Minisplit sizing

[KevinJ]

Hi all-

I'm looking for a little advice on sizing 2-3 minisplits for my 1940's Colonial. Primarily so I can check numbers against bids, also incase I install the heads and have a pro do the lines. It's roughly 1800 sq feet, with new windows, spray foam in the attic, cellulose in the walls and a full basement on slab. I did my own manual J and I came out to 30K cooling load.

We currently have 2 10K window AC's (window in kitchen, 2nd floor landing), which do a really good job keeping the house cool. Yesterday was 90+ and the house was 73 no problem. The problem is obviously noise, and that the air doesn't get in to the bedrooms. So the nursery and master both have their own window unit.

I really don't want to oversize the units from reading threads on this site. Knocking down the humidity would go along way for comfort. Initially I was thinking 4 heads (Kitchen, living room, master, 2nd floor), however it seems like that would be way overkill and I should be able to get away with 1 per floor. Is this accurate?

I was thinking of doing a 24K 3 zone hyperheating Mitsu, with 12k heads in the kitchen and top of stair landing. This would also allow me to add a third 6k in the master if I don't feel like it's keeping the room cool enough.

In your experience, will this work out pretty well? How about bedrooms if doors are closed when guests are over? I don't want to have to be putting in window units to supplement. Thanks.

 

[WorthFlorida]

The first thing is you better have your wife, if there is one, sign off on this. The heads all over the house are not very attractive. They don't go away in the off season like window units.

You'll definetly will need two heads on the second floor. Heat from the first floor rises and the second floor will always be warmer. As far as the first floor probably two for it a good size area. What is nice is each head has its own thermostat for better control of the cooling needs. I know very little about split units but this many heads is going to take a lot of copper lines and that is where it gets expensive. Also, how will they be running through out the home? What you may need to look into is the maximum length of the line from the compressor the furthest head allowed. As far as sizing the Mitsubishi qualified dealer should come up the correct BTU's needed.

 

[Sponsor]

 

[Dana]

With 1800' of conditioned space it's highly unlikely to have a design cooling load of 24,000 BTU/hr, let alone 30,000 BTU/hr in a tight house in MA (or even in FL!) That's a ton per 720' of space, which would indicated wretched thermal performance for the house itself. Typical square foot per ton ratios on carefully done Manual-Js on a range of house sizes on homes in the southeast (mostly the Gulf Coast states) can be seen in this graphic based on data compiled by a home energy consultant in Georgia:

If you look at the cluster near 2000' the middle is about a ton per 1000'-1200' or so. Only the very worst houses are in the ton per 700' range. These homes are in locations with significantly higher latent loads and significantly higher 1% outside design temperatures than anywhere in MA.

In all my experiences with local MA mini-split installers, only couple have bothered to run a Manual-J, and in those cases it has mostly been pro-forma- all assumptions at code-min R, with code max air leakage. It's important to look at both shading factors and the true SGHC of the windows when running the calculations.

Are you planning to heat the place with the mini-splits as well?

 

[BillTheEngineer]

Looking at your chart are you suggesting that possible appropriate sizing is in the 1000 to 1500 sq-ft per ton?

Not looking to hijack the thread but I am also looking at putting in a mini-split system in my home. Mainly to cool but possibly heat too.

I have a friend that uses a mini-split to heat with but based on his utility bills and mine it would seem less expensive to for me to heat with gas than the mini-split. I have not done much research to determine where a crossover temperature, but understand the shoulder seasons is where the mini-split should do its best.

I am assuming a manual J for each zone would be the proper way to go when sizing the indoor units. Do you have recommendations on how to design a system (things to not do and things to do)?

 

[Dana]

Looking at your chart are you suggesting that possible appropriate sizing is in the 1000 to 1500 sq-ft per ton?

Not looking to hijack the thread but I am also looking at putting in a mini-split system in my home. Mainly to cool but possibly heat too.

I have a friend that uses a mini-split to heat with but based on his utility bills and mine it would seem less expensive to for me to heat with gas than the mini-split. I have not done much research to determine where a crossover temperature, but understand the shoulder seasons is where the mini-split should do its best.

I am assuming a manual J for each zone would be the proper way to go when sizing the indoor units. Do you have recommendations on how to design a system (things to not do and things to do)?

It's possible for the appropriate size in Allison Bailes' graphic to be anywhere from a ton per 600' to a ton per 2000', for the houses smaller than 4000' (but note there is an outlier that comes in at about a ton per 3300' in a ~3300' house!) The problem with dumb rules of thumb is that they can be pretty far off the mark, and usually are. A Manual-J on the actual house using aggressive assumptions, giving credit wherever warranted for reducing the load (as per the Manual-J instructions) is going to be a lot closer. But the middle ground for 2000-4000' houses appears to be about a ton per 1300-1400'. Houses under 2000' are closer to a ton per 1000-1200', but there are surely going to be outliers that vary widely from that number.

And that is for Gulf Coast style latent loads at 1% outside design temperatures well into the 90s , a good 5-10F warmer than Long Island's 1% outside design temperatures.

If you oversize a mini-split head to where it's minimum modulated output at +47F (as spelled out in the efficiency test submittals) is substantially more than the LOAD at +47F it won't make it's HSPF numbers. For maximum efficiency it has to be modulating, running nearly continuously rather than cycling on/off whenever it's in the 40s F or warmer.

To size a mini-split head for a specific room or zone one needs to first calculate the load at the 99% outside design temperature, then consult the extended temperature capacity charts to verify that it covers that load. On L.I. the 99% outside design temp is +15F which is pretty close to the +17F . the other outdoor temp used for efficiency & capacity testing. For your location you can just go by the "nominal" heating capacity of the mini-split. (It's required that it at least be able to deliver the nominal rated output at +17F outdoors, 70F indoors in order to use that modulation level for efficiency testing at +47F.)

When using multi-head systems, keep in mind the at the compressor has a minimum output too, which is considerably higher than the minimum output of any individual head. If the whole house load at +47F is much lower than the minimum modulation of the compressor, you be cycling both the compressor and the ductless heads, taking an even bigger hit in efficiency. The "mini-split head for every room" approach invariably oversizes most of the ductless heads and the compressor, leading to much lower as-used efficiency.