Why a 24k BTU Heat Pump Delivered 8k BTUs Last Winter
A heat pump rated for 24k BTUs can drop to 8k BTUs at sub zero temperatures. Cold climate certified units hold their rating, so you size them normally.
A heat pump rated to produce 24k BTUs of heat can deliver 8k BTUs once the temperature drops below zero. That gap is behind most of what went wrong in the northeast this past winter. Temperatures sat sub zero for a long stretch, a lot of heat pumps simply could not keep up the heating demand, and tenants were left with high power bills and cold apartments. The read most people took from it is that the units were not big enough. That is half true, and the half that is missing is the half that costs money.
The Real Numbers
Start with the rating on the box. A unit rated to produce 24k BTUs of heat carries that rating at a specific outdoor temperature, and that temperature is nowhere near zero. When it gets really cold, some heat pumps work worse. The same unit that delivers 24k BTUs in mild conditions drops to 8k BTUs when it is really cold.
That is a third of the capacity the design assumed, and it arrives on the coldest day of the year, which is the day the building has no margin to give.
Inside the apartment it reads two ways at once. The unit runs continuously and still loses ground against the heat loss, so the space never reaches setpoint. Whatever backup exists, usually electric resistance, covers the difference at a much higher operating cost than the heat pump it is filling in for. Cold apartments and high power bills are the same failure viewed from two sides.
Last winter that failure ran long enough to turn a lot of people off to heat pumps entirely. The equipment took the blame. The sizing method that selected the equipment did not.
The Capacity Derating Problem
Here is the mechanism. A heat pump does not create heat, it moves heat, and in heating mode it moves heat out of outdoor air and into the apartment. As outdoor air gets colder there is less heat energy in it to move, and the compressor works across a larger pressure difference to move what is there. Capacity falls as outdoor temperature falls. That relationship is built into the machine.
How far it falls is the part that varies between units. On a conventional heat pump the drop is steep, which is how 24k BTUs becomes 8k BTUs. Frost on the outdoor coil adds to it, because the unit periodically reverses to clear the coil and stops delivering heat to the apartment while it does.
None of this is visible in the nominal capacity. Two units can list the same rated output and behave completely differently at the design temperature, because rated capacity and cold weather capacity are separate pieces of data on the submittal. A load calculation run against the nameplate number is sizing the building for a condition it will not be in on the day the sizing matters.
Some heat pumps are built for the cold. Mitsubishi Hyperheats and Carrier Infinity Series are two of the big name ones. Even when it hits sub zero temperatures they will heat exactly what they were rated for. Rated capacity and delivered capacity stay together through the range the building actually sees, so the number the design used is the number the tenant gets.
ENERGY STAR Cold Climate certification is the shorthand for that behavior. It is a line you can write into the equipment schedule and a rater can verify, rather than a claim somebody has to confirm by digging through performance tables.
Why Oversizing Is the Wrong Fix
Follow the undersizing theory to its conclusion and the answer looks obvious. If the unit loses two thirds of its capacity in the cold, buy one three times larger. On paper the building gets through the design day.
That path is expensive in ways a takeoff does not show. The equipment costs more, and so does everything sized around it. Larger units mean larger circuits and larger distribution, and on an all electric building they put pressure on the service size, which gets priced by somebody else several weeks later.
Then the building runs oversized for the rest of the year. The design day it was sized for is a handful of hours. The rest of the heating season is mild by comparison, and a unit carrying three times the capacity it needs short cycles through all of it. Short cycling costs efficiency, costs comfort, and costs equipment life. The tenant pays that operating penalty every month for capacity that earns its keep on two days each winter.
The derating does not go away either. An oversized conventional unit still loses the same fraction of its capacity at design temperature. You pay for a bigger machine and then pay again to insure against the same behavior.
When Extra Capacity Still Makes Sense
Adding headroom is not automatically wrong. There are conditions where the margin is the right call, and the difference is whether it was specified on purpose or bought as insurance against a spec nobody checked.
When the equipment is not cold climate certified. If the project is committed to a conventional unit for cost or availability reasons, the derating is real and the design has to carry it. Sizing against the capacity at design temperature rather than the nominal rating is not oversizing, it is correct sizing against the right number.
When there is no backup heat. A building with electric resistance backup has a fallback, expensive but functional. A building without one has the heat pump and nothing else, and coming up short produces a cold apartment rather than a high bill.
When the unit serves an unusually exposed space. Top floor corner units with high glazing carry more heat loss than the building average. Sizing every unit to the average and letting the worst ones ride is how a building that models fine produces a run of complaint calls every January.
When the design temperature sits below the published data. Cold climate performance is published down to a stated temperature. If the site design temperature falls below where the manufacturer data ends, the capacity at that condition is an assumption, and assumptions deserve margin.
When the load calculation itself is uncertain. Renovation work on an existing envelope often lacks reliable assembly data. Where the heat loss number carries real uncertainty, capacity margin costs less than being wrong in the expensive direction.
The Decision Framework
The specification that fixes this is short. Write ENERGY STAR Cold Climate certified into the equipment schedule, then rate the unit normally against the calculated load. That is the whole move. You do not need to wildly oversize the unit, and skipping the oversize is what keeps the cost down.
On the submittal, look past the nominal capacity to the capacity at the site design temperature. That is the number the building will live on. If the two are close, the sizing is honest. If the rated output falls off a cliff between them, the design is carrying a gap that the tenant, the backup element, or the utility bill will close.
The rest is building specific in the ways that always matter. Design temperature varies by site, heat loss varies by envelope and orientation, and the worst unit in a building rarely behaves like the average one. What holds across all of them is that equipment should be picked against the condition it has to perform in, not the condition it was tested in. Run that calculation on your building, with your design temperature and your worst-case unit, and the sizing question answers itself. It will save you money and work better.