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

HVAC – Gas furnace vs. heat pump vs. ducted minisplit

user-5946022 | Posted in Energy Efficiency and Durability on

It seems the general consensus here, on a very tight or “pretty good” house, is for ductless minisplit. I’m trying to understand the economics.

The following parameters are non-negotiable:
1. Climate zone 3A
2. The house will be piped for natural gas for, at a minimum, cooking and outside grill
3. No ductless mini-split. Ruled out due to aesthetics, so further conversation about this is pointless. I am also concerned about comfort in each room, since with a ductless mini split there would only be two or three distribution points of conditioned air in the house. Regardless, attempting to solve that problem is a risk I do not need to take as ductless are ruled out aesthetically.

This leaves the choices of: ducted mini-splits (3 heads), 3 zone split system heat pump, or 3 zone high efficiency gas furnace with piped outside combustion air intake. Any system with have an independent ERV.

There will be a base charge for natural gas service no matter what per item 2 above. The cost of natural gas is about $.35/therm, and the cost for electricity is about $.11/kwh.

Thus, from a purely economic point of view, the gas furnace seems like a slam dunk. Further, if in the future a solar array is added, gas for heating makes it more plausible the solar could support the balance of the house in the heating months when the solar collects less.

What am I missing here?

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

    I don't think you're missing anything. Natural gas is a cheap fuel.

    Of course, a heat pump also provides cooling. If you install a gas furnace, you'll still need to buy some type of equipment to provide air conditioning.

  2. user-5946022 | | #2

    Thanks - I should have been clearer - the options are all with paired heating and cooling. So
    - Ductless mini split to heat & cool via electricity
    - Split system heat pump to heat & cool via electricity, with gas backup when the temp drops below 37
    - High efficiency piped combustion gas furnace with 16 SEER or greater electric air conditioner
    Each system would have 3 zones. I need to run the numbers when all design data is finalized but am guessing an approx 2 ton ac unit, with 45k btu furnace.

    When I do rough numbers of initial install costs vs. operating costs, I come up with >40 year payback for the ducted mini split. I don't think they will last that long (and the owners certainly won't).

  3. user-626934 | | #3

    $0.35/ therm? Really? National average is a little under $1/therm. If so, no heat pump system is going to compete based on current operating cost.

  4. user-5946022 | | #4

    Yes, it has gone down considerably, and that is just the average. They get down to $.29/therm if you renew at the correct time. The per/therm rates they file with the PSC are in the $.40-50 range, but then they send out all these deals offering low /therm costs if you switch, and your current marketer will generally meet or beat any competitor offer.

    If you calculated TOTAL cost/therm, it might get close to $1/therm, but almost 50% of the annual cost is in the monthly "base charge" and the "customer service fees", which you need to pay no matter what if you have natural gas service to the house. After that it is just cost/therm.

    I just checked my own cost/therm for the past few years, which is as follows:
    2016: $0.345/therm
    2015: $0.519/therm
    2014: $0.485/therm
    2013: $0.509/therm
    2012: $0.579/therm
    2011: $0.649/therm

    So perhaps I should use an average of $0.55/therm. But even at that rate, natural gas seems to beat out any heat pump or split system. I figure the system will have a useful life of no more than 20 years, and at that time they can re-evaluate based on current conditions. But for the initial investment, it seems the payback needs to be max 10-12 years, and I'm not seeing it.

    The other consideration is in this area, most of the electricity is produced by coal fired plants. Fracking is not great, but coal fired seems even worse...

  5. charlie_sullivan | | #5

    I am puzzled when you say >40 year payback. That implies that the equipment+installation cost is higher for minisplits compared to the other options. I would think that the up-front cost would be lower, based on not needing a full duct system, and not needing the gas furnace.

    It's very hard to know what cost to use for natural gas. I think you are right that it's better to use an average number over a longer time period rather than just using the numbers for this year. But there's no guarantee that that's the right thing either. There is simply no way to be certain about what number to use there. But it is certain that there is risk in committing yourself to one particular fuel for 15 years. If you don't like that risk, you might want to avoid that choice.

    Of course, it's also hard to predict what electricity prices will do. But at least you have the option of PV there. I don't know what state you or in or what the net metering rules are there, but in most places, you are allowed to average out over a year, not just over a month. So you don't need to worry about matching the winter vs. summer electricity production and use. You just need the total for a year to come our right.

    Your discussion of coal fired electricity as a negative makes me think you might be interested in climate impact as well as cost. In that case the PV net metering option something to seriously consider, and something that might sway you toward heat pumps. You might also be able to choose your electric supplier in order to avoid coal. Another thing to consider there is what the mix on the grid will be in 10 years, rather than just what it is now.

  6. user-5946022 | | #6

    Yes, equipment + install for minisplits is a bit higher than for a gas furnace + 16 SEER AC. Maybe it is just what the contractors are used to. So the higher install costs would need to be made up for in lower operating costs, but I think the ducted minisplit and heat pump operating costs are higher.

    I will look into the net metering on PV. Right now PV is NOT in the budget. Costs would need to come down.

    Yes, climate impact is usually a consideration. No option here on power suppliers, only on gas. And in 10 years the power company hopes to be giving us more power from nuclear plants...But for this discussion I'm focused on the economics.

    One thing that is clear to me on the economics of all of this - providers are consistently increasing the portion of the bill that is service charge, and base charge, while dropping or keeping the unit cost the same. Those service charges and base charges make up 50% of the annual natural gas bill, and about 25% of the power company charge. Unless someone goes off the grid, none of the efficiency measures impact those costs.

    My thought that another advantage of the gas furnace would be if someone did go off grid you would not need to divert as much of your PV output to heat if you had gas...but probably not something to worry about in the next 20 years unless the cost of those Tesla storage batteries goes way down along with the cost of the PV dropping.

  7. GBA Editor
    Martin Holladay | | #7

    You wrote, "Providers are consistently increasing the portion of the bill that is service charge, and base charge, while dropping or keeping the unit cost the same."

    That's an argument in favor of an all-electric house. Lots of people live without natural gas. Far fewer are willing to live without grid electricity.

  8. charlie_sullivan | | #8

    Just in case you haven't looked at induction cooktops, the reign of gas as the only option for serious cooks is over. Induction gives the same fast response advantage. If you go that way, you could avoid the installation and ongoing base charge for gas is propane works for you for the outdoor grill.

  9. JC72 | | #9

    When spec'ing the zoned system keep in mind you really should have a variable speed blower and variable stage compressor. These items will be more expensive than the run of the mill single-stage compressor and single speed blower. You should also spend the money to have it properly designed (Manual S, T, J & D) with at least some rigid duct in order to gain max efficiency. Rigid duct is more expensive btw. IMO once that's done you'll have more information with regards to making your decision. Good luck!

    Note I live in zone 3A, have a zoned system, it has none of the above attributes and performance is mediocre at best.

  10. Dana1 | | #10

    I'm a bit surprised that a 3 zone mini-split, even mini-ducted would come in higher than a 3-zoned gas fired + split AC system. Last fall I saw a 6 zone ductless system come in substantially less than a 2 zone fossil burner + split AC ($15K instead of $25K, both perhaps overpriced due to being a high end vacation-island location.)

    An alternative to wall-blobs or ducts that can still work well in tight low-load houses is ceiling cassettes. SFAIK there aren't any half-ton ceiling cassettes out there, but there are some 3/4 tonners. In low load houses all gas burners are oversized for heating, and would sub-optimally oversize the ducts. A ceiling cassette mini-split AC solution with a micro-zoned hydronic solution running off a condensing gas HW heater can work pretty well, once you have the room by room load numbers. Without knowing the loads it's impossible to optimize the HVAC choices. Hire an engineer or RESNET rater, (and NOT an HVAC company) to run the load numbers using aggressive, rather than conservative assumptions on air tightness, etc. People who are only selling the accuracy of their calculations are far more likely to get it right than HVAC contractors, who are usually too cautious and use super-conservative assumptions, fly by the seat of the pants using rules of thumb, or have a financial incentive to choose one option over the others.

    Based on the statements about nuclear content of the future grid mix and climate zone I'm assuming you are in Georgia, and the statewide monopoly utility Georgia Power is your utility. The Georgia legislature unanimously approved allowing third party ownership of solar in their state. As electric rates rise (as they must, and already have,) to pay for the still unoperable new nukes at Vogtle, at the same time that rooftop PV pricing continues to fall, it's likely that solar companies will be willing to sell you power for less than grid retail for the privilege of putting panels on your roof before 2020. At the same time, natural gas is trading near historical lows, with no way to go but up. Doing the financial analysis on the dubious assumption that gas price will be flat or continue to fall, and that carbon emissions would never be taxed is a bit risky, as tempting as that might be. With electricity markets the long term trends are downward, due to the year on year reductions in the levelized cost of wind & solar, but in Georgia the regulators have already committed to allowing GP to recover (or actually partially PRE-cover) the substantial sunk cost of the new nuclear plants. When rooftop solar becomes cheaper than grid-retail (which is soon) the increasing retail cost of electricity on YOUR grid may make behind the meter PV a no-brainer type of investment,

    Many analysts in the investment banking sector believe rooftop solar will become cheaper than natural gas before 2030. The potential fly in the PV ointment is how the state regulators opt to tread net metering and the value of solar when that time comes. Just about every state in the US is undergoing analysis and revision of solar policies, now that rooftop PV has crossed $3.50/watt on the way down. The grid necessary to support large centralized power generators is larger and more expensive than the grid needed for a widely distributed generation model. Whether the regulators opt to protect the monopoly utility business model (as recently happened in Nevada) or open up the whole can of worms for re-defining the grid (as is currently under way in New York), makes a large difference in whether the ratepayers get stuck paying for the stranded assets of the monopoly, or whether they get to enjoy the benefits of electricity price deflation inherent in distributed solar & wind.

    If you haven't yet, you may want run the napkin-math carbon numbers:

  11. Reid Baldwin | | #11

    When I looked up natural gas costs from my utility, there was a fuel charge and a delivery charge that were both proportional to the amount used. Make sure you are including all of the proportional charges.

    Utilities would like to make most of the charges fixed instead of proportional since, in the short run, most of their costs are fixed. Regulators don't let them go too far that way because of the way that would impact consumer behavior. That encourages people to either use more or get off the grid.

    Now, it is much easier to stay on the electric grid and drop the natural gas utility. In the future, Combined Heat and Power (CHP) units may make the reverse more practical. So far, however, there are not many commercially available CHP units.

  12. user-626934 | | #12

    CL - based on your comments, it sounds like the $0.35/therm is just the fuel charge. In order for you to do an apples to apples comparison based on current costs, you would want to compare the "all-in" $$/therm (including the delivery charge as well as apportioning some of the monthly service charge to the heating system). Of course, you'd want to do the same with the $$/kWh cost for electricity.

  13. Expert Member
    Dana Dorsett | | #13

    In a zone 3A climate you may be missing the boat if you're focusing on the marginal cost of heating rather than air conditioning. At base 65F for a presumed balance point the heating degree-days and cooling degree days are comparable. But a tight code min house will have a balance point of about 60F, and a better than code house would be below that.

    To put it in perspective, play around with this tool a bit, for your ZIP code, looking at a few relevant base temps and years:

    Taking as an example, for 2015 in Atlanta GA 30301 :

    Base --------HDD--------CDD

    65F---------2204---------2151 < < < ever so marginally heating dominated

    60F---------1502---------3274 < < < twice as many CDD as HDD

    55F----------978---------4575 < < < more than 4x as many CDD as HDD

    If your location is similar and you're building a better than code house, why would you be focusing on the marginal cost of space heating rather than cooling? You'll be dealing with more cooling MMBTU/year than heating MMBTU/year, and while you'd be able to deal with some of that cooling load by opening up the windows, your cooling costs will likely exceed your (fairly low) heating costs. By building to a better standard you will have "moved" to a cooling dominated climate.

    A cooling coil in an oversized gas furnace air handler is a LOUSY choice for cooling compared to a right sized modulating heat pump with a decent turn-down ratio. The fact that it might be marginally more expensive to run the heat pump than gas for heating hardly matters, since it's less than a third of the annual load (assuming tight code-min.) Mitsubishi's MVZ series air handlers come in several sizes starting at 1-ton, and they can be configured as a zone on a MXZ series multi-split:

    To go smaller you'd need to be mini-ducted.

    IIRC Daikin's small air handlers start at 1.5 tons, unless you go mini-ducted.

    Without knowing the loads it's impossible to optimize equipment recommendations for cost, efficiency, or comfort but a generic 50-70,000 BTU/hr condensing furnace with a 2-3 ton cooling coil probably isn't the sweet spot for anything except up-front cost.

  14. charlie_sullivan | | #14

    Dana makes a really good point about cooling costs vs. heating costs. And that's not even counting the latent load in that humid climate.

  15. user-5946022 | | #15

    I was not focusing only on heating costs - I'm focused on the total cost to condition air - heat and cool. Apparently I just did not word the question clearly. Yes, it will be tight and above code min.

    The cost of a standard 2 ton/45k btu split system is sufficiently less than heat pump or ducted mini-split, that I cannot see the savings in operating costs ever compensating for the differential.

    Chris - Yes, before finalizing I will do a Manual J, S D, etc, load calc, and all of that...and regardless of the type of system the trunk lines will be hard duct, with the branch lines flex, properly installed...but the cost differentials in the different type systems remain regardless of all of that.

    Martin & Charlie: I understand the idea of no natural gas, but that is not going to happen in this situation - the house will be piped for natural gas, and there will be no ductless mini I have to deal with it from there.

    Reid -yes- there is a slight differential in base cost based on use, but it is no more than $3-4/month. The majority of the costs are customer service fees, billing fees, and a "basic" base charge.

    John - please help me understand your logic - if the house is going to have gas service and power service, then the base charges and service charges will be there regardless of usage. So the only cost impact is the incremental cost of usage, which leads me to calculate the operating cost based on unit cost of fuel/power used....

  16. charlie_sullivan | | #16

    Setting aside the ducted minisplits for the moment, because they have high cost of reasons that may be related to the local skill base not to fundamentals, you are saying that a heat pump system is substantially more expensive than A/C plus a gas furnace. Is that because you are using gas backup for the heat pump, and so you are installing a gas furnace in both cases? What if you use a heat pump that is good enough at low outdoor temperatures that you don't need a gas backup? I don't know off the top of my head a particular unit to recommend, but am just trying to understand what the factors are that make the costs come out this way.

  17. user-5946022 | | #17

    Charlie - yes, you are correct. Heat pump with backup gas heat. Supposedly required in this area...
    And the heat pump operating efficiency on cooling does not seem to be much different than a regular split system, so not sure what the advantage would be...

  18. Expert Member
    Dana Dorsett | | #18

    In recent experience on one project quotes for a 1.5 ton ducted mini-split solutions came in substantially cheaper than a gas burner + split system. The Fujitsu 1.5 ton ducted mini-split proposal was about $1000-1500 less than the 1.5 ton Daikin, but both were cheaper and more appropriately sized than any of the the gas + non-modulating split systems proposed.

    Without the load numbers in hand it's hard to say what it really costs in dollar terms. But right sized modulating systems are going to be a heluva lot more comfortable than the typically oversized split + gas furnace. Note the as-used efficiency of 1 & 2 speed AC with zoned with dampers is usually considerably less than what it is using the same air handler & condenser on a single zoned system, making it harder to do apples-to-apples with mini-splits/multi-splits.

    Unless this is the largest better-than-code house in Zone 3A it's unlikely that you'll need anywhere near 45,000 BTU/hr of furnace if it only needs 2 tons of cooling (assuming it even needs 2 tons.) Even 3000-4000' code-min houses in Maine don't usually need that much furnace (though there is often 3x that much installed) but they usually need 2-3 tons of cooling. As error prone as rules of thumb are, in zone 3A a code min house would typically run no more than 10 BTU/hr per square foot of conditioned space, and cooling loads will run about a ton per 1000'. Better than code homes with a bit of optimizing on solar gain to pull down the peak cooling can be substantially less, a ton per 1500', and ~7 BTU/hr per square foot heating. A 2-ton Fujitsu multi-split and some 7RLF or 9RLF cassettes can probably cover the your house reasonably, and more comfortably than old school zone dampered stuff.

    The original statement that led me to thinking the focus was primarily on the marginal cost of heating came from:

    "Thus, from a purely economic point of view, the gas furnace seems like a slam dunk."

  19. charlie_sullivan | | #19

    Might be worth probing further on the "supposedly required around here." That might be out of date and/or unreliable information.

  20. user-2890856 | | #20

    Nanny state Charlie . Doubtful it has gone away .

  21. Dana1 | | #21

    As long as you have the extended capacity tables (if necessary) to demonstrate that it has sufficient capacity at your 99% outside design temp to heat every room to 68F under those conditions it usually doesn't need "backup", even in areas where codes have traditionally held that heat pumps need backup (an artifact of lousy low temp capacity of ducted air source heat pumps of a few decades ago.)

    This came up on a deep energy retrofit I got involved with (in climate zone 5A) a few years back, a 3-story house heated with one ductless head per floor (with oversized mini-splits). When I pointed out that according to the Manual-J numbers and capacity data the mini-splits were already oversized by a bit more than 50% at the +5F outside design temp, and that couple of 1500 watt space heaters could heat any given floor at design temp even if the thing crapped out completely, the topic was quickly dropped. YMMV.

    A mini-split has to be able to deliver it's full "rated" capacity in an HSPF test at +17F outdoors, +68F indoors. +17F is below the outside design temp in most zone 3A locations, which means you probably wouldn't need to consult extended temperature tables, and rely simply on the summittal sheet's rated capacity.

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |