GBA Logo horizontal Facebook LinkedIn Email Pinterest Twitter Instagram YouTube Icon Navigation Search Icon Main Search Icon Video Play Icon Plus Icon Minus Icon Picture icon Hamburger Icon Close Icon Sorted

Community and Q&A

Signs of a properly sized furnace?

Jeff Watson | Posted in General Questions on

The most common theme I see when people talk about oversized furnaces is reduced runtime & short-cycling.

But what does short-cycling mean quantitatively?

a) 5 mins on, 5 mins off, repeat
b) 5 mins on, 10 mins off, repeat
c) 10 mins on, 5 mins off, repeat

According to the Manual J heat load, my furnace is only 5% oversized. When I set the stat to warm up more than a couple degrees, it does take a while (e.g., more than an hour). I’ve measured almost 100 degrees from my warmest supply register & 68 degrees at that room’s return (rise of 32deg).

However, with the stat at 65degF when it’s 20degF outside, I experience scenario (b) above where it’s only on for 5 minutes and then off for around 10 minutes. This is for maintaining the set-point. With the Manual J indicating a design temp of -4degF for a 70degF set-point, I would’ve thought the furnace would be off for longer or that it’d be on for longer than it is.

Is this short-cycling or some sign that the furnace is still not sized correctly?

GBA Prime

Join the leading community of building science experts

Become a GBA Prime member and get instant access to the latest developments in green building, research, and reports from the field.


  1. GBA Editor
    Martin Holladay | | #1

    There are several ways to determine whether a furnace is oversized.

    One way is to perform a careful heat loss calculation for your home, and to compare your design heat loss with the output capacity of your furnace. If you feel uncertain of your own abilities to perform a heat loss calculation, you can hire an energy consultant to do the work. For more information, see How to Perform a Heat-Loss Calculation.

    Another way to determine whether your furnace is oversized is to observe its performance on the coldest day of the year (or a day when the outdoor temperature is near the design temperature). Under these conditions, the furnace should operate "right out straight" -- for 60 minutes an hour.

    For more information on this topic, see How to Tell If Your Air Conditioner Is Oversized.

  2. John Semmelhack | | #2

    Martin's advice is right on.

    In addition, if you don't want to wait for another -4F day/night, here are some rough calcs for you: ACCA Manual S allows up to 40% oversizing on furnaces. So, in a theoretical situation with your design temperature, you could have a design heat load of 30,000Btu/hr at -4F outside and 70F inside and a 42,000Btu/hr output furnace and just barely meet the "right-sizing" guideline of Manual S. Ignoring solar gains, internal gains and assuming a 100% accurate heat load calc, this 42K output furnace would run for 43 minutes per hour at -4F.

    At the other scenario you mention (65F in, 20F out), your delta-T is now 45 (versus 74 at design temperature), so your heat load (again ignoring internal + solar gains, etc.) at this scenario would be ~18,000Btu/hr and the 42K output furnace would need to run ~26 minutes per hour.

    In your actual situation (65F in and 20F out), 5 mins on and 10 mins off would give you 20 mins of runtime per in reality your furnace is probably slightly more than 40% oversized (not the 5% the load calcs show). It's likely you have a mix of inaccurate load calc, internal gains and solar gains that account for the difference.

  3. Nick T - 6A (MN) | | #3

    Another element to consider in this case (an already installed system) is the controls. The input/output needs at this point won't help you....

    If the thermostat is set to allow for 5 CPH ! (cycles per hour) which is a common default (since many hvac people don't change thermostat settings based on install). Set this for heating and cooling to a very low number to help with cycling (and increased dehumidfiying in summer). Recommend a setting of 1 or 2. Although if temperatures dictate usually it will still cycle a bit more than the setting itself.

    Some thermostats allow changing a deadband (allowable decrease in temperature before furnace turns on). Could be set very aggressive (.25-.1°)causing it to turn on/off to maintain a tight temperature.

    It might be also placed in a location that is getting influenced by furnace airflow/registers. Causing it to turn off thinking it is at setpoint, however shortly after airflow has stopped it believes it is back below setpoint. Could also try to direct more of your heating flow to the exterior of your home, further from the thermostat to reduce the spikes in temperature.

    Also having a lower then design internal temperature (65 vs 72°) could also be reducing run times possibly by reducing your heat loss.

  4. Keith Gustafson | | #4

    What fuel does your furnace use?

    Most furnaces are oversized, not only because of all the biases driving that[HVAC installers rarely get call backs because the house warms up too quickly] but also they do not make furnaces small enough for a modest sized moderately insulated house, unless they are modulating gas models. Oil furnaces are rare below 70k BTU

    If your furnace takes a long time to recover from a setback, I would think you are doing as well as can be expected, and the waste from being oversized is not that big. An older house may have had a furnace 3 or 4 times oversized if it was spec'ed[and then replaced by the same] before storm windows and blown in insulation, and that can be a real pig

  5. Expert Member
    Dana Dorsett | | #5

    Use the steady state efficiency of the furnace or boiler as a measuring instrument, and log fuel use against heating degree days (base 65F). Barring other factors such as auxilliary heating (say from a wood stove), or using gas to shower all 17 of your children daily, or setting the T-stats to 55F and heading to Cancun for a couple of weeks, using a mid-winter fuel use is a reasonably accurate way to estimate the heat load, far better than timing the duty cycling on the equipment for a couple of hours.

    For example, say you have:

    * 92% efficiency condensing gas furnace


    * Between two meter reading dates a weather station near you on logged 887 heating degree days between the EXACT meter reading (not billing dates)


    * You used 176 therms (100,000 BTU per therm) during that period


    * The D.O.E. rated output of the furnace is 112,000 BTU/hr


    * Your 99% outside design temp is about +7F (see: )

    The 65F base temp is the approximate heating/cooling balance point for typical sub-code to code-min (but not super insulated) houses kept at 68-72F indoors. The heat load increases approximately linearly with the difference between the outdoor temp and 65F.

    So with the fuel-use & heating degree numbers, you can reduce that to BTU per degree-day, then per degree-hour:

    ---------176 therms (source fuel content) at 92% efficiency delivers 0.92 x 176= 162 therms of heat into the house.

    ---------Over 887 HDD that's 162/887= 0.18 therms per degree-day

    ---------at 100,000 BTU/therm that's 18,000 BTU per degree-day

    ---------so over a 24 hour day that's 18,000/24= 750 BTU per degree-hour.

    Then solve for the BTU per hour at the 99% outside design condition:

    --------65F - 7F= 58F heating degrees

    --------58F x 750 BTU per degree-hour= 43,500 BTU/hr

    The output of the furnace is 112,000 BTU/hr, the heat load is about 43,500 BTU/hr so the oversizing factor is 112,000/43,500= 2.57x oversized. That's not a super-accurate number, but it's within 10% (5% at worst.)

    Is that too much?

    AFUE testing presumes an oversizing factor of 1.7x, which is enough for reasonable recovery times from overnight setback, and small enough to not short-cycle itself into an early grave or low efficiency. At 1.4x (the ACCA Manual-S limit) you still have plenty of margin even for the coldest night of the past 50 years.

    Boilers take a much bigger efficiency hit from oversizing than hot air furnaces, but both incur equipment longevity issues when short-cycling. At 1.7x both boilers and furnaces hit their AFUE numbers, which is pretty close to their steady-state 100% duty-cycle efficiency. But going up from there boilers start slipping in efficiency and at 3x oversizing it can be a double-digit percentage hit (particularly for cast-iron boilers without heat purge controllers.) From a comfort point of view at 3x oversizing the room temperature swings necessary to keep it from short-cycling.

    How ridiculously oversized is 2.57x oversizing, as in the example above? At 750 BTU/degree hour and 112,000 BTU/hr of furnace output it means you're covered for 112,000/750-= 150 heating degrees. At base 65F that would be an outdoor temp of 65F-150F= -85F (yep, 85 below zero!). That is a temperature not seen since before the last ice age in any location that currently has a 99th percentile bin of +7F.

    What is a short-cycle?

    It depends on the equipment. Any equipment that running burns shorter than 5 minutes and more than 5 burns per hour is going through a heluva lot more flue purge & ignition cycles, but low mass equipment (modulating condensing boilers or hot air furnaces) would still be hitting it's efficiency numbers, cast-iron boilers would be slipping bit. More than 10 burns/hour with burns shorter than 3 minutes is going to be both a longevity and efficiency disaster, even with low thermal mass equipment, since some amount of heat is thrown away with every flue purge. For cast iron boilers you'd ideally want to see burns of at least 10 minutes, and fewer than 3 burns/hour under any load condition.

    Nothing induced short-cycling more than micro-zoning the heating system, since the heat load of any one zone is going to be well under the output of the equipment even if it's oversizing factor is 1.0. With hydronic boilers you also have the issue of the heat emittance capacity of the radiation on each zone, so it's more than just a single short burn when only one call for heat, but often several swings between the high & low limit settings on the boiler before the thermostat is satisfied. There are many ways of dealing with those issues, but be aware that any money you might think you are saving by only heating part of the house can easily be eaten up by short-cylcling wear & tear and efficiency loss, unless design measures have been taken to keep that short-cycling well bounded.

  6. Nate G | | #6

    Higher-level thoughts: Given that a "right-sized" HVAC unit is deliberately and often dramatically oversized 99% of the time, that seems to strongly favor modulating units or deliberately undersized</em

  7. Expert Member
    Dana Dorsett | | #7

    Nathaniel G: A right sized HVAC unit isn't deliberately oversized- it's right-sized. Oversizing by a modest factor gives some flexibility (and is even required by local codes in some places), but that's still oversizing. It's not insane to deliberately oversize by 1.4x, but it IS insane go oversize by 2x or more. It's the "....or more..." that is the root of poor comfort, poor efficiency, and excessive equipment wear. And while 2x+ oversizing is common, it's still fewer than 99% of all installations. (It might be more than 75% though.)

    There are limits to modulation- most modulating condensing boilers have a turn down ratio of only about 4:1. If you oversize one of those by 4x (yes, it happens), you've re-invented the non-modulating boiler. Most multi-stage or modulating furnaces have a 2:1 turn down ratio or less.

    Deliberate undersizing is a violation of IRC code, and can lead to some unpleasant circumstances. eg: The other day at my house the binned hourly temperature average for one 24 hour period was +5F, which happens to be the 99% temperature bin. That means for at least one continuous stretch of 12 hours the actual heat load was higher than the design heat load. The low temp that night was -7F, fully 12F below the 99% temperature bin, with a peak heat load ~20% higher than the design heat load. If the system was exactly right-sized, it would have lost some ground (not much) then slowly recovered. If deliberately undersized it would have lost quite a bit of ground and NEVER recovered.

    A lot of mini-splits have turn-down ratios of only ~3:1, though others do much better. Undersizing a modulating unit guarantees that it's almost always modulating during the average load condition. But mini-splits operate at higher efficiency at part load than when running flat-out- undersizing also guarantees that it spends an inordinate amount of time at highest speed/lowest efficiency. Oversizing a mini-split by 25-50% gives you the extra margin for the exceptional cold snaps while still running in a modulating mode most of the time (even those with a turn-down ratio of only 3:1), and the operating efficiency average is somewhat higher than if it were exactly sized to the 99% outside design condition load. But above that efficiency doesn't go up- it flattens, and eventually falls at extreme oversizing. But even at 2x oversizing comfort goes down, due to the wind-chill& noise of the higher cfm head (even at minimum speed), and on/off cycling leads to much bigger room temp swings than when modulating.

  8. Nate G | | #8

    I was referring to the design temperature. During the vast majority of the time when the actual outside temperature is warmer than the design temperature, the furnace is oversized. Deliberately oversizing it makes it even more oversized during the vast majority of the time it spends producing less heat than it's rated to provide at the design temperature.

    The idea I was toying with was basically to have a unit optimized for the "common case" of operating at higher than the design temperature, with a supplementary backup unit capable of producing enough heat to cover cold snaps that go below the outside temperature. Mini-split and wood stove, for example. Or Mini-split and baseboard resistance heaters. Or mini-split and supplementary window air conditioners, for that matter. The idea is you wouldn't need to use the supplementary unit more than a few times a year, and every time the supplementary unit was off, the primary unit would be closer to its optimal efficiency range and duty cycle.

  9. Jeff Watson | | #9

    Thanks for the answers -

    I guess the cumulative run-time per hour is the answer I was seeking. With a 5min on / 10min off schedule, I guess I assumed the constant on/off inherently implied short-cycling. I figured a properly sized unit would be on for 20 and off for 40 as one cycle as opposed to four 5on/10off cycles.

    Note: my load calculation was performed by an energy consultant, I hope it's accurate since it was far from free.

    John S - pretty clever in your estimate that I'm 40% oversized! Truth is, the basement doesn't have registers but the basement was included in the load calc. I just verified that the basement heat loss accounts for 42% of my total heat load! Pretty close to your 40% oversized estimate. I'm sure my actual load is higher than simply excluding the basement because the rooms don't account for heat loss through the floor that I suffer with an unheated basement.

    This is an older house & I've done many energy-reducing upgrades. And now that I have a supposedly-properly-sized furnace after doing those upgrades, just want to verify if I'm truly cured from the oversizing virus.

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |