I hope this is the correct  for this question. I'm looking to using the fujisu rls series for both heating and cooling. I have heard great things about thier system. Since I'm in ny are they a good chose as the sole heating system? I have heating guy telling me I should get howater system as backup.

Thanks,
Zygote

What's the total heat load estimated to be? (The answer is critical.)

I ran the numbers on a 3-family apt that's getting an R20 foam overcoat (in addition to the dense-packed cellulose in the walls) in 01610 a few weeks ago where they went with three small Mitsubishi Hyper Heat mini-splits- confidence is high, and design condition (+4F) heat load is low (~10KBTU/hr for the most-lossy unit.) With such low loads the "backup" is a pair of oil-filled 1500W space-heaters per apartment, which would keep up at design condition even if the Hyper-Heats were broken. With ~R30 whole-wall R values on the exterior the room-to-room imbalances are low, and a single indoor coil works for each apartment- a no-nonsense mini-split, not a multi-split, not ducted.

But if your heat loads are very high or R-values low (leading to room-to-room temperature imbalances) it might take something bigger than a Fujitsu-mini. It may take a variable speed R410A refrigerant Goodman or similar and a ducted-air heat distribution. Model or calculate the design condition carefully, and take care not to overshoot by more than 25% at your design-temp or it'll cycle more often and the interior temp will vary.

At NYC's 97.5% design temp of 15F almost all of the better mini-splits will have a COP of 2.6 or better.. At the NYC winter average of 30-33F it'll have COP north of 3, and competitive with natural gas burners at NYC utility rates.

If there is good distributor and contractor support for the mini-split (meaning, they'll be able to fix it in less than 48 hours no matter what), it probably doesn't need much in the way of backup. Fujitsu is a quality vendor, and their mini-splits are among the best-in-class, but the local support is everything- when it's 10F outside you can't be waiting for the parts to arrive from Japan should something break.

Read the specs carefully- the BTU output they spec for the RLS & RKX series at 47F, where it has a COP of about 4.0-4.1. At the 15F design temp it'll only be putting out ~ 2.6/4.0 or 2/3 of the rating in the spec, and a 0F it's probably running a COP of 1.8, putting out only 45% of the spec. The ASU30RLX may be rated for 32KBTU/hr @ 47F, but don't count on more than 21K @15F, and probably about 14K @ 0F, should it stay that cold for long. But if your heat load at 15F is 18K, that's probably OK- a single 1500W space heater would add ~5K should there be a severe cold snap. The biggest RLS is good for 18K @ 47F, but only 12K @ 15F and only 8K @ 0F. If it hits your 15F output dead-on your calculated heat load it'll be OK, since heat loss calcs trend only upward from reality, and the heat load can easily be 25% lower than the calculated number. (Better models can deliver closer numbers, but only building-science nerds use them, not heating guys.)

http://www.fujitsugeneral.com/wallmounted9-12RLS_specs.htm#specs

Mitsubishi typically specs 47F, 17F, and 5F output numbers for their mini-splits, which makes the interpolation easier for different design conditions, but amongst the 5 biggest manufacturers of R410A refrigerant heat pumps the outputs scale with temperature pretty much the same.

Dana1,

My heating guy did some math and come up with 45k per floor. I'm thinking putting up four units. 9RLS unit for each bedroom and 12RLS for living/kit. Will it be enough for at 14F?

You'd need the room-by-room analysis to know whether that would do it. The 9RLS would deliver ~ 7KBTU/hr @ 14F, the 12RLS would deliver ~9.5KBTU/hr @ 14F. In a 3-bedroom that would add up to 30K, well under the 45K estimate.

But unless these units are very large with 12' ceilings, lots of windows, and no wall insulation I suspect 45KBTU/hr @ 14F is grossly overestimated. What is the total square feet of exterior-wall area (and what is the R-value) and what fraction of that is window (at what U-value)?

If the heat loads are really only 1/2 - 3/4 that 45K estimate, you can probably do it with a single Goodman DSZC18 series per apartment:

http://www.goodmanmfg.com/Portals/0/pdf/SS/SS-DSZC18.pdf

(Look at 15F columns of the heating output tables for the different models, starting on p20. The DSZC180601A* /CA*F4860*6A*+TXV/ MBE2000**-1 combination delivers ~36K @ 15F, with a COP of 2.66.)

or, the Mitsubishi Hyper Heat multi-splits, which can go much higher. The PUZ-HA36NHA is good for 38K at 17F (and down to 5F) with a COP of 2.1 @ 17F. (3.6 @ 47F)

http://www.pacificairconditioner.com/files/Mitsubishi_H2i_Hyper_heating_inverter_brochure.pdf

The Fujitsu 42RCLX multi-split is similar, wiith about 28K @ 17F:

http://www.fujitsugeneral.com/PDF_06/halcyon06_brochure.pdf (See p29)

There's some independent test data to indicate that you get slightly better heating performance than spec if you oversize inverter-type heat pumps by 1.5-2x, since the COPs are higher when the compressor is running slower, but by how much will depends on the particulars. Most have short-cycle prevention built into the controls to deal with oversizing or extreme low-load, but if you go too big it'll be less comfortable.

Well i would be recommending going with one from the company of sabro or Dancom....

my house is semi detach brick. the ext wall is 10-12"thick brick with 2" CC foam. 9' ceiling and cellar will be around R50-60. The living room has a 10'x5' casement Crystal 200 windows U i believe is around .34... and rest of the room has double hung U also around .34.

How many square feet of glass total, and how many square feet of exterior wall?

The ceiling height means nothing without the length- we need area, and the U-value of the other windows means nothing without their size, and how many.

Dana1,

The total around glass 137.72 sq ft and total ext wall 1152 sq ft.

thanks

With 2" of cc foam is ~R12. U=1/R, so your U value for the walls is ~ 0.085 BTUs per degree-F per square-foot. At a design temp of 15F exterior, 70F exterior you get a delta-T of 55F. The direct conducted heat loss for the 1152' of wall is:

...1152 x 55 x 0.085= 5.4KBTU/hr

With 138' of U 0.34 window you get:

...138 x 55 x 0.34= 2.6KBTU/hr

for a total of ...

...8.0KBTU/hr

But that's a baseline- there are also infiltration losses that need to be added, and interior heat sources (~ 0.5K per live human, + plug-loads) that need to be subtracted from that load. Assuming you have reasonably tight windows and have made at least some attempt at air-sealing the total heat load may be as high as 12K-15K, but isn't likely to exceed 20K @ 15F, and that is probably VERY generous padding. Seriously, 20K is fully 2.5x the conducted losses through walls and windows, and sufficiently oversized for whatever the real heating load is that it should perform slightly higher than factory-spec COP. At 45K it would be ludicrously oversized. If the real heat load turns out to be over 20K it probably means you have very high (and fixable) air leakage.

I can't imagine how they came up with a 45KBTU/hr heating load at 14F unless they were figuring ~U 0.5-0.6 for foot thick brick walls (no foam) and bigger infiltration factors, which might be pretty typical for buildings of similar size & vintage in NYC.

The quality of the installation is far more important than the brand, but local distributor & factory technical support count too. I've read online accounts of bad experiences with almost all vendors, but most of the time poor performance & reliability comes down to improper installation rather than fatally flawed design or manufacturing flaws. In my area there's pretty good support for Fujitsu & Mitsubishi, somewhat less for Sanyo, LG, or Daikin, but it's moving target- by the time I hit "Submit" that may already be out of date.

Dana1,

I believe you are correct. The 45k btu was base on no insulation. I'm replacing all the windows also and one skylight in the living area. I instructed my contactor to pay extra attention to air sealing on windows and between floors. So base on the number you used, four units of SLR will be enogh for my home then?

Each RLS12 should be able to deliver 9K+ @ 14F, each RLS9 should be good for ~7K, so if there is one of each for every 138' of window and 1152' of insulated exterior wall that would be 16K, 2x the known conducted heat loss, which should give you some margin.

If you have four units in that space you'll have PLENTY to spare, and could even tolerate a failure and still be able to stay warm.

But it's probably cheaper to go with a multi-split with one compressor unit and multiple interior units than going with 4 separate mini-splits, or one mini, and one multi, depending on the ease of installation. Adding an interior unit to a multi- is less than $2K, whereas each mini (even the smallest ones) are probably more than $3.5K at NYC prices.