Heat runs to unfinished basement? Pacific NW area
I realize the basement is considered part of the conditioned space, but I was a little startled at an HVAC guy proposing to cut heat runs into the basement (which we do not plan ever to finish). It’s a 1901 basement with concrete walls to ground height and wood joists above that. Previous owners had put fiberglass insulation in the joists, which was a really bad idea as one can see from old water stains. That’s long gone, but when we first lived in the house we used to get water coming up through the floor drain during storms (no leaks through the walls that I am aware of). Drainage work done during a remodel took care of the floor problem, and I would now say it’s relatively dry. We are just getting around to properly air sealing it. I am embarrassed it took us so long, but natural gas prices have been low enough that I hadn’t worried about the bills and didn’t realize our usage was as high as it was (furnace is on its way out as well).
Anyway, I had always assumed that an unfinished basement should inherently be, well, cellar temperature or thereabouts, well above freezing, but not warm. It’s been convenient for us to keep beer down there, for instance. Plus I don’t want to overwork the heating system by adding square footage to the officially heated area. But is it inefficient to have the basement much cooler than the rest of the house?
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Replies
Irene,
Q. "Is it inefficient to have the basement much cooler than the rest of the house?"
A. No. Keeping your basement cool will lower your energy bills compared to heating your basement.
Thanks, Martin. I also realized today that a little bit of the basement wall is under the front porch, and some of it doesn't even have siding -- at that point the wall of the house is one board thick. (I'm honestly kind of surprised it hasn't rotted, but it looks in pretty good shape from what I've been able to see.) Should something be added on the outside? It wouldn't be awfully fun working under the porch, but it's a big enough space to be manageable on a dry day. I've occasionally thought of trying to convert it into lawnmower storage or something.
Irene,
If your basement includes an above-grade framed wall that is uninsulated, you can insulate the wall from the interior -- just like any above-grade wall.
The first step is to take steps to seal air leaks. The second step is to install a layer of insulation. And the third step (in most cases) is to install a layer of 1/2-inch drywall as a thermal barrier (for fire safety).
I know I can insulate the inside, but I was wondering if the part without siding or anything should be protected on the outside in some way.
Irene,
If the wall in question faces a crawl space that is protected by a porch roof, then I assume the wall is never rained on. Under the circumstances, what you are worried about are air leaks.
If access is possible, you could cover the boards on the exterior with a layer of plywood or OSB with taped seams. If access is impossible, you could seal air leaks from the interior by installing spray foam on the interior side of the board sheathing.
I'll probably do both -- though I'm not a fan of spray foam. I've been using silicone caulk and will likely fill in with Roxul batts.
If the furnace is going to be replaced it's a once in 25 years opportunity moment to right-size it for higher comfort, less noise. Even with cheap gas, at cheap electric rates and temperate outdoor temperatures of the PNW a right-sized heat pump can beat it on operating cost.
Unbalanced un-sealed duct systems create significant parasitic losses- a heat loss that only occurs when the air handler is running. If there's any duct leakage in an uninsulated not-air-tight basement those losses can be quite large. But to establish an upper bound on the heat load and furnace (or heat pump) sizing, run a fuel-use heat load calculation, which uses the existing furnace as the measuring instrument, as outlined here:
https://www.greenbuildingadvisor.com/blogs/dept/guest-blogs/out-old-new
Getting a handle on the actual heat load and not oversizing by too large a fraction is really important for comfort, and sometimes efficiency as well. Oversized furnaces will often satisify the thermostat quickly, before the rooms at the furthest end of the duct runs are fully heated. They are prone to temperature overshoots too. The ideal sized furnace would run very long almost continuous runs during the coldest hours giving you a steady warm summer breeze effect rather than the hot-flash followed by the chill.
ASHRAE recommends no more than 1.4x oversizing for the load at the 99% outside design temperature, which is enough margin to cover even the coldest cold-snaps and have reasonable recovery ramps if overnight setback strategies are being used. If going with a heat pump solution sizing it for EXACTLY the 99% heat load and a constant room temperature (no setbacks) would yield higher efficiency. There are lots of trade offs to be made and details to work out, but without a good handle on the heat load numbers you're doomed. The last thing you want to do is blindly replace like-for-like. There are far too many 2 stage gas furnaces out serving homes with a design heat load that less half the LOW stage output, rending the 2 stages meaningless.
The heat loss through the floor to an ininsulated basement is usually a double-digit percentage of a home's heat loss, even without parasitic losses from ducted hot air heat. A typical 8" thick poured concrete wall has a U-factor of about 0.7 BTU/hr per square foot per degree-F temperature difference . When it's 50F in the basement an 40F outside (roughly the mean outdoor temp for January in Seattle), the above-grade walls are losing 7 BTU/hr per square foot. If you have a 200' foundtation perimeter with 2' of above exposure that's 400 square feet x 7 = 2800 BTU/hr. If it's warmer than that in the basement (usually is) and colder than that outside (certainly is at the 99% outside design temperature) the loss is higher still. And that's just the above-grade portion. There is still heat loss to the soil- it adds up. That's why current code spells out either R30 between the joists when the floor is over unheated uninsulated space, or R15 continuous foundation insulation.
If you insulate the foundation walls to R15 the basement temperatures will rise (even though you're not actively heating it), but the heat loss numbers for the basement as a whole will drop by ~90% despite the fact that it's idling at temperatures of 55-60F during mid winter (sometimes higher) rather than 50-55F.
Thank you, Dana! Yes, I agree. It sounds as though you probably remember (though other readers may not) that I was posting on someone else's thread last week about trying to figure out the right size for a heat pump (https://www.greenbuildingadvisor.com/community/forum/mechanicals/104477/minisplit-sizing-dilemma-replace-oil-furnace , comment 13). To recap, one contractor actually did a Manual J and recommended a 42K Mitsubishi heat pump, while another, after a much more cursory visit, recommended a 36K system. My back-of-the-envelope calculations suggest the second guy might be right, but I don't really know how he got there. The furnace that's in there now is 66K, originally rated 80% efficient, for a yield of 52.8K BTU/hr. So the 42K (which goes up to 54K for heating) is actually rated as more powerful than this furnace was in its glory days.
So, what do the fuel use heat load numbers say the load would be at your +26F outside design temp?
If you're planning to insulate the basement walls you can knock at least 10% (sometimes 25%) off whatever the fuel use load numbers are and still have margin. My not-so-superinsulated sub-code 2400' 2x4 framed antique + 1600' of insulated but not directly heated basement has a heat load of ~35-37K , but that is at an outside design temperature of +5F, fully 21F colder than your design temperature. Before insulating the basement and doing a modest round of air sealing it was ~42-44K.
At a balmy +26F the load on my house is less than 25,000 BTU/hr, and yours probably is too (or could be with an insulated foundation and fixing the most egregious leaks.)
I accidentally edited out how I got the degree day numbers (from DegreeDay.net) -- but it seems I downloaded different numbers for the first set of calculations than for my "corrected" set, so that also confuses the issue. Not sure whether I put in wrong dates or chose a different location or what.
Deleted answer because I skipped a step in the calculations.
Well, I just realized my previous calculations were off because I forgot to allow anything for the water heater usage. But what I came out with was 28,158 BTU/hr at 65°F and 29,444 BTU/hr at 60°F, on the assumption that the furnace is 80% efficient. The 1.4 factor yields 39,421 and 41,222 respectively. Corrected calculations below. EDIT: Edited my answer because I skipped a step and also used different degree day values (from DegreeDay.net) than the first set of calculations did.
21,606,541 total usage for period (216.117 therms). Allow 1,537,733 BTU for hot water usage (average of June-August usage -- 15.381 therms).
21,606,541 - 1,537,733 = 20,068,808
20,068,808 (0.80) = 16,055,046
16,055,046/976.9 = 16,435 BTU/degree day at 65°F
16,055,046/806.9 = 19,897 BTU/degree day at 60°F
16,435/24 = 685
19,897/24 = 829
65-26=39
60-26=34
39 (685) = 26,751
34 (829) = 28,186
The 1.4 factor yields 37,451 and 39,460 respectively. Phew. I think I've got it right this time.
If it's a 2x4 framed house with an uninsulated basement, the base 65F result will be more accurate (unless you normally keep the house at 65F or cooler.) The base temp is the heating/cooling balance point, which will be lower for higher performance homes, lower for higher performance homes.
So, it looks like you're in the 27,000 BTU/hr @ +26F range based on history, and will almost certainly be under 25,000 BTU/hr after insulating the foundation and doing some air sealing of the house (and ducts, if using the same ducts.) And don't forget that's really an upper bound- there are both parasitic losses from air handler driven air infiltraion, and it's possible/likely that he burner isn't really still hitting it's 80% efficiency. (If abandoning the ducts in favor of a ductless approach, the heat load could easily be 10-15% lower than the fuel use calculation indicates, because the parasitic load of air-handler driven infiltration goes away.)
With heat pump solutions do NOT use the 1.4x multiplier for sizing!
That oversize factor is appropriate when overnight setback strategies are being used. Using overnight setbacks with heat pumps (particularly modulating heat pumps like mini-splits) ends up using more electricity than a "set and forget" approach. A 1.4x oversize factor takes bite out of the as-used HSPF of 1 or 2 stage heat pumps by reducing the average duty cycle over the season. Mini-splits with very high turn down ratios can sometimes do a bit better with modest oversizing, but not multi-splits due to the high minimum output of the multi-split compressor. Both modulating and non-modulating heat pumps will do worse when using overnight setback stratigies, and worse still if oversized to the point where the recovery ramps from setbacks are fairly quick.
And don't forget this is the "before" picture, the load before the foundation gets insulated and some progress is made on air sealing. Deducting even 10% from the calculated 26,751 BTU/hr load for the "after" picture yields 24,076 BTU/hr. Odds are pretty good you'll be even lower than that, but for now call it 24K.
If going with condensing gas a small ~30KBTU-in 2-stager like the Goodman GMEC960302, is about right. That furnace delivers ~28.8K at high fire, ~20.2K at low fire. That nicely brackets the likely loads both before & after insulation & air sealing upgrades, and even covers the "before upgrades" calculated load based on an unrealistic 60F heating/cooling balance point. The high-fire 28.8K output would be a (28.8/24=) 1.2x oversize factor for an eventual 24K-ish load, and a 1.25x oversize factor for a 23K-ish load, either of which is fine. (But you might have a hard time convincing an HVAC contractor of that.) It's entirely possible that the upgrades will bring the load down to the low-fire output of that furnace, at which point it would be roughtly the ASHRAE 1.4x oversize factor, which is also fine. It probably won't come in under 20K without doing more extensive building upgrades, but under 25K is almost a given.
Wintertime hot water heating is usually a bit higher than summertime use due to cooler incoming water temperatures, and more layers of clothing to wash, etc. In foggy-dew western WA the wintertime solar gains are pretty low, and don't affect the fuel use load much. For most locations it makes sense leave the hot water energy use in the calculations to counterbalance the error in the other direction from solar gains, but in your area it's probably more accurate to make some correction for hot water use (the way you did), due to the very low solar gains of the short & gray daylight hours that typify of the region in winter.
For the record, what were the beginning & end dates of the fuel use period in the load analysis?
12/16/16 and 1/18/17. So you're saying even the 36K Mitsubishi is likely to be oversized? We're planning on a ducted system using the existing ducts.
Note that the average past usage/HDD method of load estimation uses average wind. Load can be *much* higher on a windy day. Also note that a 99% design temperature leaves a lot of hours at well below design temperature.
If you aren't going to adjust for these, plan on either some supplemental heat or occasional discomfort.
There are lots of 36K Mitsubishis- the model numbers matter. But yes, it's highly likely that ANY of the 36K Mitsubishis will be suboptimally oversized for your load.
A 3 ton PVA-A36 & PUZ-HA36 combo can deliver 38,000 BTU/hr @ +47F, 29,000 BTU/hr @ +17F, but only has about a 2.2:1 turn down ratio which means even at 40F (your January average temp, a temp where your load is between 17-18K before building fixes, even less after) it probably won't be modulating much. It's minimum output at +47F is 18K:
http://meus1.mylinkdrive.com/files/PVA-A36AA4_PUZ-HA36NHA5_ProductDataSheet.pdf
The 2.5 ton PVA-A30AA7 & PUZ-HA30NHA5 has only a 1.8:1 turn down, but delivers 32K @ +17F (which is more than the 3 ton PVA-A36 & PUZ-HA36 at that temp, and more than you need):
http://meus1.mylinkdrive.com/files/PVA-A30AA7___PUZ-HA30NHA5_Product_Data_Sheet-en.pdf
The 2.5 ton PVA-A30AA7 & PUZ-A30NHA7 combo has a significantly better 2.8:1 turn down ratio (34K of capacity @ +47F that can still throttle back to 12K @ +47F), and still delivers 20,700 BTU/hr @ +17F. This is probably the best choice in the series for both your current and anticipated loads:
http://meus1.mylinkdrive.com/files/PVA-A30AA7___PUZ-A30NHA7-BS_Product_Data_Sheet-en.pdf
Among the MVZ series multi-split compatible air handlers...
The MVZ-A30AA7 can deliver 34K @ +47F, and is probably more than adequate at +26F depending on the compressor, but the the +17F numbers likely depend on the multi-split compressor it's married to- you'd have to consult the engineering manuals. The MVZ-A36AA7 delivers 40K @ +47F, and is probably way too much.
High winds in western WA are somewhat rare, but any concerns about undersizing for extreme conditions can be addressed with heat-strip backup, which is an option for all of the PVA and MVZ air handlers. The tighter the house, the less impact wind speed has on load overall, and there is still likely to be some padding from other errors in the fuel use numbers.
Short of deeper analysis of the "after air sealing and insulation" picture that would include an aggressive Manual-J it looks like the 30K PVA-A30AA7 & PUZ-A30NHA7 combination is your best fit within the large air handler Mitsubishi lineup. Without actually looking up the extended temperature capacity tables I'd estimate at +26F outdoors, 70F indoors it has ~27-28K of capacity, enough to cover the before-upgrades load.
There is a fair amount of "noise" error in a single month's fuel use load calculation. It's worth running the numbers from mid-December through mid-March (month by month) to see how consistent it is.
Dec-Jan 2018:
25,116 BTU/hr at 65°F
28,254 BTU/hr at 60°F
Jan-Feb 2018:
25,389 BTU/hr at 65°F
29,546 BTU/hr at 60°F
Feb-Mar 2018:
24,570 BTU/hr at 65°F
26,656 BTU/hr at 60°F
In your house, ignore the 60F base temperature numbers. Unless you keep the place at 65F or cooler the balance point is higher than 60F. If you keep it 68-70F on average, the 65F base temp.
Looks like you're reliably in the 25,000 BTU/hr range, and will be lower if you insulate the basement. These are totally credible numbers for a 2x4 framed house with no foundation insulation.
The PVA-A30AA7 + PUZ-A30NHA7 combination would cover the current 25K load @ +26F. Add a EH03-MPA-M heat strip kit (the smallest available for that air hander, ~3kw= ~10,000 BTU/hr) and you'd be covered down to about 10F outdoors even before air sealing and insulating the basement (or during the rare high winds at 26F). After insulating & aire sealing basement you'd be covered down to the low single digits.
https://meus.mylinkdrive.com/files/Electric_Heat_Kit_Install%20Instructions_2016-10-19.pdf
We keep it at 66 during the day and 55 at night, so it's possible the 60° value could be relevant. I've been including it in any case for completeness' sake, to show an upper bound. I'd prefer the Hyper Heat model (these model numbers make my eyes cross, but I think that's the PUZ-HA one) rather than monkeying with heat strips, but it's a pity about the turn down ratio being lower. The most recent contractor I talked to is not keen on the idea of going below 36K, despite my calculations, which is annoying as I like this guy and he has a pretty good price and good reviews.
With those temperature settings base 60F is more appropriate to use to derive the BTU/degree-hour constant, and it would be prudent to add 3F heating degrees to reflect the load at a code-minimum 68F indoor design temperature. (Have you been doing that?)
If you're committed to insulating the foundation the 2.5 tonner with the biggest turn-down ratio combined with the heat strip option would still be the better option, due to the modulation range around your average winter load.
If basement insulation is years away (or maybe never) the "...AA7" version of the PVA-36, the PVA-A36AA7 coulpled to the PUZ-A36NKA7 outdoor unit has a 2.4:1 turn down and is probably the best option. It can dial back to 17K out @ +47F, and delivers 28.8K @ +17F, and will cover the larger pre-insulation load at a 68F indoor temp:
http://meus1.mylinkdrive.com/files/PVA-A36AA7___PUZ-A36NKA7-BS_Product_Data_Sheet-en.pdf
The PVA-A36AA4 & PUZ-HA36NHA4 is a hyper-heating unit. It's turn down ratio is 2.2:1, but it has substantially more output at +17F (38K max), and isn't a terrible choice. With the -HA36 you won't need the heat strip option in your climate. But it's minimum output at +47F is 18,000 BTU/hr, fully 6000 BTU/hr more than the 2.5 tonner, a pretty big step:
http://meus1.mylinkdrive.com/files/PVA-A36AA4_PUZ-HA36NHA4_Product_Data_Sheet.pdf
The minimum modulation at +47F is about the same but the HSPF efficiency would still be slightly ahead of the PVA-A36AA7 + PUZ-A36NKA7 non-hyper heating version. If it's going to be a 3 tonner that's about the best you're going to do.
In the hyper heating versions there is no modulation range benefit to stepping down to a 2.5 tonner, since it's minimum modulation @ +47F is also 18,000 BTU/hr. The fact that it's minimum modulaton is that high means it wouldn't necessarily beat the 3 ton HA36, and might not even beat the non-hyper-heating 2.5 tonner with the 2.8:1 turn down ratio, and even has a slightly lower tested HSPF than the non-hyper heating 2.5 tonner:
http://meus1.mylinkdrive.com/files/PVA-A30AA4_PUZ-HA30NHA4_Product_Data_Sheet.pdf
So if you DO drop back to 2.5 tons, that isn't the one!
Ah, okay, this gives me more to chew on for sure. I have not been adding in the extra degrees (that would mean instead of using the 34-degree difference between 26 and 60 to multiply the degree days at 60, I would use 37 instead, correct? in other words, divide the previously calculated figures at 60 degrees by 34 and multiply by 37?).
I really have made progress on the air sealing -- have done about a quarter of the basement, and would have done more by now if I hadn't had a guest -- and have bought the first pack of insulation, so I do think this is all going to get done.
You have been amazingly helpful.
So corrections on figures I gave before:
Dec-Jan 2017: 30,673 BTU/hr
Dec-Jan 2018: 30,747 BTU/hr
Jan-Feb 2018: 32,153 BTU/hr
Feb-Mar 2018: 29,008 BTU/hr
So, looks like you currently have about a 30K load if it needs to be heated to a code-min 68F (you can keep it whatever temp you like, but the heating plant needs to be able to cover 68F at the 99% outside design temp.)
The PVA-A30AA7 + PUZ-A30NHA7 isn't going to cover that without heat strips. It puts out about 34K max @ 47F, 20K max at 17F. A quick & dirty linear estimate (which will undestimate the actual output which makes it conservative) is a 14,000 BTU/hr drop over 30F degrees, or 467 BTU/hr less capacity for every degree below 47F. The 99% outside design temp is 21 degrees cooler than 47F, so it will be capable of at least 34,000 -( 21 x 467=) 24,000 BTU/hr @ 26F. (It's actually not linear- reality is closer to 26-27K & 26F, but if falls off more quickly below that temp.)
http://meus1.mylinkdrive.com/files/PVA-A30AA7___PUZ-A30NHA7-BS_Product_Data_Sheet-en.pdf
If you're actually keeping it at 66F indoors the output will be more since the difference between indoor & outdoor temperature are the primary factor in capcity. The submittal sheets are the capacities at 70F indoors. It would likely cover your as-used load but for code compliance you'd need the heat strip kit. You'd want heat strips anyway for any right-sized heat pump to be able to cover the additional load of cold snaps.
The 3 ton hyper-heating version would have no problem covering a 30K load at 26F- it;ll put out 32K all the way down to +5F. so it's not an insane choice:
The http://meus1.mylinkdrive.com/files/PVA-A30AA4_PUZ-HA30NHA4_Product_Data_Sheet.pdf
So it really depends on your plans. The bench tested HSPF of the 2.5 tonner is 10.0, and it's right-sized for your as-used load prior to insulating & air sealing the basment, and would hit those numbers. It's also right sized for the full code-min indoor design temp load for the "after" picture, and it would use the heat strips less. Since the heat strips would be needed less than 1% of the time and it's only covering the shortfall of the heat pump (which is still delivering the lion's share of the heat) they don't really affect the as-used HSPF as you might think. The much lower 12K output @ 47F is probably pretty close to or slightly lower than your actual load at +47F (you can run those numbers too), so it will be modulating at high efficiency & comfort almost all the time.
The tested HSPF of the 3 ton hyper-heating option is 11.0, 10% more heat per kwh than the 2.5 ton unit, but the 18K minimum @ 47F is well above your actual heat load at 47F, which means it will probably be cycling a lot even at 40F outdoors, which cuts into as-used efficiency. If you insulate & air seal the basement lowering the load it may be cycling a lot even on fairly cold days. It will still chew through less electricity than before the basement is insulated, but efficiency hit from cycling all the time means there's a good chance it will use slightly more than the 2.5 ton non-hyper-heating option.
Insulating basement walls with a fiber-insulated studwall requires putting at least some rigid insulation between the studwall & foundation as a capillary break for the wood & fiber, otherwise it can wick moisture to the finished wall. An IRC code min insulation level requirements for zone 4C would be R15 continuous insulation (3" of polyiso or 4" of EPS), or R19 in a 2x6 studwall. An R13 studwall up tight against an inch of cheap plastic faced or foil faced Type-1 EPS (~R4) trapping it to the foundation will hit the same performance level, but without wicking moisture, and it'll keep the studs warmer/drier too. An inch of EPS under the bottom plate of the studwall as a thermal & capillary break works too.
The model you're saying is not an insane choice is exactly the same one the contractor wants to install, so that makes us and the contractor potentially way more on the same page. I am still worried that the assumption of 80% efficiency for the current furnace could be a weak link here, though.
If the furnace efficiency in it's current condition is lower than 80%, your actual heat load is lower than 30K, which nudges it toward the 2.5 tonner.
If you are in the process of lowering the heat load by insulating and air sealing the basement, it starts tipping pretty strongly toward the 2.5 tonner, despite the lack of Hyper Heating, and the need for heat strips.
Retrofitting a basement with wall insulation on a typical 1.5 or 2 story house usually lowers the whole house load by 15%, give or take, sometimes more. That pushes a 30K load down to ~25K, and the 2.5 tonner covers it.
If the load is really 28K (due to overestimating the furnace efficiency) those improvements cut that down to ~24K and the 3 tonner is almost never modulating.:
If the load at 68F in/26F out is 24K, the load at 40F (your mean outdoor temp in January, give or take a couple) will be about 15K, which less than the minimum output of the 3 tonner, so it'll cycle a lot even in the dead of winter, and nearly all the time during the shoulder seasons.
If your load at 26F is and remains 30K, the load at 40F is about 19K, which is more than the minimum modulated output of the 3 tonner and it will be able to modulate with load at least at mid-winter.
So if the contractor was expecting a design load in the 30K range it's not insane. But if the near-future reality of a better insulated version of the house lowers the load by 15-20% (probably will) the 3 tonner is not optimal, and the 2.5 tonner is nearly perfectly sized.
With a better description of the basement wall dimensions and the total square footage of above-grade exposure on the exterior of the foundation walls it's possible to estimate the magnitude of the reduction in load using a quick & dirty I=B=R type load calc.
Arrrrggggghhhh. I just realized that I don't know where I got the 26-degree design temperature. The 99% temp for my area is 28 and the 99.6% is 24. As far as I can make out, the use of 24 is required by code per http://www.seattle.gov/DPD/cs/groups/pan/@pan/documents/web_informational/p2574215.pdf. The guy who did the Manual J calculation (and came up with the 3.5 ton recommendation) used 24 for heating and 86 for cooling (99 and 1 would be 28 and 82; code specifies 24 and 83).
Adding a couple of heating degrees is less than a 1500 BTU/hr difference not enough to make a difference in the sizing, but what was used for an interior design temperature in the Manual-J?
The local code document linked to specified a heating MAXIMUM interior design temp for heating, but doesn't specify a minimum heating design temp, or at least I didn't find it. (IRC specifies 68F at some height off the floor and distance from windows: https://up.codes/s/temperature-control .)
The maximum allowed indoor design temperature in the local code document is 72F:
-------------------------
R302.1 Interior design conditions. The interior
design temperatures used for heating and cooling load
calculations shall be a maximum of 72°F (22°C) for
heating and minimum of 75°F (24°C) for cooling.
--------------------------
The Manual J had 70° for indoor heating. The local code document also said "R302.2 Exterior design conditions. The heating or cooling outdoor design temperatures shall be selected from Appendix C." Appendix C has the SeaTac temperature as 24°.
EDITED: Ack, sorry, I think I misunderstood your question.
If 68F is allows as an indoor temperature by code (usually is), that's 2 heating degrees lower than what was calculated at 70F. The load will scale pretty much linearly.
Dropping the outdoor design temp to 24F from 26F is adding 2 heating degrees. In your fuel use load calculations you've already made the adjustment upward to what it would be at 68F, so you just need to add another 2F heating degrees to reflect the 2Fdrop in outdoor design temperature from 26F to 24F.
If the Manual-J used 26F as the outdoor design temp, the load number at 70F indoors/ 26F outdoors will be the same as 68F indoors/24F outdoors.
You're right, of course, the 24 degree thing isn't enough to account for why the Manual J was so far out of whack, nor even for the 36K suggestion. I was just annoyed that I confused myself over the business. It's all getting pretty depressing. Thanks again.
So the good news is we found a different HVAC guy who seems to get what we're talking about and is willing to do the PVA-A30AA7/PUZ-A30NHA7 system. Bad news is we're looking at close to $15K (and, having taken so long getting bids, are getting into the busy season).
Fully modulating big air handler systems aren't cheap.
A 2 ton Carrier Infinity Greenspeed with a bit of heat strip would be a good option. At +47F it can still throttle down to the 10,000 BTU/hr range, but 25,000 BTU/hr or more @ +25F. In my are they are usually more expensive than a mini-split solution, but may be comparable to or cheaper than the 2.5 ton PVA/PUZ -A30.
For comparison you might want to find out what the 2 ton 25VNA024 Greenspeed would actually cost. They have a 2.5:1 turn down ratio. (I don't think make a 2.5 tonner.) Depending on which air handler option used a 2 ton Greenspeed puts out something like 25-27,000 BTU/hr @ +25F. Take a look a their handy tool page- click on the "Heating Capacities" tab, select 25VNA024 in the box, then compare output graphs using different air handler options under the "Indoor Unit" tab on the lower left:
http://www.tools.carrier.com/greenspeed/
If a 3 ton Greenspeed cost $15K, a 2 ton should be at least a BIT less. The HSPF efficiency of a right-sized Greenspeed is really quite good, usually north of 11.0, sometimes as high as 13 (it depends on the air handler combination) which is better than the HSPF 10 delivered by the PVA/PUZ-A30.
As with Mitsubishi units, the key thing is to size it correctly.
My sister recently got a Carrier system similar to what you're describing but larger. That was also $15K, but I think after some rebates we're not eligible for (she's in a different town). I'm suspecting this really is the going rate around here.