I am in the planning stage of building a new home, and because I won't have access to natural gas at my building site, I am looking into geothermal as an way to heat the home. Upstate NY, zone 5 (but nearly zone 6) well insulated 2400 sf house with 1600 sf basement (ICF). I am planning on radiant heat and either no AC, or a ductless unit added afterwards. Since the costs of the system are so expensive, I am wondering if people would/could recommend a DIY kit. Since I have plenty of acreage and will already have bulldozers and backhoes on site, I feel like I should be able to at least install the loops myself (plus avoid any mark up from an installer). What manufacturer/kits would you recommend?

Kits aren't common, but google does show some companies selling them. You'd want to be sure you are comfortable with what you are doing yourself and who you are hiring to do other parts, and make sure they are comfortable with the arrangement. For example, an HVAC company that only getting paid for 4 hours of final installation and isn't making any profit on the unit won't want to take responsibility for problems if it doesn't work well.

Another option is this new hydronic air-source heat pump that has performance to rival geo:
http://www.greenbuildingadvisor.com/community/forum/green-products-and-materials/46941/any-one-have-experience-chilltrix-equipment

If you do geo, do plan to use it for any a/c you do. Running it in that mode puts heat back in the ground which helps improve the winter performance. If you are doing a water-water system, you then need a fan coil unit to deliver the cooling. Chilltrix sells some very nice ones with low electric consumption and low noise.

I will do some more reading on that chillix system,

However the main reason I have written off mini split systems is our winters are simply too cold. We had 10-15 days this year where we didnt get above 0 F.

On another note it seems like people have written off geothermal? I thought everyone was calling it the cats-meow.

The shift has been that
1) mini-split heat pumps with cold-climate capability have become widely available at very low prices. Their COP at 0 F and below is not nearly as good as a geo system that doesn't care about the outdoor temperature, but they work and on a seasonal average can have COP 2.5 to 3, which is as good as a mediocre geo system, if not as good as a good one.
2) PV costs have come down so far that spending $$ for the last bit of efficiency no longer seems worth it vs. buying some extra PV.

I still like geo, but it does require skill in design and installation..

Will the mini split systems work with radiant slab heating?
I am not very familiar with COP either. What is a good score? I will also do some reading on my own. But, I am in NY and one of the only good thing in this state is that they will really subsidize the costs of installing Solar PV. 70% of the costs between state and federal.

COP- Coefficient Of Performance. Electric resistance heat has a COP of 1, spend 1 dollar get 1 dollar worth of heat, COP of 3 means for every 1 dollar spent you get 3 dollars worth of heat.

"Will the mini split systems work with radiant slab heating?" That Chilltrix system does radiant, for summer cooling you would need a air handler to send chilled water to.

ChrisJ

Here's a guide to what counts as a good COP--very subjectively and I'm sure some will disagree, but to get you started:

COP = 2: 2X better than electric heat so that's good. Pretty awesome actually, but it's possible to do better.
COP = 3: Very good, a system to be proud of.
COP = 4: Living the dream--what we'd hope a geo system would achieve.
COP = 5: Very rarely achieved in practice.

A mini-split is not a hydronic system- you won't heat your floors with them. In a zone 5/6 boundary climate using best-in-class air-to-air single head cold climate mini-splits you'll get a seasonal average COP between 3.0 and 3.5, not more. Even prior-generation Mitsubishis were hitting around 3.0 for fleet averages in an in-situ monitored cluster in a zone 6 climate in Idaho a few years ago. The newer FHxxNA Mitsubishis are 15-20% more efficient than the FExxNAs tested there, and the newest Fujitsu xxRLS3H series has a modest efficiency leg up on the FH Mitsubishi units (while not exactly leaving them in the dust on efficiency, the Fujitsu units have substantially more heating capacity per rated cooling ton.)

The as-installed COP of many professionally designed & installed geothermal systems don't really beat that by much, even though a meticulously designed GSHP system using state of the art equipment might deliver 4.5 in this climate. As a DIY, don't count on even making 3.0 with geothermal unless you are a true engineering genius. It's more complicated to get right than it looks, and experience counts. By contrast, mini-splits are a pre-engineered "system in a can", where the number of ways to screw it up are far more limited. (Though the more idiot-proof they make them, the more creative the idiots become. :-) )

It's always a bad idea to pick the heating equipment type or size before calculating the loads. There's no definition for "well insulated" , and the design details of the house actually matter. Design the house first, and be prepared to change details where it's cost effective to lower the loads. Only when you know the loads should you be considering whether geothermal vs. air-source vs. something else, based on the cost and efficiency of what's actually out there that fills the bill. There are "well insulated" 2400' houses with full basements that have heat loads north of 50,000 BTU/hr @ 0F, and others with heat loads under 25,000 BTU/hr.

There are many housed heated with mini-splits in MA (zone 5) that are Net Zero Energy with PV that fits on the roof, but it takes some careful design and implementation on the house to get there. If that's a goal, figure on "whole assembly R" (with all thermal bridging of framing factored in) of at least R28-R30 for the walls, R60+ for the attic, R15+ for the basement walls, and R10 under the basement slab with a limited amount of U0.28 or lower window.

Download BA-1005, and read at least the first chapter:

http://buildingscience.com/documents/bareports/ba-1005-building-america-high-r-value-high-performance-residential-buildings-all-climate-zones/view

I would challenge any claim of an air sourced heatpump to rival a ground sourced HP in terms of performance.

Especially the new variable speed technology introduced in GSHPs delivers real world COPs in the high 4s, while dual stage units are running in the lower 4s, including pumping and fan power.
The second issue is whole house distribution of the single head air sourced units. Also COP rating for air sourced units need to be discounted in the real world by about 25-30%, since certain operations (such as heating your house with strip heat during defrost mode) is not included in the rating.

True, the state of the art GSHP systems will beat any air source heat pump efficiency, in a system that is well designed and well implemented.

And how many of the state of the art GSHP units are available in kits, or suitable for DIY system design & installation by folks such as logbuilder7?

Like I said, experience counts! A docjenser designed & installed system could possibly edge into the high-4s , but that's probably better than the local average in my area, and WAY better than the average DIY GSHP design & installation.

Mini-splits do not even have heat strips for heating the house- they defrost by reversing the cycle. Some cold climate mini-splits have pan heaters to clear defrost ice from the bottom of the pan to avoid damaging the blower, but that's pipsqueak power applied intermittently compared to supporting the full heat load with strip power during defrost.

At typical oversizing factors mini-splits outperform their HSPF numbers slightly, though if undersized they can and will underperform their HSPF. The oversizing factor is one of those few things that you actually have to get right with a mini-split, but it's not hard math. The HSPF test parameters are based on a 1-speed heat pump model in a zone 4 climate, and don't fully reflect the phenomenal gains in efficiency from fully modulating systems at 1.25-1.5x oversizing. At 1.5x oversizing mini-splits nearly hit the climate zone 4 tested efficiency even in climate zone 6.

The Fujitsu xxRLFCD series mini-duct minisplits don't have the distribution problems of a wall-blob type, and will still beat an seasonal COP of 3 in a zone 5/6 boundary location. They have lower efficiency (HSPF in the 11s) than the cold climate wall blobs (HSPFs in the 13s & 14s), but are still more efficient than the cold-climate wall units of only 4-5 years ago. They're fully rated down to -5F outdoor temps, which means you can size them properly for the warm edge of zone 6. Even though they won't specify the output at temps lower than -5F, they don't just crap out and die there- they keep going and putting out decent heat into negative double digits.

That's not to say that this THE solution for the as-yet-unknown loads, only to point out that the notion that you're stuck with point source heating when going with mini-splits in cool climates hasn't been true for awhile now. The distribution and defrost heat strip and tested shortcomings are bunk, as applied to state of the art ducted mini-splits, though they might be true for other potential heat pump solutions. (The equipment is still pretty cheap in comparison to even standard fossil burning hydronic solutions too.)

But I'll grant that can't easily get radiant floor heating out of a mini-split, ducted or otherwise! :-)

Cheap air source chillers (Chilltrix or others) aren't really suitable for the climate- most don't even have output ratings below +25F or so. That's fine in climate zone 3, or marine zone 4, but for zone 5/6 fuggedaboutdit.

Concentrate on the building envelope first, run the load numbers, and report back. With a fairly simple building shape and some careful decisions on window sizing/performance/placement you'd be able to hit Net Zero fairly easily following the whole-assembly R guidance in table 3 of in the BA-1005 document. With a Net-Zero-Capable building envelope the design heat load will come under 25,000 BTU/hr or even under 20,000 BTU/hr, making the HVAC systems cheaper and simpler to decide on.

Dana articulates the advantages of modulating compressors in minisplits. When that was available in minisplits and not in GSHPs, the advantages of GSHPs weren't that great, but as @docjenser points out, that technology has arrived in GSHPs. So that means that these few advanced GSHPs have the advantages of minisplits Dana describes, plus the advantage of a high-heat capacity high temperature source.

From my point of view, the variable-speed compressor products on the market are not very interesting because:
1) They are big, too big for a really well insulated modestly sized house, and
2) They are only available with forced air distribution. I am a fan of hydronic distribution, and have my GSHP set up that way, so the variable speed compressor units on the market now don't fit my needs, so I tend to (mistakenly) forget that they are available for people with ducts and big houses with enough duct leaks that they need 2.5 to 5 tons of heating.

The Chiltrix is rated to -13 F, by the way. COP 2.66 producing 104 F water at that point. If someone would put that compressor in a ground-source water-water unit, it would get double that, but 2.66 is not bad for -13 F!

chrs: Show me the documents supporting those claims regarding Chilltrix! (I may have misjudged that one, lumping it in with some others...) And the output capacity at -13F outdoor temps & 104F output water is... ???

I don't have any arguments with GSHP other than the installed price and as-installed-performance issues, which vary by quite a bit with installer competence- and is never slam-dunk. To presume that a DIYer newbie GSHP designer would be able to get anything anywhere NEAR optimal or nameplate efficiency out of a self-designed self-installed GSHP system takes quite a stretch of the imagination.

A DIYer would have to come up to speed on many aspects, and should probably go into the biz if they actually get that kind of performance out of their first system.

By contrast, DIY installation of a mini-splits can be done by modestly handy not super-technical people with only a small amount of hand-holding or outside tech-time, and they pretty much hit their performance numbers, barring a major glaring screw up.

But designing out the load in a new house does more for comfort & operating cost than any heating system can, independent of it's efficiency.

Dana, I can point you to the data Chiltrix has online, but I can't vouch for how accurate it is. Scroll down on this page to plots and data tables.

http://www.chiltrix.com/chiller-technology.html

One nice thing about a hydronic heat pump is that it's easy to measure the heat output from flow and temperature--much easier than measuring the output of a minisplit. So there's a good chance the data is accurate if we trust it's not fudged. And it would be easy for an engineering minded owner to verify that it's performing to spec.

Posted By Dana1 on 07 Oct 2015 02:14 PM
Mini-splits do not even have heat strips for heating the house- they defrost by reversing the cycle. Some cold climate mini-splits have pan heaters to clear defrost ice from the bottom of the pan to avoid damaging the blower, but that's pipsqueak power applied intermittently compared to supporting the full heat load with strip power during defrost.


The distribution and defrost heat strip and tested shortcomings are bunk, as applied to state of the art ducted mini-splits, though they might be true for other potential heat pump solutions.

So what is the source for the defrost heat? By definition this means they are blowing out cold air, at least they used to do this, without strip heat. This has to made up by additional heating BTUs, which are not included in the rating (AHRI-ISO). The state of the art ducted mini splits or multi head units have significantly lower COPs than single head units, and are more expensive, getting closer to geo in price.

Posted By chrs on 07 Oct 2015 03:46 PM
Dana, I can point you to the data Chiltrix has online, but I can't vouch for how accurate it is. Scroll down on this page to plots and data tables.

http://www.chiltrix.com/chiller-technology.html

The table, clipped from that page:



Note, those are wet-bulb temps, which are necessarily lower dry-bulb temps the outdoor air temp.  In US climate zones 5A-6A (the northeastern quadrant of the country) wintertime w.b. temps are typically a degree or three below dry bulb temps, but it varies quite a bit with weather. So add a couple of degrees to those numbers.

Also, on the efficiency numbers, that's not the all-in power use. You have to add in the pumping and control power use to come up the full-system COP, just the way you'd have to with a GSHP system to be able to make direct comparisons with mini-splits, which have only one point of power draw.

Looks like with 104F output water it  can deliver about 12,000 BTU/hr @ -13F wet-bulb, or about -10F dry bulb which is more than nothing, but it takes a fairly small and superinsulated house for that to be the primary source of heat. A 1-ton Fujitsu matches that on capacity at that temp.  It puts out a decent amount of heat at a w.b. temp of +14F though, which would make it suitable for a number of US climate zone 4 locations.

docjenser: Yes, the source of defrost heat for mini-splits is indeed the indoor air.  During defrost mode they throttle back the blower speed to the minimum speed on the indoor head to limit the potential comfort problem.  I thought you had a Mitsubishi GExxNA at your house, and had observed it's behavior in defrost mode(?). Perhaps I'm remembering someone else complaining about how it was spending all it's time in defrost mode and never heating when it was +5F outdoors, which would be atypical for the FE or FH series units (not sure about the GE.)

SFAIK there has never been a mini-split with resistance heating strips for either defrost or supplemental space heating during defrost. That's a common feature on traditional ducted heat pumps though.

Debunking more disinformation:

The 10.5-ish HSPF of an older FE series Mitsubishi (displaced by the FH series about 2 years ago with HPSFs in the 12s) was an implied COP of about 3.1 in a zone 4 climate.

The NEEA field-monitored a fleet of 10-12 of the 1-ton units in-situ in occupied homes in Idaho Falls ID (climate zone 6A) at a seasonal fleet average COP of 2.97, not far below what the HSPF test was predicting for a much warmer climate. In-situ testing of individual units in western WA (zone 4C) were delivering averages in the 3.5 range, well above the nameplate HSPF efficiency for the climate.

The HSPFs of the Fujitsu RLFCD series mini-duct units are in the 11s & 12s and implied COP of 3.2 or better, in a zone 4 climate. It's not a stretch to think they'll break a COP of 3 in a zone 5 climate at 1.25-1.5x oversizing.

State of the art for wall-coil type mini-splits in 2015 is HSPFs in the 13s and 14s, an implied COP of ~4 in a zone 4 climate.

The hardware cost for a 1.5 ton mini-duct cassette mini-split + compressor good for ~20K of output at climate zone 5 outside design temps isn't higher than the hardware cost of a 1.5 ton cold climate mini-split of comparable output, it's comparable or even a bit lower, and even with the cost of ducting nowhere near the installed cost of a GSHP system of similar capacity.

Below zero F, the spread between wet bulb and dry bulb gets really small, and in fact -13F wet bulb at -10F dry bulb would require negative humidity :).  In other words, -11 is the warmest dry bulb you could have at -13 wet bulb, and at 50% humidity you'd have -12 F.  Realistically, -13 wet bulb means -12 to -13 dry bulb, for typical winter humidity.  So Dana's 3F correction for that was at least triple what it should have been.

Next is the comment that one needs to add pump power.  But the Chiltrix CX30 has the hydronic pump on the output side integrated, so presumably they included that power in the measurement, unless they specially rewired it for the test or something else fishy.  You do need to add fan power, but the ECM fans on their fan coils are pretty good--~10-20 W.  Three of them could add ~5% to the power consumption, and bring the 2.66 COP down to 2.5.

The reduced output at -13 F is a problem in some applications; it is possible to gang two units together to meet higher loads.

Also, the data table is for operation at maximum capacity. Like minisplits, the Chiltrix can and does modulate, and could get similar COP boosts from doing so. You can get an idea of the benefit you might expect by looking at how the COP goes up with lower water temperature, which would be part of how it got better efficiency running at part load. It's not fair to compare the Chiltrix COP running full bore to minisplit average COP over a range of operating conditions.

We'd need either monitored energy use in similar houses, or a full efficiency map including different compressor speeds to do a conclusive comparison between Chiltrix and minisplits. But just an apples-to-apples comparison at full load would be a good start. I don't see that for the new units, so perhaps the best we can do is look at the 12RLS and extrapolate by a factor of the HSPF ratio 14/12. Measured full bore operation of the 12RLS from http://www.nrel.gov/docs/fy11osti/52175.pdf is
outdoor temp COP____ 14/12 COP
-3 F ________1.9 _____2.2
28 F _______ 2.25 ____2.62
48 F _______ 2.6 _____3.0

For the same temperatures, adjusting for wet bulb and discounting for fan power, the chiltrix gets 2.56, 3.64 and 4.6. Those aren't real-world conditions for either, but the improvement by operating a lower compressor and fan speeds applies to both. Yes, the minisplits are better proven by now, but with the data we have, the chiltrix looks pretty great. If anyone can point to newer minisplit COP data vs. temperature and compressor speed, we can try to compare better. Ideally we'd have chiltrix data for part load too, but we don't.

Good catch on my late-day laziness on the narrow spread between wb & db temp at sub zero temps (what was I thinking? :-) )

I would fully expect modulation of the compressor to improve performance considerably, but without something to sync modulation of the other system pumps/blowers for optimum efficiency it may not be quite as dramatic a modulation improvement as with mini-split technology.

The output capacity at +5F w.b. is still within range of super-insulated houses in a zone 5 climate with outside design temps in the high single digits or low double digits (say, Boston MA or Rochester NY.) With my newfound awareness I just pointed out a guy near Poughkeepsie NY in a small lower-load house toward the Chilltrix, but his water temp requirements may be too high to be effective in his application (TBD). (He's currently running his retrofit radiant floors off an oil-fired HW heater.)