Minisplit or furnace?
Hi, I know there is lots of discussion about this; however, our proposed new construction seems to be in unclear realm of which to use.
Our current design has 1550 square feet with 3 bedrooms and 2 baths upstairs and a large open space and 1/2 bath down. We live in Indianapolis, climate zone 5, but close to climate zone 4. We are using 2×6 with dense pack cellulose and our builder has built energy star homes, so we will be getting a good tightness level and blower door tests during construction (NOTE: We are not interested in using exterior foam board and double stud walls are too much more expensive for our budget).
I put everything in Marc Rosenbaum’s Heat loss calculator (pic attached) and it says 31,400 btu’s needed. So this house is not a super-insulated home requiring only 25,000 or less btu’s, but it is not a normal home that needs 50,000-80,000 btu. For background, a gas hookup costs $12/month plus the therms, and electric is trying to raise the base rate to $11/month and from 6.7 cents to 9.4 cents per kWh (they keep trying to raise the rates so likely will get it passed at some point) (aslo, electric here is currently coal generated but they are switching to more gas generated). I would like to install PV at some point, but I am not sure if the budget will permit that, or if so when.
Given all that, I would appreciate recommendations regarding if mini-splits would work or if we should go the gas furnace route. Cost is also a consideration, as if to get mini-splits we need to use so many indoor units that it becomes significantly more than a gas furnace we would lean toward gas.
Thanks,
Mark
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Replies
Mark,
You haven't told us the cost per therm for natural gas in your area.
Currently, your electricity costs are quite low.
If I were you, I would install one ductless minisplit downstairs, and one upstairs, for a total cost of $7,000 to $8,000. For a little bit more, you could choose a ducted minisplit for upstairs, and you could install ducts to each of the bedrooms.
And every paycheck, I would put some money in a jar to save up for a PV system.
When comparing the cost of a furnace-based system to a minisplit system, remember that the minisplit system provides air conditioning as well as heating.
Therm costs are: 0.3289 for the first 46 therms and 0.2311 for all additional therms. Thanks for the quick reply. My primary concerns are whether the 2x6 walls will be enough to prevent cold rooms and if the 31,000 btu's projected need is really accurate enough for a 36,000 btu to cover the heat most of the time. It is nice to hear that you think two indoor units are all we need versus the 4 we thought were maybe necessary. Here is a rough sketch of the floor plan, if that helps. Thanks for your input.
The 10K+ combined infiltration & ventilation numbers seems insanely high to me, as does a 20 BTU/ft^2-hr heat/floor ratio at a 65F delta-T in code-min new construction. You also haven't subtracted out internal heat sources like the refrigrator, water heater, 200-250BTU/hr per sleeping human etc.
Unless it's leaking air well beyond code-max, reality is more likely to be a heat load less than 25K, and maybe even under 20K. For reference, my 1.5 story sub-code 1920s antique has a measured heat load of ~15 BTU/ ft^2-hr heat/floor at a delta-T of 65F based on fuel-use analysis against heating degree-day data. At the water temps I'm running in the heating system I'm radiation constrained to only about 18BTU/ ft^2-hr, yet the place has sailed through nights with lows hitting 70-75 delta-Ts without losing ground.
The 1.5 ton Fujitsu cool-climate mini-ducted minisplit (HSPF > 11) would likely cover the whole-house load to the 95% outside design temp, if not fully to the 99%. But it should cover the load for the second floor with margin- the1-ton mini-duct might be enough too. A 3/4 ton or 1-ton wall-coil type would cover the more open first floor.
If you break down the heat load by floor, what do you come up with?
Do those $/therm numbers include delivery, or is that just the energy cost? (My energy-only portion of the gas bill is only about half the total bill.)
What are your electric rates like (fully delivered), and does the utility offer time of use rate billing?
I do not think there is time of rate billing, although I could be incorrect about that. Here are two total bills from a low and high month, although they are likely to go up somewhat significantly in the near future:
Usage 1163 - charges 107.26, tax 7.51, total 114.77
Usage 271 - charges 32.69, tax 2.29, total 34.98
The delivery for the gas bill is the $12/month, so total is $12 plus the therms.
I am not sophisticated enough in my understanding of heat loads to break down the numbers by floor. While I have read a lot on this site and other places, I am still quite limited in my understanding and have no application experience, hence my questions. From your comment it sounds again like we are fine with 2 mini-splits. Given the sketch I put up, will one upstairs distribute heat enough if doors are closed at night with only 2x6 walls? Thanks for your help!
From the sketch I'd estimate that the first floor can use a simple wall-coil type head and get reasonable distribution. Mitsubishi -FH12NA or maybe even an -FH09NA would probably handle the first floor zone. A Fujitsu -9RLS3H would definitely handle it.
But the second floor might want a mini-duct cassette mini-split, and ducts mounted in a lowered ceiling above the stairwell, with supply ducts to each room. The returns don't need ducts if you use wall-cavities as jump-ducts into the stairwell/hall area, and a ceiling grille for the return side of the mini-duct cassette. The Fujistu -18RLFD or -12RLFCD with it's own compressor (not a multi-split) is the likely contender here.
In a zone 5 climate any one of those would have a seasonal average coefficient of performance of greater than 3, and it looks like your electricity runs about 10 cents/kwh. Assuming an average COP of 3, that's 3 x 3412/BTU/kwh= 10,236 BTU/kwh (an effective HSPF of 10.2, which is conservative- it'll probably do better), or 1,000,000/10,236= 98kwh/MMBTU. At 10 cents/kwh that comes to $9.80/MMBTU
Burned in a 95% efficiency boiler or furnace the heating equipment delivers 95,000 BTU/therm, or 1,000,000/ 95,000= 10.5 therms/MMBTU. At an average delivered cost of say, 28 cents/therm that's $2.94/MMBTU. That's dirt cheap- even incredibly cheap (to cheap to be actually true.)
The reason I have a very difficult time believing that as a retail price is that average wholesale price at the distribution hubs for natural gas is usually more than that, often by quite a bit. Pull out a wintertime bill, and divide the total bill by the therms. Even if your gas distributor locked in to some ridiculously low energy price contract, it's unlikely to be that low or stay that low for long. The average wholesale price is usually higher than that, let alone the retail price.
As of 5 minutes ago the WHOLESALE natural gas was trading at $2.71/MMBTU, and has run in the $2.66-2.72/MMBTU range on today's daily market trading: http://www.bloomberg.com/quote/NG1:COM About the only way to beat that price on a retail basis going forward is if the gas well on your property!
I went through the same type of calculations that Mark did and came to the conclusion that I should put in a conventional ducted gas furnace and air conditioner. I will be building near Flint, MI which is in zone 5 near zone 6. I am still considering whether to use a heat pump instead of the air conditioner to give myself some flexibility during shoulder seasons.
At my current utility rates, I calculated that a heat pump (whether air source of ground source) would need a COP of 6 to be cheaper to operate than a gas furnace. I estimated that a mini-split system would average 2-3 over the course of the year. Had it come out a little closer, like 3-4 COP to be equal operating costs, I would have leaned toward mini-split since current operating costs are not the only factor to consider. Giving up on mini-splits saved me the process of convincing my wife that mini-splits look ok and that not installing ductwork would not be irresponsible.
I also used Marc Rosenbaum's heat loss spreadsheet. For my 2400 square foot house, I come out with 34,500 btu/hr at 70 degrees delta T. Of that, 11,000 is attributed to infiltration/ventilation. I assumed ACH50 of 2.0 and 60 cfm ventilation. I hope to do better than that but didn't want to count on it in the calculation. The house will have 2x6 walls with 2" of exterior foam. I did not factor in any internal heat gains or solar gains, so I expect our heating load will very rarely actually reach that level. I have not yet talked to the HVAC contractor about specific furnace and AC units. Am I likely to have trouble finding units at the correct size?
Ventilation & infiltration numbers in load calculations are typically overstated by a factor of at least 2. The "heat exchanger effect" of infiltration means that the air coming in is never at the outdoor temp, and the air escaping is never the interior temp. If your infiltration was a large round hole it might reach those numbers, but you could actually just put a sock in it, eh? :-)
You can also ventilate with an HRV rather than an exhaust fan to reduce that number by 80-90%. But even with an exhaust ventilation scheme, unless the incoming air is a dedicated port there is still significant heat exchange on the incoming air.
34.5K of load at a delta-T of 70F for a 2400' house 2x6/R20 house with 2" of exterior foam and code-min windows isn't reality, but 26-28K might be.
In Flint a better-class ducted mini-split would average a COP over 3, the best wall-coil head units would be getting close to 3.5. Even the older Mitsubishi FE12s (HSPF ~10.5) were getting a fleet-average (10-12 units in occupied houses) of about 3 in a zone 6B climate in large in-situ monitoring project performed by the NEEA a few years ago. The newer cold-climate mini-splits test in the HSPF 12s & 13s (some hit 14) the ducted Fujitsus are in the mid-11s:
Ducted minisplits in the HSPF 11s:
http://www.fujitsugeneral.com/PDF_06/Submittals/18RLFCD%20Submittal.pdf
http://www.fujitsugeneral.com/PDF_06/Submittals/12RLFCD%20Submittal.pdf
A few random wall head units north of HSPF 12:
http://usa.mylinkdrive.com/uploads/documents/4561/document/MSZ-FH12NA_MUZ-FH12NA_Submittal.pdf
http://www.mehvac.com/media/953333/msz-fh15na_muz-fh15na_submittal.pdf
http://www.fujitsugeneral.com/PDF_06/Submittals/12RLS3HSubmittal.pdf (HSPF=14.0!)
http://www.fujitsugeneral.com/PDF_06/Submittals/15RLS3HSubmittal.pdf (HSPF=13.4)
The FE12 that averaged 3.0 in a 6B climate in real-world monitoring:
http://www.mylinkdrive.com/uploads/documents/1642/document/MSZ-FE12NA-MUZ-FE12NA_Submittal.pdf (HSPF=10.6)
An assumption of sub-3 average COPs in Flint would be a gross underestimation of current models- even the better-class ducted ones. It's been too easy to sell them short.
Dana,
I came up with my 2-3 estimate using date from http://www.neep.org/initiatives/high-efficiency-products/emerging-technologies/ashp/cold-climate-air-source-heat-pump which lists COP at different temperatures and different loads for a variety of units. I used Fujitsu AOU24RLXFZH which is a multi-zone system. The cold temperature performance ends up dominating the average since a number of factors combine to make the energy use during the few days at those temperatures very high. If the conclusion had been close, I would have spent more time trying different units. I don't doubt that I could have found a unit that exceeded 3.
Dana - Thanks for the advice. In my current big, old, leaky house I used 254 therms in January and paid $216 for the gas (not including $15 in tax). Dividing that out and using the calculations you have above, it looks like gas would come in at around 8.90/MMBTU, or about 10% cheaper, although if the mini-split ends up being more efficient it is probably close to a wash.
You mention using a mini-split on the first floor and a mini-duct cassette on the second floor. Any estimate of the pricing with install of something like that? That sounds more expensive than 2 mini-splits heads with one exterior unit.
Thanks for all your input.
Mark
In your situation it looks like gas furnace is the best choice. Just make sure the ductwork on the gas furnace is inside the thermal envelope of the new construction. Heating the attic or crawlspace/basement like a good portion of new construction does is insane.
A gas furnace is only a better choice if you're not planning to have air conditioning.
Two mini-splits are often cheaper than a 2-head multi-split solution, and is usually more efficient. A co-worker of mine recently installed 3 mini-splits (two 1-tons and a 3/4 ton) in his house for about $12K total, but was also quoted a multi-split solution for about $15K. There were several other proposals & quotes, many of which were simply insane, such as a ductless head per room (the installer needed to make a boat payment?), or units that had poor efficiency or insufficient capacity at his outside design temp, or merely overpriced due to the high early summertime demand. It pays to know what you need, (getting the load calculations dialed in correctly), knowing the equipment, and get competing bids
A ductless 1.5 ton cold-climate mini-split runs about $4.5K in my neighborhood, a 3/4 ton runs about $3.5K. Going with a ducted version would add between $500-1000 depending on the complexity of the ducts and where/how the cassette is installed. You should be able to get the whole shebang for under $10K in new construction, with sufficient foresight and planning.
What you get out of a mini-split solution that you don't get out a comparably priced condensing gas furnace + split AC is right-sizing and quiet modulating variable speed air handler. It's quieter and more comfortable. While you could get variable speed gas & split AC systems, that ends up being more expensive than mini-splits. And by having separate equipment for each zone, the upstairs/downstairs temperature balance issues go away, and there are no single-points of failure- you'll always have heating/cooling on at least one zone.
In the financial analysis we didn't attempt to calculate the air handler & other power of a gas system either. With heat pump equipment with HSPFs north of 11 it's highly likely that the marginal operating cost would be lower, since all power is calculated into the HSPF number (as well as some that isn't, since HSPF testing assumes a faster roll-off in capacity below 17F than actually occurs with cold climate mini-splits, and has assumptions about resistance heating back-up that is not actually there or needed if you size the mini-splits correctly.)