We Are the 99% — Design Temperatures & Oversized HVAC Systems

22 Comments Read/write comments

hvac design temperature thermometer ashrae acca manual j

We're making progress! With the all the emphasis on energy codes and energy efficiency programs like ENERGY STAR New Homes, more homes are getting Manual J heating and cooling load calculations these days. The intent is that the heating and cooling systems installed will be sized properly because oversized systems have problems (poor dehumidification, short cycling...). But just because an HVAC contractor does a Manual J, that's not a guarantee that the system is sized properly.

It's getting hot!

One of the ways that many HVAC systems get oversized is by someone using the wrong design temperatures in the load calculation. Here's the skinny:

  • Homes lose heat to the outside in winter and gain heat from outside in summer.
  • The rate of heat loss or gain depends on the temperature difference between inside and out.
  • ASHRAE (a bunch of engineers who know a thing or two about HVAC) has a table of outdoor design temperatures for winter and summer.
  • ACCA (the trade association for air conditioning contractors) bases its Manual J load calculation procedure on the ASHRAE design temperatures.
  • Contractors running load calculations regulary override the recommended design temperatures.

The temperatures inside and outside your home are constantly changing. Sometimes the amount of heat your home loses or gains is low, and sometimes it's high. Even over the course of a day, the loads change. For example, we might hit a low of 70° F at night in the summer and a high of 96° F, and all through the day, the loads keep changing.

If we're trying to keep the indoor air at 75° F, the temperature difference is close to zero in the morning and rises to about 20° F in the afternoon. As the temperature difference (ΔT) changes, the cooling load changes. As it turns out, though, most air conditioners installed in homes cannot change the amount of cooling they provide when the temperature conditions change. Yes, some systems do have variable capacity, like mini-split heat pumps, but most are fixed capacity.

The 99% and the 1%

Enter ASHRAE's design temperatures. It turns out that if you have HVAC equipment with fixed capacity, this temperature will provide optimal performance. But what exactly is the ASHRAE design temperature?

Winter: 99% design temperature. This is the outdoor temperature that your locations stays above for 99% of all the hours in the year, based on a 30-year average. Turning it around, the outdoor air where you live is going to be colder than this temperature for only 1% of the hours in a year. That happens to be about 88 hours per year. In Atlanta, the 99% winter design temperature is 23° F.

Summer: 1% design temperature. Your location will go above this temperature only 1% of the hours in a year, again, based on a 30-year average. Here in Atlanta, that number is 91° F, so we go above that temperature for only about 88 hours in a year.

hvac condensing units oversized cooling system design temperaturesOccasionally, you'll run across something called the 97.5% and 2.5% design temperatures. It's not what you may think. Those numbers were defined differently and are not used anymore. Because the definitions were different, the temperatures are pretty close to the 99% and 1% design temperatures.

The difference between the outdoor design temperatures and the indoor design temperatures (70° F and 75° F) is the ΔT that should go into the load calculation.

In defense of HVAC contractors

HVAC contractors don't like to get called back because of comfort complaints, and oversizing solves many comfort complaints. Of course, HVAC contractors don't have control over the building envelope and usually don't know how much infiltration a house will have. Phil Mutz, one of the smart HVAC guys at Moncrief Heating & Air Conditioning in Atlanta, recently wrote about this issue:

Homes are rarely built as designed, meaning blocking/insulation/sealing almost always fall short of design. When the home is hot mid July months after a customer has moved in, the customer is calling and yelling at the HVAC guy... not the insulation guy.

With the testing and verification required in Georgia now, we should get a lot better at controlling some of these other issues for all homes. Programs like ENERGY STAR have required inspections, testing, and equipment sized to Manual loads for years now, so we do know this can work. As we get better data about the envelope, the loads in new homes will match the load calculations better, and HVAC contractors will stop being afraid to put in systems that they think are too small.

Check the numbers

Still, anyone getting a new home or doing extensive remodeling should insist on accurate Manual J load calculations and equipment that matches the loads. If you're a builder, homeowner, or HERS rater checking the Manual J reports, one of the first things you should look at are the outdoor design temperatures. The should match the 99% and 1% design temperatures from ASHRAE, which are also given in Table 1A in Manual J.

If the 1% summer design temperature is 91° F, as in Atlanta, a Manual J report showing that the load calculation was done with 97° F will result in an oversized air conditioner. Get your equipment sized for design loads, not extreme loads. Maybe it's getting hotter in recent years, but it's still best to use design temperatures based on 30-year averages — not the temperature it got to last summer for 3 hours.


Related Articles

Mr. Bean Proposes a Radical Change to Energy Codes The article above focuses on air temperatures, but mean radiant temperatures are critical, too. The first 3 articles here give some info about MRT.

Naked People Need Building Science

Thermal Comfort: Everyone Wants It but Few Know the ASHRAE Standard

Calculating Cooling Loads, by Martin Holladay at Green Building Advisor


Photo at top by Steve 2.0 from flickr.com, used under a Creative Commons license.



It gets worse. You get a LOT of days that are in the 70's when it's raining/humid and an oversized system has no way to deal with the humidity due to short runtimes. HVAC contractors and customers blame it on the weather. Customers expect their HVAC to keep setpoint no matter how hot it is outside. Redefining comfort as a DIFFERENCE between outside and inside temps would help with undersizing complaints.  
Thermal storage of buildings is also not considered in manual J calculations. Think about it, the peak heat of the day takes time to penetrate the building materials. Thermal mass of a building in effect evens out the "peak load".

John Barba

I use this example in class if students are worried about the 97.5% or 99% number: let's say you do a heat loss using the correct number and you get a heating load of 57,421 BTUH. What are the odds of you finding and installing a boiler or furnace with an absolute net output of 57,421 BTUH? Slim and none!  
An installer will pick the brand he likes, and the size up that covers the load. If that boiler or furnace has a new output of 70,000 BTUH, they now have nearly 13,000 BTUH extra, just in case - so stretching out the design temperature becomes unnecessary.  
Great article! Thanks!


So we should design our HVAC systems to be oversized 99% of the time? Why not 2 stage units that will run on first stage 84% of the time, dehumidify much better, and provide much better comfort. As you stated, loads change constantly, we don’t buy cars where the gas pedal is an on/off switch.  
On old houses, we have no clue what the Air Changes, System loses, Insulation values or any other inputs to the calculation are, so load calculations become nothing more than performance art. We agree “proper sizing” has nothing to do with energy savings so why is it included in energy codes? Hmmm. 
“Manual J” is a registered Trademark used to restrain trade, stifle competition and innovation, protect old awful legacy software products, and make ACCA (a non profit) millions a year in fees, kickbacks, and software sales? Hasn’t heat transfer been around since Fred Flintstone?  
Yea I know, it’s not about the money.  
Keep em coming, 

M. Johnson

I think you are on the right side of this argument, and perhaps can be supported by new ways to estimate the appropriate sizing. The ways I can think of are after the fact, but I think still useful. 
Consider measuring the on-time vs. off-time of an installed AC. This is something an AC technician finds difficult to do, but with some equipment the homeowner can do easily. And the homeowner has a stronger interest in the matter.  
According to the theory as I understand it, the AC when properly sized *should* run for more than one hour without stopping on these hottest days of the year. Therefore one that runs 3/4 of an hour, I judge to be 1/3 oversized. I have experimented with this approach in the Houston TX region which has *many* days which are pretty much like the design temperature AC pros use. What is wrong with this approach? 

Allison Bailes

Bob: Yeah, partial-load conditions can be difficult to deal with. 
John B.: Very true. The difference between the 99% and 97.5% is pretty much irrelevant. 
pj: You're right that that having equipment that can change capacity to meet the loads is ideal, and we mostly spec variable refrigerant flow mini-splits. Most of the equipment being installed is fixed-capacity, however. Also, I mentioned in the article that Manual J works best when you have accurate inputs.  
M. Johnson: Nothing is wrong with that approach, and it's the best way to prepare for replacing a system in an existing home. Make sure you look at the runtime when the outside temperature is close to the design temperature.

Geoff Hartman

pj - What are you? Maoist? "“Manual J” is a registered Trademark used to restrain trade, stifle competition and innovation," What? You want to build your own software tools, you go right ahead. No one is stopping you. ** Proper sizing does affect energy use, comfort, and building durability. AND - In old houses you can still inspect, test, and verify to get a pretty good estimate for the building performance, which is WAY better than a wild Guess. 
M Johnson - THat's a good point. Manual J does not account for thermal lag entirely, so actual performance is often better than calculated and hence equipment IS oversized, even at peak design conditions. AGAIN, another reason why oversizing by nudging the numbers and inflating inputs is unnecessary. However, in some cases run time does not equal complete conditioning of the building, but only that Sensible (dry bulb) temperature was achieved AT the Thermostat. half the building could still be 80F, and 80% RH, but as soon as the T'stat is satisfied, the system cycles off. Haven't we all seen lots of homes where the AC cycles for 10 minutes, but the temperature and RH are never really satisfied, or uniform throughout the home. So maybe the system is actually undersized, but cool air blows near the T'stat. 
Good design is Good design. Use the recommended Manual J guidelines, for appropriate applications, and Do Not size up. BUT, unless you are willing to run room/zone loads, AND apply the results to proper distribution (duct, etc.) design, AND then verify the performance (AC checklist and air flows) I submit that you are wasting a lot of time and often (mostly) will still result in inferior performance and comfort.  
My advice? Do it right. Run a complete design. Follow the sizing and installation guidelines closely. Test and Verify. - Oh, and charge a reasonable fee to compensate you for all your fine work and to define the value for your clients. 
Good Luck.

John Ring

I do a lot of work with clients who complain about excess humidity in the house during the mid and cooling seasons. My belief is that the only way to solve these problems is by finding out what is actually going on. 
To do this I first check the ducts for leakage and make them as tight as possible. Then I put a run-time monitor on the compressor set to five minute increments, and temperature/humidity monitors set up to record readings every 15 minutes: one outside the house in a shaded location and a minimum of three in different locations throughout the house. I like to leave the monitors in place for at least two weeks, and three weeks is even better. 
When you take these readings and put them together on a chart you have a very clear picture of what is going right or wrong. I find that much of the time it is just over-sized equipment, but in a surprising number of cases the improperly sized/balanced duct system is at least partially the problem. And in other cases at least part of the problem is attributable to things like lack of a moisture barrier on the crawl space floor or a gap in the ceiling around a bathroom fan. 
This approach, though somewhat costly, has been very successful. 
If you don't have a complete picture of the problem, it's tough to come up a good answer.

M. Johnson

Let me respond to something wrong with using a 2-stage AC instead of a determined effort to size correctly using Manual J or similar sizing. There probably is a tendency for the AC tech to size LARGER just because it is a 2-stage AC. Many times because 2-stage ACs are generally sold in 1-ton increments, the round-up error is on average 1/2 ton larger than with a single stage. Most 2-stage ACs have a 1st stage 80% of the total capacity, so oversizing this type can leave you with a 1st stage just as large as the oversized single-stage AC you could have bought for less money. 
I like the 2-stage concept in general. One of my two systems is a Trane produce where the 1st stage is 55% of the total. And when choosing it, the AC tech suggested round-up to 4 tons, but based on knowing how little the old one ran (3.5 ton and oversized), I firmly said I wanted the 3 ton size and showed him why. 
I had done a Manual J model before, and was scared to go as small as it said. Knowing the amount of runtime on the old system, gave me added confidence the Manual J model was basically correct.

Allison Bailes

from Mike Legge: 
"I have no knowledge in the sizing of of HVACs, but it seems to me that one is trying to hit a moving target. Why are not all HVACs variable to fit the problem? It is like Goldilocks that at any instant depending on the weather the unit will be either too big, too small or just right . Further, the decisions are being made on past records while, if climate change is for real then the "oversizers" will be on the right side of history. Whilst, a whole pile of people do not countenance climate change, it seems to me that what you are trying to do is predicated on your concern about energy saving and carbon reduction. I'm sure that it is not about just saving shekels for your clients. Your blog is the best!  

Jamie Clark

You should also point out that Manual J has a built in 15% fudge factor. Here in KY it's pretty humid so I always push the 2 stage equipment. 90% of what I sell is geothermal and the colder coil does an excellent job of removing moisture.

David Butler

The reason for using 99th and 1st percentile design temps is mostly for the reason Bob said.. thermal lag. All homes have it, some more than others. An HVAC system in a high mass home can safely be designed to slightly lower percentiles.  
Also, there has to be some give at the top. There are serious downsides to installing a system that will maintain the setpoint when outside temps hit 10 year highs. But builders and hvac contractors need to "man up" and explain this to homeowners.  
And as Jamie pointed out, even equipment sized per Manual J/S will have excess capacity, sometimes quite a bit, given the small # of sizes, as John B pointed out. Sizing a furnace is less critical -- ACCA and energy Star allow up to 50%. But egregious oversizing will cause unacceptable temperature swings. 
BTW, the previous 97.5/2.5 design temps were seasonal rather than annual percentiles. However, that didn't work well for areas that don't have 4 seasons, as ASHRAE realized as they became a force on the international stage.

David Butler

@Mark, you make an excellent point about cycle timing. If the existing system has clean filters, coils and is properly charged, cycle timing is the most accurate way to size a replacement system. Doing loads on existing homes involve more guesswork. A blower door and IR camera may sound like overkill, but they do remove some of the guesswork. 
@PJ, just how do you go about sizing a variable or multi-stage AC if you feel load calcs are such an awful legacy? I believe Geoff nailed you on that. Your arguments are getting a little tiresome. 
@John R, nice comment. 
Allison wrote: 
> HVAC contractors don't like to get called back because of comfort complaints, and oversizing solves many comfort complaints 
Well, that may be what they believe, but what I believe is that most comfort issues are the result of poor design and installation practice. In particular, ducts that are unbalanced, restricted, and more often than not, both. Oversizing does nothing to fix these problems. That's the irony.

Brett Moyer

Great article Allison. 
Correct me if I'm wrong, but one of the benefits of doing Manual J calcs is to get a cfm per room, which helps to determine the amount of heat/coolth that goes to each room. 
Now, for us builders (who are admittedly not HVAC pros), is there a simple chart that breaks down duct type/size/length per cfm? 
It would be nice to be able to reference this chart on the job site to verify that the HVAC contractor is installing the appropriate ductwork.

geoff hartman

Brett - No, it does not work that way. sorry = you really do need to do the proper design and installation. Variables change with each installation.  
why do we keep trying to use a "rule of thumb" to substitute for science and best practices. 
The reference chart is the design specs provided by your design contractor. If you don't have this information, you are SOL and guessing at best. 
Right is right! (John Tooley)

geoff Hartman

Brett - yes, the primary reason for Man J, after sizing, is to get the cfm. You need to complete Man J, S, T and D correctly to get these numbers. 
when your rater tests, verifies and completes QA checklists, they mus have correct cfm, or the process is not going to get the correct results.

David Butler

And even if you have the room-by-room CFM targets, there's no way to correlate to duct size. The ability of a duct system to deliver the correct airflow depends on the relative friction rates of each duct pathway, the blower curve, and the overall duct system static.  
Also, there are many different duct design solutions that will produce a good result, and many more than won't.  
If you really want to check your mechanical contractor's ducts, you can invest in a good balometer (or anemometer) and a manometer, or just pay a 3rd party T&B; technician. But as Geoff said, without the target CFM's, there's nothing to verify.

Allison Bailes

Geoff H.: You're absolutely right. Without going through the full HVAC design process, the results will be less than stellar. 
John R.: You're doing a great service to your clients by gathering all those data. When they need to replace their system, they'll be armed with the data they need to get the right system. 
Mike L.: Variable capacity systems are great, but they're generally more expensive and not what the industry is used to right now. I'm glad you like the blog! 
Jamie C.: Well, if I point out everything in the article itself, I won't get as many comments like yours. ;~) You're absolutely right, of course, that MJ8 results in oversizing even when you do everything correctly. 
David B.: You said, "In particular, ducts that are unbalanced, restricted, and more often than not, both. Oversizing does nothing to fix these problems." You know more about all this than I do, but I wouldn't say oversizing does nothing to solve these problems. A bigger blower can push more air through a restricted duct system. But yes, my remark probably overstated the benefits of oversizing. 
Brett M.: You're right - Manual J gives you not only the Btu/hr you need for each room but also the air flow in cfm. As both Geoff and David said, though, there's no simple chart you can use to see if it's correct. You need to know the static pressure in the system, the blower's fan curve, the friction rate... There is something called a ductulator which some contractors use for sizing ducts that relates flow, friction rate, and duct size. 

geoff hartman

CAUTION: Duct Calculators are not all equal. Some are set up for only one type (Smooth metal, etc.) of duct, and they are still useless until after you know what he design static is, based on fan curve, duct design, materials, etc. 
These can be even scarier than improper load calcs. I see technicians and energy auditors pull these things out and pick a random static (0.1? 0.08?) for calc. then they look at the duct size, exclaim "this room needs xxx CFM, slap CFM & static into the calculator, and Vuala! A required duct calc.  
Garbage in ='s Garbage Out 
Complete the whole process. Use software! you can't do this in your head. Trust the results, but step back and do a "sanity check" to make sure the results are reasonable.

David Butler

Allision, I was talking about oversizing the condenser (the real problem), not the blower. When the 13 SEER rule took effect in 2006, one of the things manufacturers did to improve efficiency was to reduce the horsepower of the blower motor. Of course, if the condenser is oversized, the matching furnace or air handler is will have more blower capacity, but remember, the larger condenser also requires more airflow to keep the evaporator happy. So if the ducts are restricted, then its a zero sum game.

Thomas Anreise

Great one, Allison. I thoroughly enjoy "occupying" your blog when I have time. With regard to Brett M's question about verifying proper duct sizes, having the contractor's duct design (if there is one) would be the way to do this, assuming they have used reasonable inputs to create it. The SpecPro software has a page in the report for duct sizing, which tells the installing technician what size duct needs to be run to each location. Of course, duct size is only one component of proper design and installation (flex duct compression, distribution box design, and fitting selection all have enormous impacts on airflow as well.)


I agree with Geoff on all counts.


To comply with the IRC 2009 code a Manual I,for load, a Manual S for the size of the equipment, and a Manual D for the sizing of the ductwork is required to be submitted when you build to the Uniform Construction Code in New Jersey. Sadly this is never done. If a system is not installed according to these design standards then this is a code violation. I expect this to be the new hot button for home inspectors. I just wish homeowners would get this information when they buy replacement equipment. I just got a permit application where the house was getting 10 K in insulation and a high efficiency furnace the same BTU rating as the old unit. I wonder how it will work out.