Laying out mini-splits for a new home
I have a new home in the design stage and have been working with with my HVAC sub and the wholesaler to layout and locate mini-splits for the project. The house will be 2550 sq.ft. on a slab with R40 walls, R75 ceiling and very tight. We are in zone 5, floor plan is attached. We will be installing an ERV for fresh air.
The HVAC guy and the wholesaler want to put a head in each bedroom as well as the entry, office, family room and great room and run them off of 2 separate outdoor units. Their concern is to provide uniform temps throughout the house.
I am concerned that we are putting more equipment in than is necessary. Most of the projects I have seen discussed in these forums have used fewer heads.
Are there any standards for laying out these systems? How are bedrooms normally handled? Any general advice? (especially from those who have lived with this type of system)
Your comments will be appreciated.
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Replies
Terry,
You are being given bad advice. In a well insulated house, installing a separate head in each bedroom is gross oversizing.
Start by requesting a Manual J calculation (heating and cooling load calculation).
You'll probably end up with a couple of ducted minisplits instead of ductless minisplits.
For more information, see:
Rules of Thumb for Ductless Minisplits
Minisplit Heat Pumps and Zero-Net-Energy Homes
Practical Design Advice for Zero-Net-Energy Homes
Are Seven Heads Better Than Three?
The house almost certainly has a heat load less than 25,000 BTU/hr, possibly less than 20,000 BTU/hr. A single better-class 1.5 ton mini-duct cassette might be able to swing the whole thing, but only a room-by-room heat load calculation based on the true U-factors of your high-R house will tell for sure. The Fujitsu 18RLFCD mini-duct cassette is as efficient as some of the best wall-coil type mini splits of just 4-5 years ago (though the progress in the wall coil head types have continued as well.)
Any room with a heat load less than ~5000BTU/hr would not be a candidate for even the smallest multi-split heads, and it takes a load north of 7K to even start thinking about a 3/4 ton head (some of which can deliver 11,000BTU/hr @ -15F.)
But a head in every room will certainly help the contractor cement the relationship with the distributor, and help the distributor pay of his gambling debts! :-)
In general it's best to higher a energy nerd/engineer to do the load calculations, not HVAC contractors. Even though some HVAC contractors will do a decent job of it, most will be nervous about undershooting when assessing a true high performance house. HVAC contractors are looking at a potential down-side to under-sizing (they don't need the 5AM call from the irate shivering customer), and a financial incentive for oversizing. Conscious or otherwise, and HVAC contractor is more likely than not to overestimate the size of the loads, whereas an independent professional makes/breaks their reputation on the accuracy of their numbers.
Oversizing of capacity by up to 50% is OK with modulating heat pumps (and even improves average efficiency somewhat), but beyond that it's just more expense, more wind-chill from higher than necessary air volumes, and less comfort.
Martin and Mr D,
Thanks for your responses. It confirms what I suspected in that my HVAC guy and the distributor are not experienced with using these systems in a new home, especially a well insulated new home. I trust these guys to do their best work but this is new territory for them.
My original concept was to use a ducted cassette placed in a sealed attic space above the master bath to serve the 3 bedrooms and the bath, a wall cassette placed in the great room near the windows to handle the rest of the house and a possible wall cassette in the family room if needed. They both felt that concept would not keep the house comfortable and the temperature consistent.
To their credit, they did run a Manual J calc. which shows a heating load of 27,354 btu/hr. This was for an earlier iteration of the plan that was about 225 sq.ft. larger with almost identical room layout. I think that calculation was probably a little on the high side. I would have put this in the same ballpark as you Mr. Dorsett, 20-25 kbtuh.
I will take your advice and get an engineer involved, now to find one experienced in this type of system and house.
27,354 BTU/hr does indeed seem pretty high for a heat load even on the 2775' version of the house with R40 walls. That's ~10 BTU/hr-f^2, a number that even some code-min houses with minimal window area might hit. If you have a lower than average glazing fraction yours could easily come in under 20K in the footprint-reduced 2550' version. (I don't see great swaths of glass in your floor plan drawing.)
Run an I=B=R load calc as a sanity check. A typical zone 5 outside design temp is ~0F (not sure what yours is.) Assuming that R40 is a center -cavity number (not including thermal bridging) the U-factor of your walls is going to be something like 0.027-0.028, but that's probably on the high side of reality. (With a better description of the wall framing & stackup that can be better estimated.) With an inside design temp of 68F (68 degrees above the outside design temp) that means every square foot of wall loses U0.028 x 68F= 1.9 BTU/hr.
The R75 ceiling has a U-factor of about 0.015, maybe less. Every square foot of ceiling is then worth about 1 BTU/hr.
Measure up the windows and applied the manufacturer's published U-factors. Same for the exterior doors. The heat loss through the partition to the garage will be less. It's not clear what type of foundation, or how it is insulated.
A mystery number is the air infiltration rates, and if they used some default number on their Manual-J it may skew it to the high side. In a tight house with active ventilation you have control over that number, and can change it dynamically if needed.
Build a spreadsheet model, calculating the loads on a room-by-room basis, summing it up only at the end. You can then start subtracting off the room-loads by the anticipated electrical power use in those rooms at night, and the number of sleeping mammals (250BTU/hr per sleepin human, 3.4 BTU/hr per watt for your alarm clock or idling power on your DVRs, average refrigertor power use, etc.) In the end you'll have a pretty good base number to work from, but will still have to take a WAG at the infiltration losses. Most heat load calculation methods grossly overestimate the infiltration losses since it ignores the "heat exchanger effect". If you just call it zero and up-size the equipment capacity by 1.25-.1.5x from your I=B=R methods spreadsheet total you won't be ridiculously oversized, nor will you be cold.
An aggressively done Manual-J would get you more precision, but even there, look carefully at the infiltration loss numbers, and discuss that with whomever does the calculation.
Thanks Mr. D. I did my own spreadsheet (attached) heat-loss in preparation for sitting down with the HVAC folks. If I have done this right, the house has a peak heat load of just over 17kbtu.
I used the EPA guidance for infiltration. I wasn't sure how to handle the floor heat loss so I winged it but I think this is far more realistic than the one that they did. I only have the summary sheet from their Manual J calcs. We will go over it in detail to make sure we have the right inputs this time.