  Q = 500 × 600 × 5 = 1,500,000 BTU/hr = 125 tons ✗

Same chiller. Same flow. 58% capacity loss.

Why delta-T drops

Three-way valve bypass

Coil control valve bypasses water around the coil.
Bypass water returns at supply temperature (44°F).
Mixes with coil leaving water (56°F).
Mixed return: lower than 56°F → delta-T drops.

Fix: convert to two-way valves (no bypass path)

Fouled coils

Dirt, scale, or biofilm on waterside of AHU coils.
Reduces heat transfer coefficient.
Less heat moves from air to water per pass.
Water returns cooler than design.

Fix: chemical cleaning or mechanical brushing

Excessive pump speed

Variable-speed pump set to low DP setpoint.
More GPM than coils need at partial load.
Water moves through coils too fast.
Not enough contact time for full heat absorption.

Fix: reset pump DP setpoint based on valve positions

The pump energy penalty

When delta-T drops, you need more GPM for the same cooling:

Design:  600 GPM at 12°F ΔT = 300 tons
Low ΔT: 1,440 GPM at 5°F ΔT = 300 tons

Flow increase: 2.4×
Pump power increase: 2.4³ = 13.8×

Pump power follows the cube law. Small delta-T reductions cause massive pump energy increases.

In practice, the pump maxes out before reaching 1,440 GPM. The system starts a second chiller instead — doubling chiller, condenser pump, and cooling tower energy.

Real case: hospital

Problem: two 300-ton chillers both at 90% load
         (three years ago, one chiller handled it)

Investigation:
  ΔT = 5°F (design = 12°F)
  14 of 22 AHUs had three-way bypass valves
  Several coils not cleaned in 4 years

Fix:
  Converted to two-way valves
  Chemically cleaned coils
  ΔT recovered: 5°F → 10°F

Result:
  One chiller handles the load
  Second chiller = peak days only
  Annual savings: $47,000

The check

1. Measure supply water temp (at chiller leaving)
2. Measure return water temp (at chiller entering)
3. Subtract: ΔT = return - supply
4. Compare to design ΔT (typically 10-14°F)

ΔT within 2°F of design → system OK
ΔT 50%+ below design → investigate valves, coils, pump speed

Two thermometers on the chilled water pipes. The number between them tells you whether your chiller plant is working — or wasting energy moving water that is not absorbing heat.

Calculate your system's delta-T and capacity with the delta T calculator on CalcEngineer.

Quick Tip: Why Your Economizer Misses Free Cooling Hours — Dry-Bulb vs Enthalpy Control

Evgenii Konkin — Thu, 22 Oct 2026 03:55:44 +0000

An economizer uses outdoor air instead of the compressor when conditions are favorable. But the controller decides when "favorable" means — and the wrong control strategy leaves free cooling on the table.

Two control strategies

Dry-bulb control:
  IF outdoor temp < setpoint (55-65F) → open dampers
  Simple, cheap, works in dry climates
  FAILS in humid climates (ignores moisture)

Enthalpy control:
  IF outdoor enthalpy < return air enthalpy → open dampers
  Accurate, captures both temp + moisture
  Requires humidity sensor ($300 per location)

The problem with dry-bulb

Return air: 75F, 45% RH → enthalpy = 27.5 BTU/lb

Outdoor conditions:
                              Enthalpy   Dry-Bulb    Free
  Condition                   (BTU/lb)   Says...     Cooling?
  ──────────────────────────────────────────────────────────
  45F, 50% RH                  13.2      YES (< 55)   YES ✓
  55F, 60% RH                  19.8      NO (= 55)    YES ✗ missed!
  58F, 70% RH                  23.1      NO (> 55)    YES ✗ missed!
  62F, 80% RH                  27.0      NO (> 55)    YES ✗ missed!
  65F, 85% RH                  28.7      NO (> 55)    NO ✓
  70F, 70% RH                  29.4      NO (> 55)    NO ✓

Three conditions with free cooling available — dry-bulb controller misses all three.

The dangerous case

60F, 95% RH → enthalpy = 26.1 BTU/lb

Dry-bulb controller: "60F > 55F → no economizer" ← correct!
But barely — enthalpy is close to return air.

Now: 60F, 40% RH → enthalpy = 19.5 BTU/lb
Dry-bulb controller: "60F > 55F → no economizer" ← WRONG!
8 BTU/lb of free cooling wasted.

Same temperature, different humidity, opposite decisions needed. Dry-bulb cannot distinguish them.

Hours captured per year

Climate          Dry-Bulb Hours   Enthalpy Hours   Missed
──────────────────────────────────────────────────────────
Phoenix (dry)         2,800           2,900          100
Charlotte (mixed)     1,400           2,200          800
Atlanta (humid)       1,100           1,900          800
Houston (very humid)    600           1,200          600

The math

800 missed hours × 15 tons avg × 1.2 kW/ton × $0.12/kWh
= $1,728/year in compressor energy that was not needed

Enthalpy sensor cost: $300 × 2 = $600
Payback: 4 months

The rule

In dry climates (Phoenix, Denver): dry-bulb control is fine — humidity rarely matters.

In humid or mixed climates (anywhere east of the Rockies): enthalpy control captures 600-800 more free cooling hours per year. The sensor pays for itself in one season.

The decision is: is outdoor air enthalpy below return air enthalpy? Temperature alone cannot answer that question.