I apologize upfront, I am a complete noob at all this...

We are building a home in Albany NY and started to think about HVAC systems. Some details on the house: about 1600sqft 1st floor, 800sqft 2nd, full basement that will be finised. Foam insulation and insulated slab/foundation (I think the 2" foam?). We plan on having a woodstove in the basement. My question is would a ductless mini split (I think that is the term, Mitsubishi makes them) with solar panals to offset the cost of the electric be as efficient as installing a geo thermal system? Would the mini split system be able to heat a house in upstate NY?

We have had a few estimates on the geo system, from about 35k to 60, depending on vertical (I would lean this way due to soil conditions) or horizontal. I'm not sure about the mini splits but I believe they are less expensive, however I expect the solar panals to bring the cost of this up to about the cost of the geo...but at least with the solar panals would be benefiting the house even when the heat or AC is not running. I also started to look into solar hot water and it seems this is a definite option too.

I know with the geo NY gives 30% of the installation cost but am not sure what incentives there are for solar.

Any help is greatly appreciated. Thanks

The first and very key thing to start out with is a very precise heat load calculation, which may require using DOE2/BeOpt rather than a generic Manual-J type heat load tool if this is a high-R house. If the $35-60K estimates given were for 6 tons of geo, and the true design condition heat load is under 2 tons, you can imagine what that does to the ultimate cost of geo, and the viability of heating with ductless/mini-splits. In new construction it shouldn't be too hard to hit 2 tons or less on a house that size, and at $8-9K/ton (average southern New England local pricing) for geo that wouldn't be too bad, if it's sized dead-nuts-on the heat load. But subsidies can still skew the numbers in either/both direction(s).

In MA when the heat load drops under ~2.5-3 tons-heating the competing subsidies for geo and photovoltaic panels ( PV) are somewhat balanced. Above 3 tons geo almost always wins, under 2- tons the PV + ductless wins, but it's a moving target. That target has been moving steadily toward the PV + ductless solutions though, since the costs of PV have crashed at a time when the local & federal subsidies for PV has increase. And over the past 5-10 years the absolute cool-weather efficiency of mini-splits have been on the rise. Five years ago I was a mini-split skeptic for this climate zone, but I've become something of a fan. Still, MA has some of the fattest subsidies for residential PV of any US state at the moment, and YMMV.

It's not all apples-to-apples- heating with ductless in an Albany climate REQUIRES very low heat loss from rooms doored-off from the zones with the ductless heads, but it's done fairly often. Also note, there will be some number of hours when it's below freezing and the air temp is pretty much at the dew point temp, at which point the output of the ductless is pretty-near zero. In a high-R house with at least some thermal mass most of the time you'd be able to ride it out without becoming uncomfortable, but local code or lending institutions may require backup at 100% of load, even if you literally never use it.

And the average efficiency of ductess in this climate will lag that of geo substantially. At a minimum you'd expect to average a seasonal coefficient of performance (COP) of ~2.5 for ductless, ~3.5 for geo (though best-practices and state-of-art can improve on both). But with very low heat loads the PV can often more than make up for the efficiency difference in total annual power use.

Mitsubishi, Daikin, and Fujitsu all have units capable of reasonable output at Albany's 99th percentile design temps of ~0F (+/- 2F, depending on your exact location.) Mitsubishi may be the biggest player locally (they invented the mini-split, after all) but don't rule the others out. Under the hood they're all pretty similar, but the control algorithms so key to as-used efficiency varies.

Bottom line- calculate the heat load at 0F with a very sharp pencil/tool, then you'll be able to assess what you really need, and how to get there. Unlike fossil-burners, every additional ton of geo costs a bundle, which rationalizes high-R/low-U and blower-door verified air tight construction, since a ton of load-reduction has zero operating cost and has a higher comfort factor than any heating system. Even if it costs $10K to pull the heat load down another ton,with exterior foam & ultra-low-U windows, it's worth it in the long run, even if it's only defraying $9K in geo cost, since the insulation should be good for a century (if not more) of zero maintenance, whereas there will be periodic repair/replacment of heating system components.

Only when you know the true heat load can you sketch out the un-subsidized costs of different approaches, and work-in the subsidies from there to see what makes sense.

Start shopping for subsidies here: http://www.dsireusa.org/incentives/index.cfm?state=NY

There may also be other local or utility based subsidy for different aspects of the house or heating system under demonstration programs, etc. see: http://www.nyserda.ny.gov/

I'm far from an expert on these matters, but have you also thought about supplementing the ductless minisplits with electric heating, either in floor or baseboard? It could really help when the temp and dew point drops and the minisplits have a hard time keeping up. Electric heat is cheap to install (well, it CAN be) and it could really help give the minisplits a boost when the temps are frigid. It's more expensive to operate, but it wouldn't be used but only on the coldest days and it would be partially offset by the pv energy. Doesn't raise the upfront cost of the minisplits route a lot but might make it work better. Just a thought.

If you find that PV + heat pump makes sense, I would look at a "to water' heat pump that would allow you to produce the heat during the day and store it for later use. Battery storage is expensive.

To my way of thinking the difference between using a geothermal system and a minisplit comes down to the heat load of the house. For a house with a small heat load, use a minisplit; with a larger heat load you may be better off with a geothermal system. You say your house will have "foam insulation", but that is only a small piece of the insulation design. It's not hard or expensive to build a superinsulated house which you can heat inexpensively but it does take knowledge of what is important, a well thought out insulation strategy and attention to detail. You will need to minimize thermal bridging (which any amount of foam in a stud cavity will NOT do), minimize air leakage and install a heat recovery ventilation system. Unfortunately these are things that the vast majority of builders are still ignorant of, but there are more an mores good builders who "get it" as well as a growing army of energy auditors who understand the term "built tight, ventilate right". If you are already set in your ways, go with the geothermal - you'll spend unnecessary money but it is more adaptable and will probably have fewer issues in a semi well insulated and not-too-tight house.

Done right, you can heat this house easily with a minisplit (check out South Mountain Company's Eliakim's Way project). And, the cost of upgrading your construction plus the heating system will probably be below the 60K for your geothermal system.

Posted By strategery on 04 Jul 2012 05:16 AM
I'm far from an expert on these matters, but have you also thought about supplementing the ductless minisplits with electric heating, either in floor or baseboard? It could really help when the temp and dew point drops and the minisplits have a hard time keeping up. Electric heat is cheap to install (well, it CAN be) and it could really help give the minisplits a boost when the temps are frigid. It's more expensive to operate, but it wouldn't be used but only on the coldest days and it would be partially offset by the pv energy. Doesn't raise the upfront cost of the minisplits route a lot but might make it work better. Just a thought.

The dew point is only relevant to mini-split performance when two critical conditions are in effect:

1: The dew point is very close to the outdoor temp, and

2: It's below freezing, which requires the defrost cycle to bring the coil up by a larger difference in temperature to shed the frost.

The rest of the time output capacity it's a roughly linear function falling with temperature, with a change of slope near the freeze point.  In general the colder it is outside, the less moisture there is in the air, which means fewer defrost cycles.  But when most of the heat being pumped is the heat of vaporization + heat of fusion of the moisture in the air, and most of that needs to be returned to get rid of the frost, the capacity falls.  But with a 20F delta between dew point & outdoor temp is more than enough to keep the output pretty much on it's output capacity/temperature curve. 

When it's below freezing, the lower the dew point the better it is for output capacity.  (Or perhaps better stated, the bigger the difference between the dew point and the outdoor temp, the better it is for output capacity.) eg: Looking at the fine print on the specification sheet, a 1.5 ton Mitsubishi (MUZ-FE18NA) can still put out 19,300BTU/hr at an outdoor temp of +17F if it's at arctic-dry dew points, but when the wet-bulb temp= +15F, dry bulb temp= +17F (dew point= +10F,  or 7F below the outdoor temp) that gets derated to 11,700BTU/hr, a 40% hit in output capacity.  At even higher dew points it's output will surely be less, but they don't publish the actual curves of output vs. dew point  at fixed outdoor temps in the generic spec.

While some will find this output-uncertainty problematic, on a high-R house with sufficient thermal mass to give it a reasonably long time constant in terms of degrees/hour when being underheated, the temporary BTU/hour shortfall over the coldest overnight hours is mostly irrelevant, since during the warmer part of the day both the output and efficiency rise dramatically.  Even at current code-min in MA most houses could get by quite comfortably with heating plant output well under the Manual-J calculated steady-state heat loss numbers, but it still has to be able to do better than heat loss at the average outdoor temp on the colder days.  The time constant of a house is determined both by it's thermal mass and it's R/U values- the better insulated and more massive, the longer the time constant, the less problematic temporary heating output shortfalls become. 

On new construction it's much cheaper to go high-R/moderate-to-high mass than when retrofitting, and in an Albany climate it can cost less than the difference between higher-tonnage geo and lower-output ductless. It's not a simple calculation- how you get there matters, and it's further obfuscated by subsidy.  Finding the optimal financial crossover point is far more complicated than the optimal comfort point- high performance/low heating & cooling load building envelopes win hands down over highest efficiency mechanical systems (up to about R40 whole-wall values in this location, which usually results in design condition heat loads under 24KBTU/hr for moderate sized homes, if R40 is used as a shorthand for R10/R20/R40/R60 for slab / foundation-wall / above-grade wall / attic-or-roof type construction with low to moderate window/floor ratios and

Posted By jonr on 04 Jul 2012 07:41 AM
If you find that PV + heat pump makes sense, I would look at a "to water' heat pump that would allow you to produce the heat during the day and store it for later use. Battery storage is expensive.

Unless this house is intended to be off-grid ANY energy storage for space heating won't have a financial rationale.  High efficiency air-to-water heat pumps rather than mini-splits more than doubles the system cost even before adding in the cost of storage tanks and low-temp hydronic radiation, not to mention the additional complexity it adds to being able to air condition (essential even in high-R houses in Albany, given their moderate to high latent loads in summer.)

Grid tied PV has a much easier economic rationale when net metered in an electricity market as expensive as NY, made even easier if other subsidies such as SRECs or tax credits can be applied.  The lifecycle cost per kwh of rooftop PV even at the current un-subsidized costs is at parity with the residential retail rates for much of NY & New England, using reasonable discount rates & near-zero electricity price inflation in a present value financial calculation.

As luck would have it, this bit o' bloggery about what it takes in a similar climate shows up in a timely fashion.

Thanks everyone for the help. The higher geo quote was for a 5 ton system (we initially thought of running radiant floor heat through the basement) and the lower was for around a 3 ton system (without the radiant heat). It seems, on average, this house would be between 2 and 3 tons.

As I said I'm quite new to all this so here is what I'm picking up...

1) Insullation, Insullation, Insullation...money spent here will be less spent on heating/cooling. I understand the insullation basics but one thing that I'm trying to figure out is "thermal bridging". Is this essentially where the studs touch the outside sheething and then sheetrock and there is no insullation? Is this why rigid foam is used on the outside of the house as well? Is Tyrap used on top of the foam or inbetween the foam?

2) Air leaks are very important to control. I've poked around this forum (unbelievable amount of information!) and found a link to a company that makes gaskets for the sill plate etc. and will pass this along to the builder. Is there a specific type of caulk to use as well? How crazy do I need to make myself trying to control every leak?

3) Is it possible to have the house too "tight", if the mini splits are used is there any outside "fresh" air used?

4) Make sure to take #1 and #2 are taken into account when sizing the heating/cooling system. Sizing of system seems to make a difference on the type of system and this size house could be on the border line of 2 - 3 tons?


We were thinking about putting in some electric baseboard heaters (offset by the PV) to help on cold days along with a wood stove. Also, we intended it to be grid tied...


Thanks again for the help!

Very interesting link and similar to the South Mountain Company's Eliakim's Way...

I just priced out a quote on SunPower's website for a 7.79 kW solar system for $35k after state and federal credits...for a 5.59 kW system about $22k

In order:

1> Yes, the rigid foam on the exterior adds R to the entire assembly, providing a thermal break over the framing. With fireblocking, headers, plates, and studs typical framing comprizes ~20-30% of the entire wall area. At a 25% framing-fraction in a wood-sheathed 2x6 wall even if you had some magic insulation that could make it R500 center-cavity, the thermal bridging reduces that to ~R22.  With typical blown cellulose/fiberglass/open-cell foam that's ~R20 center cavity the thermal bridging cuts that to ~R13-R14 "whole wall" (= framing R factored in.)

To hit R40 with an exterior foam approach takes about 4.5" of  exterior polyiso sheathing on the outside of the structural sheathing, which changes the details of how windows & doors are installed & flashed pretty significantly, and how you install the windows determines whether the housewrap goes between the foam & studwall vs. foam & siding. This has to be designed it, thought out ahead of time, not just hacked in.

Figure out how to get the sub-slab insulation continuous with the foundation wall insulation continuous with the wall insulation continous with the attic insulation with the least amount of thermal bridging.  For foundations, using insulating concrete forms (ICF) and putting 3" of type-II EPS or 2" of XPS under the slab butting up against the interior-side foam of the ICF works.  Designing the above-grade wall's insulating sheathing to be co-planer with the exterior ICF foam makes that transition pretty easy too.  Roof-to-wall insulation and air sealing details will vary, depending on whether it's a cathedralized ceiling vs. trusses vs. joisted & accessible attic.

2> Gaskets are useful, (necessary in some instances), but not sufficient for air tightness.  Caulking or gluing each sheet of the structural sheathing to the framing as you go helps- at the very least caulk/glue/foam around the perimeter of each sheet, if not at each stud. (Using low-expansion gun foam fills in any knots or imperfections a bit better than other methods too.)  A bead of glue between any doubled top/bottom plates of the studwalls helps too.  There's no rocket science involved, and it's not expensive, but it requires thinking ahead as to what components will constitute the primary air barrier for floors/walls/slabs and how they are going to be made continuous at the transitions to end up with it air tight on all 6 sides of the cube, and how that will work at windows & doors.  Low expansion gun foam (and a pro-type foam gun) that screws onto the 20-24oz cans is a good investment, as is a powered caulking gun (preferably for the larger size tubes, since most acoustic sealant type caulks come only in the larger tubes.

3>There is no such thing as a house that is too tight.  Ventilation via accidental infiltration doesn't put the ventilation air where it needs to be, and is a HUGE energy drain.  Plan on installing a  (~$2000-3000) energy recovery ventilation system (ERV).  Mini-splits do not exchange air from interior to exterior.  Define the primary air barrier, and be vigilant about sealing it as you go (use the wooden sheathing as the wall component of the primary air barrier, since it's far more rugged than housewrap or exterior foam, but I'd seal up any of those layers too.)

4>The heat load of the house is determined by the difference in interior/outdoor temps, and the U-factor of all exterior surfaces. The units of U-factor are BTU-per hour-per-square-foot-per-degree difference. For wall and roofing area the U-factor is 1/R (where R is the whole-wall, not the center cavity R). U-factors for windows are on the label or specification. As the designer you have control over those numbers, and can tweak them to almost any arbitrary limit. But the bigger the window area, the more expensive that becomes.  In an R20 whole wall (a 2x4 cellulose wall with 2" of exterior foam) type house with (probably code-max in Albany) U0.30 windows the window losses and wall losses are roughly equal, but if you bump the walls to R40 the U30 windows rapidly dominate the heat loss number unless you cut back on total window area.  Using freebie tools like BeOpt are good for figuring out the most economical approach to hitting your heat loss and energy use targets.

The heat load shouldn't change with radiant vs. other, and 5 tons is an INSANE number for even a code-min house your size, given Albany's cool but non-artic ~0F outside design temp.  A code min house would probably come in at or under 3 tons, but with some careful design tweaks it usually won't break the bank to bring it under 2 tons.   And it's at that point where mini-split & multi-split solutions start looking really attractive.

Oversizing a ductless by 25-50% for the actual design condition load turns out to have a benefit to average operating efficiency.  These are fully modulating systems that operate at highest efficiency when the compressor speed and blower speed is in their lower 1/3 of  the range, and by oversizing you guarantee than it's almost never running in it's lowest efficiency flat-out-max mode. But since the modulate with a high turn-down ratio (greater than 1:3 for most), it still never short-cycles.  Placement of the interior heads where they have he least visual impact and lowest wind-chill-on-skin aspects takes some thinking too, and is more problematic when you go every bigger (since the minimum cubic feet per minute eventually becomes too much draft), but putting a a 2.5 or 3 ton 2-3 head multisplit isn't huge overkill for a 2 ton load at design temp.

You may be required by lenders or code to put in the electric baseboard as backup, whether you ever use it or not.  (This will vary with local codes & lenders.)  But mind you, if you have  2-ton load at design condition that is a load that could be met with seven 1.5kw space heaters.

In most moderately high-R homes the rooftop PV won't support much of the space heating load in winter, since that's when the sun is weakest, but the June/July peak-output time can make up for it.   Marc Rosenbaum has been able to keep power use low enough to make his mini-split heated place on Martha's Vinyard approximately net-zero-energy on an annual basis, but output falls behind load during the dim December/January time frame.