HVAC system design process

 I just finished a job for a customer that had extremely leaky air conditioner ducts in their attic.  This customer didn't want to change the entire system to a new one, they just wanted us to seal up the ducts so they didn't leak air before the air could be used in the house.  The measured leak rate was 75% of the air was leaking to the outside.  We sealed the supply ducts (air going into the rooms), bringing the entire leak rate to about 9% - quit a reduction.  An interesting thing to consider is that not only did 75% of the heated or cooled air never get to the house, but the fact that the air was being pumped outside meant that the same amount of air was being pumped from the outside into the house.  Thus instead of re-heating already warmed air from in the house, the heater had to heat newly introduced cold outside air.  (Or the other way around in the summer months.)  The overall efficiency was reduced by about 90% assuming that the efficiency of the heating and/or cooling unit is 100%.  While this seems outrageous and "impossible" I find it to be the normal situation - it is just about the same as every other house that I have tested. The vast majority of homes in California are wasting something like 90% of the energy to heat and cool their homes because of poorly installed Heating, Ventilation and Air Conditioning systems. 

After we finished the job of sealing up the ducts, my customer had a change of heart and decided that since I had previously told him that his 4-ton system is too large and that a 2-ton system would work, he asked me to replace the existing packaged unit with a new, smaller one.  Here is my letter in response to that customer.  You might find it interesting and applicable to your home:

We can install a 2-T system, but you are now fighting back upstream!  We can't put a 2-T unit on your house without changing the duct work because it won't give you adequate comfort and economy.  The HVAC and house make up a "system" where all of the parts interact with each other.  If doesn't work to twiddle one part and not the others.  I am concerned that is we just replace the existing unit with a smaller one the system (house plus HVAC) won't work properly and will give you unacceptable results.  A smaller system will cost in the neighborhood of $6,000 without doing anything to make the house match the new mechanical system (which is not a good idea). We are not going to install the insulation until such time as you decide how you want to proceed.

I would be happy to design a system that will work for you, creating an HVAC system that will function properly. However, it is going to be fairly expensive.  I just completed a bid for a very similar project. It came to about $9,000 to replace the ducts, replace the diffusers, and remove and replace the HVAC system.  Yours will probably be similar.

If this ($9,000 to $10,000) is within your budget, I would be happy to create a design and provide you with a proposal.  We can't really "back up" from the work that we have just completed, but we might be able to salvage some of our work - it will depend upon the results of a detailed engineering evaluation.  While we might be able to salvage some of your existing ducts and other parts, we will need to modify them which means that we probably won't save much (or any) money.

We will not be able to replace your "packaged" system with a smaller packaged system because there are no packaged systems on the market that meet your needs.  (Packaged means that the heat pump and/or furnace is in the same package or box as the air handler).  You will need a split system where there is an "outside unit" containing the compressor, and an "inside unit" (most likely in the attic) that contains the heat exchange coil and the air handler.  It is not currently practical to get to a properly sized unit using natural gas for the furnace because they don't make any furnaces that are small enough.  I recommend that you use a heat pump for heating, which actually takes less parts and is approximately the same cost per delivered BTU for heating than natural gas. A heat pump has an additional advantage in that it is electrical and therefore the energy used can be offset with solar production if desired.

We can change your system from a 4-T to a 2-T system.  Actually, we could go to 1.5T and it would probably be better yet.  Consider this as evidence that 1.5 T is adequate:  Your 4-T system apparently "does the job" of keeping the house at a desired temperature.  It was also leaking 3/4 of the energy before getting to the house. Therefore, you were actually delivering cooling equivalent to 1/4 of that size, or 1-T.  Of course you paid for 4-T of cooling and only got 1-T for your money - but it seems to have worked.  This doesn't count the additional cost for cooling (or heating) extra outside air pulled into the house by the leaky ducts. Therefore, even a 1.5T system will be over sized.  I expect it to calculate out at about 2-T, which includes a large margin for "picking up" a cold or hot house and for rare loads such as the highest temps while have a house full of people at a party and things like that.

The problem is that now that we fixed the ducts, they will no longer leak 3/4 of the air - and therefore things might work pretty poorly because of too much air, too much cooling, too much back pressure in the ducts, etc. I have no way of predicting how it will behave because it is incorrect in almost all ways and the tools for doing these designs don't give insight into how an incorrect design will work.  It will probably work, and you will probably get the heat and cooling that you want, but it won't necessarily be quiet, comfortable or efficient. Maybe it will, or maybe not. My bet is that it will be fine, just expensive (not as expensive as before, but more expensive than necessary).

Let me describe the design process a little bit so you can visualize what is going on. I'll briefly walk you through the design process and maybe you'll see how the system interacts. (Note: a ton (T) of energy is the amount of energy required to melt a ton of ice.  It is about 12,000 BTU and about 3.5 kWhr. A kWhr is a 1000 Watts for an hour and currently costs about $0.16 at the baseline rate and about $0.24 for the average PG&E customer because of the rate tiers in use by PG&E.)

1) The first step is to determine the energy loads necessary to keep each room at the desired temperatures.  This means we need to know the final configuration with regard to insulation, windows, electric lights, number of people in the room and other heat producing equipment.  We also need to know what "desired temperatures" means.  It generally means 72F in winter and 76F in summer.  You might not want to keep the rooms at those temperatures, but the system should be able to do so just in case you or someone else wants the rooms at that temperature.   This is a room-by-room evaluation that includes average hourly weather conditions throughout the year; including energy through the windows, walls, ceiling and floor (heating from the sun, etc).

The outcome of this is the peak heating and cooling load for each room, as well as the total for the entire house. A "perfect" system would provide this much heat/cool energy to each room at the peak times, which would then add up to the total for the house (and therefore the HVAC system).  Unfortunately that would require a zoned system that could control temps in each room independently, but that is not practical in a house with a central AC system.  There are ways to do it, but they are expensive.

2) We then determine what air temperatures should be put into the rooms to achieve the desired effect.  This is a bit tricky and is a subject of some debate.  I have come to the conclusion that the cooling air should be 59F in Davis, and between 98 - 103F for heat.  There are a lot of reasons for my choices, but I don't have time to explain them here.  I can explain them to you at some later time if you are interested.

3) Given the amount of energy needed to cool and/or heat each room, and the temperature of the air entering the room, we need to calculate the required air flow into each room in the units of CFM (cubic feet per minute).   Adding these up gives us the total air flow required from the air handler (the HVAC blower or fan that moves the air throughout the house).

4) Given the air flow requirements, the shape of each room, and the location of the air supply diffusers (supplying air to the room), we can select the air diffusers. The correct diffuser ensures that air in the room is well mixed to avoid hot and cold spots or drafts, and is directed so that it doesn't impact occupants creating drafts and uncomfortable air movement.  Each room (or section of big rooms) requires specific diffuser design and selection to allow correct air flow to achieve the goals of proper heating/cooling and comfort.

Once the diffusers are selected, the rest of the system design follows to ensure that the air into the rooms is correct.   In your house this is the beginning of the problems.  It is almost certain that none of your diffusers are designed to achieve any of these goals.  We might accidentally be able to use some of them, but generally that is not an expected outcome.  Without getting these close to correct the system will not function very well. Therefore, we will likely need to change the diffusers.

5) At this point in the design process we will know the amount of air required by the air handler and will know the amount of BTU's that the compressor needs to be able to deliver at the design outside temperature extremes (28F in winter, 105F in summer) with the design loads (lights, people, cooking, heat producing appliances, etc).  This sizes the system.  I am guessing at 2-T, but that is only a guess.  In addition, these units are only made in 1/2T increments, so you can't get exact.  I tend to round up to the next size.  The PG&E classes that I have been taking suggest rounding down because there are so many other "rounding ups" that happen.  They are probably correct, I haven't the nerve to do that yet. I am unaware of any solid science that points one way or the other.

6) At this point it is possible to select a system and the ducts - they need to be selected together in a "circular" fashion.  I do a tentative (preliminary) duct layout that shows where the ducts will be run, how they will be routed, and where "Y's" , elbows and things are to be located.  Based upon that, I can determine the anticipated pressure drop from the air handler to the diffuser and back through the return ducts (the static pressure drop of the piping system).  That results in a selection of duct sizes for each section of the duct work that gives the correct pressure drop and air flow to make each diffuser work properly.  At that point I know how the ducts will be run and how big they have to be.

7) Then comes to time to select the system.  This is a bit of a problem because there are NO systems sold that meet the needs of a house in Davis.  The air handler has to be sized to efficiently move the needed amount of air, at the correct pressure, with a minimum amount of fan energy.  Step 6 and 7 go around for a while to find a combination of parts that more or less optimizes the various needs.   The air flow from the air handler is not arbitrary, it is based upon the needs of the rooms.  The fancy "high efficiency" units with multi-speed fans won't work because changing speeds changes how the diffusers work, which is only correct for one speed.  The design requires just one fan speed to get air to move properly in the room, any other speed will be uncomfortable and won't work properly. This means that a properly designed system is much less expensive than one that is sold as high efficiency because of multi-speed fans.  Unfortunately we will be using the same air handler and set of diffusers for both heating and cooling applications.  They work differently in these two modes so we will attempt to find a "happy medium" that gets close to correct for both modes of operation.  Neither will be optimal.

8) Now that the air flow is known, and the desired air temperatures are known, the compressor and the heat transfer coil can be selected.  We need to select a compressor unit that provides enough heat (or cool) to change the air flow over the coils to the correct temperature.  For example, if the house is set to 72 degrees, and the entering (supply) air is 100F, then the compressor needs to provide enough energy to raise the temperature of the air from the return (at 72F) to the diffuser temperature of 100F.  Since we know the air volume though the air handler, that tells us the size of the compressor.   A similar situation exists during cooling where we bring 76F house air down to 59F supply air.  However, we need to be careful to cool it too much because that will cause water to condense from the air, unnecessarily using a lot of additional energy and drying the air in the house, which is already too dry in Davis during the summer months.

We are then finished!!! (Finally).  Then all we have to do is purchase the parts and install the system.

The point of this rather long and detailed discussion is to show that you can't just select a compressor and hope it will work.  It has to be matched to the air handler and coil.  And you can't just select and air handler and coil, they have to be matched to the ducts.  And you can't just select any old ducts, they have to be selected to match the diffusers.  And you can't just select diffusers, they have to be matched to the room. And the room depends upon the insulation, orientation and use.   There is a very specific flow in the design process - you get bad results when you just change something in the middle of it.

HVAC systems are very seldom designed, they are constructed based upon "rules of thumb" that don't work and have never worked. There are very detailed design standards and specifications, but they are seldom used in the field.  The Rules of Thumb might have been close before the late 60's at a time when houses weren't insulated, were very drafty, and had single pane windows.  Maybe the rules of thumb worked, but generally they didn't.  The main rule was to put in a large enough system to provide enough energy to overcome whatever was going on.   That got the house warm and cool, but not necessarily comfortable and certainly not efficiently.

Once we have gone though all of this effort, the system will not work as desired!  The reason is that since there is only one air handler, and only one thermostat, the only room that will be the desired temperature will be the one with the thermostat.  All of the others will be slightly above or below the desired temperatures, and that will change as the day goes along and the sun moves.  However, the air flow into each room will be correct to mix the air and avoid drafts.  The only solution to adjusting the temperature of one of the other rooms is to adjust the thermostat, which of course messes up all of the other rooms. That is unavoidable with this kind of system.  My personal solution is to close the doors on the rooms I don't use, and open those on the rooms that I do use.  This keeps the used rooms about the same.  I could also just open all of the doors and it leaves all rooms about the same.  With a well insulated "typical" house this isn't much of a problem. In addition, in Davis the system will be oversized in the heating mode and the diffusers will tend to cause slight drafts (but at least the drafts will be warm and comfortable).

I would hate to just stab a 2-T unit on your house without first doing a design for the entire house to select correct diffusers, ducts, duct runs, etc to get a house that works properly.  There is a very big risk that whatever we choose will not work without fixing the other things.  Unfortunately, the work that we just did to seal up the ducts is going to make it very difficult to deal with since they are all glued together.  Last week they were not connected together so they were easy to fiddle with.  Not now!  We will likely be forced into removing most or all of the existing ducts and replacing them with new, properly sized and well insulated, ones. They will work much better than what is there at this time, so it is actually a good thing to do.