Ductless Downside
homedesign
| Posted in Energy Efficiency and Durability on
How should we deal with Private Rooms?
1. Don’t worry about it
2. Seek clients who are not too concerned about comfort
3. Keep bedroom doors open (especially when occupied)
4. Provide Radiant Panels and Don’t worry about Cooling
5. Preheat the Ventilation Air and Don’t worry about Cooling
6. Undercut the door and pray
7. Provide a Ductless in every Private room instead of the Common Space(s)
8. Provide Transfer Fans between Public and Private Rooms
9. Don’t Do Ductless
10. Other
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Search and download construction details
Replies
John,
The right answer, of course, is "It depends." I'll comment on your numbered list.
1. "Don't worry about it." Nope. This is an issue that needs to be considered and addressed.
2. "Seek clients who are not too concerned about comfort." Nope. Clients will find you. You have to listen to the clients you have, and address their needs and concerns.
3. "Keep bedroom doors open (especially when occupied)." Obviously not. But there is nothing wrong with telling clients that temperatures will be more even if they remember to leave their bedroom doors open when the bedrooms are unoccupied.
4. "Provide Radiant Panels and Don't worry about Cooling." This may be a good strategy in some climates -- for example, in colder areas of Canada.
5. "Preheat the Ventilation Air and Don't worry about Cooling." Nope. I don't believe that ventilation air flows should ever be large enough to use ventilation air for space conditioning. Needless to say, the need for cooling is climate-dependent.
6. "Undercut the door and pray." Nope. I've never heard of anyone advocating that strategy.
7. "Provide a Ductless in every Private room instead of the Common Space(s)." Nope. Unless we are talking about a mansion with rooms like those in Xanadu in "Citizen Kane," this approach would be a waste of money. A strategy using ducted minisplits would obviously be preferable.
8. "Provide Transfer Fans between Public and Private Rooms." The jury is out on this approach. Some experimental homes are now gathering data on this method.
9. "Don't Do Ductless." For the worried designer, this is a perfectly reasonable strategy.
10. "Other." The best approach depends on several factors, including climate, room size, room orientation, area of windows, shape of the house, envelope airtightness, insulation levels, and client expectations. In other words, the designer has to think and listen.
Brooks, in your climate spec the best ducted system. For my area, going Asian can lower energy costs. And there are Asian ducted systems and multi-head set ups.
My bet is Dana could have useful advice for you John.
Unless a room has a design condition load of 5000BTU/hr or greater providing a separate head/zone for that room would be both expensive and unwarranted, and won't be very efficient. Using mini-ducted cassettes to split the output between adjacent rooms can often work.
Using a radiant cove heater right-sized for the room load, controlled by both occupancy sensor & thermostat works for many rooms, particularly lesser-used rooms such as laundry rooms or even bathrooms (given a sufficiently low load.) Cove heaters are far more responsive than radiant panels- they have near-zero thermal mass and nearly 100% of the heat transfer is via radiation rather than convection, whereas with radiant panels a large fraction of the heat transfer is convective. The average radiant temperature in a space has a bigger effect on human comfort than the air temperature- which is why humans can be comfortable in direct sun in a T-shirt on a 10F very calm day. Don't oversize radiant cove heaters by more than 1.5x if you can- you don't want it to feel like a broiler, and it doesn't matter how long it takes the room air to come up to temp.
It takes a lot of cfm to support a significant heat load at a 3F delta-T, so the transfer fans are only applicable in situations where there is substantial overheating on one side of the wall (say, drawing the 90-100F air near the top of a cathedral ceiling directly above a woodstove.) At about 0.018 BTU per cubic foot per degree F, to support even a 1000 BTU/hr load with air only 3F warmer takes( 1000/ (3F x 0.018)= )~ 18,500 cubic feet per hour, or ~300cfm, which is already a pretty substantial flow. Supporting a room load of 3000 BTU/hr at a 3F delta-T takes ~900cfm.
Bumping up the temp of the zone fully supported by the ductless a few degrees to achieve a comfortable temp in adjacent doored-off rooms via conduction through the partition walls works pretty well with low-load rooms.
Latent cooling is more important to human comfort than sensible cooling, and HRV systems that only exhaust air from the doored off rooms, with the ventilation air to those rooms being provided by jump ducts or door cuts gives those rooms 100% of the latent cooling benefit. That probably won't cut it for a west-side bedroom with a high sensible cooling load though- but by minimizing west facing glazed area and providing exterior shades that can still work in some locations/climates.
Bottom line, if you intend to do point source (or multi-point sources) heating/cooling, calculate the room-by-room loads DURING the design phase, and reduce them to the extent possible, and make your peace with the inherent temperature deltas going forward. In high-R houses those temperature differences will often not be too great.
To crudely estimate the approximate delta, assume the U-factor of a partition wall with half-inch gypsum on both sides to be about U0.4 BTU per degree-F per square foot. It'll be about U0.2 for a ceiling or floor that is fully conditioned on the other side. Say a 13' x 12' bedroom with only attic above and crawlspace below has a heat load at the full 70F interior design temp of 2750 BTU/hr, and shares 25' of 9' tall wall+ door with a fully conditioned space. One sleeping human deliver about 250BTU/ reducing the load to about 2500 BTU/R. That's 225 square feet of space, or 2500 / 225= 11 BTU per square foot of partition. At a U-factor of 0.4 for the partition that would impart a delta-T of about 11/0.4= 27.5F, not very acceptable. But if the unoccupied heat load for the room were reduced to 800 BTU/hr, the ~250BTU/hr of one sleeping human lowers that to 350 BTU/hr, or 550/225= 2 BTU per square foot on the partition wall, and 2/0.4= 5F, which in most cases WOULD be an acceptable delta at the 99% outside design condition.
In general rooms on an exterior corner would have higher heat loads due to the greater exterior surface area, making them more likely to be candidate for auxiliary heating (or even a ductless head), and rooms with just one of the narrow sides of the room on the exterior may be amenable to dealing with it by merely reducing the heat load with higher R values better performance smaller windows. The colder the climate, the tougher that approach gets though.
Dana,
Are heaters triggered by occupant sensors effective? The two examples you suggest, a laundry room and bathroom, are usually occupied repeatedly over the course of a day but for very brief periods of time. Surely the user's experience will be one of entering a cold room, perhaps beginning to feel it warming, then leaving before it has become really comfortable. The rooms where it might work better, say family or dining rooms, tend to be part of a larger living space in most newer houses, so I'm struggling to see where the sensors would work.
Dana, your post is excellent.
You have provided some very good food for thought and discussion.
It reminds me of a recent GBA blog by Marc Rosenbaum....
(similar subject)
https://www.greenbuildingadvisor.com/blogs/dept/guest-blogs/practical-design-advice-zero-net-energy-homes
Marc's blog convinced me to take his course ... and I'm certainly glad I did.
I will return with questions/comments
I think ductless is a good direction for not-so-big houses ...
I am just trying to brainstorm on the "Closed Door Dilema"
Malcolm- the comfort comes on within 10 seconds with radiant cove heaters- a bit like stepping in front of a sunny window in a cool room. (It won't pre-warm the toilet seat for you, but insulated toilet seats are available for those folks to whom it really matters.) That will clearly not be good enough for the 5% of all people that Lstiburek estimates can never be made comfortable, but it's really not bad for most of us.
And the higher-R the house, the less it matters- a 65F room where the average radiant temperature is 73F 10 seconds after the the radiant cove comes on feels warmer than a room held at a constant 70F room with an average radiant temp of 70F, but with some directions (floors & exterior walls/windows) being colder. Radiant floor heating is usually more comfortable at any temperature due to the fact that the humans are closer to the higher radiant-temp surface (the floor) than to the cooler ceilings or walls. That's also why big low-temp panel radiators or even wood stoves are more comfortable at any room temp than fin-tube baseboard heating. Direct radiation has a direct effect on human comfort greater than the absolute air temp.
While a radiant cove heater isn't nearly as cushy as radiant floor or panel radiators, the comfort-temperature comes on far more quickly than any system that needs to elevate the air temperature before comfort sets in. Since windows are often the coolest radiant temp in a room, it's usually better from a comfort point of view to mount the radiant cove above the biggest window, raising the radiant temperature from that otherwise coolest radiant temp direction.
Most ducted-air and baseboard hydronic heating systems in code-min houses (or sub-code, like mine) see a bigger delta than 5F between the warmest and coldest rooms on the system, but only the real purists micro-zone the hell out of it or spend the time tweaking the balance to as perfect as it possible can be. With a high-R house you can usually achieve results at LEAST as good as the code-min central heating paradigm with point-source heating if you pay attention to the layout and the loads of the doored off areas. It's not nearly as diffiicult as people seem to worry about.
The past couple of years I've heated most of my 1920s house with a wood stove. With the door closed 24 hours the bedroom with a single ~12 square foot U0.5 window that shares the biggest partition wall in common with the fully heated living room space stays within 10F of the living room with the wood stove even at 0F outdoor temps, with the door open to pre-heat the room during the non-sleeping hours it stays within 5F. Another bedroom with 3x the window area, half the common partition wall area and 4x the exterior wall area needs to have the door open to the common space to stay within ten degrees at 0F outdoor temps. But it's not as if the prior ducted heating kept it within 10F with the door closed either- it needed to be micro-zoned to for it to have a stable and controlled temp with the door closed. (And will be when the hydronic retrofit hits the top of the project list.) But it's not critical I'm happier to sleeping in a 62-65F room than a 70F room anyway (YMMV), so when it's hitting single digits out the work-around is to just leave the door open. That work-around would not be necessary if the room had even R25 whole-wall R (instead of R10) and U0.25 windows (instead of U0.5), I would HAVE to keep the door closed most of the day if I wanted it to run cooler than the common space off the living/dining area (which I would.) I'm sure a lot of people wouldn't be happy until the micro-zoning project was completed, but that's pretty far down the priority list at this point (according to "the boss", and she's always right! :-) ) In a high-R house it wouldn't even be ON the list.
Dana,
"the comfort comes on within 10 seconds with radiant cove heaters- a bit like stepping in front of a sunny window in a cool room."
Perfect . In that case it sounds like a great solution. I'll look into radiant cove heaters for my next build as backup for the wood stove. Thanks.
My thought is sort of a combination of John's #8 & # 9. I'll use a single mini split with several distribution fans (5 works well). Each of the fans will output it's air to it's own duct to a "private" area with return air flow through "jump ducts" to common areas. All of the distribution fans will turn on whenever the mini split's compressor is active (controlled by a current sensing relay). The mini split's indoor unit and the distribution fans will be in a smallish centrally located room that has open grills to common areas at it's top. The total flow through the distribution fans will be slightly greater than the mini split's air flow. Using 5 of the Panasonic "whispergreen" fans allows 450CFM for 35W of power usage. To avoid warranty issues the room the Mini split is in needs to be larger than I'd like but if I choose to locate the fans and Mini split in a laundry room I can meet that requirement as well.
Jerry: At a specific heat of only ~0.018 BTU per degree per cubic foot of air it can take a LOT of fan to distribute sufficient heat to rooms with any significant heating load.
Say the room load is ~2800 BTU/hr. A single Whisper-Greens pushing 80 CFM is 4800 cubic feet per hour, x 0.018 becomes 86 BTU per degree of temperature difference (delta-T). So you can anticipate a delta-T approaching 2800/86= ~33F at the 99% outside design condition, which may be tolerable to some arctic dwellers, but not too many of the rest of us. Conducted heat though the partition walls might reduce that to a 15F delta-T, but that's still a pretty big delta, if tolerable to some.
It pays to do least the napkin-math on any solution before embarking on a path, starting with the heat load calculations.
Dana, far be it from me to try and out-arithmetic you, but I'm not sure those calculations make sense. You're treating the air as a simple BTU delivery mechanism, but not considering that the supply air is replacing whatever air is in the room. A 12*12*8 (1152 cu ft) room will have its entire contents flushed out about every 15 minutes with that 80 CFM fan. Will that entire room full of air really be cooled down by 33 degrees in 15 minutes before the next full air change? It's not about heating the room, it's about moving already heated air in so that the slightly cooled air moves back towards the heat source instead of stagnating.
Dana,
The entire output of a 12,000 BTU/h heat pump (Fujitsu 12RLS2H) is moved with 500 CFM, that is the maximum fan output. The small room may well be superheated but if the airflow out is greater than the airflow through the mini split the airflow out of the room must be moving all the heat coming from the mini split . The "whispergreen" fans can move 110CFM at 9.1W so 5 of them can move 550CFM. . Which should be enough to distribute the output of a 15RLS2H as it's maximum fan throughput is 530CFM.
Resharpen pencils all of you
"Resharpen pencils all of you" I have no idea what this means. My point is that if one wants to "redistribute" the output of a mini split, it can be done simply and at a small energy cost. BUT to be successful one must capture ALL of the output of the mini split BEFORE it is mixed with lots of air, thereby reducing the air's temperature. As Dana correctly points out moving air to move heat requires both volume of air and temperature difference. Unfortunately there is no "small" high efficiency ducted heat pump system that operates at a COP of 2 at -13deg f outside temperature. One can achieve almost equivalent results by dedicating about 8 sq ft to a closet in which a mini split head and 5 "whispergreen " fans are located. When I asked the Fujitsu North America rep about this concept, he informed me that their warranty requires that the indoor "head" be 6 feet from the opposite wall. To sidestep the warranty issue I may use the laundry room instead of a dedicated closet.
Jerry,
You aren't the first person to be mystified by one of A.J.'s comments.
Concerning your minisplit plan: your idea -- to create a hot, indoor sauna room with a ductless minisplit, and then try to distribute the hot air from this closet to several locations with fans -- won't work.
The minisplit is designed to heat a room. It pulls return air from the room in which it is located and turns off when the return air is at the thermostat setpoint. The device isn't designed to bring a room to 90 degrees or 100 degrees. This application would void the warranty. It isn't a furnace.
If the closet is at room temperature (say, 72 degrees), then Dana is right -- you aren't going to be heating adjacent rooms with a Panasonic fan.
I don't think we should be locating the mini-split in a closet....
however, I am not convinced that transfer fans are of little use...
I assume that most of us are talking about high performance enclosures...so...
Considering Marc Rosenberg's guest blog
https://www.greenbuildingadvisor.com/blogs/dept/guest-blogs/practical-design-advice-zero-net-energy-homes
If an open door "helps a lot"
then it follows that a transfer fan could also "help a lot" ... eh?
If 150 BTU/hr is "a lot", then sure, it helps a lot. Mind you, that's less than the heat output of one sleeping human under the covers. That's less than 10% of the heat load of a room with a 2000 BTU/hr load.
Depending on the configuration, a 3F delta-T with 15-18 square feet of open door can buy you more than 100 cfm, and a 10F delta-T buys a heluva lot more cfm. Taking a WAG at it (no real data, just as there was no real data behind Marc's WAG of 50-100cfm @ 3F delta), the heat transfer curve is probably an exponential of about 2 with delta-T- double the delta-T and you get 4x the heat transfer, due to the combination of bigger delta-T and higher flow rates that are induced.
The exponentially higher flow rate with increasing delta-T makes the max delta self-limiting at a dramatically lower point than with a transfer fan. A transfer fan buys you the same cfm independently of delta-T, which is merely a linear increase in heat transfer with temperature difference,
Don't shoot from the hip on this, do the heat load math first, then the napkin math on the delta-T and fan cfm. Don't forget to include the conducted heat through partition walls too, as well as occupant heat (figure 250 BTU/hr per human for a bedroom).
Martin,
I'm surprised at your refusal to accept reality! The "closet" has air inlets above the mini split and fans leading to ducts below the mini split. The fans have a flow rate greater than the flow produced by the mini split's internal fan and these 'circulation' fans operate whenever the mini split's compressor is running. With these conditions the mini split will see input air from above it at the average temperature of the building. The mini split will add or remove heat and eject the heated or cooled air out it's bottom where the fans will take it and send it through ducts. The operating temperature of the mini split will be the temperature of the return air flow from above. The controls on the mini split will function completely normally as it senses the air temperature entering it from above and adjusts it's output accordingly. The mini split is in fact designed to add it's rated output in heat to the air stream generated by it's fan! For a 12000BTU/h mini split with a 500 CFM fan the air through the mini split is heated by about 22 degrees f! By capturing ALL of the output of the mini split BEFORE it is mixed with room air and it's delta t reduced efficient "redistribution" is achieved. FWIW if any part of the system I've proposed is operated near it's limits it is the fans which may see input temperatures of 90+f for substantial time intervals.
John,
Until the HVAC industry recognizes the need for small forced air ducted heat pumps that work at reasonable COP in sub zero outside conditions the mini split in a closet with fans is the moderate cost path to evenly heated and cooled high performance houses.
The algorithms built into mini-split controls for controlling both the compressor speed & blower speed for optimal efficiency will get totally messed up if you begin introducing other flow from other sources, and that's reality. Not only is it likely to run at lower efficiency, it may even cut into capacity, or even sense it as an error condition and shut down.
If you want ducted solution, use a mini-duct cassette and keep the duct runs as short & straight as possible, and balance temperature/flow with louvers. At least that way you won't be screwing with the feedback the thing needs to run at max efficiency & capacity. Sticking a standard wall-coil into a closet is the opposite of a solution- it's stealing defeat from the jaws of efficiency-victory.
Dana,
I respect your expertise but I'm sorry, this time you are flat out wrong! Those "algorithms" are all based on the input of a temperature sensor and that sensor is not in anyway affected by the presence of added airflow. The added airflow has the effect of moving a much larger volume, the whole house, past the mini split which simply doesn't know that it is within a closet and fans are doing the moving. If there were a competent ducted solution I'd be glad to use it, there is NONE!
Dana,
I grant that the efficiency is reduced by the added energy used by the fans and control. The penalty is about 40 watts which at full output is quite small, under 5%, but under minimum load it may approach 20%. I seriously doubt that using resistance heaters to balance a central mini split results in lower total energy use.
Jerry, When are you going to start building so we can have a real world example to evaluate?
I made a simplistic simulation and this is what I found:
Given a 10*10*10 room with two R15 walls, an R50 ceiling, an R20 floor (facing unconditioned air below), and two R1.5 indoor walls, indoor temps of 72 degrees, outdoor temps of 25 degrees...
The room cools off logarithmically, approaching about 65.8 degrees and basically reaching that within 20-25 minutes.
Add a transfer fan that blows in 50cfm of 72 degree air, and the room still cools off with a similar curve, but it levels off at about 67.4 degrees, and it takes a few less minutes to get to the leveling off point.
Make the fan 100cfm, and it levels off at 68.3 degrees.
I experimented with other R values, other CFM rates, other temperatures, and basically the theme is the same: the fan does help, but maybe not enough to be worth the hassle. The trend seems to be that you need around 100cfm to cut the temp differential roughly in half, and doing much better than that gets into ridiculous cfm numbers -- the cfm-to-temp-differential doesn't scale linearly.
Here is the code if anyone wants to scrutinize it:
without fan https://gist.github.com/mackstann/1d8b3fc2b01341e91244
with 50cfm fan https://gist.github.com/mackstann/3841417a2814ca37ece2
Here's an example Not-So-Big ...
High Performance Enclosure & Windows...
One Ductless Mini-split ....
Climate may vary (probably not Fargo)
What happens when we add partitions... and doors?
1. No Interior Walls ... One Ductless Mini-Split ...No Problem
2. Begin to add partitions but no doors... just a 32" wide x 9 ft tall opening.
Is there a potential comfort problem in the Northeast corner Alcove?
3. Fully partitioned with doors (closed at night)
Likely not-so-comfortable in extreme weather
I am with Malcolm Taylor, Jerry, you need to build your closet idea and then report back. personally I agree with Dana that a ducted split is designed to be a ducted split.
As to all the hoopla about cold rooms, I have built many a home that does not include ducts. And many of these homes have unheated rooms that are warm enough depending on some things but basically most rooms are liveable.
Much of what we build involves tradeoffs. Using one or two mini splits to heat a home is a tradeoff that favors lower installed costs over good but not perfect distribution of tempered air...
Budgets. Every project has a budget. And then the list of wants and needs. I am very much open to building homes using mini splits or not.
It simply is a choice to be made. And then it's time to build.
aj
Nick Welch, you get my "sharp pencil award." We need you here along with Dana. Both of you do a fantastic service crunching the numbers for the rest of us to examine and critique. Thank you Nick and Dana. Dana the postulating bro and Nick the puter programming back checker bro. I like it.
As to any "arm chair pontificator" that doesn't understand "sharpen you pencils all of you" not to worry just keep posting your nonsense with your dull pencils and be happy. ;) Really love yaa all as you're fun to read anyway.
aj
So this fall I did a small experiment that is relevant to this discussion
The minisplits were installed purely for AC, not believing we could get the house insulated well enough to do without a boiler.
I have attached a sketch of the floor plan
2800 foot house circa 1970, 'Deck House" clone raised ranch
750 square feet of glass[!]
spray foamed 2x4 walls
iso foamed roof ~R30
Current fuel use without minisplits 600-650 gallons of oil yearly heat and hot water, family of 4, 69 deg thermostat 24x7
3 Mitsu minisplits, 2x 9k, 1x24k
Because of internal politics I was not able to continue any testing past Dec 1st, bu tit did get into the lower 20's F so I was able to get a good idea of what heating with the minisplits would be like
I was hoping to compare what the cost of heating was like compared to the Buderus condensing oil boiler, but really short of some accurate metering of the units themselves this proved impossible. What I did find out is that I really need to get the shovel out and insulate the foundation walls, as the downstairs, despite being smaller, used much more heat than the upstairs
The downstairs experiment did not make it through november as the bedrooms[doors open] drifted about 2 degrees below the setpoint on colder nights[high 20'sF]and SWMBO objected
The upstairs stayed completely comfortable through November, with no noticeable temperature variations at all, again doors open
When I say no noticeable variation, I mean less than 1 degree anywhere, and I suggest examining the floor plan. The large living room unit has a spot behind it in what we call the kitchen ante room, that is nearly 25 feet away, directly behind it and it was perfect. look also how far and around the corner the master bath is from the bedroom unit, and it was not the least bit cool, and it has 10 feet of outside wall, and 8 feet of 2 foot overhanging the outside[with R9 foam only] as the entire outside overhangs 2 feet in every direction
Attached also is a pic of the NW side of the house, in case you thought this was not a stupid 'why did they build a house like this in New England' house.
So my conclusions are that with an open door lifestyle, this setup in this house would most likely be perfect for 80 percent of the winter, probably more and virtually 100 percent with additional foundation insulation and some level of air circulation down stairs. I have no reason to believe that the upstairs would fail to perform at any outside temperature. We in fact had a oil line freeze the first winter at approx zero degrees, and the units performed well for the 12 hours they were required, although I was not paying close attention to details
Now, I think this leads me to surmise that most of the objections to the use of minsplits in a PG or Passive house are not well founded.
Hi Keith, thanks for posting your experience ... it me reminds me of Marc Rosenberg's post(see attachment)
I am thinking that Multi level homes seem to have more Issues .... (because cold air sinks and "pools")
and just adding a Mini on every level does not completely overcome the stratification problem...
Funny I remember reading that post and thinking that there must be a leak in that room
One of the things that to me justifies aiming for very low leakage numbers is that while we kind of mentally assume that air leakage is evenly distributed, it is much more likely that it is uneven, so you will get a cold room that you can never quite trace the cause of. Some detail didn't get detailed after lunch like it was supposed to
[Edited for stupidity: Whoops! somehow missed the above comment. Thanks for the simulation.]
Nick, regarding your comment above (#11) about a room-to-room transfer fan completely exchanging a room's cooler air with warmer air: I've been wondering exactly the same thing. Have you (or anyone else) gained any insight into this question?