musingsheader image
1 Helpful?

Just Two Minisplits Heat and Cool the Whole House

Carter Scott has built 18 homes in Massachusetts without any heat in the bedrooms

Posted on Aug 17 2012 by user-756436

Carter Scott was one of the first builders bold enough to build a cold-climate home heated by only two ductless minisplit units (one in the downstairs living room, and one in the upstairs hallway). Skeptics predicted that the unheated bedrooms would be cold and uncomfortable. Yet Scott was confident that the home’s excellent thermal envelope — with high-R walls, triple-glazed windows, and low levels of air leakage — would keep the homeowners comfortable even when the bedroom doors were closed.

Scott owns a construction company called Transformations in Townsend, Massachusetts. He built his pioneering two-minisplit house in Townsend in 2008; the inclusion of a 5.7-kW roof-mounted photovoltaic(PV) Generation of electricity directly from sunlight. A photovoltaic cell has no moving parts; electrons are energized by sunlight and result in current flow. array made it into a zero-energy house.

The skeptics’ “cold bedroom” predictions were unfounded. “We have since built several houses in which the upstairs minisplit unit isn’t even being used until the outdoor temperature drops below 20 degrees,” Scott said. “Typically the response from homeowners is, ‘Wow, these houses have even indoor temperatures’ and ‘these houses are quiet.’ And the fact that there are no utility bill makes people excited.”

Carter Scott's first net-zero energy house

A ductless minisplit is a type of air-source heat pumpHeat pump that relies on outside air as the heat source and heat sink; not as effective in cold climates as ground-source heat pumps. that can provide space heating as well as air conditioning. Most of the ductless minisplits sold in the U.S. are manufactured in Japan or South Korea; the best known brands are Daikin, Fujitsu, and Mitsubishi.

The first zero-energy house built by Scott's company, Transformations, was designed by architect Ben Nickerson. Dubbed the Needham model, it was built at 18 Coppersmith Way in Townsend in 2008. Here are the home's specifications:

  • Area: 1,232 square feet
  • Basement insulation: Basement is unconditioned; basement ceiling is insulated with 3" of closed cell spray foam plus R-30 fiberglass batts (total R-50)
  • Wall framing: 12-inch-thick double-stud walls
  • Wall insulation: Flash and fill: 3" closed-cell spray foam plus 9" cellulose (R-50)
  • Sloped ceiling insulation: Flash and fill: 5" closed-cell spray foam plus 13" cellulose (R-64)
  • Windows: Paradigm triple-pane kryptonA colorless, odorless inert gas, often used with argon in fluorescent lighting and sometimes used as gas fill in high-performance glazing.-filled low-eLow-emissivity coating. Very thin metallic coating on glass or plastic window glazing that permits most of the sun’s short-wave (light) radiation to enter, while blocking up to 90% of the long-wave (heat) radiation. Low-e coatings boost a window’s R-value and reduce its U-factor. windows
  • Siding: VinylCommon term for polyvinyl chloride (PVC). In chemistry, vinyl refers to a carbon-and-hydrogen group (H2C=CH–) that attaches to another functional group, such as chlorine (vinyl chloride) or acetate (vinyl acetate).
  • Design heat load: 10,500 BtuBritish thermal unit, the amount of heat required to raise one pound of water (about a pint) one degree Fahrenheit in temperature—about the heat content of one wooden kitchen match. One Btu is equivalent to 0.293 watt-hours or 1,055 joules. /h
  • Space heating: Two Mitsubishi Mr. Slim ductless minisplit units (one 12,000 Btuh unit downstairs, and one 9,000 Btuh unit upstairs); installed cost, $5,250
  • Mechanical ventilation: Lifebreath 155 ECM energy-recovery ventilator
  • Domestic hot water: Sun Drum solar thermal system with electric resistance backup
  • PVPhotovoltaics. Generation of electricity directly from sunlight. A photovoltaic (PV) cell has no moving parts; electrons are energized by sunlight and result in current flow. system: 5.7-kW roof-mounted array (Evergreen Solar PV modules) and Fronius IG 5100 inverterDevice for converting direct-current (DC) electricity into the alternating-current (AC) form required for most home uses; necessary if home-generated electricity is to be fed into the electric grid through net-metering arrangements. (cost before incentives: $33,000)

In 2009, this home won the second prize ($15,000) in a utility-sponsored contest called the Zero Energy Challenge. The home has since received plenty of media attention, including attention from GBA; see, for example:

Consultants from Building Science Corporation help refine the specifications

Scott gave a presentation on his recent projects at the Sixteenth Annual Westford Symposium on Building Science in Westford, Massachusetts, on July 30, 2012. Scott is currently working with the DOE’s Building America program, and the program hooked him up with energy consultants from the Building Science Corporation. He’s now building homes at several locations around Massachusetts.

In Devens, Massachusetts, Scott will build 8 superinsulated single-family homes with prices starting at about $330,000 (not including PV). In Easthampton, Scott is building 33 new superinsulated homes at a development called Easthampton Meadow. The homes will range in size from two-bedroom homes to four-bedroom homes, and will be priced at $284,000 to $330,000 (not including PV).

The Easthampton homes will have all of Scott’s usual features — thick R-40 walls, R-60 ceilings, triple-glazed windows, ductless minisplit heating systems — bringing the homes to HERSIndex or scoring system for energy efficiency established by the Residential Energy Services Network (RESNET) that compares a given home to a Home Energy Rating System (HERS) Reference Home based on the 2006 International Energy Conservation Code. A home matching the reference home has a HERS Index of 100. The lower a home’s HERS Index, the more energy efficient it is. A typical existing home has a HERS Index of 130; a net zero energy home has a HERS Index of 0. Older versions of the HERS index were based on a scale that was largely just the opposite in structure--a HERS rating of 100 represented a net zero energy home, while the reference home had a score of 80. There are issues that complicate converting old to new or new to old scores, but the basic formula is: New HERS index = (100 - Old HERS score) * 5. 40 without PV. If a homeowner wants to buy a PV system — an available option — any of the homes can reach HERS 0 (or less).

When the temperature is below zero, Mitsubishi ductless minsplits still perform

For Carter Scott, heating a house measuring 1,700 to 2,000 square feet with two ductless minisplits is no longer an experimental method. It’s standard operating procedure — one he’s used on 18 houses.

Although he has considered using Fujitsu or Daikin minisplit units, Scott continues to specify units from Mitsubishi. He usually specifies the MUZ-FE12NA outdoor unit and the MSZ-FE12NA indoor unit; this pairing is rated at 12,000 Btu/h. Even at an outdoor temperature of 0°F, these units can put out 10,000 Btu/h of space heat at a COP of 1.8. Mitsubishi minisplits will still deliver heat when the outdoor temperature drops to -13°F.

Scott pays his HVAC(Heating, ventilation, and air conditioning). Collectively, the mechanical systems that heat, ventilate, and cool a building. contractor less than $6,000 to install two minisplit units in one of his homes.

After Scott heard that the owner of a home he had built (a custom home in Princeton, Massachusetts) had turned off the upstairs heating unit because the downstairs unit adequately heated the whole house, he decided to try an experiment. At his next house, Scott installed just one ductless minisplit. “We thought that maybe we could just use one unit downstairs. But I didn’t think of cooling. The cooling didn’t rise to the second floor, and the house was hot upstairs in the summer. We went back to the house and installed a minisplit unit upstairs. So if you want AC on the second floor, you need an AC unit up there.”

Since he learned his lesson, Scott always includes a minisplit unit on each floor of his two-story homes.

Anyone interested in following in Scott's footsteps needs to remember that the success of his two-minisplit approach depends in part on compact rectangular designs. Stretched-out houses on a single floor, designs with ells, or designs that include a bonus room over the garage aren't amenable to the two-minisplit solution.

On the other end of the spectrum are the one-minisplit homes; examples include Marc Rosenbaum's house in Massachusetts, the Montague Urban Homestead (also in Massachusetts), Larry and Jill Burks' Up Hill House in New York state, and John Semmelhack's house in Charlottesville, Virginia (see Comment #15, below). The best candidates for this approach are flexible homeowners who leave their bedroom doors open during the day and who have no need for second-floor air conditioning.

What about ventilation?

Carter Scott's first zero-energy house in Townsend included a Lifebreath energy recovery ventilator — an effective but costly ventilation solution.

As a builder of spec homes on his way to becoming a production builder, Scott has always focused on affordability. That's why the standard ventilation system included with his homes is often a simple exhaust-only system. However, he offers a variety of ventilation options to customers who are willing to pay extra.

“The baseline ventilation system we offer is a Panasonic exhaust fan in each bathroom,” said Scott. “They cost us $250 each installed. The next step up — an available option — would be a Panasonic ERV(ERV). The part of a balanced ventilation system that captures water vapor and heat from one airstream to condition another. In cold climates, water vapor captured from the outgoing airstream by ERVs can humidify incoming air. In hot-humid climates, ERVs can help maintain (but not reduce) the interior relative humidity as outside air is conditioned by the ERV. [energy-recovery ventilator] that exhausts and supplies from the same location. That’s $500 installed. A little better would be the Fantech VHR 704 HRV with one exhaust location and one supply location. Better than that would be the Fantech SHR 1504 HRV, which could exhaust three bathrooms and supply all of the bedrooms. The next bump up would be the Lifebreath ERVEnergy-recovery ventilator. The part of a balanced ventilation system that captures water vapor and heat from one airstream to condition another. In cold climates, water vapor captured from the outgoing airstream by ERVs can humidify incoming air. In hot-humid climates, ERVs can help maintain (but not reduce) the interior relative humidity as outside air is conditioned by the ERV., which is about $2,500.”

Subsidies, incentives, and tax credits make PV systems affordable

Before incentives or tax credits, Scott’s PV systems cost him about $5 a watt to install. However, all of his customers benefit from at least one type — and in many cases, several types — of tax credit or subsidy. Homeowners who aren't interested in owning a PV array can also take advantage of a variety of leasing options.

In addition to the 30% federal tax credit on photovoltaic installations available to all U.S. taxpayers, Massachusetts homeowners with PV systems can take advantage of Solar Renewable Energy Credits, or SRECs. The mechanics of the SREC market are complicated, and the value of these credits fluctuates. According to Scott, until recently the SREC market valued residential PV production at 50 cents per kWh — meaning that PV system owners could recoup their investment extremely quickly. The market value of SRECs for residential PV systems is now down to only 28 cents per kWh; although that number is lower than it was in the past, it still provides a quick payback to owners of PV systems — often ten years or less.

The Energy StarLabeling system sponsored by the Environmental Protection Agency and the US Department of Energy for labeling the most energy-efficient products on the market; applies to a wide range of products, from computers and office equipment to refrigerators and air conditioners. program in Massachusetts also offers generous incentives to builders like Scott who achieve the Tier 3 level of performance; Scott receives an $8,000 cash incentive for each high-performance house he builds.

“Solar systems are very cost-effective now,” says Scott. “When it comes to maintenance, PV is pretty simple. Solar thermal is more problematic than PV.”

Propane is sometimes cheaper than natural gas

At one of his zero-energy homes in Townsend, Scott installed an on-demand natural gas water heater; it’s the only gas appliance in the house. During their first year in the house, the homeowners only used $50 of natural gas, but were billed $152 — $50 for the gas, and $102 for the service charge of $8.50 a month.

Once Scott realized that the homeowners were paying more for the service charge than they were for fuel, he concluded, “This dog don’t hunt.” He now avoids hooking up his homes to natural gas lines; he’s concluded that most homeowners with a single gas appliance are probably better off buying propane. Even though propane costs more per gallon, propane dealers don’t charge a monthly service fee.

Letting go of solar thermal

As Scott continues to aim for high performance at the lowest possible cost, he has tweaked a few of his specifications. These days, Scott is installing more open-cell spray foam in his walls, and less cellulose, for two reasons: the spray foam provides a better air seal, and his insulation contractor offers open-cell foam for the same price as cellulose.

To reduce costs, he has switched from triple-glazed vinyl windows from Paradigm to triple-glazed vinyl windows from Harvey. When I asked Scott about the quality of Harvey windows, his answer made it clear that the windows’ best feature was the low price. “The quality of Harvey triple-pane windows is improving,” Scott told me. “They now come with a hole midway up the jambs, which we pin to the studs. This keeps them opening and closing above 85 degrees. We are also careful not to put more than 1/2 inch of low-expanding foam to seal around the windows. Harvey will sell me an R-5 window at an unbeatable price. In quantity, I’m getting the windows for just a few dollars more than our old double-pane windows.”

In the past, Scott tried a variety of approaches for heating domestic hot water, including solar hot water systems. His favorite water heater is now the Navien 180 on-demand tankless unit fueled by propane or natural gas. “An $8,500 solar hot water system doesn't make any sense compared to the $1,800 Navien. The fuel use cost with the Navien is so low that there really isn't much money to be saved doing anything different.”

Ground-source heat pumps are too expensive

Before he discovered ductless minisplit units, Scott built three homes with ground-source heat pumps (GSHPs). Now that he knows about minisplits, however, he has no intention of installing another GSHP.

When I asked him why, his response was simple. “The first ground-source heat pump system I installed cost $38,000. The second one cost $40,000. The third one had a direct-exchange loop and cost me $22,000,” he said. “But I can install two minisplits for less than $6,000. And the whole system efficiencies are about the same, as far as I can tell. Even if the ground-source units have a slightly higher COP, it's not enough to warrant the extra money.”

Last week’s blog: “Living Without Electricity Bills.”

Click here to follow Martin Holladay on Twitter.

Tags: , , , , , ,

Image Credits:

  1. All photos: R. Carter Scott
  2. Fine Homebuilding

Aug 17, 2012 5:58 AM ET

Edited Aug 17, 2012 6:00 AM ET.

Time for Micro & Nano-Splits
by homedesign

The Mini-split concept makes a lot of sense.
Moving heat in and out thru small copper pipes instead of ginourmous ducts.
The problem is the smallest available system size.
I know there are multi-head systems (like Daikin) out there ...but the piping can get pretty complex running back and forth to a central unit.....and also the smallest heads are not-so-small.

What I would like to see are Micro-Split Systems (or Nano) small enough to handle an average size bedroom .. with a small inside head, a small outside unit and short run of tubing.

Similar to task lighting...task heating and cooling.

Aug 17, 2012 6:06 AM ET

Response to John Brooks
by user-756436

Perhaps you're right that we need smaller units -- and perhaps the manufacturers will respond.

However, it's worth pointing out that:

1. Carter Scott's clients aren't complaining, even with no indoor units in the bedrooms.

2. Oversizing is advantageous with ductless minisplits, because the oversized units operate at a higher efficiency than right-sized units.

Aug 17, 2012 6:15 AM ET

Edited Aug 17, 2012 6:17 AM ET.

by homedesign

and Kudos to Carter for concentrating his research and development on Affordable & Buildable

Most of the Passivhaus and even the Building America stratagems are Not-So-Affordable and Not-So-Buildable

We need Stratagems for Everyman & Woman

Aug 17, 2012 6:22 AM ET

Response to Martin
by homedesign

Most homes are not as simple as Carter's and many people like to shut their bedroom doors at night and for privacy.

Aug 17, 2012 6:38 AM ET

Edited Aug 17, 2012 6:39 AM ET.

Bedroom doors
by user-756436

Carter Scott is a production builder. He is not pre-selecting his customers. He is building entire neighborhoods of houses for any interested purchaser.

I can assure you that the homeowners are shutting their bedroom doors at night. Probably not during the day, but definitely at night.

No builder wants to take a risk that home buyers will be uncomfortable or unsatisfied. Everyone hates callbacks. Carter Scott knows that these homes work, and he's willing to sell to all comers.

Aug 17, 2012 7:08 AM ET

Single mini downstairs
by UphillLarry

Add us (Up Hill House) to the list of houses with a single mini downstairs. Yes, we're a simple rectangular shape, but we call that modern. Yes it gets hot upstairs in the summer, but not unbearably and not more than handful of days in the summer, ceiling fan works wonders. And yes, it was perfectly warm upstairs in the winter with single unit, although it was not a particularly cold winter. We made the right decision for us, but I can see how it's not for everyone.

Aug 17, 2012 7:39 AM ET

Comfort is subjective
by Marc Rosenbaum

Carter's work is terrific and inspiring. Before others take his one-up, one down approach and apply it wholesale to their own work, it's worth pointing out that in addition to compact form, one aspect that keeps temperatures even in remote spaces is amount of glass - window area matters on cold nights with doors closed. Carter's home shown in the photos has very modest window area, especially in the second floor rooms, and that's a component in the success.

Aug 17, 2012 7:44 AM ET

Response to Larry Burks
by user-756436

Thanks for your post. I've added your house to the list of successful one-minisplit homes.

Aug 17, 2012 7:49 AM ET

Response to Marc Rosenbaum
by user-756436

Thanks for your comment; I value your input, and your point about window area is well taken. Designers always need to approach heating system design with whole-house thinking. There's no such thing as a cookie-cutter approach that will work with all homes.

Many of us have been holding our breath and crossing our fingers as we watch Carter Scott building more and more two-minisplit homes, wondering whether he will eventually encounter a grumpy homeowner (or a grumpy building inspector) who lands him in hot water. So far, so good. We are all gaining useful information from pioneering superinsulated homes like those built by Carter Scott.

Aug 17, 2012 10:18 AM ET

Edited Aug 17, 2012 10:59 AM ET.

Similar approach
by user-1044812

We built a smaller 2 story home for a local land trust that arranges permanently affordable housing. For a number of reasons, we did not install a HRV, but did use a 2 head minisplit - with a twist - because we were not confident that both upstairs bedrooms would be comfortable, we installed a ceiling cassette head unit that is ducted. We placed it in the lid in a back to back closet between the 2 bedrooms, making the duct runs super short. Our heating load was low enough that a single head unit would have been enough, and thought that using an HRV would help distribute the heat - but convincing local officials was another matter. In the end the 2 head minisplit plus electric baseboard in the bathrooms (hopefully never to be used) was what we installed.

I look forward to more experiences like Carter Scott's so that we can forge ahead with smaller, more efficient systems. Of course it all starts with a good design and a great envelope....

Aug 17, 2012 10:52 AM ET

List of successful one-mini-split homes
by homedesign

do you have a link to "the List"
Are there any from Mixed-Humid or Hot-Humid ?
I know there is Pasivhaus example in Louisiana ... but I read somewhere that it may be having comfort/performance problems.

Aug 17, 2012 11:38 AM ET

Response to John Brooks
by user-756436

When I referred to "the list," I was just referring to the last paragraph of the section of this blog under the subhead "When the temperature is below zero, Mitsubishi ductless minsplits still perform."

Clearly, it's not a very long list.

I invite any GBA readers with information on successful one-minisplit homes to add to the list.

Aug 17, 2012 12:41 PM ET

size of that rectangle
by duluthdesigner

Can you provide the size of the house's footprint (not including garage of course)? Is there a bedroom on the first floor or only on second floors? Great work, and timely. We've just sent out the construction docs for a similarly designed home to be built in Iowa, and we've specified mini-splits for each floor (including the basement).

Aug 17, 2012 12:52 PM ET

Edited Aug 17, 2012 1:01 PM ET.

Response to Rachel Wagner
by user-756436

The article discusses several homes. The very first zero-energy home built by Carter Scott in 2008 was a model called the Needham. It has a master bedroom downstairs and two bedrooms upstairs.

The Needham is 1,232 square feet. Its footprint is 24' by 32'.

The Farmhouse model is 24' by 36'.

Aug 17, 2012 2:24 PM ET

Wall mount mini-splits in central Virginia
by John Semmelhack


We've been pretty successful so far here in Central Virginia with wall-mount mini-splits for both heating and cooling, and we've had VERY hot summers in 2010, 2011 and most of 2012 so far. I'll outline them briefly here (one is my own house and I consulted on the other two), but they all share many of the key attributes of Carter Scott's houses: compact shape, extremely well insulated, very air-tight, and all have ducted ERV systems.

My own two-story house in Charlottesville, VA (finished in 2008) has a single wall-mount mini-split (Mitsubishi 12kBtu) located on the first floor. Though we are certainly under shaded in the summer (I'll get around to remedying this one of these days), we typically experience temperature stratification of just 3-4 degrees in summer and 1-2 degrees in winter between the floors. We have 3 bedrooms upstairs, 2 of which are closed at night (children sleeping). We also benefit in the summer from a double-height, lofted space above the living room that the master bedroom overlooks. We run a high-efficiency ceiling fan on low-ish speed in this double-height space during the summer to help improve mixing between the floors. Finally, each bedroom also has a high-efficiency ceiling fan for improved night-time comfort. We use about 7,000kWh per year (all-electric house) and have been well into net-positive energy territory since we installed a 6.2kW PV system in April.

The next house is the Lankford (certified) Passive House, built by Jobes Builders (, also in Charlottesville. This one is a little different, in that it has about 2,100ft2 total on three floors and two wall-mount units (connected to a single outdoor unit). We installed one unit on the top floor and another unit on the lower floor, nothing in the middle. While the house is not occupied (it's for sale - we measured temperature difference between various rooms (doors open) and floors at just 1F on multiple 100F days last summer.

The last house is Daniel Ernst's recently finished (not yet certified) Passive House in Steeles Tavern, VA, and it may be the cream of this crop. Daniel has a main 2-story house with a mother-in-law wing connected to the main house on the lower floor. Total floor area is about 2,500ft2. While the mother-in-law wing has it's own wall mount unit, the 2-story main house has just a single wall-mount mini-split (Mitsubishi 9kBtu) located in a small double-height space + stairway. Temperature stratification throughout this summer has been no more than 3-4F between floors, according to Daniel.

Hopefully, we’ll get more information on all of these houses up on GBA sometime soon!

Aug 17, 2012 2:26 PM ET

The nice thing about superinsulated houses is...
by user-1004076

... the time constant created by the stored heat in the interior building materials and the R-value of the exterior surfaces is sufficiently long that the mini-split need not carry the whole load at the temperature extremes to keep it comfortable.

And that's a good thing, since there will be many times during the course of a winter that the outdoor dew point is close enough to the air temp that a mini-split will be spending as much time in defrost mode than actually heating, for a severe hit in output capacity (but not necessarily efficiency) during those critical hours. If you read the fine print next to the asterisks in the specs you'll find that even at a balmy +17F the Mitsubishi's are typically rated at only ~60% their nominal output when the outdoor dew point is as high a +10F, and if the dew point was even higher, say +13F (as often happens in MA) the output capacity would quickly fall to below half.

Sure, in arctic dry air they can deliver the full rated BTUs even at 0F, but the air in MA isn't usually that dry- it's an issue, but only for homes with short time constants.

The 99% outside design temp for Townsend MA (Carter's first Zero Energy house) is about 0F, so there will definitely be times when there is a BTU shortfall below the instantaneous heat load from the mini-splits, but with R50 walls the place could coast for hours at 0F without losing much ground, even if the mini-split was only delivering a third or half the load during that period. And as soon as the air temp rises well-above the dew point the full capacity resumes, making up for the shortfall that occurred during the coldest overnight hours.

With less-well insulated houses with much shorter time constants, a significant capacity shortfall at or near the outside design temp quickly becomes comfort problem.

It's usually better to spend the ~$35K difference in heating system cost between ground source heat pumps & mini-splits on building out a higher-R/lower-loss building envelope, since it's more comfortable, uses less power, and those low temp dew-point & capacity shortcomings of mini-splits that would otherwise be problematic become a non-issue.

[Editor's note: For a further discussion of this issue, see information provided by Marc Rosenbaum in Comment #66 on 8/24/2012.]

Aug 17, 2012 2:49 PM ET

Response to John Semmelhack
by user-756436

Thanks for sharing your experience specifying ductless minisplits in Virginia. I have added your house to the list.

Aug 17, 2012 4:38 PM ET

by James Steel

Great article! It's nice to see a model that's affordable and reproducible. But I'm very curious what Carter Scott is using for ventilation in these tight houses. HRV/ERV? Small duct system to each bedroom? The $6000 minisplit system is encouraging. What's his take on an affordable ventilation system for the spec market?

Aug 17, 2012 4:41 PM ET

scott homes recently built a
by user-945928

scott homes recently built a house outside of Olympia WA that isn't PH (probably not far off, actually) that i believe has a single DHP.

one of the PHs we're consulting on in seattle is a 1,500 sf home that recently broke ground - planning a single DHP as well. the house R-values are about the same as the carter scott NZ home (R-45 wall, R-58 roof, R-34 slab). surprisingly, it performs better w/ same specs if it were located in boston (even though 25% more HDDs than here) - meaning could probably get away w/ less glass to bring cost down slightly. the upgrade to intus windows and sub-slab foam (4" more than they normally build) end up being the only big difference in terms of their typical envelope.

Aug 17, 2012 4:48 PM ET

Great blog, great posts. Love
by wjrobinson

Great blog, great posts. Love to see this hit a hundred useful posts. Personally would like to see inexpensive storage of the work Splits do developed.

Aug 17, 2012 5:24 PM ET

Edited Aug 19, 2012 4:32 AM ET.

Ventilation options for James Steel
by user-756436

Good question. I have edited the blog to add information on ventilation.

Carter Scott lets his customers decide what type of ventilation system they want. He said, “The baseline ventilation system we offer is a Panasonic exhaust fan in each bathroom. They cost us $250 each installed. The next step up -- an available option -- would be a Panasonic ERV that exhausts and supplies from the same location. That’s $500 installed. A little better would be the Fantech VHR 704 HRV with one exhaust location and one supply location. Better than that would be the Fantech SHR 1504 HRV, which could exhaust three bathrooms and supply all of the bedrooms. The next bump up would be the Lifebreath ERV, which is about $2,500.”

Aug 17, 2012 5:58 PM ET

Edited Aug 17, 2012 6:16 PM ET.

Great article Martin! We
by user-600754

Great article Martin! We have built 8 super insulated homes in CT utilizing DHP's and have had great results with them. I would agree with you on most of your points except for your comment on tempertaure stratification of single level homes. I also used to think that was the case until we built a 2400 sq/ft single level home with 2 mini splits. Its temperature stratification was amazingly low; as good if not better than the 1600-2000 sq/ft two story homes we had built prior. While we dont have any hard data from monitoring rooms temps on an hourly basis, the customers have been very happy, which is really what maters the most. This home has R45 Walls, R80 Flat Ceiling, R45 framed floors, a high efficiency ERV, and infiltration rate of 1.1 ACH50. While a two story home typically has a better surface/ volume ratio, it proves that you dont have to stick to a square box in order for a super insulated home w/ DHP's to work.

We currently have two homes entered into the 2012 Connecticut Zero Energy Challenge that utilize DHP's, you can follow their progress here:

Aug 17, 2012 7:48 PM ET

Oversized minisplits?
by user-984364

Martin, can you say more about the advantages of oversizing? I put in a 1-ton single head at the top of the stairs to keep our 2 hot weeks liveable. I sized it at what the heat load calc said, even though I knew it wasn't optimal distribution. I didn't want to go bigger, because the smaller units had higher SEER ratings. What's this about more efficiency from a larger unit?


Aug 17, 2012 11:14 PM ET

Oversize OK?
by user-1087436

Like Eric Sandeen, I'm surprised at Martin's remark about oversizing. On 8/1, in a separate thread, Keith Gustafson replied to a reader who wanted to put 7 mini-splits in a house:

"Minisplits 'think' a lot, and they never shut off. They want to run at a very low level to be efficient. This means that if they are oversized for the space[as any 9k unit is for a normal bedroom] they are blowing cold air on you when you no longer want it. So you turn it off. then it is stuffy in the morning. Having a more central unit and making it work for a living is better for comfort, especially if you have some means of air exchange room to room."

He's talking about cooling, obviously. But does this apply to heating as well?

Wonderful article and postings, by the way. I'll definitely be printing this out.

Aug 18, 2012 4:45 AM ET

Edited Aug 20, 2012 4:23 AM ET.

On the efficiency of oversized minisplits
by user-756436

Eric Sandeen and Gordon Taylor,
I'll explain the efficiency story to the best of my knowledge. These inverter-driven ductless minsiplits have variable-speed compressors and variable-speed blowers -- features lacking on most conventional heat pumps and air conditioners. These features allow the ductless units to provide heating and cooling at part-load conditions -- more efficiently, in fact, than under peak load conditions. My understanding is that peak efficiency occurs when the compressor is operating at 1/3 of peak load speed.

This applies to cooling performance as well as heating performance, as far as I know. According to this document, "MSHPs [mini-split heat pumps] can reduce the compressor speed to meet low cooling loads and have higher COPs under such conditions."

Check out the graph below. It comes from page 10 of the document linked to above. Notice the COPs when the outdoor temperature is at 27 degrees F and 36 degrees F: this unit has the highest COP when operating at low fan speed, and it has the lowest COP when operating at high fan speed.

Ductless minisplit COPs at part-load conditions.jpg

Aug 18, 2012 7:34 AM ET

Edited Aug 18, 2012 7:37 AM ET.

Oversizing is GOOD with Mini-splits?
by homedesign

I have only glanced at the document you referenced.
Are you saying that with Minisplits.......
other than the initial cost...
Oversizing is "better" than Right-sizing?

'Course I am wearing my Hot-Humid Glasses....
and the first thing I think of is short cycling...
and less water removal

Aug 18, 2012 8:12 AM ET

by JoeW519

RE: conditiioned air distribution w/min-splits -- have y'all seen Allison's latest blog?

He also has a webinar on min-splits coming up

Aug 18, 2012 1:48 PM ET

Response to John Brooks
by user-756436

Yes: with minisplits, oversizing is better than right-sizing -- especially if energy efficiency makes you smile. Oversized units running under part load conditions will have the highest COP.

An oversized unit is also more likely to give you the heating output you need at low outdoor temperatures. Of course, if you are sizing a minsiplit for a house in a cold climate, you always want to look at the unit's output at the design temperature.

If you are a production builder, however, you may not see any advantages to oversizing.

Aug 18, 2012 9:46 PM ET

Edited Aug 19, 2012 12:33 AM ET.

by user-1087436

So. Here we are again in Upper Nerdistan, and I'm lost. DHPs, according to Martin and the referred document, operate at peak efficiency when over-sized for existing conditions. But existing conditions change, and DHPs, it appears, are built to adapt to those conditions. Design temperatures, if I understand correctly, are extremes--average extremes--likely to be encountered in a given climate zone. Most of the time, the minisplits will not be operating in extreme conditions. Therefore, it would seem, if they are "right-sized" for the extreme conditions, they will, ipso facto, be over-sized for the day-to-day conditions that will be encountered. If that is the case, why would you over-size the unit for the design (i.e., extreme) temperatures, when most of the time they will be over-sized anyway?

Aug 19, 2012 3:38 AM ET

Response to Gordon Taylor
by user-756436

The main reason that it's useful for the residents of Upper Nerdistan to know that ductless minisplits operate most efficiently under part-load conditions is to remove sizing anxiety.

Sizing anxiety may or may not be useful when it comes to sizing conventional furnaces and air conditioners. (That's a topic for another blog; it turns out the issue is not as clear-cut as many Upper Nerdistan residents think.) But fortunately, when it comes to sizing minisplits, there is no anxiety.

Get your minisplit somewhere in the ballpark, or choose one that is too big if you want, and you're fine. Isn't it nice to be anxiety-free? The minisplit does all the work for you. It even operates better when it's too big. I think that is very useful information to know.

Aug 19, 2012 1:05 PM ET

Ok, we love Splits and the US
by wjrobinson

Ok, we love Splits and the US of A. What is the split situation with US manufacturers? Toss some stimulus crumbs their way instead of Solyndra maybe ... Or not.

Aug 19, 2012 3:02 PM ET

It's all very cool.
by user-1087436


Aug 19, 2012 3:09 PM ET

Edited Aug 19, 2012 7:17 PM ET.

Martin, I believe you're misinterpreting COP data.
by user-1030217

As alluded to by Gordon, a correctly sized system will operate a part-load conditions most of the time anyway. That's the advantage of having variable capacity. But COP is a steady-state efficiency metric, which mean it ignores cycle losses. Intentional oversizing increases cycle losses since these units have a limited capacity range (the minimum capacity is as high as 50% for smaller units such as the MUZ-GE09). Oversizing effectively reduces operating range.

Although COP's may be higher at part capacity, in practical use, you won't realize those high COP's unless the unit runs long cycles. Oversizing works against that. Intuitively, we understand that the initial few minutes of each cycle are inefficient as the system overcomes thermal inertia. The shorter the cycle, the more those initial minutes weight on total efficiency. If you double up on capacity, then the system will operate at the minimum most or all of the time (depending on turn-down ratio). So when the load is less than design (most of the year), cycles become shorter and efficiency suffers.

Unfortunately, there's no way to estimate this trade-off, but it's a myth to think that an oversized unit will necessarily be more efficient. Perhaps to a point, but certainly not if double-sized. Also, when sizing to the heating load, it's important to consider what happens in cooling mode. You could easily end up with 3 times more capacity than the design load. Not good when you consider what happens with cycling losses at part load conditions.

My own design guideline is to avoid selecting a system with a minimum capacity exceeding 60% of design load. This most often comes into play with multi-zone (single compressor) systems, since the compressor's minimum now becomes the minimum capacity for each zone during single zone calls, even though the individual head may have a lower minimum when multiple zones are calling.

Finally, I think you're confusing VRF with inverter technology. Just because a heat pump has an inverter driven compressor doesn't mean it has variable refrigerant flow. VRF is a method for managing refrigerant distribution to multiple heads in commercial multi-zone systems, such as Mitsubishi's City-Multi model line.

Aug 20, 2012 4:28 AM ET

Edited Aug 20, 2012 4:29 AM ET.

Response to David Butler
by user-756436

Thanks for your detailed comments. And thanks for the correction on variable refrigerant flow; I have removed the phrase from my earlier post.

I feel confident on the truth of the statement that ductless minsiplits operate more efficiently at part-load conditions than peak-load conditions. However, you have suggested several other factors that call into question whether an oversized unit will save energy over the course of the year. I intend to do more research on this issue and report back.

In the meantime, I always welcome more comments.

Aug 20, 2012 8:05 AM ET

by Marc Rosenbaum

I don't intentionally oversize minisplits even though they are variable output. That's because if they are sized to handle the peak load, they'll be running at less than peak output almost all the time, and therefore take advantage of the efficiency gains available when the compressor and fans ramp down from peak capacity. As has been stated above, these products are variable in nature, but nonetheless have minimum output below which the compressor is off.

Somewhere above is a discussion about dewpoint and capacity. I hope what follows is clarification rather than the opposite!
Heat pumps are rated at 47F and 17F. Most of the minisplit products lose capacity as the outdoor temperature drops, and all do at some point. The principal reason is that the work that needs to be done to extract heat from progressively colder air increases, and the output of the unit decreases (some units use more power as the temperature drops, to keep capacity up, and others use less power but drop off in output - in both cases COP drops). There is a secondary reason, which is that there is a point where the condenser coil runs below freezing and therefore has to occasionally run in defrost mode to keep the coil from excessive icing.
Mitsubishi, in engineering literature for their S series City Multi (true VRF product), has specific information about output reduction vs. outdoor temperature, as well as output reduction due to defrost. The maximum defrost output reduction is 12% at 32F wet bulb; at colder temps the defrost capacity reduction is less, because there is less moisture in colder air, so it rises to a 5% output reduction. So the outdoor temperature is the dominant capacity determinant (at least on equipment I'm familiar with, so not to ay others don't operate differently).
My personal observations living with a Fujitsu minisplit is that this is borne out, and the unit is in defrost mode more around freezing outdoor temperatures than when it's really cold out (say, 10F). At any time i doubt the unit is in defrost mode more than 10% of the time.

I hope this helps and I haven't inadvertently pushed anyone's Flame button :-)

Aug 20, 2012 8:21 AM ET

Response to Marc Rosenbaum
by user-756436

Thanks very much for your helpful comments. I especially appreciate the information on defrost output reduction.

Aug 20, 2012 11:35 AM ET

Re: Ventilation
by James Steel

Thanks for adding information on Scott's ventilation strategies to the blog. I suspect that an exhaust only strategy comes up way short. But a conversation best had in another thread. Thanks again for the info.

Aug 20, 2012 5:00 PM ET

In-re oversizing and dew-point/capacity
by user-1004076

Thanks to David Butler for the confirmation on the limits of oversizing. I've never attempted to formally model it, (with insufficient performance data to work from), but it's pretty clear that the standby losses will eat your efficiency lunch when it's duty-cycling at min-speed most of the time rather than modulating. Being 1.25=1.5x oversized at the design condition, most would still be modulating near their low end of the average winter load giving a slight boost to average efficiency over a unit that just barely made it at the design load, but beyond that point it slides off an every steepening cliff.

The defrost derating is a function of how close the outdoor dew point is to the actual outdoor temp, which will vary widely by climate and from day-to-day. Even though they'll spend more time in defrost when outdoor temps are near to but above the freeze point than at colder temps, capacity derating isn't as severe. The raw efficiency is substantially higher at 35F then at 17F, and when it's still above freezing the heat pump is reaping the full benefit of heat of vaporization of that water, while only giving up the heat of fusion of that water during defrost. While the amount of moisture in the air is less when it's colder, there are be conditions where the dew point is close enough to the air temp to put a dent in capacity that it wouldn't otherwise experience.

Mitsubishi H2is have "rated capacity" at 17F of about 60% of the "maximum capacity" at that temp (which is usually close to the nominal nameplate output). The "rated capacity" is dependent on a specific wet-bulb temp at which it is valid- there's a little asterisk and number at the various "rated capacity" values at different outdoor temps, and next to the +17F is a "*3". (Clipped from a spec I have on my computer) the "*3" conditions are:

*3 Rating conditions (heating)-
Indoor: D.B. 21.1º C (70º F), W.B. 15.6º C (60º F);
Outdoor: D.B. -8.3º C (17º F), W.B. -9.4º C (15º F)

A wet-bulb temp +15F, and a dry bulb temp of +17F corresponds to a dew point of +10F, which is not a rare mid-winter dew point in southern New England.

Putting it in real terms, the popular MUZ-FE18NA is nominally a 1.5 ton, and can deliver more than ~18KBTU/hr in dry air even at +0F. (They're rated for 21.6KBTU/hr at +5F)

But even at +17F, with a dew point of +10F it's only rated for 11.7KBTU/hr.

While heavily derated conditions don't occur every day, they DO happen multiple times every winter, and can sometimes persist for hours. But those hours are usually when most people are in bed. If say, your outside design temp is +10F and your design heat load is 18KBTU/hr @ 10F, the unit makes pretty handily on nominal numbers, but there will still be a shortfall at even the +17F heat load when the dew point is that high.

In a very-well-insulated house you'd sleep through it and never notice, whereas in a small barely insulated summer cabin with the same +10F heat load you might feel the chill first thing in the morning. The lower thermal mass and shorter time-constant of the smaller less insulated house will see a more rapid drop in temp when the shortfalls occur.

After daybreak the capacity improves rapidly with outdoor temperature due to the greater separation from the dew point. Dew points move much more slowly than air temp at the start of the day, but more than the raw outdoor temp it's the separation between the outdoor temp and the dew point that gives it the increased capacity. The wider the separation in dew point and outdoor temp, the more closely the capacity approaches the "maximum capacity", which is it's bone-dry-air capacity. But in freezing-fog conditions where the outdoor temp has pretty much hit the dew point, the output crashes.

[Editor's note: For a further discussion of this issue, see information provided by Marc Rosenbaum in Comment #66 on 8/24/2012.]

Aug 21, 2012 10:37 AM ET

One lingering question,
by user-1087436

getting back to the Carter Scott heating strategy. I ask this because I can see a skeptical HVAC contractor (or home owner, or building code official) asking the same thing, and I'd like to have an answer ready. Suppose you have a minisplit head mounted high on a wall in a hallway outside a series of bedrooms. It's daytime, and the doors to the bedrooms are open. The house, of course, is compact and well insulated, and the air handler has a heating capacity adequate for those bedrooms. You know the strategy is going to work, because it has worked for others. But the question remains, and we all know a skeptic is going to ask it, how can that minisplit head sense the heating/cooling requirements of those bedrooms when its only connection to them is a series of 30"x80" openings?

Of course one answer is going to be: Have faith in the Second Law of Thermodynamics. And the second answer is going to be: Put electric baseboards in the bedrooms for cheap insurance. Still, I was looking for a bit more than those answers, and if somebody has something else to offer, I'd love to hear it.

Aug 21, 2012 10:51 AM ET

Edited Aug 21, 2012 2:27 PM ET.

Response to Gordon Taylor
by user-756436

Like most heating systems, a ductless minisplit is controlled by a wall-mounted thermostat. It will turn itself on and off (or will modulate its output) based on the temperature of the air near the thermostat. If you want, you can place the thermostat in a bedroom or in a hallway; it's your choice. [Later edit: This isn't true for most ductless minisplits - they actually monitor the temperature of the incoming air. Sorry for the mistake. See further comments below.]

When a house has superinsulated walls, triple-glazed windows, and low levels of air leakage, the temperature variations from room to room are less than in a conventional home. Let's imagine that a hallway is maintained at 72 degrees F. That means that the partitions and doors (whether open or closed) are also at 72 degrees F. These partitions aren't insulated. The thermal mass in the house (including the framing and drywall) act like a flywheel. Everything inside the thermal envelope of the house tends to stabilize at the same temperature.

Of course, heat loss (especially heat loss through windows) at night affects the temperature of each room, and heat gain (especially heat gain that happens when sun hits a window) also affects the temperature of each room. Proper window sizing, window shading, and glazing specifications can minimize problems from these types of heat loss and heat gain.

Aug 21, 2012 1:20 PM ET

A hallway may not be the best location for a ductless unit
by homedesign

A couple of months ago I was having an email exchange with David Butler..... where I proposed locating a single Minisplit in a hallway near the bedrooms....

David Butler's advice :"Hallways are not the best place to put a ductless unit. Better to put in a room with large load. "

Aug 21, 2012 1:49 PM ET

Wall thermostats?! Remote controls!
by user-1004076

Wall thermostats are probably the least common way to control a ductless these days. Most use infra-red remotes (similar to television remotes) that communicate with the head, and the head tracks the temperature of the incoming air to determine the room temp.

Most manufacturers support some sort of wireless or wired wall thermostat as an option, though it's usually company-proprietary, not a generic third party T-stat. (Recent Mitsubishi models support Honeywell RedLINK wireless thermostats, and it's at least theoretically possible to use a non-Mitsubishi version.) Most people seem to be comfortable with using the remote though.

With R40 exterior walls and ~R2 partition walls, in a blind room (no windows) the temperature difference between the heated hallway and the bedroom could be roughly calculated by the R-ratio against the hallway temp & outdoor temp. But the heat loss out the windows could easily be 2x that of the rest of the wall, making the calculation more complicated, but still a number that can be roughly calculated on a bigger napkin (ignoring mass effects and solar gains, etc.). The simpler model balances the heat flux between the hall & R2 partition at the hall-to-room delta with the heat flux out the exterior wall/windows to solve for bedroom temperature at a fixed outdoor temp.

But of course, in an occupied bedroom you would also have the ~400-500BTU/person-hour (awake, not particularly active) or 250-350BTU/person-hour (sleeping) either supporting the heat load or adding to the cooling load of that room. In a high-R home and a low-loss bedroom it makes a small but measurable difference in where the room temperature stabilizes.

Bottom line at all but the temperature extremes (or in some cases high solar gains), the room to room differences will be pretty small. It's OK to bump the setpoint of the ductless up/down a few degees during colder/hotter weather to bring the doored-off areas into a more comfortable range. Most people can tolerate a 5F or greater delta between the warmest & coolest rooms in a house, and most of the time adjacent spaces will have deltas much lower than that.

Aug 21, 2012 6:34 PM ET

Edited Aug 21, 2012 6:54 PM ET.

pay attention to location of largest loads
by user-1030217

If a mini-split head is installed in a hall with little if any intrinsic load, it's going to short cycle. In fact, it will hardly ever run unless there's a way to induce air changes with adjoining rooms. If a remote stat is installed in a bedroom, the unit will over-condition the hall, especially when bedroom doors are shut. In fact, the blower should never be separated from the sensor by a closed door.

While it's true that temperature swings will be smaller in a super-insulated, super-tight home, it's important to look at where the loads actually occur. The key to minimizing temperature variation is to deliver conditioned air in the rooms with the largest loads.

A lot of folks advocate the use of ERV/HRV's to help distribute conditioned air, but this strategy makes no sense with a ductless mini-split as primary heat/cool source, since the ventilation system does little to drive internal air changes. In fact, it creates additional loads in rooms with fresh air supply vents. Moreover, ERV/HRV blowers are extremely inefficient air handlers. It takes a lot of power to push air through the dense heat exchange core. Even models with ECM blowers consume 5 to 10 times more power per CFM than a conventional ECM air handler.

A ducted head is the most logical and effective approach to addressing bedrooms loads. But it's important to pay attention to available external static since these heads have very little fan power. For example, Mitsubishi's SEZ-KD ducted head can only handle up to 0.2"WC. I listed some design tips in my comment on Allison's mini-split blog (

Aug 21, 2012 6:50 PM ET

Edited Aug 21, 2012 6:51 PM ET.

is this what clients really want?
by user-1030217

Dana wrote:
> Most people can tolerate a 5F or greater delta between the warmest & coolest rooms in a house

That's pretty much status quo for conventional building practice. It's easy to save energy if we push occupant tolerance limits. However, in my experience, people who build (or retrofit) high performance homes are usually motivated by a desire to *improve* comfort, sometimes even more so than realizing a positive payback. On the other hand, I've had clients who willingly push their tolerance limits to great effect. Bottom line: I would never specify a central ductless heat pump unless the client fully understands and embraces the trade-offs.

In practice, temperature variation may be acceptable in cold weather since people generally prefer a cool bedroom for sleeping. But in areas with significant cooling loads, especially homes with upstairs bedrooms, it's hard to get around the need for ducted air distribution. Especially if any bedroom windows are exposed to afternoon sun in summer.

Aug 21, 2012 10:18 PM ET

Edited Aug 22, 2012 1:20 AM ET.

Getting back to Carter Scott...
by user-1087436

This is all great stuff from the best in the biz, and worth studying at length. But could we get back to the Carter Scott houses? Martin has written that CS uses two minisplits, one on each floor of his two-story houses. Now presumably, if his second floors contain bedrooms, they must have something resembling a hallway. And I have to assume that his 2nd minisplit is located there, which David Butler says is not a good idea. If so, how does Carter Scott handle the problems we're talking about? Does he do anything special with thermostats, for example? Or does he just open up the bedroom doors and let the air handler take over?

Aug 22, 2012 12:01 AM ET

Boost fan?
by zimmerdale

A few Passive House projects (my bungalow remodel included) are experimenting with using the combination of an ERV to exchange air and a ducted minisplit to condition air. They share the same duct work. I would like to add a small Fantech inline boost fan to supplement the wimpy minisplit fan and allow for more manageable duct sizes. Any experiences or thoughts on syncing a secondary fan with the built in fan?

I am using the Samsung 1-ton ducted model EH035 and the UltimateAir RecoupAerator.

In Kansas, we have significant latent loads during the summer, so I will have stand alone dehumidifier. Several HVAC contractors have also recommended the option of increasing air circulation--which would not be part of the ERV cycle, so as not to increase indoor humidity. The boost fan could be used for this task as well, without running the minisplit.

Our house will be very similar to the ones Carter Scott is building--compact design with around 2000 ft2, heat load of 8000-10000 Btu/hr, Passive House levels all around.

Perhaps I should have just used a single or double-head ductless minisplit, or waited for a magic box, but I have the ducted model already, so I want to try to use it.

Aug 22, 2012 12:14 AM ET

Edited Aug 22, 2012 12:22 AM ET.

@Jason Miller
by user-1087436

New question at Q&A?

Aug 22, 2012 3:41 AM ET

Edited Aug 22, 2012 3:46 AM ET.

answered at EV blog...
by user-1030217

@Jason, I posted extensive reply to your question which you double-posted in Allison's mini-split blog.

@Gordon, I can't answer for Carter, but the fact that his homes are in a heating dominated climate mitigates the issue somewhat, as I noted in my last comment. Also, if upstairs unit is located in a loft area or balcony that communicates with downstairs (as opposed to a confined hallway), at least the over-conditioning issue would go away.

Aug 22, 2012 5:07 AM ET

Edited Aug 22, 2012 5:09 AM ET.

Ductless minisplit locations
by user-756436

I am attaching an illustration from one of my Fine Homebuilding articles that shows the locations of the minisplit indoor units in Carter Scott's first net-zero-energy home. The upstairs unit is in a hallway at the top of the stairs, pointing down the stairs.

Carter Scott - ducltess minisplit indoor unit locations.jpg

Aug 22, 2012 9:02 AM ET

Response to several objections to Carter Scott's approach
by user-756436

Several GBA readers have questioned aspects of Carter Scott's approach:

  • John Brooks has noted that a hallway is probably not a good location for a ductless minisplit unit.
  • David Butler has noted, "If a mini-split head is installed in a hall with little if any intrinsic load, it's going to short cycle."

These comments are all fine and good, up to a point. However, Carter Scott has completed 18 homes using this approach, and the owners of the homes are apparently happy.

Register for a free account and join the conversation

Get a free account and join the conversation!
Become a GBA PRO!