Ducted minisplit efficiency vs. ductless
We have a 1979 single floor ranch style house with baseboard heat. We are planning on adding minisplits for air conditioning and energy savings while heating. One of the options presented by a local HVAC company was four ductless heads and a fifth zone being a ducted unit ( Mitsubishi SEZ-KD12NA4R1.TH) in the ceiling running into a bathroom and three bedrooms. The bedrooms and bathroom are at the end of a hallway where it appears tricky to aim the airflow. My question is this…. Would I be better off efficiency wise using a ducted setup like this, or adding more ductless heads? How much of an efficiency loss is having one zone out of five being ducted? This is in an average insulated house, and we will use more energy in heating than in cooling. We are in Zone 4-5. Thanks for any insight you can provide.
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
I encourage you to consider more than efficiency - add thermal comfort, humidity and in some cases, the ability to operate at low enough outdoor temperatures.
The answer is that it depends on the results of your room by room Manual J.
The smallest ductless indoor unit is 6k btu, these units are sized for heating, not cooling. So when sizing for cooling loads, in small rooms they can be off by a factor of 2 in size (since cooling loads in zone 4 on the hottest days are about half the heating loads on the coldest days). If you have lots of rooms like this grouping them together is a great idea.
>"The smallest ductless indoor unit is 6k btu, these units are sized for heating, not cooling."
The 6K is it's rated cooling capacity, not heating. A 6K Mitsubishi head puts out more than 7K in heating mode, some put out more than 8K:
Almost no bedroom loads are high enough to be ideal for a 6K head.
>" One of the options presented by a local HVAC company was four ductless heads and a fifth zone being a ducted unit..."
>"How much of an efficiency loss is having one zone out of five being ducted? "
Rather than an efficiency LOSS, having the zone ducted could be an efficiency GAIN over adding multiple ridiculously oversized ductless heads, and could also be less prone to comfort problems.
A 5 zone Mitsubishi -5C42NAHZ is good for 48,000 BTU/hr all the way down to +5F ( what is your 99% outside design temperature?) , but it's minimum modulated output @ +17F or higher is 24,000 BTU/hr, which may be MORE than your actual design heat load. When a single 6K head is calling for heat it's cycling the compressor on/off at a ~33% duty cycle or less, since it can't deliver more than about 8 KBTU/hr even at it's maximum speed. It only gets worse if bumping it up to the -8C48NAHZ, which has a minimum output of 27K,, though the non cold-climate version can modulate a bit lower:
When a ductless head on a multi-split is ridiculously oversized for it's room/zone load and one or more others are more reasonably sized for their respective loads the rooms with the oversized heads can be overheated /overcooled even while nominally "OFF", since whenever any zone is running there will be refrigerant flowing in all heads/cassettes.
Before soliciting any more proposals, run a room by room Manual-J heating & cooling load calculation. You might even run more efficiently with a 2-ton 3 zone system with only one right-sized ductless head and two right sized ductless zones that modulate most of the season rather than cycling the compressor a and a bunch of ductless heads into lower efficiency and lower comfort.
To get this right, as a rule of thumb the maximum compressor capacity at your 99% outside design temperature should be no more than 1.5x your whole-house heat load at that temp (1.2x-1.3x is better), and no individual cassette no more than 2x oversized for the whole house load. As an example, my ~2400' house +1600' of basement built in 1923 has a load of about 36,000 BTU/hr at the local 99% outside design temp of +5F. With 48K of output @ +5F that would make the 5C48NAHZ 48K/36K= 1.33 x oversized for the design load, and a reasonable fit for my house. With a mixed ductless/ ducted zones the 56,840 BTY/hr max output of the -8C48NAHZ would be 56,840/36,000= 1.58x, the design load, bigger than absolute maximum I should even consider.
The minimum compressor output of 27K @ 17F is 27K/36K= 75% of the design load, barely less than m load at +17F, and HIGHER than the load at my ~25F mean January temperature. That means it really only modulates the compressor during the coldest hours of the day on the coldest month of the year and during cold snaps. The rest of the time it would be only cycling on/off.
Bottom line- both the heads/cassettes and the compressor have to have enough load to operate efficiently, and a single ducted cassette 1.25x oversized for it's load will be more efficient than three ductless heads, each 2-3x oversized for their respective loads.
I have a system very similar to the system proposed to you.
If your bedrooms and bathrooms are under 125 sq feet I would strongly lean towards the ducted setup for comfort. Even the smallest Mitsu wall unit (6,000 BTU )may result in the short heat cycling and humidity removal issues (as pointed out by others)
If you went with wall units instead of ducted, then the bathroom and hallway might be OK if the bathroom door is mostly open. A small heat strip could be added in the bathroom if needed.
That said, my Mitsubishi FH 6,000 BTU wall unit can run at very low fan speed and IMO would keep most average insulated >125 sq feet bedrooms very happy summer and winter (but as Dana pointed out) possibly at the expense of outdoor unit efficiency.
Would your ducts be in unconditioned attic? If ducts are going to be in unconditioned attic then wall units may be most efficient no matter what. Depending on ductwork lengths and quality of installation.
On a hot summer day in my unconditioned attic the SVZ air handler and ductwork would increase the intake to exhaust temp by 3 degrees F when in fan only mode...that’s significant energy loss.
I just spent 3 days taping and wrapping my ductwork to gain some more performance.
Thanks for all of the comments everyone. I should have been more clear about the ducted portions location. It will be in the attic that is not part of the conditioned space. This was what got me a bit concerned about efficiency.
Even so, the bigger problem is likely to be WAY sub-optimal oversizing and with it the potential for short-cycling any 5-zone compressor (unless it's a fairly large or lossy house.)
The non "-NAHZ" (not cold climate)versions have a better turn down ratio, but crap out fast on capacity as temperatures drop so it really matters what your 99% outside design temperature and whole-house heat load is. The max capacity of the MXZ-5C42NA at +17F is literally HALF the capacity of the cold-climate MXZ-5C42NAHZ at +5F.
Which version was in the proposal?
Got a ZIP code? (For weather data and design temperature estimation purposes.)
Short, insulated duct runs or an insulated dog-house over the cassette in the attic isn't necessarily a big efficiency hit if the supply and return duct system are well designed to not end up with air-handler driven outdoor air infiltration. The air sealing is always going to be an issue when you start punching holes in the attic floor, but it's possible to do it right. Burying the ducts (and cassette-doghouse) in insulation can bring the conducted losses to ho-hum levels.