Options to rectify attic system
I’m in a 1930’s 2 story Colonial in zone 5 (greater Boston area). I have a furnace in my unconditioned attic, along with all of the ducting, which is leaky and has very little insulation. The attic floor is also filled with insulation gaps and air leaks. This appears to be absolutely awful for ice dam formation, so I’m trying to rectify the situation. I’d like to be energy efficient, but I’m worried much more about the risk of major damage from the warm roof in winter. I’ve put a thermometer up there, and when the furnace kicks in, it gets close to 60F when it’s ~25F outside. There’s a separate complete system (both are heating and A/C) in the conditioned basement.
I’ve done a lot of reading here and BSC reports and Building America information, but am struggling to figure out the right way to move forward.
It seems like the most straightforward option is to insulate at the roof deck and attic walls with close cell spray foam. I’m a bit worried about the current state of the roof, since there has been moisture damage in the past , but it has been dry since we moved in around a year ago. I’ll be having roofers out to estimate this week, but even if the roof condition is good, it will be a pretty big cost to foam the whole area and replace the furnace, as the current furnace is uses atmospheric combustion, which seems to be a no-no in an unvented attic.
It seems like I also have the option of removing the attic furnace and using a ductless mini-split and insulating the attic floor with loose cellulose. My wife isn’t crazy about this option because of temperature differentials (the second floor is just 3 bedrooms and a bathroom, and one of the bedrooms is quite small). We’re also a bit concerned about the resale value decreasing with this option.
I’ve also seen good reports of Buried and Encapsulated Ducts, but am not sure how the furnace could be handled. Could I have a platform built above the cellulose that I would put on the attic floor? It seems that even if this does work, I couldn’t use a high efficiency furnace in a vented attic as the condensate could freeze.
Is there some way to heat/cool the whole house from the basement furnace? There’s currently only floor ducts on the first floor. The only possibility or running ducts from downstairs to upstairs that I could think of is building a duct chase in part of the open separation between living room and dining room soffits to carry it across the house perpendicular to the floor joists, but that sounds like it’s going to get expensive and would look a bit out of place.
Are there other options I’m missing, or am I being too hard on some of these options?
Thanks in advance for any thoughts!
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Replies
Regarding the ductless mini-split option, some people use a centrally located head that does a decent job when the doors are open and electric resistance heaters in the bedrooms for when the doors are closed. The resulting average efficiency depends on how often the doors are open.
Have you looked into a ducted mini-split for the upstairs?
Another option to look at would be small-diameter, high velocity ducts from the basement to the upstairs. These are sometimes feasible to retrofit where regular ducts aren't feasible. They do result in higher pressure drop in the duct system. You can't just hook up your existing furnace to them. They need a furnace designed to work with the more constricted ducts. The purveyors of these systems claim that the higher air velocity improves temperature distribution.
If you determine that the roof does need to be replaced, consider a layer of nailbase rigid foam on the outside and fluffy insulation on the inside to bring the attic into the conditioned space.
Important information necessary to finding the most optimal solutions is the room by room heating & cooling loads for the upstairs, excluding the portion of the loss attributable to the duct losses to the unconditioned. It should be calculated using aggressive assumptions about air leakage & R-values for the "after air sealing & insulating" condition.
If only heating (and not air conditioning) is the goal, it may be cheaper & better to run a hydronic heating system (possibly even driven by the water heater) rather than continuing to use the ducts, which only aggravate the air sealing & insulating problem. With the ducts & air handler out of the way sealing the upper floor ceiling and insulating the attic is dirt cheap (and subsized by MassSave), whereas insulating at the roof deck is expensive (and un-subsidized. It's pretty easy to be north of $10K insulating at the roof deck to code minimums, but odds are pretty good that with at least R30 on an air-tight attic floor (less than a grand, after rebates) the upper floor loads would be well-within the burner output of just about any water heater, and it's likely that it could even be micro-zoned on a room-by-room basis for less than the cost of insulating at the roof deck. For the record, what do you have for a hot water heater?
But let's figure out the room loads first, which will give a much better idea of what we're dealing with.
If AC is a must, run the cooling load numbers too.
BTW: Insualating framed cavites with closed cell foam is expensive and provides very little marginal performance compared to open cell foam despite the much higher center-cavity R values, due to the thermal bridging of the rafters & studs. Save the high R/inch foam budget for continuous foam above the roof deck or sheathing the walls, where it delivers the full-R. Closed cell foam uses 2x the polymer per R of half-pound open cell foam, and is blown with envrionmentally unfriendly HFCs, whereas open cell foam is blown with water.
When talking about "attic walls", does this mean there are kneewalls, or are you talking about uninsulated fully exterior walls?
The fact that you need roof work anyway makes it good timing to consider insulation at the roof. An economically and environmentally preferable way to do that would be with reclaimed foam above the roof decking. If you go with spray foam under, specify Lapolla 4G foam to avoid the major climate impact of the blowing agent used by others.
Otherwise, ducted minisplits are a good option to consider. Another way to avoid ducts is a chiltrix air to water heat pump system.
http://www.chiltrix.com/
You can then deliver heat with any mix of conventional radiators (silent, which is nice for bedrooms) and fan coil units that are sort of like mini-split heads but can be smaller and slimmer. The fan coils can heat and cool.
The only drawback of Chiltrix is that it's hard to find a qualified installer.
Edit: As Dana pointed out, the hydronic system need not be a heat pump like Chiltrix--it can be simply a combustion boiler.
Thanks for all of the feedback!
I just performed the room by room Manual J calculations with no duct losses and assuming that I fix some of the other issues like lack of wall cavity insulation. This is what I came up with:
Bedroom 1: Heat=3577, Sensible Cooling=1265
Bedroom 2: Heat=3289, Sensible Cooling=1089
Bedroom 3: Heat=2185, Sensible Cooling=990
Bathroom: Heat=1115, Sensible Cooling=311
Hallway: Heat=939, Sensible Cooling=574
TOTAL: Heat=11105, Sensible Cooling=4229
I haven't done these before so hopefully they are reasonable -- the area is around 700 sqft for reference. I'm not sure why no latent cooling numbers were populated. When I did this with the current conditions (albeit a lot less carefully), I got a heating load of ~52,500--duct loss alone is more than the total system load if I can tighten things up!
I have done a quick look at ducted mini-splits, but it seems like that keeps the problem of equipment in the attic, so I'd have to insulate at the roofline. Is that a wrong assumption?
I hadn't thought about giving up cooling, and am not sure what to expect in terms of indoor temperatures in that case given that there will be cooling in the first floor. Is there any way to get an idea of that? I could see an argument that with a tighter house, the cool air from downstairs would make it upstairs, or that the hot air can't make it out through the walls so it gets rough upstairs.
Regarding the hot water heater, I'm not sure what the important numbers are, but it's a 40 gallon natural gas heater with an energy factor of 0.59, and is atmospheric. Looks like it was manufactured in 2013 (before I bought the house).
I'm not sure of the terminology, but there are no knee walls--I'm talking about the sides of the attic that are not sloped. This might be called the gable walls? Right now they appear to be something like exposed 1x10's laid horizontally as the only thing visible behind the studs. There are also old single pane windows, one on each side, which I'd have to either remove or replace if I extend the thermal envelope.
On the topic of open vs. closed cell foam, I think that the rafter bays are around 7.5", which seems to get me up to around R-27, even if I don't count the space needed for a vapor barrier. Would the foam just continue past the rafters? Or would we put some rigid foam or lumber to extend the depth of the bays to get the insulation up to code levels?
It is seeming like things would be easier if it turns out to be close to needing a new roof, but I'm not sure what the roofers will find. Most of the roof looks good to my untrained eye, there are just a few sections that look questionable.
I'm attaching a floorplan with the loads by room in case that helps. The ceilings are 7.5', so I don't think that there's much room to lower them for ductwork or anything along those lines.
I had the roofers come and the roof's apparently in good shape, so I can't see it making sense to rip up the roof to apply rigid foam above the deck.
I'm opening up to the idea of mini-splits, but this isn't going to be a tight home--the walls are currently 2x4 studs with no insulation and will get dense pack cellulose, but that's far from the levels I see people on here comfortably living with a single point source. The only feasible place i see is building something from the left wall to out to around the '9' in 939, pointing towards the bedrooms. Would this be reasonable, or is there another location or pair of locations that might make sense?
Insulating at the roof deck is looking like the most promising solution, but one concern is around where the rafters meet the edges of the house. It's a colonial house with a very simple roof structure, but there is about 1.5' diagonal on the front and back of the house. In these areas, I wouldn't be able to fit as much insulation there, and I'm not sure if that would affect formation of ice dams. I'm not sure of the terminology for this, so I'm not sure if this has been written about before. Is there a recommendation of how to handle these areas? Would insulating at roof floor or ceiling be better in this case?
Mike,
You'll have to either accept some compromises somewhere, or bite the bullet and invest in one or more expensive solutions. You have a pretty good handle on all of the issues you face.
If the space at the perimeter of your attic is too tight for an adequate thickness of insulation, the best interior solution is to use closed-cell spray polyurethane foam. However, if the insulation layer is less than required by code (usually R-49), there is no guarantee that you won't have an ice dam. If you can afford the work, installing a thick layer of rigid foam on the exterior side of the roof sheathing is always the best solution to ice dams.
You may have enough space to put a 3/4 ton ducted mini-split head in the closet between the MBR and the largest bedroom, with enough total load that it could run reasonably efficiently.
Assuming the left side of the image is a stairwell, there may be enough space to install another mini-duct cassette in the hall closed (is that what that is?) or the closed in the larger of the two smaller bedrooms, with a duct running on the left wall near the ceiling to serve both. It's not enough load to keep it from doing more cycling than modulating, so it's effective COP is likely to average maybe 2.5.
Individually zoned hydronic air coils running off a condensing hot water heater can be sized reasonably for the room loads, but you'd still likely need some sort of air conditioning. See:
http://www.htproducts.com/fan-coil.html
A 3/4 ton or 1-ton ceiling cassette type mini-split head mounted in the hall ceiling could probably cool the whole house. You'd have to leave the bedroom doors open during the day during the sensible load peaks, but it would dry the place out and keep the temps well bounded after dark even with the doors closed. See:
http://portal.fujitsugeneral.com/files/catalog/files/9RLFCC2.pdf
http://www.ecomfort.com/manuals/suzslz09submittal.pdf
The ceiling cassette could also be used for auxilliary heat or shoulder season heat- more if the bedroom doors can be left open.
An alternative to the hydronic coil solution would be flat panel low-temp radiators running off the hot water heater, and for plumbing simplicities sake sizing them proprotionally to the load and running it as a single zone. The panels may be more expensive than the coils, but you would save on electrician labor. With 120-125F water the radiators put out about 1/3 the amount of heat as their 180F rating, but it's likely you have plenty of wall area to be able to size them correctly.
There are alternative vendors other styles, but for example, an Ecostyle B-24.56 (24" tall x 56" wide) or B-36.40 (36" tall x 40" wide) could handle the MBR load, and another for the large bedroom, a B-24.40 (24"h x 40" w) could handle larger of the smaller bedroom loads, and a B-20.24 (20"h x 24" w) could cover the smallest bedroom. You could ignore the stairwell/hall load, let convection up the stairwell from the lower level heat it, or let it run cool, but if you really wanted to a B-24.16 or another B-20.24 would cover it. See:
http://www.qhtinc.com/wp/wp-content/uploads/2013/01/EcostyleTechnical-Booklet-011713.pdf
The B-24.56 s run ~$475 at internet pricing, B-24.40s run -$350, B-20.24 ~$200. There are cheaper options from other vendors. (Myson has a good range of lower-price flat panel rads.)
The heat load numbers are a bit higher than expected (not insanely high), but that is probably due to higher-than-reality air leakage assumptions, combined with failing to subtract the heat output of the occupant(s) and electrical plug loads from the room load numbers (?), or perhaps are they the "before wall insulation" load numbers? What were you using for inside and outside design temperatures?