Ducted mini-split serving multiple floors
![](https://images.greenbuildingadvisor.com/app/uploads/2018/07/24185413/picture-222599.jpg)
I am attempting to break my house design into zones such that I can use one mini-split for each zone. I am having trouble simultaneously satisfying two pieces of advice from this forum. In “Rules of Thumb for Ductless Minisplits,” Martin says that one ductless head can serve a maximum of about 1100 square feet. That is based on distribution, not capacity. I realize that it is a rule of thumb that might not apply to a super-open floorplan, but if I violate that rule, I should have some justification. Since heating dominates the sizing for me, I am favoring the Fujitsu products, but the smallest seems to be 3/4 ton. In “How To Buy a Ductless Minisplit,” Dana is quoted as saying that a zone needs a design heat load of at least 6000 btu/hr for a 3/4 ton head. Since the 3/4 tone Fujitsu units have a heating capacity of 14,000, the minimum zone is probably more like 9,000-10,000.
My floorplan has several natural regions within which air should move around easily. However, some of these regions are less than 6000 btu/hr design heat load. If I attempt to combine these regions with adjacent regions, the total area exceeds 1100 square feet and there are some airflow barriers. In order to combine regions, I may need to look at ducted units. If I go with ducted units, can I combine a region on the main floor with a region directly below it in the basement? The first floor rooms would get floor registers and the basement rooms would get ceiling registers. Are there rules of thumb for how many registers are acceptable and the maximum distance from unit?
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
![](https://www.greenbuildingadvisor.com/app/themes/greenbuildingadvisor/dist/img/modules/detail-library/thumb1.png)
![](https://www.greenbuildingadvisor.com/app/themes/greenbuildingadvisor/dist/img/modules/detail-library/thumb2.png)
Replies
Reid,
Dana gives excellent advice for designing ductless and ducted minisplit systems, but I'd like to point out a few facts: (a) Most systems are designed and installed by contractors who aren't as smart as Dana, and (b) Most homeowners are very satisfied with the comfort provided by their minisplit systems, even if the systems weren't installed according to Dana's rules. The reason is simple: these systems have the ability to ramp up and down as needed, to handle partial loads and higher loads at or above design loads, and to maintain a fairly constant indoor temperature.
In short, you can bend the rules.
a) That's a tragedy, given that I'm actually pretty stupid (you can even ask my wife!)
b) Ridiculous oversizing factors are more likely to create comfort problems than modest undersizing.
b+) The look at the MINIMUM modulated output of the units @ 47F relative to your calculated load, when sizing them, not just the max output relative to your design load. When the thing starts cycling on/off at low load is when room temps can take big swings- the higher the outdoor temp at which that begins, the less likely it is to become a problem. If it's starting to cycle on/off at 25F it's more of a problem than if the min-mod output balances with the load at 50F+. For instance:
A 3/4 ton Fujitsu has a min-modulated output 3100 BTU/hr @ 47F (it is listed on the HSPF /SEER submittal sheets), which isn't terrible. If your load at 47F is 2500 BTU/hr there's a potential issue, but not a huge one. If your load at 47F is 1000 BTU/hr there's more potential for wide temperature swings since it's probably not modulating continuously until it's below freezing, but that still might be OK in a high-R house. But if the load at +47F is only 500 BTU/hr YOU and the lights become the space-heater for the room, not a problem at +47F, but means that it'll still be cycling on/off rather than modulating at a much lower outdoor temp, something like +5F (or even colder), and the room temperature swings over those cycles may be significant.
A 3/4 ton cold climate Mitsubishi can throttle back to ~1700 BTU/hr @ 47F, making it a better choice for low-load rooms if you're going with a dedicated mini-split for that room. But the same 3/4 ton head married to a multi-split compressor is limited by the minimum output of the compressor, which is universally higher than 1700 BTU/hr, and would have to cycle the head on/off at a higher blower speed when serving just that zone.
The power of the air-handlers on mini-duct cassettes vary by vendor (a LOT!) Of the mini-duct units the Fujitsu units are a bit easier to design around, since they have more oomph than most of the competition. They can also be mounted vertically, which is handy in some installations.
Among the general rules for ducts, keep them straight and short, and use only long radius ells, which are much lower impedance that those with sharp corners. Use hard-piping rather than flex duct (important if the run is going to be very long.) Even if the design estimates say a 5" round duct is big enough, bump it up to 6" or even 7", oversizing them uniformly over the entire duct system. Install operable vanes for tweaking flow balance between rooms.
While it's traditional for supply duct registers to be located under windows as counterbalance to the biggest heat-gain/loss point in the room, that makes for unduly long duct runs. That tradition goes back to the days when windows were single-pane and often left open a few inches overnight to mitigate against carbon monoxide risk from coal fired furnaces. In a high-R house or even a code-min house there isn't much down-side to putting the registers somewhere else to minimize duct lengths to ensure higher air flows.
Putting a mostly below-grade basement on the same zone as an above grade floor almost always leads to the temperature being way off in the basement, varying by season. If you dial it the flow balances perfectly for January during mid-day it might be bit off at midnight, and it'll be quite a bit off in March, crazy wrong in July. With thermostatic registers or daily/weekly vane tweaking it can be made to work well enough for most people, but it won't be fully automatic.
Unlike the mini-duct cassettes, mini-splits with bigger air handlers (like some of the Daikins or the newer Mitsubishi H2i air handler units) have plenty of power for pushing air, and HVAC folks used to other types of equipment can treat it in a "business as usual" fashion, with somewhat less sensitivity to the duct design.
Thank you for the detailed answer.
Based on the guidance here and in other threads, I have come up with a mini-split proposal for my house plan. My goal was to flesh out a min-split proposal enough to make a decent comparison with the baseline system in terms of initial costs, operating cost, and comfort. The baseline is a 95% AFUE forced air natural gas furnace paired with a central air conditioner. My proposed mini-split system has three zones:
Zone 1 - family room, kitchen, dining room, guest suite: Design heat load = 12,500 btu/hr. Area = 1000 ft^2, There are no walls between the family room, kitchen, and dining room. The guest suite has a bedroom, walk-in closet, and bathroom which are pretty open to one another. (The door to the bathroom will probably be open except when it is being used.) The guest suite will probably be occupied only occasionally. My plan is to use one Fujitsu 9RLS3H wall mount unit in the family room. That is rated for 12,000 btu/hr down to -15F and quite a bit higher at higher at higher temperatures. My biggest concern in this zone is heating the guest bedroom when guest close the door at night.
Zone 2 - master suite, office, laundry room, pantry, half bath, and stairwell: Design heat load = 11,000 btu/hr. Area = 1110 ft^2. The master suite is on the second floor directly above these other rooms. The other rooms are connected to one another by an entryway. I expect the doors will be open most of the time. The entryway also has a doorway to the kitchen. My plan is to use one ducted Fujitsu 9RLFCD mounted vertically in the pantry. That allows short duct runs to the entryway ceiling, the office ceiling, and the master bedroom floor. I can put a return register near the floor immediately under the unit. I plan to put a duct connecting the master bedroom wall to a wall in the entryway to provide a return pathway when the master bedroom door is shut. When the master bedroom door is open, I expect air will flow through the stairwell. The unit is rated at 12,000 btu/hr down to -5F.
Zone 3 - basement, which includes bedrooms, a bathroom, and a rec room for the kids: Design heat load = 9,000 btu/hr. Area = 1080 ft^2. The kids are college age, so they will only be home occasionally. The load and area do not include storage areas or utility rooms that I will not intentionally heat and cool. The kids keep their bedroom doors shut now and I don't expect that to change. I plan a second ducted Fujitsu 9RLFCD for this space. My hope is that I can turn this unit off when they are not home. The initial question above was motivated by an idea to add ducts from this unit to the guest suite to help with heating that space.
All of these load calculations are based on Marc Rosenbaum's spreadsheet which, in addition to wall, windows, and ceiling, includes formulas for heat loss to the ground and heat loss due to infiltration./ ventilation. The house is in climate zone 5 with a design heating temperature of 0F. (We did have a number of -20F mornings last winter, but that is unusual.) In another thread, Dana said that the infiltration numbers seemed excessively high. I distributed the infiltration losses to rooms in proportion to exterior perimeter.
On zone 1, what is the design heat load of the guest bedroom? It's not insane to either bump the set point on the mini-split up 3-5F if it's going to be cold out when that room is occupied. It may also be reasonable ton install resistance electric heat sufficient to carry 100% of the design load and leave it off except when that room is occupied with the doors closed. Don't forget to subtract off 200-250BTU/hr per sleeping human.
If it's anticipated that it only rarely used overnight, a $50 oil-filled radiator type portable space heater can be a reasonable solution. Most codes require that all rooms be capable of being heated automatically to at least 68F at the 99% outside design temp, and an electric space heater does not fully meet that criterion. But if you sized the zone 1 mini-split big enough to heat the main zone to 78F or something, you can run the math on the U-factor of the partition wall to see if a 10F temperature delta could heat the guest room to 68F.
You may need to bump up to a 1 ton or even a 1.25 ton, but all of the Fujitsu RLS3s modulated down to 3100 BTU/hr @ 47F- there's not a big down-side to oversizing them. The higher cfm of the blower on the larger units can also be useful for moving that heat around over a larger zones.
http://www.fujitsugeneral.com/PDF_06/Submittals/15RLS3HSubmittal.pdf
I accidentally reversed the heat loads of zone 1 and zone 2 in the note above. Zone 1 is about 11,000, of which the guest suite is about 4500 (4000 with two sleeping humans). The guest suite and family room have a shared wall about 15 feet long by 8 feet high. What is a realistic R-value for an uninsulated interior wall? There is also a direct vent gas fireplace between the two rooms facing the family room. We considered one of the two-sided fireplaces but decided against it based on privacy for the guest bedroom.