BTU/hour to Donkeypower

The classic rule: 20 BTU/h per square foot. A 300 sq ft bedroom needs about 6,000 BTU/h; a 500 sq ft living room about 10,000 BTU/h. But this varies wildly with sun exposure (+10% for south-facing), ceiling height, insulation quality, number of occupants (+600 BTU per person), and climate zone. A room above a pizza oven in Phoenix needs more than a basement in Seattle. When in doubt, oversize slightly — an undersized unit runs constantly and never reaches setpoint.

Undersizing is obvious — the unit runs constantly and never reaches the thermostat setpoint on hot days. But oversizing is worse in subtle ways. An oversized AC cools the air quickly then shuts off before removing enough humidity, leaving you with a clammy 72°F house. The short cycles also wear the compressor faster (startup is the hardest moment) and waste energy. A 1-ton oversize in a humid climate like Florida can raise indoor humidity from a comfortable 45% to a muggy 60%. Proper Manual J load calculations matter more than most homeowners realize.

Exactly 12,000 BTU/h. One ton of AC is the cooling effect of melting one short ton (2,000 lbs) of ice over 24 hours. The ice absorbs 288,000 BTU of heat as it melts (2,000 lbs × 144 BTU/lb latent heat), divided by 24 hours = 12,000 BTU/h. Residential systems run 1.5–5 tons; commercial buildings 10–500 tons. The "ton" unit persists because HVAC contractors think in tons — "that house needs a 3-ton unit" is faster than "that house needs 10.5 kW of cooling."

Modern units achieve 12–25 BTU/h per watt of electricity (SEER 12–25). A SEER 20 unit removes 20 BTU/h of heat for every watt consumed — effectively a 3:1 heat pump ratio. That 12,000 BTU/h window unit draws 500–1,000 W of electricity depending on efficiency. The best mini-splits achieve SEER 30+, removing 30 BTU/h per watt, making them cheaper to run than resistive electric heaters even in heating mode.

A gas furnace's BTU/h rating is its thermal output after combustion efficiency losses (typically 80–96% of fuel input). A heat pump's BTU/h rating is the heat delivered including energy moved from outside — at COP 3, a heat pump delivering 36,000 BTU/h uses only 12,000 BTU/h worth of electricity. This makes direct BTU/h comparisons misleading: a 60,000 BTU/h furnace and a 60,000 BTU/h heat pump deliver the same heat, but the heat pump uses one-third the energy.

It's both. The value of 250 watts is well-defined and occasionally referenced in engineering education and humorous technical papers. It never achieved official recognition from any standards body, but it has appeared in legitimate engineering textbooks as a pedagogical tool. The unit highlights a genuine gap in the power scale — many household devices and human activities fall in the 50–500 W range where fractional horsepower feels awkward. "Two donkeypower" sounds better than "0.67 horsepower."

A real donkey can sustain about 125–200 watts (0.5–0.8 dp) of useful mechanical work over a full day, and briefly peak at 750–1,500 W (3–6 dp) during a short burst. So the unit slightly overestimates a donkey's sustained output — much like horsepower overestimates a horse. A working donkey in a developing country might turn a water pump or grain mill for 6–8 hours, delivering roughly 0.5–0.7 dp of sustained useful work.

A desktop computer (200–300 W), a bread toaster on low setting (250 W), a box fan on high (200–250 W), a sewing machine motor (250 W), and a human cycling at a moderate sustainable pace (200–250 W). It's a delightfully human-scale unit — roughly the sustained mechanical output of one fit person, or the electrical draw of one modest appliance. Your entire body at rest produces about 0.35 donkeypower of heat.

There's "manpower" (about 75 W sustained, or 0.3 dp), which was used in ancient and medieval engineering for human-powered machines like treadwheels and capstans. "Oxpower" appears in some agricultural texts at roughly 500 W (2 dp). None are standardized. Some engineers have jokingly proposed "hamster power" (~0.5 W, so 0.002 dp) and "ant power" (~10⁻⁵ W). The zoo of animal power units perfectly illustrates why SI standardisation was necessary.

It would actually make more sense for many applications. A blender at "3/4 horsepower" sounds industrial; "2 donkeypower" is more honest and relatable. A hand mixer at "1/8 hp" is awkwardly fractional; "1/3 donkeypower" is cleaner. But the ship has sailed — horsepower (and watts) are too entrenched. Donkeypower's true legacy is as a beloved teaching tool and pub quiz answer. It reminds engineers that units are human inventions, not laws of nature.