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

Furnace Sizing and Duty Cycle

pjpfeiff | Posted in Mechanicals on

I have a gas furnace and an energy monitor logging its electrical usage.  For the sake of argument, say we experienced a night at design temperature.  If the furnace is operating 1/3 of the time in those conditions, does that mean it is 3x “oversized”?  I suppose you would also have to say that conditions led to design air leakage too (if there is such a thing).  Is that likely to change the equation much?

Attached is a screenshot of the longest duty cycle I have found so far.

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.

Replies

  1. Expert Member
    BILL WICHERS | | #1

    It's a bit of a guess, but you can get an estimate with what you're proposing. In theory, a furnace will run continously at the lowest temperature it's designed to be able to handle in any given house. In reality, you never want to push things that far (we can't predict with 100% certainty exactly what a house will need, because there are too many variables that can't be controlled), so you always want to leave a little safety margin.

    If you have, for example, a 50 year record low, and your furnace is only running about 1/3 duty cycle (on for 20 minutes then off for 40), then you can safely say your furnace is oversized a good amount. I'd want to see closer to 80% or so on a record cold day, just in case the next day set an even colder record.

    Bill

  2. paul_wiedefeld | | #2
    1. pjpfeiff | | #3

      Cool, thanks. The first comment on that article basically suggests what I'm asking about, but adds that you can do it for any outdoor temp, you just need to take the the delta-T into account. So really I ought to do the calculation over a range of conditions and see how consistent it is.

      1. bfw577 | | #5

        Before I installed my mini splits I used my forced air oil furnace runtime as a sizing guide. I used the same method by logging the runtime with an electricity monitor. It was really easy to make a chart at each outdoor temp and compare to how long my furnace ran each hour. You then get a pretty accurate rough heat loss. When I sized my mini splits this information was crucial in sizing them.

        I still use my oil furnace during severe cold and we had single digits here last weekend. Below is the electricity graph of my oil furnace cycling over night. My oil furnace is 80k btu/h. You can see it ran around a half hour between cycles. A 1/2 hour runtime would be roughly 40k btu/h delivered. There are duct losses and other losses so you could probably subtract 10-15% from that. It was around 5 degrees that night so my rough heat loss at that temp is around 35k btu/h. At 32 degrees my furnace only sees 1 15 min cycle an hour so at that temp my furnace delivered roughly 20k btu/h for a rough heat loss of 16-18k btu/h.

        As a note my oil furnace is 2x oversized for my house. I heat with 1 12k and 1 18k mini split 98% of the winter. I switch to my oil furnace around 5 degrees which is extemely rare here on the CT shoreline.

  3. Expert Member
    NICK KEENAN | | #4

    Note that for a gas furnace oversizing isn't really an issue. Really what this exercise tells you is what size you need when it's time to replace it.

    1. pjpfeiff | | #10

      Yeah, this is mostly out of curiosity. But also after noticing that my water heater is 30 years old I starting thinking of a HPWH, and then of getting off gas entirely. The furnace is also 30 years old, which is apparently past its life expectancy according to Google, but it seems to be running like a champ.

  4. maine_tyler | | #6

    I notice you have the Lantern up and running.

    Paul, can you clarify what you mean by 'longest duty cycle'?

    I am trying to wrap my head around the frequency of the cycle and what affects that, as I have what I consider to be a fairly short cycle. Does the run time for a single cycle increase but maintain frequency by proportionally decreasing the off time? Or is the frequency stretched a bit?
    I have debated just replacing my old thermostat with one that has some swing as a stop gap until I can do some air sealing and maybe eventually install some heat pumps. My assumption is that with oversized equipment, you would be better off reducing frequency and increasing run time (?).

    I ordered a Lantern PM, but I can see the cycling with a temperature logger as well.

    1. Expert Member
      NICK KEENAN | | #7

      The longer the cycles the greater the temperature fluctuation. This is mostly determined by the thermostat. Some fancy thermostats allow you to set a minimum cycle time, but generally what they key off of is the temperature swing. So if a thermostat has a 1-degree swing and is set for 70F it will turn on when the temperature hits 69.5F and go off at 70.5F. If you set it for a 2-degree swing it would run half as often for twice as long. While longer duty cycles are better there comes a point where the swings are noticeable to occupants.

      Some equipment -- notably heat pumps -- will have a minimum run time and/or a minimum off time. So if the thermostat turns on, it will run for 15 minutes even if the thermostat turns off after five. And once it turns off it won't turn on again for 15 minutes.

      For electric motors starting is the most stressful thing they do. The current at startup is several times the run current. You want to avoid stopping and starting them more than necessary. Generally the bigger the motor the more of a concern this is. So with the big motors in a compressor it's a big concern, less so for the smaller motors in a fan or circulator.

    2. pjpfeiff | | #8

      Yeah, by "duty cycle" I really meant the fraction of time that the furnace is on, which probably is somewhat a misuse if the term. My furnace seems to turn on and off according to temperature like DC described, so the frequency varies.

      I'm friends with the Lantern creator so I got an early delivery with parts he already had :). I have a temp monitor too, so it might be interesting to calculate the on time with both sets of data too see how they compare.

  5. user-723121 | | #9

    If you have metered natural gas it should be easy to check your gas usage compared to heating degree days for the period. From this you can determine the heat loss for your house in Btu's per degree F. This will guide you in system sizing using the design temperature for your area.

    I once tracked furnace run time for my house over an 8 hour period in January. The house was up to temperature when I started and it was a cloudy and still winter day. The temperature remained within a degree or so for the timed period. My furnace is a known efficiency of 95%. From this exercise I found the Btu per degree F heat loss was very close to my manual calculation and RemRate.

  6. pjpfeiff | | #11

    Here is some data for the month of January if anyone interested. I have the furnace electrical usage at 1 Hz sampling. The first plot in the attachment shows a couple of typical power profiles.

    I would be interested if anyone could explain all the features in that profile, but anyway I do know that the first little step up to just under 100 W is when the furnace starts to make some noise but is not actually blowing yet. To determine % on-time (what I was calling "duty-cycle" earlier) I have a simple algorithm. First, define power > 300 to be "on" and everything else to be off. This creates an array where 1 means on and 0 means off. Then take a moving average of that array (and multiply by 100) to determine the % on-time.

    The most arbitrary part of this algorithm is the size of the window used to do the moving average, so I'd be glad to hear some opinions. The middle plot shows the % on-time throughout the month using a 4-hour window (note that I have no data for the first 11 days...it's not that the furnace wasn't on). The last plot shows the maximum % on-time for the month of January calculated for window sizes ranging from 1 to 24 hours.

    1. pjpfeiff | | #12

      Hopefully the plots are readable. It appears that submitted images get shrunk...

Log in or create an account to post an answer.

Community

Recent Questions and Replies

  • |
  • |
  • |
  • |