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

Minisplit – One indoor unit for two adjacent rooms?

SteveR_79 | Posted in Mechanicals on

Do any of the minisplit manufacturers offer a wall mounted indoor unit that will service two adjacent rooms? I’m considering a minisplit to replace the gas furnace and a/c in my 1300 sq. ft. single level home in zone 3A (mixed-humid). Without a shared indoor unit on the common wall between two bedrooms, I might need five indoor units. The multizone minisplits seem to jump from 4 to 8 zones and the price jumps too. I understand the technical problems with one wall unit serving two rooms, but have any of the manufacturers solved them.

As a side note, retrofitting an older house with a minisplit is a challenge. First off, my house has R15 walls instead of the R40 or R50 used in Carter Scott’s houses. I would expect my results to vary from his if I installed a single zone minisplit.


Most contractors I have spoken with have installed minisplits in additions only. Deciding on the number, the type (wall, ceiling cassette, ducted) and the location of the indoor units for a whole house is a new experience for them. Is there a Manual D for minisplits?

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. wjrobinson | | #1

    If you have natural gas, going to mini splits is a bad idea.
    Five heads is a bad idea.
    Copying a superinsulated project without that label of insulating is a no go.

    What is the reason for the work?

  2. wjrobinson | | #2

    The proper mini split if there is one would be a ducted mini split.

    Mini split installers are rare. If you go that route find someone with many installs done with proven good results. Where I live they do not exist. Would be nice if some installers would come forward here at GBA.

  3. srenia | | #3

    Never install them youself. Manufacture does not warrantee. Most hvac people are crooks. Not all. If you want high performance and low cost go for mini ptac heap pump systems. Will pay for themselves decades before a mini split will.

    The furnace and ac system keep if you have though!

  4. SteveR_79 | | #4

    Thanks for the response AJ and S.E. The reason for the work is that the furnace and a/c are both over 30 years old and are becoming troublesome. The duct work in the vented crawlspace is 50 years old. It needs to have the insulation removed and mastic applied to all the joints. I wasn’t smart enough 30 years ago to have done that when I insulated them. Combine all that and it seems that a minisplit is a worthy consideration. I'm also considering a new gas furnace and a/c as well as rehabbing the ducts.

    I think we're in agreement on Carter Scott's houses. There is no valid comparison between our houses.

    Tell me how to decide that natural gas is the no-brainer answer for heating. Over the past year, my quoted cost for natural gas was $0.57 per therm. After service fees, taxes, etc. my actual cost was $1.20 per therm. That’s the lowest in several years. My actual cost for electricity last year was $0.13 per kWh. Using an AFUE of 90% and a HSPF of 9.6, the cost per Btu for both systems seems real close. The minisplit is slightly higher but the furnace duct losses could negate that. If my electric costs increase by a penny or two, the comparison doesn’t seem as good. What am I doing wrong?

    You could be right that five heads are a bad idea. I don’t know and I have yet to find someone who does. That’s why I’m here. I have tried to read everything at GBA on minisplits. At this point I’m not convinced that I should or should not use a minisplit.

    @S.E. – It wasn’t my plan to install the system myself. Sorry if I implied that. But to be clear, I don’t have a problem with those who do want to install it themselves. I haven’t considered the PTAC or PTHP at this point. I’ll add that to my list of questions for the HVAC contractors.

  5. mackstann | | #5

    There are the mini ducted units that can serve an arbitrary number of rooms, but they require more work to install the ducts and probably some kind of soffit built to enclose them -- thus, more cost. Seems like the best solution though.

    I have a 900sqft single level house in zone 4c with ~R-25 attic and no insulation anywhere else. We almost get by with our 1.5 ton Mitsubishi single head. We supplement with a single space heater during the winter. I'm having air sealing and insulation work done soon and hopefully this will eliminate the need for the space heater. We do keep bedroom doors open -- that's important.

  6. Expert Member
    Dana Dorsett | | #6

    Carter Scott's houses are all in the cold-edge of zone 5 into climate zone 6, where the 99% outside design temps are at most low-single-digits, and the average mid-winter daily average temps are in the low 20s or high-teens. It takes a much higher-performance envelope in his location than it does in yours.

    Point-source heating/cooling can still work OK in a zone 3A climate with R14-R15 cavity fill, providing the doored-off rooms have low solar gain with minimal window area, and U0.34 or lower windows.

    Start with a Manual-J load calculation. In zone 3a in a 1300' house your entire heat load is probably less than 25,000 BTU/hr (might be under 20,000 BTU/hr) and your cooling load is probably something like 10,000-18,000 BTU/hr, depending on how much west-facing window area you have

    If a room doesn't have a peak load of at least 4000-5000 BTU/hr it should not have it's own ductless head. The smallest ductless heads run 6000-7000 BTU/hr for multi-splits, but for single head mini-splits you're looking at 9000 BTU/hr, as the smallest size and would need a peak load greater than 6000 BTU/hr. The rated size should be no more than 50% oversized for the load (either cooling or heating) or the efficiency will begin to suffer. Given your likely heating & cooling loads you're probably looking at a MAXIMUM of 3 heads/cassettes before it starts sliding down the efficiency scale.

    But you can do a lot with three. I could heat my 1.5 story 3 bedroom bungalow in New England with three heads comfortably, with some judicioius selection and location of the interior units. But for now there is no real comfort or financial incentive to go there, having very low cooling loads and being on the gas grid.

    If the heat load of a given room is no more 1500 BTU/hr at the 99% outside design temp it's pretty easy to get by even without auxiliary heating sources in that room. If the load is more than that you may need to have some baseboards or something to stay warm during the polar vortex coolth, but a mini-ducted solution could still work.

    It's pretty easy to split the output of mini-duct cassettes between two adjacent bedrooms, especially if there are closets between the rooms, where you can mount the cassette at the top of the closet (or between the joists, if they're running the right direction.)

    For simple floor plans the Mitsubishi SEZ./SUZ series single mini-duct solution has efficiency advantages over multi-splits, since the control algorithms are optimized between the compressor end and the cassette end. With multi-splits that take multiple and often different types of cassettes/heads it's not possible to fully optimize everything.

    Any home in sticky-humid zone 3A would do well to convert the crawlspace into a sealed-conditioned crawlspace, but if you continue to use the ducts that's even more important. With the crawlspace sealed and insulated the efficiency hit from duct leakage drops an order of magnitude, and it may be easier to insulated and seal the crawlspace well than get near to perfection sealing and re-insulating the existing ducts.

  7. SteveR_79 | | #7

    @Dana – Thanks for the detailed response. That’s very helpful.
    Your suggested loads are amazingly close. The manual J calculations for my home are in the works by a contractor. I made some calculations using three methods, the old ASHRAE method, measuring the on/off cycle time of the furnace’s gas solenoid valve and lastly using the free online program They all are in the same range. The loadcalc program gave 21,289 Btu for heating and 20,249 Btu for cooling (17,499 – sensible and 2750 – latent). The design indoor heating temp was 72 and outdoor temp was 26. The design indoor cooling temp was 75 and the outdoor temp was 91. The current furnace is rated at 80,000 Btuh and 80% efficient. It was installed in 1985. The a/c is a Carrier 2.5 ton condenser and a 2 ton coil. It was installed in 1981.

    Before the rehab of the house, the furnace would run almost constantly when the outside temp was in the low 20’s. The a/c would run constantly when outside temp was in the mid 90s. The rehab included new drywall on ceilings and walls with the intent of making it airtight. Fiberglass wall insulation was added with care based on Martin Holladay’s installation articles. The single pane aluminum frame windows were replaced with windows having a U-factor of 0.30 and a SHGC of 0.30. (There is about 12 sq. ft. of west facing windows with a 1 ft of overhang.) After the rehab, in the winter when the temp is in the low 20s the furnace runs about 30% of the time. In the summer when the temp is in the 90s the a/c actually cycles.

    You are currently heating your New England home with gas and have no a/c. But am I correct that at my current gas and electric rates (1.20 $/therm, 0.13 $/kWh) a minisplit at 9.6 HSPF and a furnace at 90% AFUE have a similar cost per Btu? If my calculations are off please tell me. The problem I see with these comparisons is trusting the HSPF and AFUE ratings.

    I am not determined to install a minisplit because it is the latest and greatest. My goal is to install the most cost effective solution whether that is a minisplit or conventional furnace and a/c. To heat my home last winter required about 400 therms. By my calculations the savings from a 96% AFUE furnace over a 90% AFUE would be $26 per year. Unless the 96% were a few hundred dollars more than the 90% it doesn’t seem worth it at current natural gas prices. Of course I can’t predict the future of natural gas prices.

    Your input about sealing the crawl space is spot on. That is in the plan after the HVAC is installed. If I replace the furnace and re-work the ducts, there will be a lot of commotion in the crawl space.

  8. Expert Member
    Dana Dorsett | | #8

    The HSPF numbers are a presumptive average efficiency at a somewhat cooler location than yours (but somewhat warmer than mine.) Something that tests at 9.6 (a COP of 2.81) will likely deliver performance a shade north of that in your climate, something like a COP of 3.0-3.15 (HSPF of 10.2- 10.7.), where it as it would deliver an average COP in the mid-2s near me.

    If you want to be conservative, stick to the HSPF number, and know if you don't oversize the equipment, it will beat that number.

    An HSPF of 9.6 means it delivers 9600 BTU/kwh. That means a million BTUs (MMBTU a handy unit for estimating heating loads) takes 1,000,000 / 9600= 104.2 kwh. If you're paying 13 cents/kwh that means it will cost about 104.2 x $0.13= $13.55/MMBTU.

    A 96% AFUE gas burner will deliver 96,000 BTU / therm into the heating duct. So it takes 1,000,000 / 96,000 = 10.42 therms per MMBTU. If you're paying $1.20 / therm, your heating fuel cost 10.42 x $1.20 = $12.55 /MMBTU.

    But to that $12.55 you have to add in the air handler power use, which at 13 cents/kwh likely adds at least another buck per MMBTU (it depends- some air handlers are much more efficient than others), and if the ducts go outside of the thermal & pressure boundary of the house, or aren't well balance on supply / return on a room-by-room basis, you could be sending 10% or more of the heat elsewhere. Even a pretty tight house will take a 5% hit due to system efficiency issues such as those, even with the ducts fully inside of conditioned space.

    So in the end a mini-split (or mini-duct) solution may be slightly cheaper to run than a condensing gas furnace, but not by a whole lot. The anticipated lifecycle of either will be about 20-25 years. The gas furnace will almost certainly be cheaper to install, especially if you are setting up mini-duct cassettes and a few short duct runs instead of 1-2 wall-coil units or using the pre-existing ducts with a 2-ton tradtional style heat pump.

    Got a good crystal ball?

    You'll need it to predict which method will be the most favorable financially over the next 25 years!

    Natural gas prices are currently near their inflation adjusted historical low, and electricity may be too. Natural gas prices are currently low only because of fracked-wells that also produce liquids (light oil, propane, butane, etc.) which fetch a high enough market price. To support that level of capital expeniture for dry-gas wells that produce no liquids would require a wellhead price north of $10/MMBTU (instead of the current $4), which would mean higher natural gas prices at the retail end too. The price of oil just bounce off $80, which is a price point at which many of the Bakken Shale production is no longer profitable. If the price of oil stays there or falls further, the price of natural gas is going to rise.

    A rising natural gas price also yields somewhat higher electricity prices, but probably not in the same proportion as retail natural gas.

    But the crashing cost of PV solar is defining the high-limit for retail electricity prices. At 2013 US installed prices the lifecycle cost of PV power (including financing), was slightly lower than your 13 cent electricity, but by 2020 it will be roughly half the cost. If you have the space for enough PV and the financial capacity to swing it, if you can net-meter at retail your 20 year cost of electricity can come in below the current 13 cents, more or less locking in your heating costs for the life of the heat pump. (In 2015 in a handful of states Solar City is going to be offering 30-year loans at 4-5% for their grid-tied PV systems for people with a sufficient FICO score. If that program takes off there will be competitors with similar deals.)

    But even though PV has been dropping for 9-10% per year for 30-40 years doesn't mean it will always hit those marks, especially if government & utility policy support is lacking. There are any number of ways it can come off the rails, but it probably isn't going to get more expensive than it is right now.

    If the anticipated tsunami of PV hooks up to the grid over the next decade electricity prices may initially rise slightly (to cover grid infrastructure aspects needed to manage it), but since it competes directly with fossil burners and has no marginal fuel costs, the wholesale price of electricity will fall, reducing the rates overall.

    There are lots of moving parts to it, but my gut tells me electricity will continue to get cheaper over time on average (as it has for 30-40 years, in inflation-adjusted terms) and natural gas pricing will always be volatile- the gaseous tail being wagged by the petro-dog. Natural gas prices are down right now, but they will likely to go up and down a number of times in the next 20 years, as the profitability of fracking for oil varies over time.

    I've been wrong on long term future energy pricing in the past, but so has everybody else who has given it a shot, eh?! What has changed is that the price tragectory on PV has even steepened as it crossed the grid-retail-parity line, and is by far the fastest "learning curve" of any new energy technology, dropping 22% in price every time the installed base of PV in the world doubles for 30+ years. The doubling rate has been shrinking now that it's directly competitive with grid-retail in many markets, now under two years. This isn't the same sort of deal as busting the OPEC cartel's grip in the 1980s or reaching such sky-high oil prices that tight-shale oil becomes profitable creating a natural gas glut. It's a whole 'nuther animal entirely: Scalable durable manufactured goods with a falling cost basis, more akin to the computer revolution than the shale revolution, where every time you have to replace your PC it costs slightly less, but has 5x the compute power. The learning rates aren't exactly the same as with computer tech stuff, but the logarithmic decay characteristic in bang/buck over time is something unmatched in most other energy technologies. Wind power is getting cheaper & better year on year too, but not as fast as PV, and only really works at the utility scale. But wind too is a long term damper on electricity price inflation.

    That's what MY fuzzy crystal ball is telling me, anyway. :-) (The investment banking world seems to agree too, but we know how good THEY are at screwing it up. Can I sell you a sub-prime mortgage derivative? :-) )

  9. SteveR_79 | | #9

    Thanks again Dana. It’s reassuring to know that my comparison of cost per Btu between the mini-split and the gas furnace is correct. As you noted Mitsubishi tests there minisplits with Zone 4 conditions which should improve the HSPF for my Zone 3 home.

    That was a great discussion on the oil, natural gas and electric prices. As for the future of energy prices, I don’t trust my gut. My gut told me that thermal coal production wouldn’t decline as fast as it did. My crystal ball doesn’t work. You are one step ahead of me with the PV system. As you say since I don’t know the future I can hedge my energy costs by installing PV. It is in the plan for 2015 and is also a reason I’m considering the all-electric HVAC system.

    When selecting the size of a minisplit when is a system oversized? If Manual J gave a heating load of 24,000 Btuh what size minisplit should be selected? Is it the obvious answer of 2 tons? The Mitsubishi MXZ-3B24NA-1 has a rated heat capacity of 25,000/24,600 Btuh at 47 F (non-ducted/ducted). At 17 F that drops to 14,000/14,000. The max capacity at 17 F is 18,800/17,000. My design outdoor temp is 26 F. The footnote for the two ton multizone heat pump states that the load is based on "Data from combination of two Indoor Units 6,000 Btu/h and one 9,000 Btu/h (non-ducted) or three 9,000 Btu/h (ducted)."

    What's the correct way to size this system?

  10. Expert Member
    Dana Dorsett | | #10

    To size a mini-split correctly you need to know it's output at YOUR 99% outside design temp. Most manufacturers will give you at least a few points along the derating curve, but sometimes you have to look up (or request ) "extended temperature" capacity tables if it doesn't appear in the other documentation.

    The head/cassette and compressor combinations will affect capacity somewhat too, but if you can find a "submittal" sheet for the combination you intend to use, that is the verified-tested setup by which they tested it for efficiency, and it will have capacity information at a few datapoints too.

    As you point out, a non-mini-ducted -3B24NA-1's ocan only deliver ~19KBTU/hr @ +17, even though it can deliver ~25K @47F: A mini-ducted solution will deliver at least 17,000 BTU/hr @ 17F:

    In the absence of a capacities table for your exact combination it's reasonable to use a linear interpolation to estimate what it's output would be at your design temp of 26F. Take a combination that uses no mini-ducts as an example.

    In the 30F difference between 47F and 17F it's capacity drops by about (25K-19K=) 6,000 BTU/hr. If you made a line between those two data points on a graph it has a slope of about 200 BTU/hr for every degree-F. So, at a difference of (26F-17F=)9F you're looking at a capacity difference of 9F x 200 BTU/F= ~1800 BTU/hr. So if you add that 1800 BTU/hr to the rated 18.800 BTU/hr it comes to 20,900 BTU/hr.

    The interpolation slope number is still about 200 BTU/F when using multi-mini-duct solutions, but you'll be looking at about 20K instead of about 21K.

    The non-ducted solution is just shy of your Manual-J estimate of 21,289 BTU/hr by about the heat output of two sleeping humans. Given that most Manual-J calculations overshoot by more than 10% even the mini-ducted solution should cover you until/unless the coldest night of the year also corresponed with a 40mph gale. You will probably be just fine with that setup, knowing that during a Polar Vortex event you might have to leave some lights on or let the dog cuddle up to you a bit closer to hold the line at whatever your indoor design temperature was. (The rated output assume 70F for an indoor design temp, if you read the fine print.)

    If that's a bit too close for (psychological)comfort it wouldn't be an efficiency disaster to go with a 3 ton (and 3 or more tons of input head) but beyond that you might start slipping a bit on efficiency. during the shoulder season due to oversizing.

  11. SteveR_79 | | #11

    Mitsubishi has a de-rating curve or a heating capacity correction factor on p.26 of the pdf in the link below. This is for the multi-zone minisplit heat pumps with nominal ratings of 20, 24, 30 and 36 kBtuh.

    If I understand the curve correctly, the defrost operation occurs when the outdoor wet-bulb temperature is less than about 40 F. Because of the defrost cycle the correction factor curve has a step at 40 F and drops from ~0.97 to ~0.77. This assumes an indoor intake air dry-bulb temperature of 68 F. At 14 F, the correction factor is ~0.59. The correction factor for my outdoor design temperature of 26 F is ~0.68.

    The correction factor is at the rated frequency but it's a bit unclear what the rated capacity is. I assumed it was the rated capacity at 47 F since that is provided. But I expected the correction factor at 47 F to be 1.0 and instead it is ~1.08. In the end I'm fine with approximating a correction factor from this curve and estimating the maximum heat at 26 F. That's assuming that I have not misunderstood the intent of the curve.

    I have talked to a couple of HVAC contractors with minisplit experience. They have been upfront that their minisplit experience installations have been single-zone units for sun rooms. They have no experience with a whole house application. They are leaning on their distributors to help plan the installation. I'm guessing the distributors have no experience with whole house installations either. The preliminary designs involve at least 5 heads using either the MXZ-5B42 or MXZ-8B48 outdoor units. It's well known that HVAC contractors hate to undersize systems, and this is no exception.

    Nick Welch commented that a 1.5 ton single head unit almost heats his 900 sq. ft. house in Zone 4c. Dana commented that "with some judicious selection and location of the interior units" he could heat his 3 bedroom bungalow in New England with three heads. With that it should be possible to heat (and cool) my 1300 sq. ft. home in zone 3a with three heads. All I need now is Dana's judicious eye. Is there an equivalent to Manual D for minisplits?

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |