GBA Logo horizontal Facebook LinkedIn Email Pinterest Twitter Instagram YouTube Icon Navigation Search Icon Main Search Icon Video Play Icon Audio Play Icon Headphones Icon Plus Icon Minus Icon Check Icon Print Icon Picture icon Single Arrow Icon Double Arrow Icon Hamburger Icon TV Icon Close Icon Sorted Hamburger/Search Icon

Community and Q&A

Ductless minisplits – Zone 5 – with basement

Brian Martin | Posted in Energy Efficiency and Durability on

Hello folks, this is my first post at GBA so bear (beer) with me. My son and I are building a new for my wife and I to live our days out in, two story bungalow style with a full basement under.

Some stats;

Exterior Walls – 2×6 16″ on center with 2′ XPS on exterior, wet blown cellulose in the bays. 3/4″ rain screen furring over foam, cedar lap siding. R19 in the bays + R10 = R29+

Ceiling – 2×10 sealed vault flash (core bond) + batt – R49

Basement Walls 8′ (6.5″ below grade/ 1.5″ above) with 1.5″ polyiso + R13 Batt in 2.4 stud wall over foam – R22

Basement Slab – 2″ XPS

Windows/Patio Door – Milarg Essence casement average U25

I have exhausted myself with spreadsheet upon spread sheet trying to come up with my own head load calculations, all from data on this site. Certainly my methodology is imperfect, but how much so I question. The below grade basement walls I used a Delta T of 20 degrees assuming ground temps of 50. Same for the basement slab.

Delta T used for all above ground surfaces was 82 degrees. Design Day = -15

I calculated in a 25% framing factor, the window U factor and ran my heat loads for each floor. I then ran heat loads by floor as I was hoping to use Mitsubishi Hyper Heat mini splits. I did it this way because of the extremely open layout of the main living floor, and the understanding that for the upstairs it appeared that more than a single head split would be overkill anyway despite the two separate rooms (bedrooms) with doors.

Unless I did something terribly wrong I shocked myself by seeing how much my weighted average R/U factor dropped as a result of the windows and doors. Here’s the data Assuming 8′ walls, and adding the floor joist bay to the basement above grade calculations, and adding the 1′ floor space between first and second floors;

Total First floor wall surface = 1,476; 906 @ R30 wall cavity, 308 @ R17 (framing factor) , and 268 @ R 4.348 (window r.o.)

I came up with a weighted average of R22.8 WOW!

Did the same for upstairs. R24.84 The upstairs has a 2′ knee wall before the vault starts. Added the 1′ of floor cavity to the upper level (had to choose one)

So, in trying to size the heating options for mini splits to within up to 50% greater BTU’s than the house calls for, I’m left with trying to calculate air leakage now, but having even more frustrations with that.

Additionally, as you can see by the attached floor plan’s, I plan to heat the house primarily with a wood stove in the living room opposite the stairwell wall. A Pacific Energy Alderlea T5. Of course for lazy days, vacations, and resale, I want to be sure that I have another “primary source” in the mini splits should for some unforeseen reason we sell the place.

To make it clear, I do plan on having an open stairwell both up to the 2nd floor bedrooms and down to the basement. I’d like to heat the basement, although I see only utilizing a portion of the basement marked laundry room/bathroom for now, and possibly the room marked family room down the road.

I came up with the following heat loads not including any air leaks being calculated yet. The house will be presumably tight, using Prosoco caulk at all window bucks, sheathing seams, top and bottom plates, rake wall plates, etc. We’ll do our very best.

Basement Wall Surface Area = 1,458 sq. ft., of which below grade is 1,053, above grade 339, and windows (egress) 66. And a slab of approx 1,330 sq ft. For the walls I come up with a weighted average R factor of just 20.06.

Again, the delta (if I’m correctly doing this) for below grade is 23 degrees, given ave soil temp of 45, same for the slab, though I know at the near ground surface the delta may be higher due to frost layer.

U VALUE’s

– Basement

Below Grade = U.0476 x 1,053 sq. ft. x dt 23 = 1,152 btu
Above Grade (weighted for windows) = U.0567 x 405 sq. ft. x dt 83 = 1,905 btu
Slab = U.1 x 1,330 sq. ft. x dt 23 = 3,059 btu

*Total Basement BTU = **6,162
* Don’t plan on heating the entire basement unless useless to try not to.
**surprising how much the floor need here.

How many more BTU’s to add for leakage? 4 egress windows.

– First Floor Living Area – Wide open other than office with french doors.

Total Wall Surface Area x Weighted Average U x d.t. 83 = BTU + leakage

1,476 sq. ft. x U.0438 x d.t 83 = 5,365 BTU

– Second Floor Bedrooms Area – 2″ knee then vault begins, full sealed vaulted ceiling

Walls = 954 sq. ft. x U.0376 x d.t. 83 = 2,977 BTU + leakage
Ceiling = 1,644 sq. ft. x U.025 (24′ o.c. 20% framing factor on R50) x d.t. 83 = 3,411 BTU = Leakage

Total Upper BTU needs 6,388 + leakage

I’m exhausted !!! But there is more….

How to heat the basement. Baseboard electric? Small pezio electric started smallest gas decorative stove. I have a 12,000 BTU unit in my carriage home. The gas with no electric assist would act as a fall back for the whole house should we be gone and the electricity goes out deeming the mini splits inoperable.

How big and where to place mini split(s) for main and upper heat. And cooling, not that Bozeman requires much cooling at night when you’re sleeping.

Two splits, one upper and one lower, and some basement heat source is my guess

But the leakage question???

Looks like I can’t post attachments here.

GBA Prime

Join the leading community of building science experts

Become a GBA Prime member and get instant access to the latest developments in green building, research, and reports from the field.

Replies

  1. Brian Martin | | #1

    Trying attachments again.

  2. User avatar GBA Editor
    Martin Holladay | | #2

    Brian,
    First of all, it's hard to make a flash-and-batt insulation job between 2x10 rafters (9 1/4 inch depth) add up to R-49. I get R-43 -- although if you are willing to install as much as 6 inches of closed-cell spray foam before you switch over to fiberglass, you might get there. But that's expensive -- and you still have thermal bridging through your rafters.

    For more information on pencil-and-paper methods for performing heat-loss calculations, see:

    How to Perform a Heat-Loss Calculation — Part 1

    How to Perform a Heat-Loss Calculation — Part 2

    The second of these two articles provides detailed instructions for applying infiltration factors when calculating heat loss calculations with a pencil and paper.

    If you want a more accurate result -- and it probably makes sense to have one, if you are planning to build a relatively tight building -- you could hire a HERS rater or energy consultant to perform the calculations with software.

  3. Brian Martin | | #3

    Thought about the rafters, and as mentioned above, 6" of foam = R42 + a 3.5" batt at R13, provides R55. So I should have stated my vaulted and sealed ceiling would be R55 (6" foam = R42 + 3.5" batt =R13). Of course there is the thermo bridging with the probability of bringing that down 15% to 20% on a weighted average. But the batt would be held perfectly secure by the drywall, this I see as a benefit. The alternative is an 117/8" rafter/I-Joist. With 5" of foam (R35) + a 5.5" batt (R19), the ceiling would perform about the same as above, but I'd have the additional cost of the 117/8 I joist and the batt would be suspended in the cavity with some mechanical assist, with an air gap of an inch or so. Given the additional expense of the larger joist/truss along with the air gap issue vs. the cost of the additional 1" of foam, my thoughts were to go with the 2x10 dimensional lumber. Now wondering if an "I" joist would have less thermo transfer than dimensional lumber. I think the answer is yes. So what to do....

  4. User avatar GBA Editor
    Martin Holladay | | #4

    Rigid foam on top of the roof sheathing is the best way to go.

  5. Alan B | | #5

    You can use Hot2000 to do heat loss calculations, its free and not hard to use, your method may be more accurate (i find it overestimates basement losses, either its wrong or i am not using it properly) but you can do upgrades virtually and see the results instantly.
    Windows seem to raise the design load quite a bit, so improvements in R value (within reason, don't spend 5 figures for window upgrades) or smaller or fewer windows make an appreciable difference.

  6. D Dorsett | | #6

    A bunch of things aren't really adding up. There is no US climate zone 5 location I'm aware of that has a 99th percentile temperature bin as low as -15F (that's more typical of the cold edge of zone 6), and I'm not buying your U-factors for the framed assemblies.

    Note, R19 batts compressed to 5.5" in a 2x6 assembly only perform at R18, not the full-loft R19, and they're more of an air filter than an air retarder. It's low enough density to lose performance at high delta-Ts to convection within the batt, even if installed perfectly. R21HDs or R23 rock wool would be a better choice there.

    With R21s or R23s in a 2x6 studwall 24" o.c., with sheathing & gypsum & siding comes in at about R14.5, R15 whole-wall if it's the full-on advanced framing package with no doubled-up headers or top plates, no narrow stud bays. Add R10 foam to that and you're at about R25, R26-ish if you want to count the air films. Figure on a wall U-factor of aboutU0.038-U0.040. If it's 16" o.c. framing it's higher, about U0.043. With R19/R18 batts and 16" o.c. framing bump that to U0.045.

    I assume you meant U0.25, not U25 for the doors & windows.

    Unless you have a lot of blocking or skylight framing or something your ceiling framing fraction is likely to be under 20%, but probably over 15%. I you have the full framing plan you can determine that fairly accurately without a lot of work. With a flash inch of CorBond and 8.25" of R4.2/inch high-density fiberglass the center-cavity R of about R41. With lower density batts you'd be in the high mid-30s.

    Using a single U-factor & delta-T for the sub-grade concrete etc. is grossly inaccurate. In a location with a -15F 99% outside design temp mid-winter temp 2' down is going to be below freezing, whereas in the middle of the floor slab the temp could still be slightly above the deep subsoil temp. While there are no really great simple models for the thermal conductivity & thermal mass of the soil (which will vary by soil type & moisture content), the approach taken here is going to be pretty far from reality.

    Most infiltration loss methods overestimate the true heat loss. Like subgrade losses it's impossible to have a method that is both simple & accurate. The "heat exchanger effect" is widely ignored in most heat loss calculations, and presume that the incoming air reaches the interior still at the outdoor temp, and the exiting air escapes to the outdoors at the full indoor temp, which is a worst-case. It would be true if the leaks were all large holes exchanging no heat with the building materials, but that's not a typical air leak in most modern homes. Both the incoming and outgoing air are exchanging heat along the paths.

    Try modeling this as best you can with BeOpt (it's DOE2 under the hood) or Hot2000, both freebie downloads. (google 'em.)

  7. Brian Martin | | #7

    Yes, sorry. I'm in Bozeman MT zone 6b and yes the Delta T I'm using is 83, assuming -15/99% and comfort temp at 68 degrees. Thank you all for your valuable input, can't tell you how I appreciate getting put on the right track here. We are in a seismic zone equivalent to California, so I'm not sure about alternative wall construction methods such as advanced framing. I do know there is a green builder here that does build double walls so I'll contact them to garnish some additional info about their methods and experience. But from what I've read on this site, it appears that double wall techniques without exterior foam board yield much higher moisture levels on the sheathing, the colder the climate to worse that condition. Now being that we are a "b" zone, perhaps the relatively low humidity here would deem those comparisons not fair. Lastly, if Milgard is stating U.025 to .026 why is it that would be different than the framed assembly, and is there a simple way to calculate surface area of glass and use a multiplier to determine the framing assembly losses for a nice fiberglass frame?

    Thank you, Thank you, and Thank you all!! Brian Martin

  8. User avatar GBA Editor
    Martin Holladay | | #8

    Brian,
    Q. "If Milgard is stating U.025 to .026 why is it that would be different than the framed assembly?"

    A. Typical window U-factors range from 0.10 to 0.40.

    A U-factor of 0.025 for a Milgard window is extremely unlikely. I'm not sure what you mean by a framed assembly. If you mean a wall, then a typical U-factor for a wall is 0.07 to 0.025 -- but you won't see U-factors that low for a window.

  9. Brian Martin | | #9

    Martin, thank you for your patience with me. The U-factor for the Milgard Essence windows that I was quoted state between U0.024 to U0.026. I just got off the phone with Milgard direct in Tacoma and specifically asked about the "complete assembly" U-factor. I was told that their method for calculation was from "center of glass" and that the more toward the edge of frame the lower the U-factor was, and that they do not provide a complete assembly rating, nor publish ratings on the NRFC website. I was told that they accomplish these low rating by use of and I89 4th surface coating that increases insulation value while not affecting VLT numbers, keeping the window bright. Also was told that they accomplish triple glaze numbers with double glazing. They also have the window in triple glaze for $20 more a window, but the light transmission is awful imo. I asked if the fiberglass frame was insulated and was told they were not, they have chambers, but are in "essence" hollow. It makes me wonder if hollow fiberglass offers a lower U-factor than the glass, questioning his statement that the closer you get to the edge of the assembly the lower the U-factor. Perhaps a solid wood window assembly would offer a lower U-facotr than fiberglass. Wondering if Marvin "Integrity" insulate their frames? Thank you.

  10. User avatar GBA Editor
    Martin Holladay | | #10

    Brian,
    Either you or the Milgard rep has made a factor of 10 error in reporting these U-factors. I think you mean 0.24, not 0.024.

    For more information on the difference between center-of-glass U-factors and whole-window U-factors, see All About Glazing Options.

  11. Brian Martin | | #11

    Shucks, BeOpt and Hot2000 are windows only programs. Mac guy here. The end result of performing the heat load analysis was to determine by floor of my home of 1,331 basement, 1331 first floor, and 650 upper floor (Bungalow with basement) so that I could correctly size and place Mitsubishi Hyper Heat Mini Splits. I simply wish there was someone with experience with those units, knowing how I plan to insulate, who could tell me what would work in my environment with a chosen delta T of 83 degrees. My guess is that 12kbtu to 15kbtu for the main level and another 9kbtu would be adequate. For the basement my wild guess is some baseboard electric or gas supplemental heating since I don't care if the basement is cooler than the living area's above.

  12. User avatar
    Dana Dorsett | | #12

    It's possible to spec a Hyper Heating mini-split for what you have, but you have to calculate the loads they need to support first. There's simply no dumb rule of thumb that would get you there. You have to calculate realistic U-factors for the assembly types, and apply them to the actual structure- there are no shortcuts that don't carry significant risk.

    BTW: The 99% outside design temp for Bozeman is -12F, not -15F:

    http://www.energystar.gov/ia/partners/bldrs_lenders_raters/downloads/Outdoor_Design_Conditions_508.pdf

    A location with a 99th percentile temperature bin that low will likely have several days per year that bottom out at -18F or lower. The fine-print on the Mitsubishi submittal pages indicate that at -18F they turn off, and won't re-start until the temp rises to -13F. While second-hand internet reporting has it that the real shut-off temp for most units isn't that high it's not something you can just ignore. The competing Fujitsu xxRLS3H series doesn't have that issue, and has higher capacity at -12F than the nearest comparable Hyper Heat.

    Wet spray cellulose comes in at R20 @ 5.5". The typical framing fraction of 16" o.c. framing is about 25%, but if you're required to install mid-level fireblocking or seismic reinforcement it rises to about 28-30%. Assuming a 28% framing fraction 2" of exterior XPS an R20 2x6 hemlock studwall with half-inch ply/OSB sheathing half-inch gypsum, and cedar lap siding comes in at a whole-wall- R of about R23.8, R24.6 if you count the air films, for a wall U-factor of U0.041, not less.

    For simplicity, calculate the basement losses as if you had 4' of basement wall fully above grade, and ignore the slab & below grade losses. Assuming it's bare-concrete on the exterior your stackup is then half-inch gypsum, 2x4 hemlock w/R13s, 1.5" polyiso, and 8" (?) concrete. Since the studwall is not structural and has few windows & doors I'll assume a 20% framing fraction. If one assumes the full R on the polyiso that would yield a whole-wall R of about R19.5-R20, but when it's -12F outdoors the average temp through the foam will be well under freezing, and you're probably looking at about R3- R3.5 on the foam rather than the Rated Rdue to the wildly non linear derating curve of the foam type. While it's average performance will be considerably better than that, from a heat load calculation point of view assume it's no better than R3.5 (R2.3/inch), which delivers a whole-wall R of about R13.6, R14.4 if counting air films, for a U-factor of about U0.07. (If you swap the polyiso for 1.5" of XPS it would drop to about U0.054 or a bit less.)

    Assuming anything less than those wall U-factors would be a triumph of hope over reality. That's just wall area, not window area, which has to be calculated separately based on the manufacturer's specs.

    If you're using 2x10 rafters 24" o.c. with simple gable roof and not a lot of blocking you're probably looking at ~15-18% framing fraction. (If it's a simple shed roof it could be less than 15%.) At 1" with the CorBond + batts on the interior you'd be pushing your luck on moisture accumulation in the roof deck, and pushing your luck a LOT for moisture accumulation in the fiber.

    The cold sheathing problem is all about outdoor temperatures and indoor humidity levels. The low outdoor humidity in a 6B climate as oppose to a 6A climate is a secondary factor. If you keep it at 30% RH indoors it will still be a problem. In Zone 6B for a cathedral ceiling you would need 50% of the total R to be on the outside of the fiber layer to have any dew point margin at the foam/fiber boundary, and with mid-density fiberglass, an R30 compressed to 8.5" (the inch of foam in the 9.5" cavity reduces it to 8.5") would run about R27, to R6 of foam, R33 total. (A high density R30C only has 8.25" of loft and would leave a 1/4" gap with only 1" of foam). Compressing an R38 down to 8.5" would yield about R32 for the batt layer. That's still a terrible ratio for dew point control, and doesn't meet IRC 2012 code-min for R, but you can fix that with an interior side "smart vapor retarder:

    An air-tight layer of Certainteed MemBrain between the fiber and gypsum would be fairly protective though- cheap insurance in the grand scheme of things. For a 1332' ceiling that's less than $200 worth of material, but the labor would run you more, call it a $500-700 cost adder, but one that would keep the drip-drip-drip of melting frost at the foam/fiber boundary from ruining your ceiling, since there wouldn't be enough wintertime moisture migration into the cavity for that much frost to form, and it would keep the roof deck dry enough too.

    At a 18% framing fraction with compressed R38s, 3/4" roof deck w/ shingle layup and half inch gypsum on the interior you're looking at a whole-assembly R of about R28 for a U-factor of about U0.036. At a 15% framing fraction it's more like R29.5 for a U-factor of U0.034. For estimation purposes split the difference, call it U0.035 unless you know the actual framing fraction. If its a simple shed roof with a 12% framing fraction or less you could be looking at U0.030-0.032, not lower. (Code demands U-0.026 to meet it on a U-factor basis.)

    Go ahead and make a room-by-room heat load calc spreadsheet using I=B=R techniques and see where it ends up. Rooms that have an individual heat load of 2000 BTU/hr or more at -12F would need supplemental space heating or their own ductless head to stay comfortable with the doors closed, but any room that has a design temp load less than 7000BTU/hr wouldn't be a candidate for it's own ductless head. Fujitsu has some mini-duct cassettes fully rated down to -5F, but it would be pushing your luck a bit to spec one of those for -12F, even worse at -20F, but with a big enough oversizing factor for the load at -5F it would "probably" deliver, but don't expect any promises from the manufacturer on that. The RLS2s were only rated down to -5F too, but people are successfully heating low-load houses with them in zone 6A.

  13. Brian Martin | | #13

    Dana, thank you! I'm sorry if somehow I misrepresented my intent with the roof. First off the home has not been built yet, and I'm the contractor, so I could just as easily change 2x10 dimensional hemlock to 11 7/8 TGI's or dimensional 2x12, most likely would be the TGI's due to cost. But back to my original plan. I thought about the following from the outside in. 40 year ashphalt shingle, tar paper, tar paper plus 4' of ice and water shield (have 2' exposed rafter tails and code calls for min 3' shield, so the extra 1' gets added), 3/4 CDX roof deck, and as per the entry on this post #3, I would use 6" (maybe 5.75") of core bond , then a R13 fibre batt, 5.8" drywall, latex paint. I have minor concerns about the 4' of ice and water over the roof ends where it meets the wall as this kinda creates a shit sandwich for the CDX, but since the bulk of the roof can dry to the outside other than the 4', and our humidity is low, and climate is sunny, I'm not losing sleep unless you suggest I do. I don't know where the measly 1" of core bond came from, but that is clearly not in the program here. Wondering if the MemBrain would still be a useful product considering the amount of core bond I'll use. An alternative would be to skip the affordable 2x10's and use a 12" product. Then, since R30 batts are 8 1/4", and with the 11 7/8" joists used in the roof I'd have room for 3.5" of Core Bond, all yielding R54.5 should the core bond be sprayed with perfection. I think I'd trust 3.5" of foam to prevent the lions share of dew drive. Seems to me the compression of the batts due to imperfect spraying probabilities would be safer the more thickness of the fibre. What are your thoughts about upspraying wet blown cellulose into the ceiling over the core bond vs. fibre batts? Certainly it would provide a much more controlled loft than fibre over sprayed foam. Although this is fun learning, how would you recommend I build my own sealed vaulted 7:12 pitched roof system, with the caveat being relatively affordable and doable for the average carpentry team such as ourselves? Your answers are extremely well thought out and very valuable to me right now. Brian P.S. I actually shimmed my 2x8 rafters in my carriage house to get the R value's to code using flash and batt. How the inspector approved that one I have no idea. I have since found a new architect for the home i am about to build on the same property. We're now living above our garage. Fun

  14. User avatar
    Dana Dorsett | | #14

    If you're looking at 3.5" of CorBond at $3-4 per square foot you should really be comparing that to continuous 3" of polyiso + 3" of EPS above the roof deck and R30 in the 2x10s and using standard latex paint as the interior vapor retarder. IIRC CorBond runs about 0.8 perms @ 1", which means at 3.5" you're looking at 0.2 perms which is protective of interior moisture drives, but means any moisture that makes it to the roof deck takes forever to get out.

    The exterior foam approach keeps the roof deck MUCH warmer and drier, with much more moisture resilience due to the rapid drying rates through 3-5 perm paint, and even with temperature derating the polyiso it would outperform the R54 all cavity fill approach due to the thermal break over the framing. It takes some long-ish pancake head timber screws to get the necessary 1.5" penetration on the rafters, but if local Yankees near me can do it, you probably can too.

    It would also be a heluva lot greener, since the ccSPF is blown with HFC blowing agents at about 1000x CO2 global warming potential, and both EPS & polyiso are blown with pentane at about 7x CO2 GWP.

    The installed costs of sheet EPS or polyiso typically comes in around 10 cents- R-ft^2, on simple roof designs, so R30 above the roof deck would be about ~$3/square foot, to which you'd have the cost adder of a nailer deck for the shingles and facia board at the edges. I doubt it's over $4 per square foot in any market.

    The dual-foam stackup is important, with the EPS on the exterior of the polyiso, which keeps the polyiso in it's higher-R temperature range. If you flip the stackup the mid-winter performance of the polyiso falls off a cliff, as I alluded to in the basement foam discussion. But with the EPS on the exterior it does OK. The colder it gets the better EPS does, so when it's -12F outside the average temp through the 3" of EPS would have it performing at about R14 instead of it's rated R12.6 (labeled at it's 75F mean-temp performance), offsetting the declining performance of the polyiso with temp.

    To get the same seasonal performance out of an EPS-only solution requires thicker foam and longer more expensive timber screws, and it's more difficult/awkward to install. A 2- layer 6" foam stackup is quite a bit easier than an 8" stackup, though both are possible.

  15. Brian Martin | | #15

    Like the idea because I'm no fan of spray foam, and don't plan to use it elsewhere in the house, but how would that work in a historic re-creation of a bungalow with a mellow 7:12 roof pitch with exposed rafter tails, it seems to me there would be a facia above the rafter tail, deeming the historic look and feel rather frankensteined. Or I'd be adding 6" of height to the currently designed 2' knee wall and with historical district rebuild restrictions on height, I'm out of luck. The additional 6" has to go somewhere. I could get away with 2" going from 9.25 to 11.25, but not 6". And, I've not got the interior head height of 6" or 4" (if I added 2" to knee wall as I would with the increase from 2x10 to 2x12. Hmmm?

    I calculated insulation using a weighted average for 18% framing factor and came up with these;

    2x10 Rafter (with a .75" shim on top) - 4.5" cor bond = R31.5 + 5.5" fiberglass R21 batt totals R52.5 (for code approval) Weighted to 46.7. The alternatives are more foam less batt OR 11.25" 2x12 with little gain and a loss of 1.25" in height to play with.

    I'd be all for the foam over roof if I was building a home of a different design, but I'm mandated in our historic district and the plans are near completion. I'd think 6" of foam over 2x10 would create some interesting re-design.

    Don't give up on me Dana, I'm not po-po'ing this....

  16. Brian Martin | | #16

    I just drew an example of how I'd frame this proposal of yours and I see that I'd simply sheath up further past the top of the rafter 6" additional. Question is what the sheathing ties into on the top of the foam (glue) and how I'd attach my 2x6 rafter tails that would carry my t&g weight with roofing atop.

  17. User avatar
    Dana Dorsett | | #17

    This is a common (and solved) problem in "chainsaw retrofits", to make the wall foam continuous with the roof foam in a variety of architecture styles. I don't know if there is a good detail drawing in the library of this site to show how to do an exposed-rafter bungalow look. (Paging Martin Holladay...)

    Short of a picture, 10,000, words...

    The wall sheathing can stop at the top of the studwall, but extend the furring for the rainscreen gap all the way to the nailer deck, cutting the roof foam edges co-planer with the wall foam. If the windows are "outie", lap the WRB up & over the top of the roof foam and tape it there, under the nailer deck.

    Stop the siding 10" below the nailer deck and install a 2 x10" facia board above that (leaving a 1/2-3/4" gap to the nailer deck), through screwed to the structural rafters with a timber screws. Then, screwing to the top of the rafter tail through the nailer deck from above as well as screwing it to the mechanically secured facia board would probably cut it mechanically. A 2x10 or 2x8 exterior rafter tail would then be structurally tied into the both the structural rafter and the nailer deck, and the nailer deck is be structurally tied to the top of structural rafters with timber screws. It would be pretty stiff, and should handle the load unless you're talking 4' overhangs or something (which I've sometimes seen in snow-country,) At 18-24" overhangs it would be pretty strong.

    The look would be similar to most bungalows, which typically have a flat board spanning the space between the exposed rafter ends, with siding below that.

    A small ~1/2" -3/4" gap between the top of the facia board and the nailer deck is in an inconspicuous place, and it top-ventilates the rainscreen gap. Install some of the roll mesh used for ridge vents to inhibit critters from using the rainscreen cavity as a condo complex in the rainscreen cavity.

  18. Brian Martin | | #18

    Ok, ran more calculations and came up with the following;
    Basement is 1331 sq. ft. with 7'10" ceilings, with 1.5' above grade. I calculated wall surface area using 5' as I'm including my 1' joist height, and will insure it is insulated the to the same R/U value as my basement interior walls. In fact the R value of the band joist area will be higher as there will be R10 on the exterior.
    First floor is 1,331 sq. ft. with 8' ceilings
    2nd Floor is 800 sq. ft. with 8' ceilings

    Basement Wall Area - 748.28 sq. ft. @ *U.054 with delta T of 80 = 3,233 BTU
    * per recommendation by Dana. I feel the U value will be lower as 20% of the wall surface area is +R10 due to my foam on the exterior of the house for the 1' I've included in the total sq. ft. in the basement calcs.
    Basement Windows - 46.72 sq. ft. @ U.29 = 1,084 BTU
    Total Basement BTU needs of 4,316 BTU

    Main Living Floor 1,163 sq. ft. @U.041 = 3,816 BTU
    Main Floor Windows 267 sq. ft. @.29 = 6,208 BTU
    Main Floor Ceiling (one story mudroom) = 150 sq. ft. @ U.0223 = 268 BTU
    Total Main Floor BTU needs of 10,291 BTU

    Upper Floor 2 Bedrooms/2 Bath - Walls 922 Sq. Ft. @ U.021 = 3,024 BTU
    " " - Windows 95 sq. ft. @ U.29 - 2,204 BTU
    " " - Ceilings 1450 sq. ft. @ U.024 = 2,482 BTU

    Total Upper Floor BTU needs of 7,711 BTU

    Now AGAIN is the golden question on ACH & I/F. I cannot find anyone to run a responsible Manual J for me here. I have a MAC and cannot download Ebot or Hot2000. So I'm doing the best I can with the info I've garnished on this site and recommendation you've all made to do this manually. My question is if one should use wall sq. ft. per floor x the I/F x ACH x delta t, or use room volume x I/F x ACH x delta T. There seems to be some valid arguments for wall sq. ft., but the chart on part #2 of running your own heat load suggests volume. Furthermore how would the basement be handled since I was suggested to use only 50% of the wall height (to which I added a 1' for the floor joist cavity)

    Basement 1330 sq. ft. x 8' wall = 10,648 of volume x I/F .012 x ACH .4 x Delta 80 = 4,085 BTU add on
    Option 2 1330 sq. ft. x 5' wall = 6,650 " " = 2,553 BTU add on
    Option 3 795 sq. ft. wall surface x 5' wall = 3,975 " " = 1,526 BTU add on

    I picked (out of my fanny) .012 for I/F as this was a "tight house number from back in the day" and .4 as an ACH as it appears it is easily attainable, perhaps I'll get down to .2 or less, but there are the bathroom 20/40 ERV Panasonic units, one in each bathroom upstairs and one downstairs in the main floor, a wood burning fireplace, dryer vent, range vent.

    Wall calc's alone are as follows;

    Basement = 4,316 BTU
    Main Floor = 10,291 BTU
    Upper Floor = 7,711 BTU

    Leakage ???

    Basement - Pick one of the options above
    Main Floor - Using Volume = 4,086 BTU
    Upper Floor - Using Volume = 2,458 BTU

    I'm thinking about a 12,000 to 15,000 BTU Mini Split on the main floor centrally located, a 9,000 BTU unit upstairs (cooling a bonus here), and either a small gas stove at an adjustable 12k to 18k/btu for a redundant system should I be out of town and the splits aren't on.
    I will also have a 1.8 cu. ft. wood stove on the main floor in the living room that I hope to use darn near exclusively.

    Thoughts?

  19. User avatar
    Dana Dorsett | | #19

    The ERV cores on the tiny Panasonic units are going to ice up and be damaged pretty quickly in your climate, and they won't warranty it for that damage. A Lunos HRV pair is probably a better option.

    A basement load of 4,316 BTU conducted plus 3-4K infiltration comes in at about 7000-8500 BTU/hr peak.

    With an FH09NA you'd cover it if the low infiltration estimate is right, but maybe not if the high estimate is right. The fact that it can modulate down to 1600 BTU/hr @ 47F, is attractive, but the competition's -9RLS3H is probably a better option- it has a bit more capacity at cold temps, but it's min-modulation at 47F is 3100 BTU/hr. If it's an all Mitsubishi show, with the FH12NA you'd be covered, and it modulates down to 2500 BTU/hr @ +47F, slightly more modulating room than the Fujitsu.

    http://usa.mylinkdrive.com/uploads/documents/4560/document/MSZ-FH09NA_MUZ-FH09NA_Submittal.pdf

    http://www.fujitsugeneral.com/PDF_06/Submittals/9RLS3HSubmittal.pdf

    http://usa.mylinkdrive.com/uploads/documents/4561/document/MSZ-FH12NA_MUZ-FH12NA_Submittal.pdf

    The main floor load of 10,291 BTU + 4.1K infiltration is a total load of about 14.5K. It's a reasonably open floor plan, but with the doors closed (or even with them open) there may be comfort issues in the office room on cold depending on just how big & lossy those windows are.

    Either the FH15NA or 15RLS3H should cover that zone with at least a bit of margin after factoring in the electrical plug loads.

    http://www.mehvac.com/media/953333/msz-fh15na_muz-fh15na_submittal.pdf

    http://www.fujitsugeneral.com/PDF_06/Submittals/15RLS3HSubmittal.pdf

    The Upper Floor's 7,711 BTU + ~2.5K infiltration comes in at 10.2K of total load. It's cut up into a few doored-off rooms, and while a single-head unit could meet those numbers you're probably going to run into distribution issues leaving you cold when it's -10F outside. The 3/4 tonners won't have quite enough capacity to cover it on their own, and even the 1-ton units might be tapped out. But a 1-ton with auxilliary space heaters as the "Hail Mary" backup to meet code is the right choice, since you can probably tolerate sleeping in a cooler than 68F room.

    The 1.5 ton 18RLFCD mini-duct cassette (splitting the output between the rooms with duct runs) would probably be a better choice, even though it's output @ -12F is not specified. (It is at -5F though.) It's minimum modulation is 3100BTU/hr @ 47F, which is the same as the 1-ton, and you're less likely to have to resort to resistance heaters.

    http://www.fujitsugeneral.com/PDF_06/Submittals/18RLFCD%20Submittal.pdf

    That's my best guesstimate based on the load numbers you've calculated.

    The heating capacity tables for the RLS3H series are essentially the same as the RLS2H, which start on p.15 (pdf pagination) of the tech manual:

    http://bangorwinsupply.com/wp-content/uploads/2014/02/RLS2H-Design-Technical-Manual.pdf

    The "TC" numbers in the boxes are "Total Capacity", in 1000s of BTUs. You'll note that the 9RLS2H will deliver 11,100 BTU/hr into a 70F room when it's -15F outside, which is more than the FH09NA will muster even at +5F according to the submittal sheets. At -13F the FH09NA will deliver something like 70-75% of it's +5F output, or about 8000 BTU/hr, which is what makes it a bit iffy even on the basement zone.

  20. Brian Martin | | #20

    So good, seems like we're moving forward here. Can I assume since you didn't call out my leakage BTU's as something that "sounds" wrong, and that since I calculated them using volume, they are on the high to very high side, that the ratio of heat loss from the leakage #'s to the wall/window/ceiling (probably quite accurate( are reasonable?

    I've been reading about calculating leakage by using wall surface area of envelope. The .25 SFBE @ CFM50 sounds like a winner to me. And CFM50 / SFBE = air leakage. In a pre construction stage, and assuming my house will be detailed to be tight, how would I use this calculation to have a litmus test to the volume numbers. And how would this be calculated for the basement? I really wish Martin or somebody could go out on a limb here and state what typical ratio's would look like for this sized home in a tight house condition. Seems like the BTU's for leakage could easily be wildly off, directing one to a larger than necessary split set up. The local heating guy with a good reputation here calculated my needs at 50 to 60k btu and although he sells mini splits, he said for this house he likes a dual stage gfa furnace with and ecm motor. If I could understand how to do a leakage calculation based on a worst case scenario of sloppy construction, average tightness, and tight, I'd be much more comfortable making a decision to go the mini split route. The quote I got (no a/c, nor do I care about a/c) was 10k for a 60Kbtu box that I could feasibly run at 30Kbtu. I understand GBA's recommendation on sizing heating systems correctly, but if the leakage is either such a high percentage of the wall/envelope numbers, or such an unknown, I understand why the pro's go overkill. I'm perplexed, I've read article after article here that it is suggested that in Vermont (similar to Bozeman in heating days and temps) that a house my size would have to be a real leaker to necessitate a 20,000 Btu heat loss total. Stumped but very appreciative of the education to this point and all time you folks take to try to guide people into their own fact finding. I'm just not cozy with my leakage numbers here, certainly not enough to gamble on mini splits that now appear much larger and more expensive than I had hoped for.

  21. User avatar
    Dana Dorsett | | #21

    The BTUs for leakage are GUARANTEED to be wildly off, even if you run blower door tests, since a blower door can only tell you how big the total leakage is, not where the leaks are. If it's exactly two leaks, one in the basement and the other at the top of the house you'll have a lot more stack effect leakage than if it's one round hole (located anywhere), or an even distribution of air leaks across the whole exterior surface.

    But if you build tight the numbers you used will be a maximum- it won't be anywhere near 2x that, and could be as little as 1/2 that. Adding the calculated infiltration losses for the whole house together comes to about 12 KBTU/hr. which doesn't add up to a huge load, and it's more likely than not going to be overstating reality if you are building with air-tightness as a specific goal. The total calculated load is about 33,000, 12,000 attributed to infiltration. Even if the calculated number is 33% under the real number (possible, not likely), that only adds 6000 BTU/hr to the peak heat load. A shortfall of 6000 BTU/hr is something that would be covered by the heat output of 3 sleeping humans, one refrigerator, and a 1200 watt space heater.

    If reality is really only half the infiltration number your total heat load is about 27K, which means you might be approaching 50% oversizing on some zone, but not 2x oversizing. With any oversizing at all against the calculated loads you're pretty much covered. I=B=R methods typically overestimate the whole-house loads by 10-25%, so reality probably really IS something between 27-30K. But it's not a safe assumption that it would overestimate the load of any individual room or zone by that much.

    As a sanity check, it looks like your upper floor level come in at about 8 BTU/hr per square foot of conditioned space, the main floor comes in at about 10 BTU/hr-ft^2, and the basement zone runs about 7-ish BTU/hr-ft^2 or less, all of which are credible numbers for the as-described better-than code house.

    As for the local contractors' estimate, a code min house that size might come in at 50K-60K, but not your house. You can test that thesis- make a copy of your spreadsheet and plug in the code-max U-factor numbers in Table N1102.1.3 rather than the U-factors we conjured for your better-than-code house :

    http://publicecodes.cyberregs.com/icod/irc/2012/icod_irc_2012_11_sec002.htm

    Oversizing a modulating mini-split by 25-50% has only beneficial effects on efficiency, since they will run at a higher-efficiency part-load modulation nearly all the time. Anything more than 50% can run into comfort issues related to higher than necessary air volumes (a bit more noise and breeze), and less stable room temps in the shoulder seasons due to more on/off cycling due the loads being lower than the minimum modulated output. Oversizing a hot-air furnace has similar comfort issues, but won't gain efficiency from oversizing the way modulating mini-splits do.

    If you did the upper floor with a mini-duct cassette it looks like you might be able to install the cassette at the top of the closet at the top of the stairs or the linen closet, and duct it to both bedrooms and the master bath reasonably. The load numbers for the smaller bath that has no window has to be miniscule since it's surrounded by conditioned space except for the ceiling. The total design heat load for that room is probably less than the heat output of one conscious human, so it wouldn't need any heating.

  22. Brian Martin | | #22

    Awesome Dana, I will digest your info here then begin digging deeper into the mini split specs. Thanks

Log in or create an account to post an answer.

Community

Recent Questions and Replies

  • |
  • |
  • |
  • |