Is a single head minisplit practical?
This is a mostly accurate plan of dear, old Mom’s house.
2265 ft²; zone 4 marine; target ACH50 = 1.5; 5000 HDD; 22°F design temp;
2×6 walls w/2 stud corners, 24″ o.c., 1½” horizontal cross-hatched walls (similar to “Mooney”), 7″ of dense pack cellulose; front wall faces north; awesome PV potential on south facing roof.
I used BuildItSolar’s load calc to get a rough idea of the heat load. I’ll get a manual J before specifying the heat pump. Depending on what guess (or SWAG) for ACH(natural) I used, I came up with a load of about 26,000 BTU/hr @ 0.5ACH (which is NOT @ 50psc.) At 0.33ACH, the load was 22,500.
Whether the load is 2 tons or 2½ tons, is a single head mini split, located high up on the wall on the back side of the pantry, and partially facing toward the bedroom hallway, practical? What can I do to make a single head work here? If not, then what?
I’m avoiding a ducted system because I can’t condition the crawlspace and don’t want to go through the attic. If I had to go ducted, space can be taken from the laundry room for the equipment.
Thanks.
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Replies
Mark,
The main variable affecting whether or not you will be comfortable concerns your windows.
We need to know:
The area of each window.
Which cardinal direction each window faces.
The U-factor and SHGC of the windows you have specified (or at least a description: for example, "triple-glazed windows with two low-e coatings and argon gas fill").
-- Martin Holladay
The BuildItSolar load calculator is really crude, an hits reliably to the high side of reality. The BIS tool makes no attempt to account for internal gains, such as mammalian occupants (Manual-J uses 230 BTU/hr per sleeping human), 24/7 plug loads such as refrigerators, etc. All internal gains need to be subtracted directly off the room load where those inputs are located.
It's important to run accurate U-factor calculations for custom framing types such as Mooney Walls etc to get close using any load tool.
At 22F most decent 1.5 tonners would be able to cover even the (higher than reality) calculated 22K load. A 2.5 ton system would almost certainly be sub-optimally oversized- the biggest unit you should be contemplating would be 2 tons of ductless, and 1.5 tons is far more likely. Better two ton compressors are usually good for ~25-30,000 BTU/hr @ +17F.
It's unlikely that a single head located as drawn would be able to adequately the remote bedrooms. It would be useful to mark the room-by-room loads on the drawing to be able to make reasonable recommendations.
But along the theme of taking a WAG: Two ton multi-splits usually max out at 3 zones, but a 3/4 ton mini-duct cassette mounted in one of the closet between the two bedrooms, a half ton head in the bedroom adjacent to the entry, and a 3/4 ton or 1 ton head mounted in the common area as drawn would probably get you there. It's also likely that a 1.5 ton Fujitsu slim duct mounted in the crawlspace using insulated hard-piped ducts could do it with a competent duct design.
http://portal.fujitsugeneral.com/files/catalog/files/18RLFCD1.pdf
http://meus1.mylinkdrive.com/files/MXZ-3C24NAHZ_Submittal.pdf
https://www.master.ca/documents/regroupements/1Submittal_AOU24RLXFZ.pdf
I would never give up privacy for the sake of EE, I just think is wrong for couples or teenagers to live with doors open to save a few bucks. Also, it goes against my architectural training to install something so ugly in a beautiful new house. IMO you have better options:
1. Install a multi-zone system where every room would have its own heat pump, and hide the heat pumps in the walls.
Furr-down hallways and closets so you could run ducts, even if you need to create coffer ceilings and cat-walks:
2. Install ducted mini-split system.
3. Install regular ducted HVAC equipment.
4. Install a high-velocity system, with 2" PVC ducts.
Anther WAG: If the real heat load numbers come out at 20KBTU/hr or less (probably will) a single 3/4 ton mini-duct unit mounted in the "possible furnace location" could probably serve all 3 bedrooms as a single zone with very short duct runs (either soffited or in the unconditioned crawl space) with a 3/4 ton- 1 ton wall coil (or floor mount) serving the common area.
https://www.master.ca/documents/regroupements/12Fujitsu_Submittal_AOU18RLXFZ.pdf
https://nonul.mylinkdrive.com/files/MXZ-2C20NAHZ_Submittal.pdf
If there is a pre-existing duct system in the crawl space that's in reasonable shape with reasonable room-to-room temperature balance, or if the local HVAC folks seem incapable of designing ducts for mini-ducts, but OK with more powerful air handlers, a 1.5 or 2 ton Mitsubishi MVZ series bigger-deal air handler can make it (married to the appropriately sized compressor unit.)
http://meus1.mylinkdrive.com/files/MVZ-A18AA4_For_MXZ_MULTI-ZONE_SYSTEMS_Product_Data_Sheet.pdf
http://meus1.mylinkdrive.com/files/MVZ-A24AA4_For_MXZ_MULTI-ZONE_SYSTEMS_Product_Data_Sheet.pdf
The turn-down ratio on the MVZ series only about 2.5:1, so oversizing it by very much would cut into modulation comfort & as-used efficiency.
Each bedroom and the front living room each have one window ~22ft². I am told the windows are 0.25 U. I used 0.45U in the BIS calc just to be conservative.
For the Mooney-like walls I used R26 (R3.7 x 7"). Without the cross-hatching, I would have used R16 to account for the thermal bridging. At the design temp the result would have been +3,000 btu/hr.
I know the BIS infiltration calc isn't using @50Pa, but it doesn't seem to be using @natural either. (1.5ACH ÷ 17 LBL = 0.09 ACH(nat)). I used 0.33 & 0.5.
I ignored the internal gains. I believe they were about 10% of the load w/2 people. ~2200 BTU/hr.
For the entire master suite + adjacent bedroom, using these assumed values, the BIS gave me a combined heat load of 5650 BTU/hr. A half ton for two bedrooms, a WIC, and an ensuite. Other than solar gain, there may not be much internal gain in the adjacent or front bedrooms for much of the time.
Dana, "a 3/4 ton mini-duct cassette mounted in one of the closet between the two bedrooms, a half ton head in the bedroom adjacent to the entry, and a 3/4 ton or 1 ton head mounted in the common area as drawn would probably get you there"
How does the air move from the closet to the two bedrooms? Does the cassette blow out in two directions?
Is there an alternative to having a wall-mounted head in the front bedroom?
Can one zone be on and the others off? Or two on and one off?
At a design temp of 45°F (common winter day), the whole house load is about 12,000 BTU/hr. Can a three-zone modulate down that far. (I don't understand how to read all the HP specs.)
Capacity
Nominal Cooling………………….…….……...…………………...... 22,000 Btu/h
Min-Max Cooling……………………….……………...……. 6,100 – 27,000 Btu/h
Nominal Heating……………………….…….………....……..…........ 24,000 Btu/h
Min-Max Heating…..………………….…………………….. 6,800 – 29,800 Btu/h Is THAT the answer? Do the zones each get part of that?
Many thanks.
Put in the regular ducts, it'll help with resale value. You could try single head mini split, then use inline fan with the ducts for circulation, let us know how that works out.
But I think with that layout you are better off with a normal HVAC setup.
Dana, Can there be a three-bedroom ducted zone AND a wall mounted coil on one mini-split? The alternate head location for the common area is above where the hallway begins.
If not, two separate minis are acceptable.
There are no existing ducts (new construction) and the crawlspace will be tight.
Is the three-port head similar to what might serve the three bedrooms?
Does this other unit need an additional air handler?
Armando, I hear you. The next family might not like a quirky system.
All multi-split compressors from first tier vendors can accommodate a mix of wall-coils, mini-duct cassettes. All zones can be operated independently, including leaving some zones off. What isn't possible is to run one in cooling mode while others are in heating mode- they're either all in heating mode, all in cooling mode, but any subset of heads/cassettes can also be off.
The minimum & maximum output capacity of the multi-split compressors stated in the submittal sheets is tested at +47F. The nominal capacity is the modulation level at which HSPF efficiency is tested, at both +17F & +47F. The minimum for a 1.5-2.5 ton multisplit typically between 5000-7000 BTU/hr min @ 47F for most newer/better designs, but in the past there were a few with minimums north of 10K.
The minimum modulated out of the heads/cassettes varies with model and vendor, but most 1/2 through 1 ton options can run at ~3000 BTU/hr or less. When only one head is calling for heat it has to be able to take the minimum output of the compressor though, which results in higher blower speed at the head and will cycle on & off if the zone load is lower than the min-output of the compressor, yielding slightly lower efficiency than when it's all modulating at minimum speed.
Splitting the output of a mini-duct cassette to multiple rooms is implemented in the duct design. If mounted in a closet space between two rooms the duct runs are inherently VERY short, which means it can usually be hacked rather than fully designed, and tuned for room-to-room temperature balance with balancing vanes. For longer runs and more rooms the ducts need to be more rigorously designed. Think about the return paths (ducted or otherwise) as well as the supplies, as well as access for changing filters, and for routing the condensate drains when in cooling mode. Fujitsu's mini-duct cassettes have the advantage that they can be mounted vertically, which means it could be mounted in a very narrow closet space type room to make it accessible, even if the ducts are run under the floor. All other vendors require horizontal mounting. Fujitsu's mini-duct cassettes also have more powerful blowers, making the duct design a bit easier.
If using a single head for the front bedroom, both Fujitsu & Mitsubishi have floor mounted coils that are thinner profile than their wall blobs, that operate at the same efficiency as the wall coils. But IIRC the smallest floor mounts they make are 3/4 tonners, whereas half-ton wall coils would be less oversized for a single-room load.
What mean's "tight", in the context of "... the crawlspace will be tight..."? Air tight? Tight headroom? ( Drinks way too much? :-) )
Using "...0.45U in the BIS calc just to be conservative " for windows labeled U0.25 is exactly the WRONG thing to do. (Are you planning to be an HVAC contractor or something? :-) ) Being conservative is what drives ridiculous oversizing factors. Being conservative you start out with a load number that's already too high, then are forced to cover that load with equipment on the next size step up, and end up with equipment more than 2x oversized that costs more up front, runs less efficiently and is less comfortable. What's worse, a crummy U0.45 window isn't even code-legal for new construction in the US zone 4C states! That's an 80% overstatement of what to expect out of U0.25 windows, so divide the 5909 BTU/hr window losses by 1.8, yielding 3283 BTU/hr. That's difference of 2626 BTU/hr, which is more than 10% of the original calculated load!
DON'T be conservative- be AGGRESSIVE, taking advantage of every plausible factor to reduce the load, as directed by the Manual-J instructions.
R16 whole-wall is basically an IRC code min wall. With most sheathing & siding options and 24" o.c. framing a 2x6 dense packed cellulose wall will beat that. A 2x6 24" o.c. cellulose wall with 1.5" deep furring will come in pretty close to the thermal performance of Case #3 in this bit o' analysis, which was a low-density R19 (=R18 at 5.5" ) plus a 2.5"/R8 Mooney configuration.
https://buildingscience.com/sites/default/files/migrate/pdf/BA-0903_High-R_Value_Walls_Case_Study_rev_2014.pdf
With cellulose you'll have about the same center-cavity R, and only slightly worst thermal bridging. It won't quite make the R21.5 modeled in that study, but it'll be at least R20, quite a bit better than your R16 assumption, so your wall losses are overstated by ~1.25x. That reduces your 3923 BTU/hr wall loss becomes 3923/1.25= 3138 BTU/hr, a difference of 785 BTU/hr.
So with just the window error and wall U-factor error you can peel off ~3500 BTU/hr from the calculated 22,250, for a load of 18,750 BTU/hr, and that's even before subtracting off the internal gains of occupants & plug loads, and any other errors of commission that were plugged into the calculator. Most homes will have at least 1000 BTU/hr of occupants, refrigerators, water heater standby, cable box/DVR etc, many will have 2000 BTU/hr. So really you're looking at less than 18,000 BTU/hr of design heat load, possibly quite a bit less, which makes most 2-tonners sub-optimally oversized.
It's highly likely that you can run the whole shebang with a Fujitsu 18RLFCD mounted in a smaller space than the "possible furnace location" using a grille to the hallway for the return path, and jump ducts between joists as the return from doored off bedrooms to the hallway return. It can still deliver 21,600 BTU/hr @ +17F, which is more than a 1.25x oversizing factor for your likely load at 22F. But keep refining the load numbers- there are likely other errors to fix, and lots of internal gains to be subtracted.
Solar gain only matters for design cooling load, since the sun isn't shining in the pre-dawn hours when temps hit the 99% outside design temp. While it makes quite a bit of difference in heating energy use, it doesn't affect the peak heat load number that you need to cover.
A buddy of mine owns an HVAC company, but that job is not for me. lol.
I used 0.45U before I got the 0.25U from the builder. Now I'm using the 0.25U.
I actually used R26 for the walls (#5 above) in the first load calculation. I've corrected that to R21.
In principle, I can see heating the bedrooms with the 18RLFCD and some ductwork. I don't yet know how to actually put that into practice.
What I don't see is how to supply the common area with that setup. The 18RLFCD is a one-zone.
The 18RLXFZ that you linked to in #4 above plus the ARU9RLF indoor slim duct for the bedrooms plus a 1 ton head in the common seems like the best route. The outdoor unit can be paired with up to 21,000 btu of indoor units.
Now for a full manual J.
Maybe part of the reason for being (irrationally) conservative in oversizing a heating system is that people fear that the heating system will fail to work well enough for some room(s) or on the coldest nights.
You might consider adding insulating cellular shades (or similar insulating window treatments) to increase the effective U-value of windows. Cheap way to add R-1, or maybe R-3 if you get side tracks to reduce air convection drafts. Even more with an insulating shutter. After all, two-thirds of mid-winter days its dark outdoors. As Dana mentioned, the coldest period is typically overnight before dawn when shades are drawn closed.
If you want some backup insurance for bedrooms, you might consider an electric radiant or resistance heater per bedroom. They are cheap, and only need to be used on the coldest nights or by those who want a warmer bedroom. Or for that matter a portable electric heater can be used later, if needed.
Code requires that every occupied room be able to hit 68F at the 99% outside design temp without auxilliary heat (such as space heaters) using an automatic heating system. That's primarily a bedroom issue, but pay attention to the room by room load numbers.
Thermostat operated resistance heating only needs to be added if the ductless (or mini-ducted) capacity comes up short after discounting for the internal gains such as human occupants, and the size of the resistance heater only needs to make up the shortfall, not 100% of the heat load.
The capacity of an AOU18RLFXZ at 68F outside, 22F inside is close to ( slightly more than) the 17,650 BTU/hr specified at 70F indoors, 23F outdoors, and it's probably a decent fit. If the ducts are run in the crawlspace there has to be an accounting for the duct losses, but not if it's all inside of conditioned space.
Bathrooms with small or no windows are often "heated" by a single occupant (figure 350 BTU/hr out for a conscious sitting person) plus the lights , but low voltage mesh radiant floor or a small electric panel radiator sized for the design load, operated under occupancy/vacancy sensor control + thermostat can be a luxurious touch using only miniscule amounts of power.
The mini-ducted 18RLFCD can heat the whole place as a single zone, with larger duct capacity going to the common areas than to the bedrooms. Joist bays or the crawlspace may need to be used for ducts to the living & dining areas, which may affect your floor framing decisions. The duct capacity to each room has to be designed properly per Manual-D, and the designer has to pay attention to the fact that it has less blower power than typical gas furnaces or split AC systems, but it's not terrible.
Note that if you want a closed off/unheated room as warm as the rest of the house and add an electric heater, this heater will supply 100% of the room's load.
That's true, if there is no other heat source in that room (such as a mini-duct register), and the electric heater has enough capacity to cover the entire load. That's why it's best sized only for the anticipated shortfall (not the full load), or run the resistance heaters only when the room is occupied.
The bathroom example typically has fairly short occupancy duration but also low loads- most bathroom heating loads are too low to merit a ductless head. For most purposes it's fine if it's running 5F or more cooler the setpoint of the ductless in an adjacent fully heated space, but it's nice to step out of the shower or bath onto a pre-warmed floor/room. The bathroom temp will rise rapidly when a bath is drawn or a shower is in progress, but without the resistance heaters it won't necessarily be able to hit a code-min 68F if bathing happens during 99% outside design temperatures. Even if it doesn't fully come up to temp before a short bath or shower is over, a panel radiator or radiant floor sized to the design heat load will improve the comfort levels in only few minutes.
For a bedroom example, if the ductless or ducted heat isn't sufficient to cover the load, the resistance heat only has to make up the difference, and if it's sized at the difference (not the full load) it can't supply 100% of the room's load until/unless outdoor temperatures are pretty mild. Setting the thermostat for the resistance heat a few degrees below the mini-split's setpoint allows the mini-split to always carry the lion's share.
Great discussion. The real question is how does one hire Dana as a consultant for HVAC design...:)
I haven't been following this whole discussion, but I want to note that when people get into wanting to supply a small amount of heat in a small room from a ductless heat pump system, a solution is to use an air-to-water heat pump such as Chiltrix. The equivalent of heads is fan coil units, and they come in smaller sizes than mini-split heads, both in physical size and in output. And if you want even lower output, you can get a small panel radiator for $110 for a bathroom or the like.
Dana posted:
On the advantages and disadvantages of oversizing. “Oversizing by as much as 1.5 times would usually leave the equipment in the modulation zone most of the season and provide some margin for lower temperatures, and run slightly higher efficiency than if exactly sized for the load. Above that [in other words, oversizing by more than 50%] and it’s pretty much downhill on both comfort and efficiency.”
More on oversizing. “At moderate to cool outdoor temperatures, modulating air-source heat pumps have a much higher efficiency at part load than when running full-blast, but the range of efficiency narrows as outdoor temperatures fall. But that high part-load efficiency means that with some amount of oversizing, the seasonal efficiency goes up, often substantially. The limitation comes to where the oversizing factor is high enough that it is always cycling on and off rather than modulating during the spring/autumn shoulder seasons. Part of the specifications on an HSFP test submittal sheet is both the minimum and maximum modulated output at 47°F. It can be important to know your heat load at 47°F when selecting the equipment. If the minimum output of the equipment at 47°F is two times your actual heat load at 47°F, both efficiency and comfort are going to suffer, since the equipment will begin cycling on and off at temps well below 47°F, with wide swings in room temperature.
Using the (crude) BIS heat load calc, do I enter 47°F as the design temp to determine the load at 47°F? If so, the load at 47° is (crudely) ~7500btu/hr (net internal gains). At design temp of 57°F, load (net internal gain) is ~3400btu/hr, or roughly half the minimum output of either the 18RL or 24RL. Is it correct to calculate this way?
Also, what happens when we have 13°F for 10 hours with a high of 22°F? Despite the 99% temp and averages, we do get outlier days once in a while.
Yes- bump the outside design temp in the calcuation tool to 47F, see where that lives relative to the minimum modulated output. If the min output @ +47F is 2x the load @ +47, in a US climate zone 4C climate (where even the January binned hourly mean temps run a balmy ~40F) the thing will be cycling much more than modulating, and subsequently taking a hit in as-used HSPF.
For the outlier days when it's 13-22F, look up the capacity @ +13F, and calculate the load @ +13F, on a room by room , zone by zone basis. What is the shortfall (if any?) Is there as shortfall in every room/zone? Sometimes it may be necessary to leave the lights on in some rooms or run a small electric space heater to keep up for the outlier days.
If sized only 1.25x for a 70F/22F design temperatures, a 48F difference , most systems will still fully cover the load @ +13F , a 57F difference. Note, 57F/48F= ~1.19x, so even with falling capacity with temperature a 1.25 oversize factor should cover it. Equipment step sizing isn't that granular, and most of the time the smallest equipment that covers the load will have some oversizing.
[started new thread] MW.