kleach: 

I've only been talking about mini-ducted systems here, not the point source wall-blobs (with the possible exception of the basement), and it's really an efficiency choice.  The wall-blobs are more efficient, but point-source, and some people hate the look, which is fine.

The mini-duct units put out about 20% less heat per kwh than the latest-greatest wall blob types, but ahve comparable or better efficiency than the wall-blobs of just a few years ago. The mini-duct cassette itself is a tiny air handler located in a ceiling , wall  or floor joist, with the ducts running to wherever they need to, just as with any ducted heating/cooling system.

Locating the outdoor unit(s) where they don't create aesthetic problems is no more difficult than hiding a propane tank or a split-system AC compressor.  If there's a will, there's a way.

Or, you could spend an extra 50 grand for a GSHP system and cross your fingers.

docjenser:

The notion that auxilliary heating with wood stoves etc in some of the homes tested skews the fleet average performance in a large way is an argument that needs to be supported.  The crankcase heaters and pan heaters would skew the COP numbers downward in a zone 6A climate, independently of it's modulation level when it's cold outside.  In some individual cases that may in fact happen that the woodstove operation backed off the modulation levels sufficiently to skew the numbers, but it's unlikely to have been a significant factor for the fleet averages.  But even if it's as much as 10% for the fleet average- doubtful, but let's go there anyway, the RLF mini-splits bench-test at about 8-10% more output per/kwh than the FExx Mitsubishi's, and are tested at a higher modulation level relative to their maximum output than the FExx are tested in the submittals. Higher modulation levels usually means lower HSPF efficiency, not less.  Then, factor in the much warmer climate of RI (higher outdoor temp = higher COP with ASHP) and the fact that it's no tcold enough in RI even at the mid-winter average temps for the crankcase heaters or pan heaters to engage more than half the time. It's really reaching to say that it won't make a COP of 3 in RI.   I don't think it'll make 3.5, but it'll beat 3.

The long term monitored performance of the mostly FExx Mitsubishis at the cold edge of zone 5 in Devens, MA were pretty similar, and the mini-splits were the sole heating source. I suppose you'd have to point out that most of them are a bit oversized by 150-200% for their actual loads at the local 0F outside design temp, which means it skews them to a slightly higher average efficiency too.  (I've been suggesting 1.25-1.5x oversizing factors for capacity at +10F to keep them modulating into the shoulder seasons for comfort).  The average outdoor temps are more than 5F colder in Devens than coastal RI. Unlike RI,  Devens it's cold enough that the crankcase heaters would run almost continuously for 2-3 months in winter sapping the efficiency despite the oversizing. The only unit that had significantly lower efficiency was on Lot #23 where the occupants kept turning the thing on/off rather than letting it modulate. See the discussion on p38, p52 of the PDF pagination.

Bench testing of an FE12 shows that even at -13F when running FULL OUT it achieves a COP of 1.8, but about  2 running full out at +10F. The average January temp in say, Westerly RI, runs about +30F, a temp at which the FE12  hits about 2.8-3 at max speed (see figure 10 in the bench test link). 

The Fujitsu 12RLS2 hits  a COP of 3 running intermediate speed at +17F, and over 3 at intermediate speed @ +30F, and over 4.0 at low speed @ +35F (see figure 5, p.18 in the PDF). This is where a Fujistsu xxRLF would be run if sized with a bit of capacity margin @ +10F.

These bench tested units were from over 5 years & 2 product revision/model generations ago. The technology has seen significant incremental improvements since then.  No matter how much they ran the wood stove in Montana or Idaho, a 2015 Fujitsu is going to do at least as well as a 2009 Fujitsu. For these reasons it's near-certainty that in an RI location it should average at least 3.0, and not less.

Posted By Dana1 on 15 May 2015 03:52 PM

docjenser:

The notion that auxilliary heating with wood stoves etc in some of the homes tested skews the fleet average performance in a large way is an argument that needs to be supported.  The crankcase heaters and pan heaters would skew the COP numbers downward in a zone 6A climate, independently of it's modulation level when it's cold outside.  In some individual cases that may in fact happen that the woodstove operation backed off the modulation levels sufficiently to skew the numbers, but it's unlikely to have been a significant factor for the fleet averages.  But even if it's as much as 10% for the fleet average- doubtful, but let's go there anyway, the RLF mini-splits bench-test at about 8-10% more output per/kwh than the FExx Mitsubishi's, and are tested at a higher modulation level relative to their maximum output than the FExx are tested in the submittals. Higher modulation levels usually means lower HSPF efficiency, not less.  Then, factor in the much warmer climate of RI (higher outdoor temp = higher COP with ASHP) and the fact that it's no tcold enough in RI even at the mid-winter average temps for the crankcase heaters or pan heaters to engage more than half the time. It's really reaching to say that it won't make a COP of 3 in RI.   I don't think it'll make 3.5, but it'll beat 3.

The long term monitored performance of the mostly FExx Mitsubishis at the cold edge of zone 5 in Devens, MA were pretty similar, and the mini-splits were the sole heating source. I suppose you'd have to point out that most of them are a bit oversized by 150-200% for their actual loads at the local 0F outside design temp, which means it skews them to a slightly higher average efficiency too.  (I've been suggesting 1.25-1.5x oversizing factors for capacity at +10F to keep them modulating into the shoulder seasons for comfort).  The average outdoor temps are more than 5F colder in Devens than coastal RI. Unlike RI,  Devens it's cold enough that the crankcase heaters would run almost continuously for 2-3 months in winter sapping the efficiency despite the oversizing. The only unit that had significantly lower efficiency was on Lot #23 where the occupants kept turning the thing on/off rather than letting it modulate. See the discussion on p38, p52 of the PDF pagination.

Bench testing of an FE12 shows that even at -13F when running FULL OUT it achieves a COP of 1.8, but about  2 running full out at +10F. The average January temp in say, Westerly RI, runs about +30F, a temp at which the FE12  hits about 2.8-3 at max speed (see figure 10 in the bench test link). 

The Fujitsu 12RLS2 hits  a COP of 3 running intermediate speed at +17F, and over 3 at intermediate speed @ +30F, and over 4.0 at low speed @ +35F (see figure 5, p.18 in the PDF). This is where a Fujistsu xxRLF would be run if sized with a bit of capacity margin @ +10F.

These bench tested units were from over 5 years & 2 product revision/model generations ago. The technology has seen significant incremental improvements since then.  No matter how much they ran the wood stove in Montana or Idaho, a 2015 Fujitsu is going to do at least as well as a 2009 Fujitsu. For these reasons it's near-certainty that in an RI location it should average at least 3.0, and not less.

The support for that argument is very easy:
If you supplement heat at the time where the ASHPs run the least efficient, e.g. at low outside temperatures, they do not turn on as much as they would do otherwise without the supplement heat. That will skew the average COP higher.
Inaddition you keep bringing up the North-Eastern study> Did you actually read through it? The results are highly questionable as well!

Dana, I had a chance to look at the field study you have cited multiple time. 8 houses Zone 5A, high performance houses, between 12-18 kbtu design heating load, with one exception between 1100 and 1700 sqf.


No all the sites had actually the full monitoring equipment installed.


3 sites had full set of monitoring equipment, but the interior T/RH recording stopped before the start of winter, due to premature battery failure caused by logger firmware problems. One site had no energy monitoring at all.


Two small houses were equipped with a single MSHP on the first floor, due to their small loads. However, during the first summer, the second floor did not cool down to set point (10°F or warmer than the first floor), even with the use of transfer fans. An additional MSHP had to be retrofitted to the second floor, correcting this issue.



One house experienced comfort issues that are instructive: the owners complained of a bedroom suite and a bonus room that were consistently cold in wintertime. A constant set point was used, but leaving doors open was not compatible with their lifestyle and schedule. Monitoring confirmed that extended winter periods with closed doors resulted in temperatures in the high 50s in the bedroom suite, and high 40s in the bonus room. This house is larger than other monitored houses (2300 ft2, versus 1100–1700 ft2 for others); in addition, it has unfavorable geometries in the problem areas. The bonus room had severe conditions, of exterior temperatures on five of its six sides. Calculations indicated that this was not an equipment undersizing issue, but a heat distribution issue. The problem was resolved by installing a 3:1 (indoor units: outdoor unit) MSHP, with indoor heads in all three bedrooms.





A larger negative, though, was that the lower efficiency of the 3:1 MSHP (per official ratings) penalizes modeled energy performance. With this new equipment, HSPF drops from 10.6 to 9.3, and SEER drops from 23 to 17.5. The peak summer operating efficiencies (energy efficiency ratios) is only 12.9, pretty much on par with a conventional A/C unit. Cold weather capacity and efficiency suffer, as it is not a cold ambient condition (Mitsubishi “H2i”) unit.
These performance differences resulted in the loss of the Massachusetts Residential New Construction Program Tier III incentive, and a substantially drop to the Tier II incentive.


2 sites were unoccupied during a large portion of the data logging, since they were model homes.



3 sites were monitored using a battery-operated multi-channel logger set to measure amperage from a split-core current transducer, logging one of two electrical legs at the outdoor unit. This is only a measurement of amperage or electrical current, and does not include power factor correction.



Standby load was not accounted for. Nor was crankcase heat.


They mentioned that on an installed per-ton basis, MSHPs are more expensive than conventional systems. Which is no so relevant to such house with are super insulated high performance such as exclusively used here



Interestingly, despite the builder claiming having much experience with these kind of installs, the units in the pictures are either mounted on the ground or on wood blocks less than a foot high, right were one expect roof avalanches to come down. As evidence of snow never being an issue, they show a picture right after snowfall, with less than 4 inches on the ground. Great.....



In summary, this is one of the poorest executed studies I have read so far, and pretty much meaningless!

The bench testing of the 12RLS2 tells you pretty much exactly how well it would perform in a RI climate if sized with any margin for the load at +10F.    The tested HSPF efficiency of the 12RLS2 is 12.0, very close to that of the 11 & change of the bigger RLFCD mini-duct units, HSPF 12.2 for the 9RLFCD. The performance will be comparable, and it should be pretty straightforward to model the performance of the 12RLS2 if you have temperature bin data for locations in RI using reasonable oversizing assumptions using reasonable estimates of the efficiency at different modulation at different temps based on the curves for the RLS2. See figure 5, p18 (pdf pagination):  

http://www.nrel.gov/docs/fy11osti/52175.pdf

The mean temp in January February is about 30F, a temp at which the 12RLS2 running anything but flat-out delivers a COP of 3.0 or better, and assuming a heating/cooling balance point of 60F (it might even be 55F, with the amount of exterior foam he's putting on) that's about half the peak modulated output that it would see at 0F or wherever the oversizing runs out of capacity.  The 99th percentile bin is around 10F, where it's doing about 2.5, 2.2 if running flat-out.  So for those two month's it may average slightly less than 3, but it'll average over 4 during the shoulder seasons since it'll be running at it's minimum modulated output at a COP of 5 whenever it's over 40F

Temperature averages for Westerly RI:





The monitored units from the New England units are in super-insulated houses with heating/cooling balance points somewhere around 45-50F, with lots of variations in plug loads and occupant behavior issues.  The fact that there's no load at outdoor temps when they are at their most efficient, and that being more than 150% oversized for the design load they spend a lot of idling time even when it's cold enough to run the crankcase heaters makes it a less than perfect comparison, to be sure. (There's a PassiveHouse near me heated with a 2.5 ton Mitsubishi multi-split where they allegedly bump it into AC mode  when it's going to be in the 30s F on sunny days rather than leave the windows open while they're at work.) Yes, they always have the advantage of always running at part load, but they also don't modulate much, and are almost NEVER running in heating mode at outdoor temps where the COP efficiency is much above 4, which makes it a bit harder to model.

Posted By Dana1 on 27 May 2015 02:11 PM
The bench testing of the 12RLS2 tells you pretty much exactly how well it would perform in a RI climate if sized with any margin for the load at +10F.    The tested HSPF efficiency of the 12RLS2 is 12.0, very close to that of the 11 & change of the bigger RLFCD mini-duct units, HSPF 12.2 for the 9RLFCD. The performance will be comparable, and it should be pretty straightforward to model the performance of the 12RLS2 if you have temperature bin data for locations in RI using reasonable oversizing assumptions using reasonable estimates of the efficiency at different modulation at different temps based on the curves for the RLS2. See figure 5, p18 (pdf pagination):  

http://www.nrel.gov/docs/fy11osti/52175.pdf

The mean temp in January February is about 30F, a temp at which the 12RLS2 running anything but flat-out delivers a COP of 3.0 or better, and assuming a heating/cooling balance point of 60F (it might even be 55F, with the amount of exterior foam he's putting on) that's about half the peak modulated output that it would see at 0F or wherever the oversizing runs out of capacity.  The 99th percentile bin is around 10F, where it's doing about 2.5, 2.2 if running flat-out.  So for those two month's it may average slightly less than 3, but it'll average over 4 during the shoulder seasons since it'll be running at it's minimum modulated output at a COP of 5 whenever it's over 40F

Temperature averages for Westerly RI:





The monitored units from the New England units are in super-insulated houses with heating/cooling balance points somewhere around 45-50F, with lots of variations in plug loads and occupant behavior issues.  The fact that there's no load at outdoor temps when they are at their most efficient, and that being more than 150% oversized for the design load they spend a lot of idling time even when it's cold enough to run the crankcase heaters makes it a less than perfect comparison, to be sure. (There's a PassiveHouse near me heated with a 2.5 ton Mitsubishi multi-split where they allegedly bump it into AC mode  when it's going to be in the 30s F on sunny days rather than leave the windows open while they're at work.) Yes, they always have the advantage of always running at part load, but they also don't modulate much, and are almost NEVER running in heating mode at outdoor temps where the COP efficiency is much above 4, which makes it a bit harder to model.

Again, you cannot conclude from bench testing under certain rating standards how systems will before in the field under certain conditions. The lab data does not assume any rain, snow blowing around, crankcase heat, standby heat, und under certain condition, defrost operation where you simply heat the outside. Thus it is not valid to take BIN data, and plug it in the performance under ARI 210/240.
I could cite you studies where the new Mitsubishi MSZ-FE12NA was running at a COP of less than 1.5 at 10F outside temp, and at less than a COP of 2.5 at 32F outside temp.
This is not about modeling, this is about real world performance. And building super insulated houses has a point of no return, since us humans need oxygen and fresh air, and heat recovery units only go to a certain percentage either.
There is not magic bullet or technology the single head ASHP use, they have the advantage (or disadvantage) to not having to overcome the energy for the fan distribution through the whole house, usually a penalty of less than 10% (7% max for variable speed), but certainly a significant comfort gain to heat distribution in every room of the house. But the lift they have to overcome is significant higher on colder days compared to ground source HP, same in cooling mode.
So from an overall energy perspective, their performance is the worst when they are needed most, not shaving the summer peak significantly, and creating a winter peak. In addition, their upfront price advantage is fading when multiple mini splits have to be installed to have comfort around the house.

This is not about modeling, this is about real world performance. And building super insulated houses has a point of no return, since us humans need oxygen and fresh air

Well said.. there seems to be few facts from actual use and lots of lab work.

Sadly I think that I can't afford to pay for a geo system for my new build. It simply is tooo expensive in RI. I think the same exact hardware is a fraction of the cost elsewhere in this country.

I can't afford to pay 80K + for a heating system.. for a 3400 sq/ foot home with a 1000 sq/ft basement that I will finish in the future.. the payback would be after I die!! I would love Natural Gas but it is not available.. Looks like propane.. the house seems to big for the ASHP that Dana pushes so hard... I would have to get multiple "mini splits" and they are ugly too..

The quotes (2) I have thus far build 2 systems not one... and it takes the price from 60k to 80k+...
I need a solution that will work.. I would rather not play the propane game with the used car salesmen.. hate that ..
I think we will put solar on the roof to offset.. since RI has a nice tariff program right now that I can get in for 15 years on.. so that payback in more in the <5 year mark..
but I still need a system to heat and cool my home. we are going spray foam, polyiso on exterior .. so good envelope.
Just don't see a viable option for geo.. sadly..

Not happy..

How about adding solar and going mini-split?

Posted By kleach on 03 Jun 2015 08:03 PM

This is not about modeling, this is about real world performance. And building super insulated houses has a point of no return, since us humans need oxygen and fresh air

Well said.. there seems to be few facts from actual use and lots of lab work.

Sadly I think that I can't afford to pay for a geo system for my new build. It simply is tooo expensive in RI. I think the same exact hardware is a fraction of the cost elsewhere in this country.

I can't afford to pay 80K + for a heating system.. for a 3400 sq/ foot home with a 1000 sq/ft basement that I will finish in the future.. the payback would be after I die!! I would love Natural Gas but it is not available.. Looks like propane.. the house seems to big for the ASHP that Dana pushes so hard... I would have to get multiple "mini splits" and they are ugly too..

The quotes (2) I have thus far build 2 systems not one... and it takes the price from 60k to 80k+...
I need a solution that will work.. I would rather not play the propane game with the used car salesmen.. hate that ..
I think we will put solar on the roof to offset.. since RI has a nice tariff program right now that I can get in for 15 years on.. so that payback in more in the <5 year mark..
but I still need a system to heat and cool my home. we are going spray foam, polyiso on exterior .. so good envelope.
Just don't see a viable option for geo.. sadly..

Not happy..

Why would you need anything but a single heatpump solution? Get more quotes.

I am being told that

"Yes there are larger units the issue is keeping the units at tier 3 for the federal tax incentive The larger units don't make the requirements. "

SO can anyone tell me the largest system size that meets the fed requirement and if it will heat/cool a newly constructed tight house of 4200 sq/ feet?  (on average)...


<!--[if gte mso 9]> Normal 0 false false false EN-US X-NONE X-NONE

I think your talking to the wrong person, this is not something that needs a commercial unit.

I sent you a PM.

Chris

I get that kleach will never go ASHP purely on aesthetic grounds, but the notion that the bench test modeling isn't reflected in the NEEA in-situ testing is just wrong. (Either that, or the engineering group at Ecotope fudged the data on both, and should be brought up on fraud charges.) Correlation between the bench modeled & field measured performance adjusted for climate was quite high, at least according to Ecotope.

I'd be VERY curious to see the alleged studies showing that a ...Mitsubishi MSZ-FE12NA was running at a COP of less than 1.5 at 10F outside temp, and at less than a COP of 2.5 at 32F outside temp." REALLY?
That might be true under very specific conditions requiring very high defrost duty-cycles, but it would be good to know the conditions under which those measurements were made, and how that would correlate to a specific climate.

BTW, kleach: Whether you go with propane heating or geothermal, going with a 3" -4" foam-over is still likely to be cost effective. (Measured against up-front system costs on the geo end, and against propane use if going with condensing propane.) The need to have a good load calculation prior to quoting a system doesn't change, but is of critical importance if going with geothermal.

Tried to send you a Private Message... Was 2 companies names to talk to about quotes.

Chris

Posted By Dana1 on 05 Jun 2015 12:06 PM
I get that kleach will never go ASHP purely on aesthetic grounds, but the notion that the bench test modeling isn't reflected in the NEEA in-situ testing is just wrong. (Either that, or the engineering group at Ecotope fudged the data on both, and should be brought up on fraud charges.) Correlation between the bench modeled & field measured performance adjusted for climate was quite high, at least according to Ecotope.

I'd be VERY curious to see the alleged studies showing that a ...Mitsubishi MSZ-FE12NA was running at a COP of less than 1.5 at 10F outside temp, and at less than a COP of 2.5 at 32F outside temp." REALLY?
That might be true under very specific conditions requiring very high defrost duty-cycles, but it would be good to know the conditions under which those measurements were made, and how that would correlate to a specific climate.

REALLY!

Here is the Here is my run down to the report posted below.

All but one house had other sources of heat, meaning that on the colder days, COP is skewed by supplying unaccounted amount of supplement heat, so the heatpump had to run less in unfavorable conditions.

They had trouble measuring the correct air flow.

They could not believe their measured (calculated COP) since it was so low during heating and cooling. They state that they have not much confidence in their data, especially the cooling.

Their peak days (90F outside) EER was around 6, their 10F COP was around 1.73.

Much more data (and honesty) than other reports. They also had up to 3 mini splits installed, and still used other forms of heat.

So the installation costs are significantly higher (since you install multiple heat pumps) compared to the claim of $4000 and a single head, and the efficiency is much less than any geo system.

Especially the lack of peak shaving capabilities, and need for conventional supplement heat and backup, and the lack of cold weather efficiency which created a winter peak in the study, are apparent key weaknesses of air source technology.
http://www.neep.org/sites/default/files/resources/NEEP%20DHP%20Report%20Final%205-28-14%20and%20Appendices_0.pdf

That's a pretty good study- thanks for that!

Looking at the low heating efficiency point for site # 9  in the Table 6.1 ( p42 in pdf pagination) between 2AM and 7AM on 7 February 2013, the venerable FE12 was only mustering an estimated average COP of 1.42 @ 10-12F outdoor temps.

Weatherspark.com data for nearby Laconia NH for that time & day indicates dew points in the -5F to -8F range, so the explanation is probably not excessive defrost activity.  I could believe measurement error, but as an engineer I don't dismiss measurements easily without knowing the source of the error.  Let's call it a "mystery datapoint" for now.

Mind you, the bench-test data for the FE12 is much less consistent than that of the 12RLS2 (compare figure 5, p18, to figure 10, p 22). Note that the bench-tested COP of the FE12 is isn't very different (and sub-2) between 5-10F at ANY modulation level, and that with either manufacturer's unit the range of efficiency across modulation at low temps quite small.  That  COP 1.42 data point does not differ wildly from expectations based on the bench testing, but it does mildly.

The narrow difference in efficiency between highest/lowest efficiency modulation levels at low temps stands the whole "auxilliary-heating-improves efficiency" argument a bit on it's head too.  Everybody I know who is using a mini-split to offset oil or resistance electricity doesn't crank on the more expensive heat until/unless the mini-split is maxed out, which happens at high speed/low-temp, which is it's WORST efficiency level.  With wood stoves & pellet stoves people may ease the load on the mini-split even during warmer weather when the mini-split might be modulating, but not folks looking to offset $4/gallon oil.  

If you read the descriptions you'll note that system #9  is offsetting oil-fired radiant floor heating  in a 600' sun-room, which almost certainly has a heat load at +10F well in excess of the ~13.6K capacity of the FE12 at that temp. It probably runs at max speed/lowest efficiency any time it's below freezing at night, which is going to negatively impact its seasonal average efficiency rather than improving it. (But it's still cheaper than heating that space totally with the oil boiler at $4/gallon.)  Most people with that type of heating zone would set the mini-split to 70-72F and the radiant thermostat to 65-68F to ensure that the oil burner doesn't come on until/unless it has to.  It's also possible that the radiant floor wouldn't keep up with the load either, but the owners were also looking to air-condition that space (good luck with that, with only a ton of AC in a greenhouse!)

Also note, the tested HSPF efficiency of the FE12 is a hum-ho 10.5.   In Q1 2014  Mitsubishi released it's second generation cold-climate mini-splits, the FHxxNA series.  The FH12 (same output specs as the FE12), tests at an HSPF of 12.5, which is a nearly 20% improvement in efficiency over generation 1.  Even worse (for Mitsubishi) Fujitsu's second generation xxRLS3H cold climate mini-splits are even more efficient, with even higher heating capacity. The  3/4 ton -9RLS3H has higher capacity at +10F than the FH12, and tests at an HSPF of 14.2.  That's 35% more heat per kwh than the -FE12NA.   (You can still buy the FE series, now getting onto a decade or more since it's initial release, but it's worth the up-charge for the FH equivalent, and well worth looking at the Fujitsu competition.)

The test location for the NEEP study was in central NH, which outdoor temperature average well below that of coastal RI, which are zones V and IV respectively, for a HSPF performance zones (Zone 6A and the warm edge of zone 5A per DOE heating /cooling climate zones).  The HSPF test is the presumptive model efficiency of a winter climate with temps that range between 17F and 47F,  and while coastal RI has a 99% bin cooler than +17F, it's average winter temp is over 30F.  The HSPF test doesn't model modulating efficiency issues very well, since oversizing increases efficiency (up to a point).  With  1.25x to 1.5x oversizing  for the loads (as opposed to the undersized units in the NEEP data) a modulating mini-split should pretty much meet and sometimes beat it's HSPF efficiency in coastal RI.   (See the map on Figure 4.1, p 25 for HSPF zones )

The HSPF efficiency of the -18RLFCD mini-ducted unit is 11.3 , which is about 8% better than the FE12s tested in central NH. If the HSPF test were completely valid that implies a seasonal COP of 3.3 in a RI location.  Even with modest oversizing wouldn't bet the farm on hitting that number, but beating 3 is a pretty good bet.

As long as the seasonal average efficiency is reasonable and it has sufficient capacity margin for the load at the 99% temperature bin, nobody really cares if there is some particular set of conditions under which it only hits a COP of 1.42.  Even the undersized FE12 in site-9 was operating in the mid-2s at RI's 30-something average winter temp (also consistent with the bench test efficiency), and had it been sized for capacity at Laconia NH's -5F 99% outside design temp it would have been loping along just shy of 3.  The Fujitsu bench tested 12RLS2 1-ton did slightly better than the FE Mitsubishi, and I would expect the mini-ducted Fujitsu to perform similarly. 

Making progress..
Got foundation in... question.. ANY difference between taking 1" XPS and doubling up vs 2" solid XPS for foundation exterior walls?
Do I have to secure the panels to the concrete somehow or just rely on tape to hold together and backfil to hold in place?  IF I need to secure them how?  Seems like I would be punching holes that could fill with water and freeze ??

I plan on doing the same thing in the basement under the slab...

Comments?

-Ken

I used one layer of 2" and they just nailed w/large washers then back filled.

I wish I had used 2 layers of 1" gluing and taping staggering the seams.

Chris

The difference is subtle, but unless it has ship-lap edges it's better to go with two layers of 1". The stuff has long term shrinkage issues, and with one layer you end up with narrow gaps with zero insulation.

Unless it's reclaimed foam going with XPS is a lot more damaging than using EPS, due to the HFC134a blowing agent's extreme global warming potential (1400x CO2), whereas EPS is blown with pentane (7x CO2). In a handful of decades as the blowing agent leaks out the R-value of XPS drops to that of EPS of equal density too.

And...


...IRC 2012 spells out R15 continuous insulation for foundation insulation in US climate zone 5A (RI included.) With 2" XPS you're looking at only R10. (R10 would meet code across the water on Long Island, zone 4A though) See:

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

If you want to compromise, go with 3" of EPS (R12-ish), but going too much below code-min isn't a great way to start.

First, waterproof the foundation with your goop of choice. Tack the foam in place to the concrete with dabs of foam-board construction adhesive, tape the layer then tack the first layer to the next with dabs of foam board construction adhesive. Then, through-screw it to the concrete with cap screws 24" o.c. before finishing it. QuiKrete Foam Coating applied over the whole thing helps keep ants & termites from tunneling through the foam.

Unless it's reclaimed foam going with XPS is a lot more damaging than using EPS

It IS reclaimed XPS.. not sure what the difference is that you are pointing out.. reclaimed vs new?
I think you are saying the reclaimed XPS is about the same cost as EPS..
I took polyiso off the list since it retains a fair amount of moisture...

So If I understand what was said.. three 1" layers offset taped, glued and through screw to concrete 24"oc with a final coating ofr QuiKrete to stop the bugs? Is that the secret formula? That is a lot of work for a buried wall...
Same for walls that are NOT buried?

=Ken

With reclaimed foam the environmental hit has already been taken, and you're just extending the lifecycle, maximizing the benefit. With new foam it's driving the market for more new foam.

With 20-25 year old XPS, derate it's performance to R4.5/inch- it's no longer performing at R5. But 3-4" of reclaimed XPS would get you to code min.

Alternatively, if you put 2" on the exterior (R9-ish) then put an inch or two of foam on the interior when you finish the basement you'll be there. It's pretty easy to retrofit 2" foam on the interior with dabs of foam board adhesive then strapping it to the concrete with 1x furring through-screwed to the foundation with 3.5" -4" TapCons, onto which you can hang the wallboard. Running electrical circuits in that sort of wall stackup is a bit of a pain though. A 2x3 studwall trapping 1" foam to the wall makes it easier to run the power, and split R19 batts (or compressed R11s) brings it up to code-min performance, assuming 2" of exterior foam.