Our fixer upper in Florida had a 3 ton air conditioner with ductwork in the attic. We ran it for a month last April. It took an hour before it would start cooling the home. It cost a lot to run it so I tore it out this spring. I discovered the HVAC installer had cut several truss webs and did a very poor job of sealing the duct. I used BEopt to size a mini split based on our remodel plans; R-23 2x4 walls required one ton of cooling and 6,000 BTU of heating while walls up at R-15 required 1 ¼ ton of cooling and 10,000 BTU of heating. My base BEopt run was the existing conditions in the home. BEopt on the base run called for a bit over 3 tons of cooling. I had a HVAC company come out to size and give me an estimate. They have been in business for over 60 years. The very first suggest out of his mouth was to install ductwork in the attic while I had it open. I responded with my experience with the old system and told him that putting ductwork in an unconditioned space was not a good thing. He responded with the suggestion I bury it in the cellulose insulation. I still maintained that I wished to have a mini split and wanted him to do a Manual J. I gave him a floor plan, thorough window schedule, R-values for walls and ceiling etc., etc. He got back to me still insisting on putting in an air conditioner outside, ductwork buried in the cellulose and a resistance heat unit. His Manual J calculations said I required 18,000 BTU of cooling and advised bumping it up to 2 tons. He said the mini split would not cool the whole house (1,440 sq. ft.). I asked to see the Manual J and said I get that AFTER the installation! So that’s where I’m at. From what I have gathered over-sizing a mini split in a hot-humid climate is not a good idea in terms of humidity control. We have an open floor plan with the great room (combined kitchen, dining and living room) making up 665 sq. ft. with a 4’ opening and hall way opening up to a 216 sq. ft. family room. The two bedrooms and two bathrooms are on the far end of the home, thus causing me concern for adequate cooling. My thought is that it’s a no brainer to put a 1 ¼ ton mini split wall mount head in the great room and see what happens. If I need more cooling a window unit would be a low cost easy solution. Any advice on this problem would be appreciated.

Oversizing AC somewhat is less of a problem if 1) it's an inverter driven compressor and 2) if you have good air sealing so that less dehumidification is needed. If you haven't checked it, I'd work on the latter.

HVAC contractor Manual-Js are all over the place and not to be trusted, especially if it's a "I'll give it to you after I install it." kind of deal. I've even seen one that included a duct loss load for a ductless system (!). It's reasonable to charge a few hundred USD for running the calculations, and maybe taking that fee off the installation price if they are awarded the contract, but even that is rife with incentives to put a thumb on the scale.

Hire an energy nerd like a RESNET or HERS rater or a professional engineer to do the load calculations if you don't trust your BEOPT output. A 1.5 ton cooling load wouldn't be unusual for a 1400' house but a 3 ton load would be. Installing leaky ducts in a hot attic really DOES add a ton or more to the load, which is probably why they went with the three tonner in the first place.

If you just punted and go with a 1.5 ton mini-split compare the MINIMUM modulated output. Lower is better, always, since the lower the minimum, the longer and more efficient the run times. It will vary a lot by model, even from the same vendor, eg:

A GE18 can throttle back to 3700 BTU/hr in cooling mode:

http://www.mitsubishipro.com/media/456672/msz-ge18na-8~muz-ge18na-1%20submittal.pdf

.... the FE18 can only modulate down to 8200 BTU/hr:

https://meus.mylinkdrive.com/files/MSZ-FE18NA_MUZ-FE18NA_Submittal.pdf

...and the FH18 dials down to 6450 BTU/hr:

http://richairductless.com/wp-content/uploads/2015/09/MSZ-FH18NA_MUZ-FH18NA_Submittal.pdf

Clearly the GE18NA will be a better choice for you than the FE18, since most of the time your loads won't be much more than the 8200 BTU/hr minimum cooling output of the FE18.

If installing the head in the bigger open floor plan zone, deally the open area zone would be located on the west side of the house, since the west side gets the late-day gain that drives the peak. If that's not the case you might be better off with a 2-head multi-split. There are some decent 1.5 ton 2- zone mini splits out there too.

http://portal.fujitsugeneral.com/files/catalog/files/18RLXFZ2.pdf

http://www.ecomfort.com/manuals/mxz2bessubmittal.pdf

The down side to 2-head multi-splits is that they generally don't modulate much below the 6000-7000 BTU/hr range, but it would fix any zone balance problems you might encounter if a west-side room with a bunch of west facing window is down the hall from the mini-split head in a single head system.

Thanks for the comments. So over sizing by 50% isn't that serious for cooling load. I have confidence in BEopt having used it several times and found it reliable. Placing the head on a West wall would be best. The great room has a North, East and South wall. The south won't work because it's only 16' and contains the door and two windows as well as faces the street. East or North would work. The adjacent family room has a West wall . The room flows into the great room with a 4' doorway and 3' hall. Note the attached floor plan.

Hold on. You need to reread the caveats. Oversizing is not problematic with inverter driven units that can modulate down, or tight homes where humidity drivers are less a concern. in Florida, an oversized traditional HVAC unit can be problematic for humidity.

Got it! Inverter driven and tight house. That's plan. I have installed a whole house fan that I plan on using when the temperatures are moderate. Will that device be problematic? What is the best strategy for its use?

A whole house fan is fine if it can be air sealed. It can also be used to "charge" interior thermal mass for future outdoor temperature extremes. Ideally it would all be automated based on temperatures and humidity and your comfort limits, but I haven't seen anything that can do that.

In a FL climate a whole house fan can sometimes work during the shoulder seasons, but in mid summer it can bring in a huge latent load that you might not want. Latent cooling (moisture) is a big part of cooling load in FL, it's not all about the sensible load (temperature).

With modulating mini-splits again, look at both the minimum and maximum output. For maximum comfort and better latent cooling it's better to be slightly undersized with something that can modulate way down than 50% oversized with only a 3:1 turn down ratio. Ideally the duty cycle would be nearly continuous at low speed/high efficiency, but you won't get anywhere near that if the min-modulation is 50% of the design condition load.

So if I over size my BEopt calculation of 15,000 BTU by 50% I would be at 22,500 BTU. With only a 3:1 turn down ratio my minimum would be 6,750 BTU while a mini split with a 5:1 ratio could run down to 4,500 BTU? Am I understanding it correctly? Is there any justification in installing two smaller ones that one larger one?

I'm saying that mini-splits all have the same turn down ratios, and you have to look carefully at the submittal sheets for both the minimum and maximum outputs to figure it out. A lower total min-modulation is better.

A pair of Mitsubishi FH09s can deliver a combined 18,000 BTU/hr cooling, but also throttle back to (2 x 1700= ) 3400 BTU/hr.

http://www.centralmaineheatpumps.com/sites/default/files/ec_pro/centralmaineheatpumps/MSZ-FH09NA_MUZ-FH09NA_Submittal.pdf

An FE18NA can deliver 18,000 BTU/hr of cooling, but only throttles back to 7500 BTU/hr, more than twice what a pair of the FH 3/4 tonners would deliver:

http://www.ecomfort.com/manuals/mzfe18submittal.pdf

But the FH18NA (newer series than the FE), delivers 17,400 BTU/hr, and dials back to 6450 BTU/hr, which is only slightly better than the FE18, still nearly 2x the min-modulated output of two FH09s:

http://richairductless.com/wp-content/uploads/2015/09/MSZ-FH18NA_MUZ-FH18NA_Submittal.pdf

A Fujitsu 18RLFCD (mini-ducted type) unit can deliver 18,000 BTU/hr of cooling, but throttles back to 3100 BTU/hr, slightly lower than a pair of FH09NA, but may have a less favorable room-to-room temperature balance.

http://www.fujitsugeneral.com/PDF_06/Submittals/18RLFCD%20Submittal.pdf

The Fujistu RLS2s and RLS3s wall coil types can all drop to 3100 BTU/hr independent of size, so there is no modulation advantage to splitting them up. The SEER 33 of the 9RLS3 may look slightly better than the FH09's 30.5, but since the min-modulation is nearly twice that of the FH09 they probably would not beat it on as-used efficiency.

The 15RLS3 has the same 3100 BTU/hr minimum modulation, but a lower SEER than the FH09s, and would have lousier distribution, since it's a single wall-coil:

http://www.fujitsugeneral.com/PDF_06/Submittals/15RLS3Submittal.pdf

The FH15NA suffers from both an even lower SEER and a higher minimum modulation:

https://meus.mylinkdrive.com/files/MSZ-FH15NA_MUZ-FH15NA_Submittal.pdf

If BeOpt is giving you 15,000 BTU/hr for a cooling load your better options are:

a: A pair of FH09 wall coils

...or...

b: A single 18RLFCD mini duct cassette.

With either of those solutions it would actually be modulating most of the time, and the conditioning is better distributed than a 1.25 ton wall coil type in the more open space. The 30.5 SEER of the FH09s clearly beats the 19.7 of an 18RLF with some margin, so you'd have to price it out to see if there's enough of a cost difference to tip it toward the lower efficiency mini ducted version.

Both of those would also be able to modulate some during your heating season, whereas the larger single wall coils (except for the 15RLS3) would be only cycling on/off during the heating season, if your heat load is truly 6000 BTU/hr. You can find those numbers in the submittal sheets too.

With the single unit why the duct cassette and not a wall coil? And please define the turn down ratio and an example how to calculate. I'm seeking to better understand the concept. Thank you.

With a single unit it's possible to directly cool areas of the house that are down a hall or behind a door from the main area achieving better temperature balance. With a wall coil it only cools the area that has good convective air exchange with the room with the wall coil.

The turn down ratio is the ratio of it's maximum modulated output and it's minimum modulated output at a given temperature. In an SEER test the test condition is at a 15F difference between indoor and outdoor temperature, typically 95F outdoors, 80F indoors, which is usually specified in the submittal sheets (often in the foot notes.) This is different from the "rated capacity", which is the modulated level at which the efficiency was tested.

eg: http://www.younits.com/media/wysiwyg/Literature_PDF/Mitsubishi/MSZFH09NA_MUZFH09NA_Submittal.pdf

For the FH09NA the cooling Capacity range is specified to be 1,700 - 12,000 Btu/h, but the Rated capacity is 9000 BTU/hr. The test conditions are found in their "Rating conditions" list, where you will find:

*1 Cooling | Indoor: 80º F (27º C) DB / 67º F (19º C) WB
*1 Cooling | Outdoor: 95º F (35º C) DB / 75º F (24º C) WB

What this means is that running at full speed at a 15F temperature difference (say 91F outside, 76F inside) it can deliver 12,000 BTU/hr at some unspecified efficiency, but if modulating at 9000 BTU/hr it would deliver the rated SEER of 30.5.

But when it's operating at a lower output level it's efficiency will exceed the rated SEER as long as it's running continuously rather than cycling on/off.

Your cooling (and heating) load varies with the outdoor temperature and the amount of solar gain the house is receiving. The heat pump has to be sized sufficient to cover the maximum load, but the maximum load is much larger than the average load. With a modulating heat pump it will run far more efficiently at part load than if it's running at full speed, but that efficiency gain is severely blunted if so oversized that the minimum output is higher than the average load. For example:

Say your peak load is 15,000 BTU/hr at 3pm on a sunny day when it's 90F outside and you're getting a bunch of gains through the window, a condition which happens only 1% of the time. If your average cooling season load is 6000, BTU/hr when it's 82F and hazy, a unit that can deliver 24,000BTU/hr but with only a 2.5:1 turn down ratio only modulates down to (24,000/2.5=) 9600 BTU/hr, such as a 2-ton Carrier GreenSpeed, is less than ideal because at your average seasonal load it will be cycling on/off rather than modulating, continuously adjusting it's compressor & blower speeds tracking the load. If it's not modulating, it's not maximizing efficiency or comfort, even if it's running at pretty decent efficiency modulating on/off at minimum speed. A Fujitsu 18RLFCD can modulate between 18 down to 3100 BTU/hr and 20,100 BTU/hr which is a (20,100 / 3100= ) 6.5/1 turn down ratio. (see: http://www.fujitsugeneral.com/PDF_06/Submittals/18RLFCD%20Submittal.pdf ) 

When the cooling loads are that lower than the 1% outside design condition (which is MOST of the time) it will automatically adjust it's output levels the 15,000 BTU/hr load at the 1% outside design condition with a decent amount of margin to spare. When modulating at 18,000 BTU/hr (it's "rated" output level) , it will be delivering it's rate SEER of 19.7, but it will run higher efficiency than that most of the time as long as it's within modulation range. When the cooling load is below it's minimum modulated 3100 BTU/hr it's efficiency will be lower than it's efficiency when the load is between 3100 BTU/hr and the average load of 6,000 BTU/hr, but at loads that low the efficiency matters less.\

Bottom line, its the minimum modulation level that limits the efficiency, which is not exactly the ratio, but a high turn down ratio gives you a wider load range to work with.  You need  a range with enough capacity on the high side to cover the 1% load with reasonable efficiency, but to maximize efficiency and comfort the minimum modulation should be an output level that's lower than the average seasonal load. When you get the sizing right it will be able run almost continuously for days or weeks on end without cycling, but at extremely high efficiency and comfort.

I was under the impression that a wall mount unit would throw air 30' or more. What about ceiling cassette located in central hall pointing in all directions? Or where would one put a ducted unit? Attached is floor plan I hope.

Floor plan