GBA Logo horizontal Facebook LinkedIn Email Pinterest Twitter Instagram YouTube Icon Navigation Search Icon Main Search Icon Video Play Icon Plus Icon Minus Icon Picture icon Hamburger Icon Close Icon Sorted

Community and Q&A

Manual J for Ducted Heat Pump / Air Handler System

whanna27 | Posted in General Questions on

I’m sizing a heat pump to install one the second floor of my house to replace the electric baseboard in each room. I got several outrageous quotes last spring (ducted $19k, ductless $17k) that didn’t include electrical hookup, drywall work, or plumbing (gas furnace, proposed ducted system was AC only and NG furnace). This was for a single zone system that only served the second floor. Decided to start planning it out and I’ll do the install myself in the spring. I’m a mechanical engineer by schooling and trade and I regularly work with AHJs and am intimately familiar with permitting. I’d consider my DIY skills well above average. Hardest part will be finding a local HVAC company to come run the lines and charge them with refrigerant.

Ducted heat pump with ducts and air handler in the attic looking at Mitsubishi and fujitsu. Anyways, long story short I’m doing my Manual J and S now and I have some questions on the Manual J as my cooling load seems very high compared to my heating load (or maybe my heating load is too low). Second floor is currently cooled with a 10kBTU window unit in the master bedroom and 5000-6000BTU window units in the remaining bedrooms. ~25kBTU of cooling overall

I thought NJ leaned towards a heating dominant climate. The internal loads seem a bit high.

Design details:

– R11 2×4 walls average air tightness 1976 construction no renovations

– first floor is slab on grade no insulation on slab. 42kBTU pellet stove for heat (1-2 bags of pellets a day when it’s below 32deg). No AC on 1st floor, stays very cool in the summer, hottest I’ve seen it get is about 78 degrees.

– low E rated windows with the sticker still on them appear to be circa 2000

– unconditioned attic with ~R22 insulation in rafters

– 2nd floor is 4 bedrooms and 2 bathrooms

– 5 occupants in respective bedrooms

– ceiling fan, TV, and light fixtures in each bedroom

– I will be installing R8 (or higher if I can) ducts in attic, extremely airtight (need to pass pressure test per IECC)

– default design temps

Do these internal loads seem high and cooling load seem high, or conversely is my heating load too low? I’d appreciate any input or suggestions, posted this on Reddit as well with no luck. It looks like given the latent load I will need a 2T Mitsubishi hyper heat.

GBA Prime

Join the leading community of building science experts

Become a GBA Prime member and get instant access to the latest developments in green building, research, and reports from the field.


  1. Expert Member
    Akos | | #1

    Your appliance loads are way too high. Do you actually have a 100W incandescent bulb and is it running most of the time? Most modern TVs use about 1/2 to 1/3 the power used by CoolCalc.

    I would update these and see where you stand. Running ducting in the attic can be done if you take some care but please don't put the air handler up there. I know it takes space, but find a spot in one of the closets or dropped ceiling in the hallway and mount it there. Above the staircase going to the 2nd floor tends to be a good spot. This is much easier to work with, much more energy efficient and better for servicing down the road. It also means not crawling around the attic for the most of the installation plus the return can be a simple filter grill.

    1. whanna27 | | #2

      Regarding light bulbs, yeah I’m all LED, but the master has 8 of them (~10W each) and also acts as my office so if I’m working from home they’ll all be on, especially in the afternoon because my room faces east. As for the other bedrooms, yeah 100W is a bit aggressive. TVs I will adjust.

      As for the air handler, I have no potential closet space, or dead space above stairs, and the ceilings on the second floor are 7’8” already so a drop ceiling isn’t an option haha. My roof is a 11 on 12 pitch so the attic is extremely spacious and it will be a dream to work up there. Easily accessible too, drop step ladder in the hallway and plywood on a majority of the rafters. Is insulation on the air handler the concern? I recall it only being R4 or something. I could beef it up with some additional insulation?

      1. whanna27 | | #3

        Additionally, what about window areas? Coolcalc only offers “low” “med” “high” window areas. How do these translate to actual sqft?

        1. whanna27 | | #4

          Gonna answer my own question here lol. I found this on the coolcalc forums. Will update my manual J with new measurements later tonight and post it here.

    2. whanna27 | | #5

      Okay, how does this look now? I adjusted internal loads to reflect more reasonable equipment ratings and added use factors (0.5) to the loads in the other bedrooms since it's unlikely they'll be used in the middle of the day (assuming only 1-2 people working from home). Also adjusted all the window areas to be "low" based on my measurements.

      The indoor design temp for the summer was dropped to 72 deg, because the Mitsubishi 18k Hyper Heat's have markedly better latent capacities at 72F than 75F and that was the only way I could get it to work with the manual S. Can't quite wrap my head around this yet, are they just that much better at cooling/heating due to the variable compressor? Otherwise I would need to upsize to a 24k and the sensible cooling would be ~200% oversized. Then I believe partial loading of the unit might become an issue since the max cooling load is half the rated capacity....I think. I think this adjustment is reasonable because I keep the bedrooms at 68-70F at night and 75F during the day. Thoughts? Any other suggestions? CFM is right on the money with the 18k unit as well, it would put the fan at the medium setting which should position me well for duct design.

      1. Expert Member
        Akos | | #6

        I think you are within the range of a 12000BTU unit, no need to oversize:!/product/34552/7/25000///0

        The CFM numbers in Coolcalc generally are all useless. Get the load from individual rooms than look at the data sheet of the air handler you want to figure out how many CFM/BTU it is. Size you ducting from there.

        The issue with air handlers in the attic is that they tend to leak a lot of air. Any air leaks there are double loss in cold climate. First you are dumping conditioned house air into the attic. Second, this conditioned air that is lost is now replaced with outside air through the air leaks of the structure. Even simple things like the drain can leak surprising amount of air. In cold climate you can also run into issues with drains freezing up and bursting.

        If you must put the air handler in the attic, I would get a couple of sheets of say 2" thick Thermax and build an air tight doghouse around the air handler that is sealed to the ceiling bellow (not the floor in the attic). You can air seal the seams with a quality tape (ZIP, 3m8067) and seal to the ceiling bellow with canned foam. I would do all duct connections inside the doghouse and run only point to point ducting to the ceiling registers. This way air handler leaks don't matter, things are easier to access for service and you get some real R value around the equipment.

        I would run the return to the ceiling bellow inside the doghouse as well. Most air handlers can be converter to bottom intake, from there a short straight run to a plenum and a filter grill in the ceiling bellow.

        1. whanna27 | | #7

          Thanks for all the info. See the attached manual S for the 12k mitusbishi. It only meets ~85% of the sensible cooling load and just barely satisfies heat loss, so the 18k it is then.

          Any chance you know what CFM the performance tables are conducted or if blower heat is accounted for in the attached engineering manual, I didn't see anything? Additionally is coil pressure drop accounted for in the ESPs given in the manual?

          The return air configuration is one of my next major decisions I need to make. I was initially planning for dedicated returns but the ductwork was going to be complicated. A central return would be easier but the only location it can go is over the open stairwell and I worry it would just suck the air out of the first floor (pellet stove down there as well) and air would get "trapped" in the bedrooms. Only of concern at night when bedroom doors are closed, but in a heating climate this is when it get's coldest and the most btus are needed. Doors have about 1/2-3/4" of undercut on them which would only amount to ~15sqin of area. Thoughts?

          1. Expert Member
            Akos | | #15

            A quick look at the data and it looks like you are using the rated numbers. For 99% design you use the max numbers, with is 15000BTU for the 12k H2i unit at 5F.

            The blower data includes everything in the cabinet, the external static pressure you need to meet is sum of the loss through your supply ducting and return ducting.

            For a simple attic install I don't think you need an SVZ air handler. In some ways it makes ducting easier as it looks like a standard furnace so you can use off the shelf bits but it is also spendy. For heating/cooling bedrooms, I've used their SUZ low static unit without too much problem but does require careful ducting. An nice in between is the PEAD which has much beefier blower.

            Central return is the norm here and works very well. Air won't get trapped in the bedrooms with door undercuts. If these are not big enough, you can create too much pressure locally and start pushing conditioned air out through the air leaks of the room. For bedroom flow rates, 3/4" is typically good enough, 1/2" might be pushing it. If you have a sensitive manometer, you can check after the install and increase the undercut if needed.

            As for drawing in polluted air from the stove, that air is already in your house. No matter where you put the return, it will mix and draw it in. No easy answer to this except be careful when loading the stove. You can also add a carbon filter to the unit which does a decent job of removing smells. It is easy to design with a larger 5" filter frame which would have room for a 4" pleated + 1" carbon filter.

            If you still feel the 1 ton unit is too small, Mitsubishi also has 1.25ton which has about the same minimum output but slightly larger heat output.


            In case you need more engineering data, it can be found here:


          2. whanna27 | | #17

            As usual, thank you for all the info. I'm in agreement on how to size the heating capacity and I am looking at that same table. I think at least, are you referring to page A-678 (for the 12k unit)? However in sizing the cooling capacity, latent cooling needs to be considered and the table on A-677/678 does not break out sensible heating vs total heating which is why I was referring to the table on page A-648. Am I looking at the right stuff? How else am I supposed to determine latent cooling capacity of these units?

            I need to look into the air handler more. I initially was looking at the low static ones but I have some pretty long duct runs from one end of the attic to the other so was leaning towards the full air handler. However without having done manual D I can't say.

            And sorry, let me clarify on the central return. The pellet stove is sealed and vented to the outdoors, there will no pollution. My concern is the 2nd floor return in the stairwell removing warm air from the 1st floor because it's a path of least resistance compared to pulling air from under doorways on the second floor. Is that a valid concern or no?

          3. Expert Member
            Akos | | #19

            Latent capacity of these units is all over the place. If you look at the cooling min BTU and min airflow, the CFM/BTU will give you almost no latent cooling. Things get better at higher modulation but never great, sacrificing latent cooling is one way they get the super high SEER.

            Trying to match both latent and sensible is somewhat futile effort. Do your best to right size the unit, if you run into humidity issues you can run the unit in dry mode or reduce the static pressure setting on the air handler to decrease CFM/ton.

            As for the return, with a wood stove providing heat to the main floor, you want the return in the stairwell. That extra bit of heat means the heat pump won't have to work as hard and it will cost you less to heat the 2nd floor. Around me wood is much cheaper than either electricity or gas, so it is definately the more cost effective way. This won't really steal heat from the main floor, just reducing the temperature stratification a bit.

          4. whanna27 | | #20

            2 follow up questions here:

            1) Can you elaborate more on CFM/BTU? I found no indication in the engineering manual of what CFM those measurements were taken at. I would surmise that the "max" capacities were taken at max CFM because basic thermo; Q=m*deltaH. However at higher flow rates and velocities you lose latent capacity, right? So in that case isn't it very important to know what CFM this data corresponds too? Also without knowing the CFM how will I be able to determine the CFM/BTU for my manual D?

            2) How do I size this system then and create a manual S to match then? Manual S is also required by NJ IECC adoption for a permit. I can use the max ratings to size the total cooling capacity but if there is no latent capacity when at max capacity then how do I ensure my latent load is satisfied? I don't think I can just assume the sensible cooling in the "rated" table is the same as the "max". It seems like there is a lot of information missing to properly size this but maybe I'm getting into the weeds a bit.

            Also the 15k PEAD and SUZ combo looks like it will probably work and I found it in stock on HVAC direct. Only concern will be a filter eating up a lot of that ESP.

          5. whanna27 | | #22

            For example:

            Goodman lists CFM in their performance data tables, and as you can see the latent capacity decreases at higher CFM. I emailed mitsubishi applications engineering to try to get some clarity here. We'll see if I hear back.

          6. Expert Member
            Akos | | #24

            "maybe I'm getting into the weeds a bit"

            Yup. Sometimes it is fun to get into the weeds and learn something new. Your Man J/S/D/T is aready more accurate than probably 95% of the ones submitted.

            The max ratings are at max CFM. The latent removal at is not great but not too bad as the unit is running around 430cfm/ton. Their data sheet at 95F shows it an SHR of 0.82.

            I haven't looked at cooling on these in detail but looking at the data sheet, their turndown is pretty bad. You'll have to configure the unit for fan off in cooling after setpoint is reached as it will cycle which can cause humidity removal issues if the fan is kept on.

            If you are going to DIY, these have a much better modulation but don't have a base pan heater:


          7. whanna27 | | #25

            AH, looking at these fujitsu manuals make me much happier. Why do you say Fujitsu is better for DIY, and can you elaborate on turndown? I was leaning towards mitsubishi cause there are a lot of mitsubishi installers in my area and I think it'd be easier for someone to come run lines and turn it on for me.

            The slim duct is capable of 0.8 in ESP as well which is solid. However the 1 ton doesn't look like it will meet my cooling load fully... But I'll be damned it's a lot closer than the mitsubishi. Sensible is 90.3% at max CFM and latent is 112.4% at max CFM. What are your thoughts on this? Are these fujitsu's also variable compressor so will these numbers be higher in actuality? Can I design my ducts to increase the sensible cooling? The fan performance curve shows CFM above 500 so I'm a bit confused because 500 is listed as max on the spec sheet. I'm really not as concerned with the cooling as the heating. Summer can be pretty mild here but humid. I don't think it got over 95-96 here last summer and if so it was only for a few days. That being said a 75 degree indoor design temp is pretty high in my opinion, I typically sleep at 68-70 and leave it at 78 when away.

          8. Expert Member
            Akos | | #26

            When you are installing the unit yourself, the brand preference of the installer doesn't matter. You can pick any unit you prefer as the only work they will be doing is refrigerant connections which are brand agnostic.

            "fujitsu's also variable compressor so will these numbers be higher in actuality"

            They are variable but the cooling numbers are what it is on the data sheet. Of course if your outdoor temperature is lower than the spec, you can get extra bit of cooling out of them, this is almost always the case at night time even when very hot outside.

            "The fan performance curve shows CFM above 500 ".

            The max number is at max rated pressure. If you have lower loss ducting, you can get some extra CFM out of the unit which can increase the heating efficiency a bit, not something I would worry too much about chasing. It is always better to design for lower static pressure, this lets you run the blower at lower speed and also quieter.

            "Summer can be pretty mild here but humid" For humid days it is best to set these units to dry mode. This reduces the blower speed and reduces the evaporator temperature for better SHR. You will still get a fair bit of sensible cooling which sometimes it can overcool the place during the shoulder season. What I have found works well in that case is to run the main floor unit on low heat to add a bit of extra sensible load to the house. This does feel wrong but costs less energy to run than a dedicated dehumidifier.

          9. whanna27 | | #27

            I just spent the last few hours reviewing all the technical docs for the slim duct Fujitsu units. It looks like the current model number is ADUH12LUAS1 and the other have been retired. This one is only 0.36ESP and 383CFM. Capacities are eh, about 90% of my cooling loads , which according to manual S is still okay...but I'm a little uncertain. Design manual is here:

            I have 3 questions before I proceed with this unit:
            1) There is a lot of information in the design manual for external/auxiliary heat and configuring the outdoor temps which aux heat kicks in and when the heat pump locks out completely. Based on these setting it doesn't appear you cant configure the heat pump to run standalone below 14F. This can't be right? Can the default settings just be left and since I don't have an aux heater installed that control signal won't be getting sent to anything?

            2) Will the no base pan heater be an issue? Do you know if they sell a kit to add one on? The only mention of low outdoor temps in the install manual is to not install the drain connector hose or something of the sort...

            3) Outdoor unit is precharged for up to 49 feet meaning no additional refrigerant needs to be added. My lines will be about 35feet, so does this mean I can run the lines myself, pressure test, vac down, check with manifold gauges, etc...? Or will refrigerant need to be removed?

            Again, appreciate all your help. Happy new year

      2. jwasilko | | #10

        When you start thinking about controls, we found that this tech note really helped optimize humidity removal. We use the Thermostat Interface with a Honeywell thermostat that can manage the AC for de-humidification. Locking the fan speed at low for 1st stage made a noticeable difference:\Application%20Note%203057%20ME%20-%20Thermostat%20Interface%20Settings%20for%20Humid%20Climates.pdf

        1. whanna27 | | #11

          That's awesome, I am considering a Nest thermostat. My only concern with using a lower fan speed is that the air velocity will then be lower and more heat loss will be experienced in the ducts since they are in the attic. I've watched several duct design courses on youtube and they seem to indicate that in harsh environments (attic or other unconditioned space) you should run at max air speed (900fpm) to reduce losses and reduce chances of condensation. Do you have any insight to this?

          1. jwasilko | | #12

            I can't give any specific duct design advice, I can only pass along the difference in comfort. Before we made that change (and also changed dip switches on our branch-box outdoor unit to lower the target coil temp) we couldn't get below 50-55% humidity in the summer. Now we can easily hit 40-45%.

            This talks a bit more about the fan speed/coil temp relationship:


        2. whanna27 | | #13

          For some reason I can't reply to your most recent comment..curious are you ducted or ductless? If ducted, where are your ducts and air handler located?

          1. jwasilko | | #18

            We're ducted. 1st floor air handler is in the basement, and 2nd floor air handler is in the unconditioned attic.

            The first floor ducts were sealed with Aeroseal when we upgraded to the heat pumps, and the 2nd floor duct work was completely replaced. It was tested to ensure it met specs. In an ideal world our ductwork wouldn't be in an unconditioned space, but it is what we had.

            We had one central return for the 2nd floor, but added returns in the 3 bedrooms and that was well worth it.

          2. whanna27 | | #21

            Thanks for the pics and feedback, That's exactly what I was looking for. I'll probably ask you for more at some point.

  2. Expert Member
    Dana Dorsett | | #8

    Rather than CoolCalc (which has both good an less-good aspects), a better (as in, "easy to use, harder to screw up") load tool for sizing heat pumps is the BetterBuildNW freebie developed specifically to inhibit HVAC contractors from chronically (and tragically) oversizing heat pumps. You need to share an email with them to sign up, but they don't resell it:

    It does allow you to edit U-factors manually, but the default values for everything are appropriately aggressive.

    1. whanna27 | | #9

      I actually just created an account the other night and started to populate a report! Hopefully I can knock the rest out tonight. I'll share here when I'm done. I was planning to verify my coolcalc report with it since I need an ACCA approved manual J report or calculation to submit to the AHJ. Will the AHJ care...I don't know haven't called and asked yet...I don't have a local office so I have to deal with the state and they can be picky about things.

    2. whanna27 | | #28

      Ended up coming out kind of close to coolcalc:
      Heating = 16,261 BTU
      Cooling = 8,465 BTU

      Much more heating dominant which is what I expected. Most of the heating load is coming from a higher infiltration rate assumption than what coolcalc assumed. Cooling load seems a tad low, but I stuck with the standard internal load assumption of 2400BTU/hr which I think is a bit low. This makes me more confident in the 1 ton Fujitsu I was looking at.

  3. walta100 | | #14

    Is this line a typo?
    – unconditioned attic with ~R22 insulation in rafters

    If the insulation is truly between the rafters and parallel with the shingles seems to me the attic would be conditioned. Perhaps you meant to say the insulation is between the joists and on the attic floor.

    If the line is true is there some reason not to move the insulation to the attic floor?

    I think putting air ducts in unconditioned attics is a very poor choice consider your options to undo this poor decision. . If not, a conditioned attic is marginally better than ducts in an unconditioned attic.

    Seems to me being as you have an existing building with a fuel usage history would give you a much more accurate load calculation than the theoretical manual J.

    As an engineer I have to ask does the math really say you can operate a heat pump for less money than NG furnace? Seems unlikely at best.

    Seems to me this building need air sealing and insulation improvements done before you replace the old equipment. If not, your equipment will be over sized when you get around to making the sensible upgrading.


    1. whanna27 | | #16

      Yes, typo. R22 is in the joists. Attic is truly unconditioned. However, I really don't want minisplits cause I think they are an eye sore, despite them being a potentially better retrofit. This is more of a work with what I got situation then spend the extra dollars to improve the efficiency of the house. If I were to improve the efficiency I'd tear it down and start again. It has an uninsulated slab on grade which sucks every last bit of warmth out of the house during the winter. Realistically I won't be making any major renovations or improvements to this house for another 5-10 years. I'm looking for a way to replace my electric baseboard heat which is astronomically expensive to run. Keeping the house at 60avg (high of 64 when occupied) degrees yields $300/month in the winter. I can't imagine what it would cost if I kept it higher. Propane/NG aren't off the table but will be cost prohibitive since I don't have any gas currently run to the house. Natural gas is in the street but the gas company makes you round up a bunch of neighbors to sign up as well otherwise they make you pay for the hookup. Plus with the price of natural gas this winter I'd say a heat pump is more cost effective. The issue is I don't really have existing usage data to look at. It's all electric heat now which is difficult to tease out what portion of that is heat vs appliances, water, etc...The pellet stove I've only had for 3 months and honestly it's doing a pretty good job at heating the whole house but goes through a lot of fuel. I've been tracking pellet usage so it will be easy to tabulate heating load at the end of the winter.

      If you look purely at $/BTU this heat pump can output 21kBTU/h @ 2000W @ 5degC outdoor and 77degF indoor (the baseboard just in my master bedroom is 2000W). At 14c/kWh would be $0.013/kBTU. Math:
      (2kW * $0.14/kW)/21kBTU = $0.013/kBTU

      As for the furnace. Smallest one I can find on HVAC direct is 30k BTU/h input 28kBTU/h output, so about 200% oversized not terrible. 30kBTU/h input would be just about 30 cuft of gas per hour. Currently NG is $16.50/kcuft in NJ. 1cuft of NG = 1030BTU. Therefore $/kBTU(output) = $0.017/kBTU making the furnace 31% more expensive to run. Math:
      ($16.50/kcuft * 30/1000kcuft)/28kBTU = $0.017/kBTU

      That being said...if natural gas were back to $10/kcuft or less...the cost would be $0.011/kBTU. However, I don't think we're headed that way anytime soon. Heat pump is still the clear winner. It all boils down to BTUs at the end of the day. The heat loss in the house is what it is so if you can obtain BTUs for a lower cost it will be a cheaper solution. So as an engineer, the math checks out to me, but I'm open to challenge or discussion :)

    2. whanna27 | | #23

      To add on to this further. Propane (delivery truck and storage tanks) is the predominant form of gas heating fuel in my area. It's about $3.50/gal now and was $4/gal last winter at $3.50/gal:
      ($3.50/gal*30000BTU/93000BTU/gal)/28kBTU = $0.04/kBTU

      Pellet stove. 8000BTU/lb, 40lb bag = $6.10, $0.1525/lb. 42,000BTU input, 72% HHV efficiency
      ($0.1525/lb*42,000BTU/8000BTU/lb)/(42kBTU*.72) = $0.0265/kBTU

      Electric baseboard. kW in = kW out, BTU in = BTU out
      ($0.14/kWh/3412BTU/kW)*1000BTU/1kBTU = $0.041/kBTU

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |