Your old furnace or boiler is gasping its last breath and it’s time to pull the trigger on something newer, more reliable, and more efficient. How do you quickly size the new equipment?
If you leave the sizing calculations to HVAC contractors, most would replace the old furnace with equipment that has a comparable output rating. That would guarantee that you wouldn’t get cold, but at least 19 times out of 20, that would be a mistake.
Furnaces are routinely oversized
Most of the installed heating equipment in the U.S. is oversized. In fact, most equipment has a heat output that is between 2 and 4 times the heating load.
There are valid reasons to oversize the equipment a bit, but not by 2 to 4 times the load. Oversizing on that scale results in lower comfort and lower efficiency. Most people want enough capacity to cover somewhat colder temperatures “just in case” there is a cold snap or record low temperatures, or if they need to keep the house at 78°F for a frail elderly parent.
With the average installed equipment being 3 times oversized, that means you’re covered. But what are the consequences of this type of equipment oversizing?
What’s the 99% outdoor design temperature?
For a particular location, the “99% outdoor design temperature” is the temperature which is exceeded for 99% of the hours in an average year. In other words, only 1% of the hours in a year are colder than the 99% outdoor design temperature.
Heating appliances can be sized to meet a building’s heat load at the 99% outdoor design temperature or at the 99.6% outdoor design temperature. Building codes in the U.S. stipulate that every room be capable of being automatically heated to a minimum of 68°F at the 99% temperature bin for that location, making the 99% approach more relevant than the 99.6% approach. So it’s useful to know the “99% outdoor design temperature” for your location.
Three times larger than necessary? What does that mean?
A building’s heat load grows (approximately) linearly with the difference between indoor and outdoor temperatures (otherwise known as the delta-T). According to building code, a heating appliance is correctly sized when it is sufficient to cover the difference between 68°F and the outdoor design temperature.
For example, consider a house in Washington, D.C. (Climate Zone 4), where the outdoor design temperature is 20°F. If your furnace is oversized by a factor of 3, you could heat the house in a location with a delta-T that was three times larger than your actual delta-T of 48 F° — in other words, in a location with a delta-T of 144 F°. With that much capacity, the heating system won’t lose ground until the outside temperature drops below -76°F — an outdoor temperature not seen in Washington, D.C. since the last ice age!
That’s ridiculous, of course — but oversizing a furnace by a factor of 3 is the norm rather than the exception.
Oversizing a little is OK
The AFUE testing protocol (used to determine furnace efficiency) presumes an oversizing factor of 1.7 times, which still gives a large margin for colder weather — more than covering the absolute record low temperatures in most locations. When a furnace is oversized by a factor of 1.7, it isn’t so oversized that it impacts efficiency, but that much oversizing is really too much for multi-stage furnaces. A typical two-stage condensing gas furnace has a turn-down ratio of less than 2:1. With most of these furnaces, the low-fire output is still 60% or more of the high-fire output.
When a furnace with a low-fire rating of 60% is oversized by a factor of 1.7, you could cover 99% of the building’s heating needs at low fire. You might as well hard-wire the furnace so that it never steps up to high-fire mode.
For comfort and efficiency, ASHRAE recommends that heating equipment be sized at 1.4 times the design heat load.
At 1.4 times oversizing, the house in the example above would have its heating load fully covered at a temperature difference of (1.4 x 48 F°) = 67 F°. With a delta-T of 67 F°, the heating system is adequate when the outdoor temperature drops to (68°F – 67 F°) = 1°F, which is 19 F° colder than the 99% outside design temperature. But that’s an outdoor temperature that may actually occur a few times over the 15-to-25 year lifecycle of the heating equipment (but not every year). When that happens it’s only for brief periods of time — short enough that the thermal mass of the house keeps it from losing much ground. So there is usually no comfort problem.
Methods for calculating a building’s heat load
That’s the target for sizing equipment. But to get to this target, it’s important to come up with a reasonably accurate design heat load number.
You could measure up the windows and walls, estimate the U-factors of different building assemblies, and run an I=B=R load calculation on the whole house, or you could even run a Manual J calculation. But those methods take time, and it’s easy to make errors when estimating the U-factors of components of an older house. It’s human nature to err to the high side when in doubt, which is also a mistake.
Fortunately, if you have access to historical fuel purchase history, you don’t have to guess.
You can calculate a building’s heat load in 15 minutes
You have instrumentation already in the house that is measuring the heat load: namely, the existing heating equipment. The way to use it for measurement purposes is:
- Take a mid- to late-winter fuel bill, and note the exact dates covered by the bill — the fill-up dates or the meter-reading dates.
- Look for a specification label on your furnace or boiler that includes the input BTU/h rating and the output BTU/h rating for your equipment.
- Download base 65°F or base 60°F heating degree-day spreadsheets covering those dates for a nearby weather station from a website called .
- Look up the 99% outside design temperature (sometimes called the “heating 99% dry bulb temperature”) for your location from a website — for example, from an online document called .
Now you have enough information to estimate your building’s heat load with reasonable accuracy, independent of the house construction details.
For example, assume that a house in Washington, D.C. (where the outdoor design temperature is 20°F) used 182 therms of natural gas between January 6 and February 8.
If the gas furnace nameplate shows an input rating of 110,000 BTU/h and an output rating of 88,000 BTU/h, you can use those numbers to determine the furnace’s thermal efficiency — in this case, 80%.
Multiply the input fuel amount by the efficiency of the equipment to determine how much heat was delivered to the building.
ENERGY CONTENT OF FUELS
Natural gas: 1,000 BTU/cu. ft.Propane: 91,333 to 93,000 BTU/gallonFuel oil: 138,700 to 140,000 BTU/gallonKerosene: 120,000 to 135,000 BTU/gallon
To calculate the net amount of heat that was delivered into the ducts (or into the heating pipes if we are talking about a boiler), take the number of therms indicated on your fuel bill and multiply it by the equipment efficiency:
182 therms x (88,000/111,000) = 145.6 therms
Then multiply therms by 100,000 (the number of BTU per therm) to convert therms to BTU:
145.6 therms (x 100,000 BTU/therm) = 14.56 million BTU (MMBTU).
Next, download and sum up the daily base 65°F heating degree days (HDD) for the nearest weather station — in this case, from station KDCA: Washington National Airport, Virginia — from the DegreeDays.net web site for the period of January 6 through February 7. (Include only one of the meter-reading dates, not both.) In this example, the sum comes to 937.7 HDD-65°F. (See Image #2, below.)
Next, download and sum the date for base 60°F. The result is 772.9 HDD-60°F. (See Image #3, below.)
14.56 MMBTU / 937.7 HDD is 15,527 BTU per degree-day. With 24 hours in a day, that’s an average of 647 BTU per degree-hour at a balance point of 65°F.
14.56 MMBTU / 772.9 HDD is 18,838 BTU per degree-day, and with 24 hours in a day that’s an average of 785 BTU per degree-hour at a balance point of 60°F.
A balance point of 65°F with design temp of 20°F is a difference of 45 F° degrees, and the implied heat load is then 45 F° x 647 BTU/F-hr = ~29,115 BTU/hr.
At a balance point of 60°F there are only 40 F° heating degrees, and the implied load is 45 F° x 785 BTU/F-hr = ~31,400 BTU/hr.
That’s a range of about 8% between the calculation based on 65°F heating degree days and the calculation based on 60°F heating degree days. Which is closest to reality?
It depends. Most 2×4 framed houses will have a balance point close to 65°F, most 2×6 framed houses will balance closer to 60°F. But unless it’s a superinsulated house, it’s likely balance point is somewhere in that range.
Comparing 65°F HDD calculations with 60°F HDD calculations
At this point, you may be thinking, “Why would the calculated heating load for a house with 2×4 walls (29,155 BTU/h) be lower than the calculated heating load for a house with 2×6 walls (31,400 BTU/h)?”
The short answer is, “Both calculations assume that you’ve used the same amount of fuel over the average outdoor temperatures during the period in question, which yields a higher BTU per degree-hour constant for the house with 2×6 walls.”
Put another way, if the better-insulated house used the same amount of fuel during the same weather conditions, its load is going to be higher when it’s really cold out. If two identical houses were built, one with 2×4 walls and the other with 2×6 walls, the 2×6 house should have used less fuel at the average outdoor temperature over the period, not the same amount of fuel. But if different 2×4 and 2×6 houses use the same amount of fuel, the incremental heat requirement of the 2×6 house per degree will be bigger. When you then use that bigger load per-degree constant to predict the load at the outside design temperature, the calculation results in a bigger number.
What about thermostat settings?
If the average indoor temperature was kept substantially below 68°F, you can account for that fact by dropping the degree-day base.
For example, if you normally keep the thermostat at 62°F rather than 68°F, subtract 6 F° from the temperature bases to get the BTU/degree-hour constant, but add 6 F° to the total heating degrees when you run the final number to be sure it meets code when sizing the equipment.
The heat load calculated from the difference in temperature from the balance point isn’t a perfectly linear BTU/degree-hour constant as implied by this calculation method. There is an offset related to the internal heat sources like electrical plug loads and warm bodies. But the error from the difference in slope between the linear approximation from a presumptive balance point method shown here and other methods — for example, an I=B=R linear model (based on the indoor temperature) or a more nuanced Manual-J calculation — doesn’t induce a large error when wintertime data are used.
If the same heating fuel is also used for domestic hot water, this calculation method exaggerates the implied load numbers, since some of that fuel was used by the water heater and sent down the drain. But some of the space heating came via solar gains that would reduce the implied load numbers. These errors tend to balance each other out to a greater or lesser degree.
If you spent 10 days on the beach in Belize during that period, with your home thermostat set to 50°F, use a different billing period.
If an auxiliary heating appliance was being used on a regular basis (say, a wood stove or a ductless minisplit), this calculation method will be too far from reality to be useful. If that’s the case, go back to I=B=R or Manual-J.
In some cases, your heating equipment may be old, decrepit, and not performing very near its original name plate efficiency. That would skew the calculated number to something higher than reality, but it would have to be pretty far off to make a meaningful difference. If that’s the case, calculate it using some lower efficiency. Even a 100-year-old steam boiler is usually still delivering at least 55% efficiency, and often 65-70%.
Unless there is an obvious large error factor that skews the result badly, move on:
For sizing the equipment, use the ASHRAE 1.4x sizing factor:
1.4 x 29,115 BTU/hr = 40,761 BTU/hr (with a 65°F balance point assumption)
1.4 x 31,400 BTU/hr = 43,960 BTU/hr (with a 60°F balance point assumption)
If reality happens to be the 60°F balance point — the 31,400 BTU/h implied load number — then using the 1.4x multiplier on the lower 65°F implied load of 29,115 BTU/h yields about 40,761 BTU/h, in which case you’re even covered for the higher implied load with ample margin. Since older equipment probably isn’t fully as efficient as it was when it was new, equipment rated at 40,000 BTU/h should be good enough.
But if you got nervous and sized it at 50,000 BTU/h of output, it would still be only ~1.7x oversized for the lower 29,115 BTU/h estimate, which means it would hit its AFUE efficiency number (even though it would be bigger than ideal). From a practical point of view, any heating appliance with an output between 40,000-50,000 BTU/h will be fine.
The highest comfort occurs when it’s cold out, when the equipment is actually running and delivering steady heat — rather than running for a while and overshooting the thermostat, with a long cooling off period between cycles. If the new equipment is multi-stage or modulating, it’s best if the lowest stage output is well under the 29,000 BTU/h load, so that the firing range is meaningful.
With boilers, use only the DOE output rating for the equipment; ignore the net I=B=R numbers. The fuel use calculation has the distribution and idling losses included — they can’t be separated out. (There are other factors that come into play when dealing with modulating condensing boilers, but that’s a topic for another day.) If the replacement equipment will be a heat pump, consult the extended temperature range tables for its output at the 99% design temperature.
Whatever the equipment type, have the load numbers and minimum / maximum output numbers in hand before talking to an HVAC contractor.
Expect pushback from contractors
HVAC contractors have become accustomed to installing oversized equipment, and may even think that equipment really needs to be that big. But you don’t have to follow them down the rabbit hole.
Have confidence in your fuel use numbers. This calculation method is better than an estimate; it’s a measurement.
If you push back, some contractors will balk or refuse to bid equipment that small. (Good riddance!) Others will want you to sign a waiver. (OK — but really?!)
Still others will understand that a lipstick-on-mirror fuel-use calculation is sufficiently close to reality that they’ll just go with it if you direct them to.
Typical arguments heard from contractors are rules of thumb such as: “It needs to be at least 25 BTU/h per square foot of living space. Your house is 2,400 square feet, so that’s 60,000. Let’s bump it to 75,000 just in case it gets cold out.”
Which reliably oversized most houses by at least 2x. Or: “It needs to be at least 90,000 BTU/h or it’ll take forever to return from overnight setback.”
Which is almost never true.
Recent feedback from a contractor insisting on a 100,000 BTU/h condensing boiler for a house with a design heat load under 30,000 BTU/h (based on fuel use calculations and later verified by Manual-J) went, “It needs to be at least 100,000 BTU/h or it’ll take forever to bring the house up to temperature after you’ve been out of power for a few days.”
Out of power for days? How often does that happen each winter (or decade)?
Every day, contractors come up with new creative reasons for oversizing. But with the load calculation in your back pocket, you don’t have to accept these arguments.
Dana Dorsett has lifelong interests in energy policy, building science, and home efficiency. He is currently an electrical engineer in Massachusetts.