Size it right before you buy it.

Roughly 20–30 BTU per sq ft for cooling in most of the US. Use the low end for shaded, well-insulated rooms in mild climates and the high end for sunny rooms, high ceilings, or hot climates like the Southwest and Gulf Coast.

One ton equals 12,000 BTU, so 9,000 BTU is ¾ ton, 12,000 is 1 ton, and 18,000 is 1½ tons. Window units and mini-splits are sold in BTU; central systems are sold in tons.

You’re cooling volume, not floor area. A 12 ft ceiling holds 50% more air than an 8 ft ceiling, and hot air stratifies up high while the thermostat sits at 5 ft — the system works longer to satisfy it.

Typically 3 to 4 tons, depending on climate. About 3 tons in northern zones with good insulation, up to 4 tons in the Deep South or in a drafty older home. This calculator’s range is the honest answer — anyone giving one number without asking about your climate and insulation is guessing.

No. Right-sized systems run longer, steadier cycles, which removes more humidity, evens out room temperatures, and usually lasts longer. Bigger only helps if the house was genuinely undersized — chronic inability to hold setpoint on design-temperature days, with verified airflow and charge.

Manual J is the ACCA-standard load calculation that accounts for your actual windows, walls, orientation, insulation, infiltration, ducts, and local weather design data. It’s the calculation a contractor should run before final sizing — these web calculators are for planning and quote-checking.

For most homes, the value sweet spot is the mid-tier (roughly 15.2–17 SEER2 range): noticeably cheaper to run than minimum efficiency without the steep premium of top-tier variable-speed systems. Go higher if you’re in a hot climate, plan to stay in the home 10+ years, or rebates/credits close the price gap.

Not necessarily — SEER is efficiency, not capacity. A 3-ton 20 SEER and a 3-ton 14 SEER both deliver ~36,000 BTU. Higher-end systems are often variable-speed, which does improve comfort and humidity control, but that’s the compressor staging, not the SEER number itself.

Duct leakage, refrigerant charge, airflow, and thermostat habits all move real consumption. A leaky duct system (25–30% losses are common in older homes) can erase a full SEER tier — sealing ducts often pays back faster than upgrading equipment.

About 2,100–3,600 watts depending on efficiency: roughly 3,600 W at 10 SEER, 2,570 W at 14 SEER, and 2,250 W at 16 SEER. Check the nameplate MCA/RLA or measure with a clamp meter for your specific unit — startup surge is much higher but lasts under a second and doesn’t move the bill.

Yes — switch Power entry to "Amps (clamp meter)," clamp a lead with the unit running, and enter amps, voltage, and phase. The watts are computed as V × A × PF (× 1.732 on three-phase) and filled into the watts box so you can sanity-check them. Power factor is an assumption in that math — compressor motors typically run 0.80–0.90, and 0.85 is the default. Don't substitute the nameplate RLA or MCA for a measured amp reading: those are maximums used for wire and breaker sizing, not the draw the unit actually runs at.

Usually runtime. This calculator uses the hours you enter — a system that’s undersized, low on charge, or fighting leaky ducts may run 14–16 hours on a hot day, not 8. Also confirm your true rate: divide the full bill (including delivery and fees) by kWh used, not just the supply rate printed in ads.

For a conventional single-stage system, no — setback saves energy because the system moves the same heat with less total runtime. The exception is extreme heat, where recovery from a big setback during peak afternoon temps can cost as much as maintaining. A modest 3–4°F setback is the practical middle ground.

A 6″ round duct at standard velocity handles roughly 100–110 CFM comfortably. In flex duct, step up to 7″ for the same airflow. A 6″ is the most common bedroom supply size for exactly this reason.

400 CFM per ton is the standard design point. In humid climates techs often drop to 350 CFM/ton for better moisture removal; in dry climates 450 can make sense. Too little airflow across the coil risks icing; too much hurts dehumidification.

Rarely. If the branch duct itself is undersized or crushed, a bigger register grille changes nothing — the bottleneck is upstream. Check for kinked flex, closed dampers, and long runs first; those are the usual culprits.

In a Zone 4 climate with average insulation, the heat load is roughly 68,000 BTU of output — that’s about an 80,000 BTU input furnace at 96% AFUE, or a 90,000 at 80% AFUE. In the South the same house might need only a 60,000; in northern Minnesota, a 100,000.

On gas use alone, a 96% furnace burns about 17% less gas for the same heat. Whether it pays back depends on your winters and gas price — in cold climates with long heating seasons it usually does; in mild climates the 80% (where code still allows it) can be the smarter buy. Condensing furnaces also need a condensate drain and PVC venting, which adds install cost.

They share a blower and duct system, not a size. It’s common for a home to need 3 tons of cooling but 80,000 BTU of heat — the furnace is picked for the heat load and its blower must still move the AC’s 1,200 CFM. Both calculations matter.

TXV systems: subcooling at the manufacturer’s target (often 8–12°F) and superheat around 8–14°F. Fixed-orifice systems: superheat varies with conditions — anywhere from about 5°F on a mild, humid day to 20°F+ on a hot, dry one — which is exactly why the target has to be calculated, not guessed. On a blend with glide, judge superheat against the dew point and subcooling against the bubble point shown in the results; using one saturation value for both can throw your reading off by the full glide — about 10°F on R-407C.

R-22 production and import ended in 2020 — service gas is reclaimed stock only, and R-407C, R-407A, R-422D, and R-438A (MO99) are the common retrofit blends for those systems. R-410A is under an HFC phase-down and new residential equipment is moving to the A2L refrigerants R-32 and R-454B. In refrigeration, high-GWP R-404A and R-507A are being replaced by R-448A and R-449A, and R-134a by R-513A and the HFOs R-1234yf and R-1234ze(E). Retrofits are never drop-ins: oil, seals, capacity, and glide all change — follow the retrofit manufacturer’s procedure.

A2L (R-32, R-454B, R-454A, R-1234yf, R-1234ze(E)) means mildly flammable — equipment listed for it, A2L-rated tools, and leak-check discipline are required. A3 (R-290 propane, R-600a isobutane) means highly flammable: charges are tiny by regulation, systems are sealed and specifically listed, and open-flame or unsealed electrical work around a charged circuit is a hard no. And no refrigerant can be charged by pressure alone: the same suction reading can be correct, undercharged, or flooded back depending on line temperature and load. Pressure only gives you a saturation temperature — superheat and subcooling turn it into a diagnosis.

Use a sling psychrometer or a digital psychrometer probe in the return airstream (or at a return grille, away from supply air). A wet-sock over a thermometer probe with air moving over it gives a field approximation. Take the reading after 15+ minutes of steady runtime.

On hot, humid days the formula can approach zero — it’s floored at 5°F here because charging below that risks liquid floodback to the compressor. If conditions give you a floored target, the load is extreme; verify airflow is at full spec before adding charge.

No — this method is for cooling mode on fixed-orifice systems. Heat-pump heating charge is verified by weight (nameplate charge adjusted for line set, tool 10) and confirmed with manufacturer heating-mode checks, because metering and pressures behave differently in reverse.

Whatever gives the saturation temperature the job calls for — for comfort cooling, about a 40°F evaporator: near 118 psig on R-410A, 69 psig on R-22, 35 psig on R-134a, or 125 psig on R-32, all for the same coil temperature. On the liquid side, condensing runs roughly 15–25°F above outdoor ambient. Pressure without the temperature it corresponds to — and a line temperature to pair with it — isn’t a diagnosis.

Zeotropic blends boil across a range of temperatures instead of at one temperature. Use the dew point when calculating superheat and the bubble point when calculating subcooling — mixing them up introduces an error equal to the glide: about 2.7°F on R-454B, ~6°F on R-448A/R-449A/R-454A, and roughly 10°F on R-407C, which is a big deal when your subcooling target window is ±2°F. Pure refrigerants and azeotropes (R-507A, R-513A) have no meaningful glide, so dew and bubble coincide.

Because a standard PT chart would mislead you. CO₂’s critical point is only 87.8°F — above that there is no condensing temperature at all, and the system runs transcritical at high-side pressures of roughly 1,000–2,000+ psig, far beyond standard manifold gauges and hoses. CO₂ systems are engineered and serviced with dedicated high-pressure equipment and manufacturer tables, so it’s deliberately left out here rather than shown in a format that looks interchangeable with the other 19.

Don’t vent the difference off by feel. Most manufacturers still spec the full nameplate charge down to their minimum line length, then final trim is by subcooling/superheat at startup. If the manual gives a per-foot removal for short runs, follow it — it’s printed for your exact model.

Slightly — it’s driven by liquid density. The rates here are the widely published R-410A family values; R-32 and R-454B tables from their manufacturers are close but not identical. Use the install manual’s table for the refrigerant in the system whenever it’s available.

Most residential equipment allows 50–80 ft total equivalent length with capacity derates and sometimes a larger line size, per the manufacturer’s piping tables. Beyond that you’re into long-line accessories and engineering — the adder alone doesn’t make a 150 ft run acceptable.

Because the coil spends capacity wringing water out of the air instead of dropping its temperature. On a muggy day a healthy system can show 14–16°F of split. That’s why pros confirm charge with superheat/subcooling (tool 7) instead of condemning a system on split alone.

Whatever the rating plate says — commonly a range like 35–65°F. Rise is a safety-adjacent number: too high means low airflow and heat-exchanger stress; too low can mean over-fired airflow or a blower set too high. Measure it and compare to the plate, not to a rule of thumb.

Delta-T says the machine is working; it says nothing about whether the machine is big enough, or whether the ducts deliver the air where it’s needed. Undersized equipment, duct leakage into an attic, and a house gaining heat faster than the system removes it all show a textbook split. That’s a load problem (tools 1–2), not a charge problem.

Hard starting, a humming compressor that won’t kick over, the fan needing a “push start” with a stick, intermittent trips on hot afternoons, and higher-than-nameplate amp draw. A bulged top or leaked oil means it’s already done — no test needed.

Match the µF rating (both sides, on a dual) and meet or exceed the voltage rating — a 440V cap can replace a 370V, never the reverse. Don’t “round” µF up or down beyond tolerance to use what’s on the truck; the motor’s start and run windings are wound for the specified value.

Usually heat or a struggling motor. Check that the motor it serves isn’t over-amping (which cooks caps), the condenser coil is clean so the compartment isn’t an oven, and connections are tight — a loose spade terminal arcs and eats capacitors.

2.31 feet for water (specific gravity 1.00). For other fluids divide by the specific gravity: at SG 1.06 (roughly 50% propylene glycol), 1 psi ≈ 2.18 ft of head.

Work through it in order: air in the loop, a clogged strainer or fouled coil adding or a worn impeller subtracting, a valve not fully open, or the pump running at a lower speed setting than you think. If suction pressure is also low and noisy, suspect cavitation from low fill pressure.

No — that’s the point of subtracting. The static pressure from the fill and expansion tank appears on both gauges and cancels out. Differential is only what the pump itself is adding, which is the number the pump curve describes.

Common hydronic design targets land around 1–4 ft of loss per 100 ft of pipe, with flow velocity kept at or under about 4 ft/s in occupied spaces for noise. If this tool shows you far above that, the fix is a pipe size up, not a pump size up.

Every elbow, tee, and valve resists flow like extra feet of straight pipe. On a typical residential or light-commercial circuit, the fittings total roughly half the straight run’s resistance — the standard field allowance. Fitting-dense circuits (lots of elbows, zone valves, check valves) can run higher; use 75–100% if the circuit is busy.

Yes — glycol mixes are more viscous, especially cold, and friction loss rises (often 10–25% on a 50% mix at low temperatures). This estimator is calibrated for water in copper; treat glycol results as a floor and confirm with the pump and fluid manufacturers’ correction charts.

Either the load genuinely is light (part-load days do this), flow is too high for the load, or a coil control valve is bypassing. Compare flow per ton against the 2.4 GPM/ton @ 10°F check: high flow with low split means the water is moving faster than the coil can cool it.

Yes — the same formula with the same factor gives BTU/hr delivered by a boiler loop or heating coil, using entering minus leaving temperature as a positive number. Only the label changes; the physics is identical.

They’re field-grade approximations for propylene glycol at typical chilled-water temperatures. Exact factors shift with concentration and temperature — for commissioning paperwork, use the glycol manufacturer’s density and specific-heat tables at your actual loop temperature.

Because the refrigerant inside it is boiling below freezing. Suction pressure sets the boiling (saturation) temperature — when saturation falls under ~32°F, moisture freezes on the coil and the line. Three families of causes get you there: the coil is starved (low charge or a restriction — superheat runs high), air isn’t moving heat into the coil (dirty filter, blower, or duct problem — delta-T runs high), or, less often, the coil is flooded by an overfeeding valve and superheat collapses. This tool reads the pattern and ranks which family you’re in; the frost itself only tells you the temperature, not the reason.

Subcooling is the separator, because both faults starve the coil and drive superheat high. With a restriction, refrigerant dams up behind the blockage, so subcooling reads high (often 15°F+) while suction pressure falls. With low charge, there’s no spare liquid to stack, so subcooling reads low (under ~6°F) with both pressures down. Two physical checks back it up: a temperature drop over ~2–3°F across the filter drier, or frost starting right at a component, names a restriction at that spot; a leak detector hit, oil staining at a joint, or a charge that weighs in light names low charge.

No — it narrows one. It pattern-matches the refrigerant circuit and the air side from the readings you give it, but it can’t see a failing contactor, a miscalibrated thermostat, a condensate switch tripping the system, duct leakage into an attic, or an intermittent control fault — which is why “readings look normal” is one of its answers, with pointers to where to look next. It also can’t hear the compressor or smell a burnt winding. Use it to decide what to test first, then confirm with the listed check before replacing anything.

Clamp one compressor lead (or the feeder) with the machine running loaded and enter the amps, the nameplate voltage, and the phase. The tool computes kW = V × A × PF ÷ 1,000 for single-phase, or V × A × PF × 1.732 ÷ 1,000 for three-phase, and shows that kW in the results before dividing by tons. The one judgment call is power factor: PF is an assumption, not a measurement — compressor motors typically run 0.80–0.90 at load, so 0.85 is the default. A true kW meter or drive readout beats the estimate; switch the entry mode to kW when you have one. And note the nameplate RLA and MCA are sizing references (maximums for protection and wiring), not the draw you should expect to measure — a healthy machine often runs well under RLA at part load.

At full load: roughly 0.95–1.35 for air-cooled, 0.55–0.80 for water-cooled screw/scroll, and 0.50–0.65 for centrifugal. Part-load numbers should be better (lower) on a VFD machine and worse on a fixed-speed machine running unloaded — which is why this tool asks for current tons and judges the reading as entered.

When chilled-water ΔT collapses (say 5°F instead of the designed 10–12°F), the plant moves the same tonnage with far more water — usually because coils are fouled, valves leak by, or flow was balanced high "to be safe." Pumps work harder, chillers stage on early, and efficiency craters even though every individual machine looks "fine." The fix is coil/valve/flow correction, not another chiller.

Surge is aerodynamic stall: at low load with high lift (a big gap between condensing and evaporator saturation temps), refrigerant flow through the impeller breaks down and reverses in pulses — rumbling, vibration, and swinging amps. Sustained surge damages bearings and impellers. Raise load, lower lift (colder condenser water, reset chilled-water setpoint up), and check the guide-vane/diffuser operation; don't let the machine sit surging.

A well-tuned conventional (non-condensing) gas or oil boiler typically tests around 78–84% steady-state combustion efficiency; condensing boilers run 90%+ but only when return water is cool enough (below ~130°F for gas) to actually condense. An estimate far below the band for your type usually means excess air, fouled heat-transfer surfaces, or both.

Pressure has to go somewhere. The usual cause is a waterlogged or failed expansion tank — with no air cushion, every heating cycle spikes pressure to the relief setting (typically 30 psi residential). A failed pressure-reducing/fill valve overfilling the system and an internal tankless-coil leak pressurizing the boiler are the other common culprits. A discharging relief valve is a symptom; replacing the valve alone fixes nothing.

A properly tuned burner typically produces well under 100 ppm CO in the flue (air-free readings are the analyzer standard). CO at or above ~100 ppm, rising CO during the run, soot, or a yellow lazy flame all mean combustion is wrong and the boiler should be shut down and serviced by a licensed technician before further operation — CO in the flue can become CO in the house through a blocked vent or cracked exchanger.

Not reliably. The EEV meters flow to hold a superheat or discharge target, so superheat stays near normal across a wide range of charge levels — the system compensates until it can't. The manufacturer's method governs: most specify recovering the charge and weighing it against the nameplate amount adjusted for line length. Pressure readings, where ports exist at all, are usually only valid in a specified test/forced-cooling mode.

Almost always condensate that can't leave: a clogged or poorly sloped drain line, a failed condensate pump, or a coil icing and then melting off the pan's edge. Algae slime in the drain is the classic — clear the line, verify the fall, and confirm with a poured-water test before blaming the unit. A head mounted out of level can spill a full pan even with a clear drain.

Usually not. Most codes are sensor or communication reports: a thermistor out of range, a fan that didn't prove rotation, a head that stopped answering the outdoor board. Look the code up in that brand's table, check the cheap causes first (power quality, loose communication terminals, a genuinely failed sensor reading open/short), and only condemn a board when its inputs test good and its outputs don't happen.

Roughly 90–105% under rating-condition-ish loads is a healthy field result for this estimate. Below ~85% something real is being lost — charge, airflow, or capacity itself. Above ~110%, suspect the inputs first (CFM overestimated or a wet-bulb read too wet/dry) before celebrating; the formula can't create capacity the machine doesn't have.

On a TXV system, on-target subcooling with low capacity points away from charge: check airflow volume and static pressure, coil cleanliness, and duct leakage, in that order. The confirming split is amps: low capacity with low amps suggests the compressor isn't loaded (airflow/heat-transfer problem); low capacity with amps near or above RLA suggests the compressor itself or a severe restriction — that's tool 15's pattern work.

The 4.5 factor (60 min × 0.075 lb/ft³ air density) assumes near-sea-level standard air; at altitude it overstates capacity unless corrected for density. With honest CFM (±5%) and careful wet-bulbs (±0.5°F), expect roughly ±10%. It's a field verdict — "about 95% of rating, healthy" — not a certification measurement.

Because that's what a clamp meter gives you in the field. Clamp the feeder (or add compressor + fan amps), pick the voltage and phase, and the tool converts to kW for the COP math: V × A × PF ÷ 1,000, times 1.732 on three-phase. Power factor is the one assumption in that conversion — compressor motors typically run 0.80–0.90 PF at load, and 0.85 is the default here. If you have a true kW reading from a meter or drive, switch the entry mode to kW and skip the estimate. Nameplate RLA/MCA are sizing references, not expected measured draw — don't enter those as amps.

Yes — it's water vapor. Melting frost on a hot coil in cold air makes a visible cloud that looks alarmingly like smoke, usually with a whoosh and a brief pause in the outdoor fan. It lasts a minute or two and the coil ends up clean and wet. Actual smoke, a burning smell, or steam that doesn't stop is not defrost — shut it down and have it checked.

Normal: occasional (not every few minutes), short (under ~10 minutes), ends with a fully clear coil, and the house holds temperature with maybe a brief aux-heat assist. Fault: ice that never clears or builds into a solid block, defrosts that run long and back-to-back, an outdoor fan that never restarts, or supply air that stays cold long after the cycle ends. Those point to the defrost sensor or board, charge, or the outdoor fan — the patterns this tool ranks.

Often not. Below the balance point (commonly 25–35°F), the heat pump alone can't match the house's heat loss, so aux strips supplement by design — and during defrost they temper the air briefly. Aux heat becomes a problem when it runs at mild outdoor temps, never cycles off, or carries the whole load with the compressor idle: that's a lockout setting, a failed compressor/outdoor unit, low charge shrinking capacity, or a thermostat configured to call aux too eagerly. Compare outdoor temp and ΔT above before condemning anything.