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

Can Solar Electricity Trump a Ductless Minisplit?

A GBA reader calculates that PV systems are now cheap enough to displace a high-efficiency air-source heat pump

Cheap and getting cheaper: The plunging cost of residential photovoltaic systems has prompted one GBA reader to wonder whether PV is a better investment than a ductless minisplit heat pump.
Image Credit: National Renewable Energy Laboratory

Ven Sonata’s query is simple: If the falling cost of installing a photovoltaic (PV) system has killed off the viability of solar hot water systems, as GBA senior editor Martin Holladay believes, does it also represent a threat to the beloved ductless minisplit for heating and cooling?

“Is it possible that heat pumps themselves are at least wounded?” Sonata writes in a Q&A post at GreenBuildingAdvisor. “PV at $3.50 [per] watt installed. Note that heat pumps are still great for all places where PV is impossible.”

Efficient, relatively inexpensive, and able to function even in below-zero temperatures, ductless minisplits have become a first choice for many builders specializing in high-performance houses. But is it possible that electric resistance heat powered by PV is now a better choice?

That’s the question for this Q&A Spotlight. Be prepared to crunch some numbers.

Wounded, maybe, but far from dead

Even though the cost of solar electricity has dropped dramatically, the numbers still aren’t there for a wholesale conversion from heat pumps to PV-powered resistance heating, writes Dana Dorsett. At least for the moment.

“Heat pumps are only wounded when the cost of electricity (from PV or any other source) falls to such a ridiculously low level that the up-front cost of the heat pump relative to resistance heating is not viable on a lifecycle cost basis,” he writes. “At 3 US cents/kWh, that may be compelling, but most of the world is paying three to 10 times that much for electricity.”

The levelized lifecycle cost of grid-tied PV, Dorsett continues, is typically greater than 10 cents/kWh for residential systems, much greater in some areas, even as large-scale arrays can produce power for less than that.

That said, there are signs the cost of PV will continue to fall…

GBA Prime

This article is only available to GBA Prime Members

Sign up for a free trial and get instant access to this article as well as GBA’s complete library of premium articles and construction details.

Start Free Trial


  1. dankolbert | | #1

    Great summary!
    Thanks, Scott - this is a really useful and comprehensible overview of a very complex (and important) conversation.

  2. wjrobinson | | #2

    Sunny summer day, PV will
    Sunny summer day, PV will produce energy for many hours to run a heat pump for many hours. The PV helps with lowering the demand on the utility.

    I like all three... superinsulate, install PV, install mini splits, and a small wood stove if you live in the rural Adirondacks where the stove emissions do less harm than say downtown LA.

  3. user-984364 | | #3

    PV is cheap, so we should waste it now?
    I guess I have a problem with what seems to be the fundamental premise here:

    PV is cheap, so we should feel free use it inefficiently?

    I suppose if you are driven by nothing but pure economics, perhaps PV+resistance is cheaper than PV+minisplit. But then coal+resistance would likely be even cheaper, soo....

    I'm pretty happy with my PV + minisplits. Making my own clean power, and using it as efficiently as I can.

  4. GBA Editor
    Martin Holladay | | #4

    Response to Eric Sandeen
    You asked a rhetorical question: "PV is cheap, so we should feel free use it inefficiently?"

    Here's the thing: You can almost always increase the efficiency of a house or a piece of mechanical equipment. The question is almost always: "Is it cost-effective to increase the efficiency of this house or piece of equipment?"

    You seem to be happy with the performance of your $5,000 minisplit with a COP of 3. But I bet you don't want to pay $20,000 for a ground-source heat pump with a COP of 3.5 or 4. Why not? It's certainly more efficient.

    You may be happy with a $200,000 house with energy bills of $900 per year. But you probably don't want to pay $400,000 for a house that has energy bills of $700 per year. What not? It's certainly more efficient.

    My point is: none of us chooses the most efficient option. Cost matters.

  5. user-984364 | | #5

    $5k minisplit vs. what?
    Cost matters, but this blog isn't about the cheapest option, it's usually about some perceived sweet spot of affordability, sustainability, and comfort.

    And sure, your $15,000 or $200,000 differences above are very significant. But we're talking about a $5k minisplit vs. a bunch of resistance heat, which so far appears to cost $0? What is the real cost delta of a minisplit over resistance heat? I don't see a resistance heat equipment cost anywhere in this post.

    I enjoy this navel-gazing as much as the next person, really. ;) But if we're just talking cheap, then you may as well do the math on coal+resistance heat, too. By the time you're installing PV, I assume you're not simply going for the lowest cost approach. I suppose "lowest cost net-zero approach" draws a reasonable box around the possibilities, though, and we can have a meaningful comparison of options with that stipulation.

  6. GBA Editor
    Martin Holladay | | #6

    Response to Eric Sandeen
    I agree that most of us are looking for some sort of "sweet spot of affordability, sustainability, and comfort."

    I was just pointing out that improved efficiency can never be an absolute goal. Every day, we choose less efficient or more efficient options from a range of available products, and one criterion we use to make these choices is cost.

    If the time ever comes when a PV-powered electric-resistance water heater is the cheapest way to heat water, we shouldn't dismiss the technology out of hand just because an electric-resistance water heater is less efficient than a heat-pump water heater. So these questions are worth asking, and these mathematical calculations are worth making.

  7. vensonata | | #7

    Response to Eric,
    The net zero houses I looked at by Marc Rosenbaum had both electric resistance panels (ceiling mounted apparently) and Daikin heat pumps. The reason? All 8 house required, at some point, electric resistance additional when temperatures were very cold. The heat pump is least efficient when it is most needed, ie when it is cold! The heat pump also only runs seasonally in cold climates, which is why heat pump water heaters, which run year round are entirely different than home heating heat pumps. Pv runs year round, least efficiently in the winter, just like the space heating heat pump.
    Resistance heat is cheap to install and always runs at the same efficiency. There is a valid concern expressed in the article about burdening the grid in winter. Heat pumps lighten the demand, as does PV to a lesser extent (unless in very sunny cold climates, such as Arizona where the Pv contributes about 70% as much in winter as in summer) In New England the Pv may produce only 20% in winter of its summer production.
    To some extent I am advocating for simplicity in building the new green home. Should we simply install 8kw Pv rather than: Heat recovery ventilators, heat pumps, radiant in floor hydronics, very expensive triple gaze windows and even super efficient appliances which cost more than they can save? When we do the math at $3 watt, Pv comes out ahead of all these other techniques. Granted it all exclusively depends on two factors: a suitable south facing roof and a reasonable grid to tie into.

  8. kevin_in_denver | | #8

    Going Off Grid in the City
    We are hearing rumors that net metering in Colorado will be sabotaged by Xcel. (We think) a compromise will take the current credit of $0.15/kwh down to $0.06.

    If that happens, Denver is sunny enough in the winter to go off grid with battery storage and a backup system of some sort. The PV will work better in winter with a very steep tilt, and shed snow better.

    An automatic propane powered generator can keep you completely without a gas or electric bill. A natural gas generator is also an option, and so is using the electric grid for backup.

    All-electric with minisplit heat pump is the conventional wisdom when there is generous net-metering. If generous net metering goes away, maybe the trend will be back to a gas forced air furnace for backup heat. Xcel allows you to turn off gas service for, say, the 6 months when you don't need backup heat, which saves the $12 service fee.

  9. user-626934 | | #9

    Response to Ven...

    First - the houses in Marc's study ( all had electric radiant ceiling panels (I think just in the bedrooms) to be used at the household's discretion. Four of the households used these sparingly during the study period (~200kWh and less), while the other four used them quite a bit (one to the tune of about 1700kWh!). The Daikin units that were used (these houses were designed back in 2009, BTW) are "OK" units, but not specifically designed for very high capacity and efficiency in cold climates. Look around at references to newer mini-split heat pump installs (including Marc's own new house!) and you won't find much mention of Daikin. Most folks who are currently designing compact "PGH" or better houses and who are using heat pumps are using the Fujitsu or Mitsubishi units that do not cut off until about -20F, and have significantly higher COPs. In addition, I'm sure that it's not universal, but most folks in colder climates who are installing mini-splits in these houses are either forgoing back-up electric resistance heat (because the heat pump can cover ALL their load, even down to record low temperatures in many cases), or supplying backup heat in the form of a couple of 1500W space heaters stored in the closet or perhaps a woodstove that's mostly there for ambiance. All of this is to say that your use of Marc's study to inform your calculations is in the right direction, but off the track a little bit. You're using the most up-to-date info. for PV systems, you should do the same for heat pumps.

    Second - if your goal is to speed up the death of net-metering, then the solutions you propose are a good way forward. Many electrical grids in cold climates have winter peaks already...where the peak comes at ~7am-8am and distributed PV is providing absolutely no benefit to the grid. Increasing the use of electric resistance heat en masse will require more power plant capacity that will require more cost that has to be paid by someone. Guess who the utility companies will try to pin these costs on...Even on my own grid in central Virginia (Dominion Power section of the PJM) we are now a winter-peaking grid. Dominion hit an all-time peak on Feb. 20 for the hour ending at 8am with a total demand of 21,651MW….up almost 1,600MW(!) from the previous record of 20,061MW that was set in July, 2011. At my house we averaged a "whopping" 2,000W of demand during that hour despite our mini-split heat pump running at about 80% of capacity and our water heater running as well (heat pump, of course!)….no PV during that time….the sun is too low, and besides, it was covered with snow and ice and had essentially zero production from 2/16 to 2/26. Electric resistance space heating would have put our peak demand during that hour over 4,000W...electric resistance space heating AND water heating would have put our peak demand over 10,000W. Someone has to pay for the peaks...

  10. GBA Editor
    Martin Holladay | | #10

    Response to John Semmelhack
    You wrote, "If your [Ven's] goal is to speed up the death of net-metering, then the solutions you propose are a good way forward."

    I appreciate your analysis. You are correct -- more electric resistance space heating would not be good for most North American utilities (although some utilities, notably Hydro Quebec, manage these loads just fine).

    In areas of the country with net-metering contracts that are favorable to PV system owners, however, individual homeowners are going to make economic decisions that make sense for their own family's electric bill rather than choosing a (possibly more expensive) path that benefits the local utility. Your opinion (like Ven's opinion) is irrelevant.

    Your prediction -- that utilities will react with new fees to make grid-connected PV less attractive -- sounds correct to me. But utilities will impose these new fees not because of Ven's analysis; they will impose these fees in response to bottom-line facts that tens of thousands of homeowners will eventually figure out. If it makes more sense for a homeowner to invest $10,000 in PV than to invest $10,000 in energy-efficiency upgrades, the homeowner will probably choose the better investment.

  11. John_Hansen | | #11

    How many in Northeast do not have A/C now?
    At the end of this post, the last paragraph states: "... in the Northeast, where minisplit heap pump incentives are driving building owners from fossil fuel to electric forms of heating, the utilities will also have to deal with new summer peak loads as those same heat pumps provide central air conditioning for the first time to all those folks!"
    That statement raises a number of questions:
    Is this incentive that you mention going only to existing building owners? Are there similar incentives for new construction? How many of these existing buildings have no A/C at this time? Is the incentive targeted at buildings without A/C now? Will the net effect be higher energy usages at the same or lower cost than in the past? How well are the incentives achieving the goals of those who are funding them? Are these incentives intended to lower the carbon footprint, or simply shift the income from fuel oil producers/suppliers to the power generators?
    Many incentives are political rather than economic in nature. Some energy efficiency retrofit programs for existing homes would never work economically if they were not funded by a mandate that some government agency created. Imagine if the consumer were faced with the cost of borrowing the money that was not offered by the incentive. Will the outcome simply be the use of more total energy but having that energy coming from different sources than in the past?

  12. GBA Editor
    Martin Holladay | | #12

    Response to John Hansen
    You have posted more questions than I can reasonably answer. But if you want to know more about one program that provides minisplit incentives, you might be interested in reading this article: Heat Pumps Get a Leg Up in Vermont.

  13. dankolbert | | #13

    AC in Maine
    Central AC is very rare in Maine, even in new construction. Maine is also, apparently, a serious outlier in terms of the rarity of forced air for heat.

  14. GBA Editor
    Martin Holladay | | #14

    Vermont is like Maine
    Most homes in my area of Vermont certainly don't have central AC, and hydronic heat distribution is much more common here than forced air.

  15. Expert Member
    Dana Dorsett | | #15

    The death of net metering...
    ... isn't necessarily a bad, thing in the face of cheap PV (on either side of the meter), as long as it comes with rate reforms that charge folks for their actual impact on grid infrastructure costs rather than basing it simply on total energy use, and pays distributed generators for the avoided infrastructure & other costs. The "Value of Solar Tariff" approach is working in Austin Texas, and has been written into Minnesota's state code to address some of these issues, but residential demand charges can be reasonable, if applied across the board, and not selectively applied only to PV owners. When people are required to pay for their actual impact on grid costs, their decision making will get better. (In this context, the heating equipment decisions.)

    The dumb per kwh block energy use approach without real time pricing information keeps power users in the dark as to their true impact, and offers no incentive for power users to "do the right thing" in terms of keeping the grid stable and affordable. If the new paradigm is demand charges assessed for pay for the grid infrastructure, along with real-time pricing on energy use, the value of heat pumps & PV to the homeowner will be enhanced rather than degraded, and the heavier grid load of resistance heating would be less attractive in terms of both peak grid-use and peak energy costs.

    Net metering was originally taken on for the convenience of keeping it simple, and it has been working reasonably well at current PV penetration rates. Getting rid of net metering WITHOUT creating an energy-price & grid-cost aware level playing field for all grid users is often rightly seen as a gouge sold to the regulators on a fairness against non-PV ratepayers cross subsidizing the wealthier PV owners. But all careful studies that have looked into the cross-subsidy aspect in the lower 48 of the US to date have found that no cross-subisidy is actually happening (indeed, PV owners are avoiding costs for other ratepayers), even if that might become the case at some point down the road when 20% or more of all ratepayers are also PV power producers.

    But that day will come, and with it, net metering as we know it really must die. By then we'll be looking at much cheaper PV than we have today (and much cheaper local storage too.) There are no technical hurdles for making the grid smart enough to handle 2- way power flows equitably. Some get that fact, others don't. The National Bank of Abu Dhabi just published a report indicating that PV is expected to be cost-competitive with $10/bbl oil or $5/MMBTU natural gas in the not too distant future (as in maybe right now, in some instances.):

    Agora Energiwende/Fraunhofer agrees:

    State regulators who seem more interested in protecting the monopoly interests of the utilities than the ratepayers will not be able to prevail forever. The train has left the station- the utilities can either get on board or get run over. Cheap PV on the utility's side of the meter can be good for ratepayers too, but sticking the ratepayers for the full cost of stranded assets built while the utilities/regulators had their heads in the sand pretending distributed PV wasn't inevitable, or that it was only a problem to be crushed in it's infancy will become a chapter in the epitaph for those utilities.

  16. STEPHEN SHEEHY | | #16

    A/C in Maine
    I've lived in New England all my life and never thought about air conditioning. Sure, a few days every summer are steamy, but not enough of a bother to think about central air. But, since my new house is going to be heated by minisplits, I'll probably use the A/C once in a while. It's a minor plus over just using resistance with more PV.

  17. Expert Member
    Dana Dorsett | | #17

    New England AC loads are mostly latent
    The sensible cooling loads in New England are indeed fairly modest, even more so in home with high-R attics/roofs and better-than-code windows with appropriate seasonal shading from overhangs. But the latent loads (humidity) in New England is still significant enough to matter, especially for those with dust-mite allergies who need to limit the interior RH to 50%.

    A $250 room dehumidifier is enough to manage the latent loads of a tight house at modest ventilation/infiltration rates. But a dehumidifier also converts that latent cooling load into a sensible load- it raises the room temperature. A mini-split in "dehumidify" mode moves that latent heat outdoors, and dehumidifies while doing at least some sensible cooling, which may or may not be needed. Either/both may be needed or desirable in different seasons, depending on the particulars of the house.

  18. user-1044645 | | #18

    Two's company?
    Interesting article and full discussion - I think we agree that solar PV for resistive heat (except in extremely well insulated homes) is almost never compelling vs. a heat pump. From the perspective of a solar installer in Maine, I'd say, if anything heat pumps HELP us sell more solar PV, especially in retrofit situation where the heat pump is displacing hundreds of gallons of nasty oil. So we see them more as extremely complementary technologies and not antagonistic. With financing in the picture, even with cheap oil, a homeowner can usually get, say, a 4kw solar PV system + 1 or 2 heat pumps and be cash flow positive. And we are in a state with NO state-level solar incentives (apart from net metering).

    Of course, we also disagree with the claim that solar thermal is dead (dead in certain - perhaps common - use cases but still great when a large volume of hot water is demanded and even better for heavy summertime guzzlers like inns and restaurants).

    - Fred Greenhalgh
    ReVision Energy

  19. Expert Member
    Dana Dorsett | | #19

    It's a matter of "at what price". (response for Fred, #18)
    At a buck-a watt for PV, are $3K/ton heat pumps "worth it", in a fully net-metered utility environment?

    At that price you'd get sufficient input even in winter to cover the wintertime heat load with resistance heating to make PV a better investment. It's hard to imagine buck-a-watt installed pricing at the residential level in New England even if the hardware were free, but it isn't out of the question within our lifetimes.

    But simple net metering buck a watt PV doesn't seem like a very likely scenario when that price point is met, given that it costs something to maintain the grid. Correctly & fairly pricing the grid for all users is the essential nut to crack going forward, and at the moment nobody seems to have gotten it right. Eventually demand charges or something even more nuanced than that will have to get built into the utility rates, even for net-zero-electricity users. But it's hard to say until those grid-use details get worked out whether the cost of the heat pump has value in offsetting both the peak power use of resistance heating plus the energy use.

    $4/gallon for #2 oil has been incentive enough for heat pumps in New England, with or without access to the sun! Oil may be more like half that this year but unless Chinese oil demand growth goes negative, it won't stay this low for long. The competition to PV is really still the retail rate of grid power, which varies quite a bit from utility to utility within Maine (or the rest of New England). Five years ago the cost of PV was high enough to really need subsidy to compete. But in 2015 retail grid-parity for PV is already here for most New Englanders even without subsidy. In another five years it will beat grid pricing everywhere in New England, and by huge margins in the higher-cost utility areas.

    The subsidy shouldn't last forever, but perhaps should be stepped down based on PV penetration levels rather than firm calendar dates such as 1 January 2017. On Oahu about 1 in 9 residential customers already has PV, and the arrays are big enough to have created mid-day backfeeding issues on some local grids, yet PV installed in 2015 in those neighborhoods will still be subsidized as much as PV installed in Maine where maybe 1 in 100 homes have PV and all PV simply offsets the mid-day grid demand peak, lowering the capacity requirements of the transmission lines & substations. Staking it to the calendar is too crude to be rational.

    It takes a clearer crystal ball than mine to figure out what the rate structures will be over the lifecycle of a PV array or heat pump. But it doesn't take a crystal ball to know that PV is going to be a LOT cheaper at the end of life of equipment installed in 2015, and that it will be very disruptive to utility business models.

  20. user-1044645 | | #20

    Sort of related
    Dana, others might be interested, Maine's PUC completed its first 'Value of Solar' study and modeled a 33cents/kilowatt-hour 25-year levelized value of solar. It'll be an uphill battle to get the State to support any new solar policies, but it is a strong affirmation of net metering. Good bedtime reading (178 pages) -

  21. Expert Member
    Dana Dorsett | | #21

    I'm sure the utilities aren't interested...
    In paying out a 33 cent/kwh tariff on PV in Maine, even if allowed to pass-through that cost to the rate payers, eh? :-) As the penetration levels increase, the incremental value of that PV will of course decline, but given the very low implementation in ME to date I'm not surprised that the PUC would come up with a VOS north of 30 cents. I haven't had time to scan the 178 pages of analysis, but it's probably worth a look:

    It'll be awhile before Maine undergoes the tectonic shift currently under way in NY, but it'll probably happen eventually:

  22. Expert Member
    MALCOLM TAYLOR | | #22

    A moral component?
    Things are getting interesting, and may become a lot more so if PV continues to realize its potential. Suppose in the future PV and some sort of storage becomes efficient enough that electricity is effectively so cheap and abundant that we can squander it by living in poorly insulated houses? Where does this leave the super-insulation movement? Does it fall back on the comfort arguement, or claim somehow that there is still a moral imperative to save energy, even if its generation isn't really polluting?

  23. dougzo | | #23

    Late to the party
    I build energy efficient homes for low income families in Utah. We are currently starting our 7th and 8th all electric, 1000 sq/ft straw bale, homes here in Moab, Utah. The summers are hot, the winters are cold and we are lucky to have abundant sun. Here is what I've found so far. We have 3 homes heated by hydronic heat loops powered by two standard shorty electric hot water heaters tied in series. These have evaporative coolers installed for cooling. We now have 3 homes conditioned by mini splits. Two houses have ~ 2kw PV systems, two have ~3kw PV, and two homes have ~4kw systems. Everyone is relatively happy with the comfort of their home. I visited one of the homes today that has 2kw PV and the electric resistance ,open loop, hydronic heat (they only keep one of the hot water heaters turned on)-- it will come close to net zeroing this year. It's a couple and they are thrifty with their energy use. The take away is we begin to split tiny hairs.. all things added in the systems (mini split vs electric and swamp cooler) come in well under $5000. I like the mini split for its simplicity (I do add a bath fan with a built in resistance heater "in case" that costs $100 extra per bath) and ability to deliver heat to the occupant almost instantly. I have thought long and hard about what is more cost effective and what is more efficient because we have limited budgets for the homes. It just doesn't seem to matter to much because we have good solar, great insulation and mass and a pretty small heating season for the home, 2 to 4 months. I still advocate to oversize the PV as the uses for that extra power are many (might power your car someday). I do worry about where net metering will go in the near future. With the Fujitsu 9rls3h drawing ~650 watts and providing 12,000 btu/h heating ( enough to meet the heat loss on our homes) I think the possibility of heating a home with PV / battery system my not be to far off... so watch your step "utility company". One thing I've noticed about using all electric + hydronic in floor heat is the ability to add the heat to the house at off peak hours (this applies to occupant need of domestic hot water but correlates with grid off peak also). The draw back, as mentioned in other articles, is not being able to deliver heat quickly to occupants. I can't wait for an inexpensive outside air source heat pump that ties into a domestic hot water heater. Cheers.

  24. GBA Editor
    Martin Holladay | | #24

    Response to Doug Nichols
    I agree with you: "The takeaway is we begin to split tiny hairs."

    If we are talking about a small, well insulated, all-electric house that has a PV array that is between 2 kW and 4 kW, you are 100% correct. These homes are easy to heat. Almost any type of heating system will work.

  25. ethan_TFGStudio | | #25

    Having lived in a small house with electric baseboards... I found that I was able to effectively zone... letting the living area get cold and keeping the bedrooms warm at night... I know that in a better insulated home, maybe this wouldn't even be necessary. But I think it is a worthy consideration when considering electric resistance in each room compared to heating a whole house with one minisplit.

Log in or become a member to post a comment.



Recent Questions and Replies

  • |
  • |
  • |
  • |