You Can’t Eat Soup with A Fork……..

Not if you are hungry, anyways—

A true story. I have a good friend who owns an HVAC company. Not much of a Ph.D., but he is known to be honest, doesn’t pull vacuums through manifolds, and claims to almost always use American-made capacitors. We’ll call him Captain Kirk.

He called me recently, quite worked up about one of his jobs. He had just replaced the A/C system for a newly acquired client, a family of four that had just bought the house. As the summer was gaining momentum, it started to get hot and muggy inside, so they figure they’d replace their A/C system with a new one, and all their comfort problems would go away.

The brand-new HVAC installed is a 4-ton, 16 SEER A/C system with a constant torque air handler motor–featuring a whopping 10-year warranty and a whole lot of disappointment for the client so far.

“Why a 4-ton?” I asked.

“Because it’s what was there,” my friend, the installing the contractor, replied.

For the last month, however, Captain Kirk has had to keep going back. Double-checking everything, and despite doing little tweaks here and there, the system would just not keep up.

“It only cools down at night when the thermostat shows it goes down to 72 & 88% relative humidity (WHAT?!). But then, even if we set it to 65, it runs all day, and the temperature in the house will go up to 80 in the afternoon!” The clients cried, “I want a 5-ton!”

LOL!! I actually heard them say that.

I agreed to try and help my friend out. So, I reached out to Mr. “I want a 5-ton” to set up an appointment.

The House & HVAC

The house is a 2-story, 2,260 square feet, 1989 built slab on grade with a vented attic. Ducts are running outside the envelope through the attic and in between floors—typical of climate zone 1.

It seemed like enough capacity for the house to me. But to be sure, I ran a blower door test and a load calculation. The blower door number came in quite high at 3,230 CFM50. That is 1.4 CFM of leakage per square foot of area at a forced pressure differential of 50 Pascals or a 1.4:1 LAIR. You can read up on the relevance of that number here.

That corresponded to a total of 186 cfm worth of infiltration (air from outside the space being forced in) under regular/average operating conditions in the load calculation estimate. The total cooling load calculation came in at 42,322 BTUH, with a .74 sensible heat ratio. That means that 31,169 of those BTUH must be of sensible cooling and 11,153 of latent.

Ok, so could the installed unit not be adequate then? I went and dug out the extended performance data from the manufacturer of this 4 ton. When compared to the cooling capacity requirements for the house, this is what I’ve got:

Something wasn’t right. This unit should have plenty of capacity to handle the load in this house. Literally thinking outside the box, I took a house pressure measurement with reference to outside. For this, I slid a piece of tubing in between the door and the frame. I took this measurement with all interior doors open and only the HVAC running. More about this procedure is explained in this video.

In this case, the pressure went down to 2 Pascals (0.008″ H2O). It's a tiny pressure, but it can have a huge impact. Now we’re getting somewhere!

If the pressure in a house drops when you run the HVAC equipment, that means air is leaking from the envelope to the outside. In this case, if the supply ductwork were leaking, it would be to the attic or the space in between floors outside the building envelope because the blower door test revealed that it is so leaky. – As a side note, in newer, well-sealed homes, the space between two conditioned floors is considered conditioned space.  In a well-sealed home (draft stops and sealed rim joists), if this space is pressurized, the air is more likely to go back into the house than to the outside.

With this in mind, my next step had to be running a delivered capacity check using MeasureQuick. For this, I averaged the supply air temperature at the vents of the two coldest rooms and two of the warmest ones (of course, one of them was the master bedroom). The TESP was quite high at 1” WC. The fan tables for this air handler don’t go up that high. So, using a TrueFlow Grid, I measured approximately 1,200 CFM. This is what I got back:

The Energy Conservatory makes a free software tool called See Stack that can be used to estimate how much leakage (or exhaust flow) it takes to change the house pressure by 2 Pa. The fact that the house is so darn leaky makes this small pressure very significant. These screenshots show the difference between no depressurizing leakage on the left and enough depressurizing leakage to make the pressure in the house 2 Pa lower, shown on the right. It takes about 384 CFM of exhaust flow–or leakage–they have the same effect on the house. Remember this was with only the HVAC running. This is even worse if any exhaust fans, the dryer, or the kitchen hood are running.

“Pffft… we went through all this trouble to realize that a duct is disconnected?! You could’ve just crawled in the attic and found it within minutes!”–you might say. I did go up in the attic. It was quite spacious, actually (for an attic), so much so that all of the duct connections in the attic were accessible. There were no disconnected ducts up there. Whatever the issue was causing this depressurization, was not accessible unless some walls were cut open. Again, all exhaust ventilation devices disclosed were off.

Remember how in the load calculation, the amount of infiltration that the HVAC had to handle was 186 CFM? Well, according to our new findings, it looks like it's closer to 380 CFM, which is double the original number. If the load calculation is adjusted to reflect this, then:

First, notice that the trued load with the HVAC driven infiltration is about 12,500 BTU/h more than the Manual J load calculation tells us–more than 1 ton! Now, look at the sensible and latent numbers.  The sensible load didn’t go up much–only about 12% or about 4000 BTU/h.  But the latent load is up by 80%! That’s a problem because the system doesn’t have that much latent capacity, hence the 88% relative humidity reading by the thermostat, especially when “it cools down to 72 at night.”

The manufacturer says that the equipment makes 47,000 BTUH total, which is 35,600 BTU/h sensible and 11,400 BTU/h latent.  Those numbers cover the sensible and latent loads of the original Manual J loads nicely. But after we add in the HVAC-driven infiltration, we’re covering just over half the latent load.

So, between the real load being higher and 26% of the HVAC’s normalized capacity being lost out somewhere we can’t see, now we’re only covering about 60% of the total load and 63% of the latent load. Things are starting to make sense. The inability to maintain set point on any days near design and very high indoor humidity fit very well with what we’ve calculated so far.

The second bar of the graph below shows the capacity reduction, and the 3rd bar shows the load increase. They are now WAY too far apart.

So, do they need a 5-ton instead of 4?

As I’m mid-briefing the installing contractor on my findings so far, he asks: “So, do they need a 5-ton instead of a 4? I just want this to go away, you know.”

No, they don’t. Installing just about any HVAC in this house without improving the shell and ducts is like eating soup with a fork.  If you go at a bowl of soup with a fork, it will get you some of the big chunks in the soup, but there’s no chance of getting much broth. In fact, if you keep eating it that way, what’s left in the bowl will get more and more watery. Leaky duct(s) and envelope are similar, you’ll get some of the “big chunks” of cooling delivered where it should, but lots of it is slipping through the cracks, and the house is getting wetter and wetter.

But what if you just install a bigger unit, a 5-ton in this case? What we’ve got with the 4 is something like this:


Replacing it with a 5-ton is like getting another fork with a longer handle. It won’t solve the problem in any meaningful way. In fact, it might make it even worse. Connecting a 5 ton to the existing duct system would increase the airflow, but it would also significantly increase the supply duct pressure. This will exponentially exacerbate the additional HVAC-driven depressurization, creating even more leakage and even more net loss of delivered capacity. Even with a nominal increase in refrigeration capacity. A classic no-win scenario.

This author loves Manual J calculations. It provides us with a framework for quantifying estimates of heat loss and heat gain in a house. But, whenever you can turn an estimated number into a measured value, you are much better off, as the inaccuracy of one large thing can destroy the accuracy of everything. And to be accurate we need to test.

A caveat on measuring house pressure with reference to outside

We discussed measuring the house pressure, but that’s not enough by itself.  There are a few reasons why measuring only the house pressure might not tell you everything you need to know.

  • Sometimes the leakage on the supply side and return side can be large but nearly equal. In this case, you might have a big increase in the load and a big decrease in your capacity, but the change in house pressure would be nearly zero – Luckily, this wasn’t the case in this instance.
  • If the house is very leaky, even large duct leaks might not create much pressure change.
  • Windy weather can make the house pressure jump around as much as 5-10 Pa, making it much more difficult to measure a few Pascals of pressure change when you turn the HVAC system on.

So, to really understand what’s going on, these steps can be taken.

  1. Use a precision micromanometer to measure the house pressure with the HVAC system and all other exhaust fans in the house off, and then again with the unit on. It should change by less than 1Pa.
  2. Do a blower door test. If the house is very tight, a 3-Pa house depressurization might not indicate very much air is leaking from the ductwork. But if it’s a leaky house, 3 Pa might mean a LOT of air leaking in.
  3. Use a tool like SeeStack to estimate how many CFM are coming in or blowing out to create the pressure you measured and your blower door results.
  4. If the house is very leaky, it might be hiding the duct leakage problem. In this case, a DuctBlaster® test is the best way to be certain how much duct leakage you have.

Back to Captain Kirk and Mr. I-want-a-5 ton’s conundrum

Armed with the mighty power of fairly complex-to-comprehend data, I sit down at the kitchen table with the homeowner and the installing contractor. We’re not 6 feet apart, but we are all wearing our masks.

I explain the interpretation of the data to the best of my abilities–hopefully better than I did here. The homeowner seems to get a good grasp on the basics of what the real problem is. I can’t tell about Captain Kirk, though. He seems mildly relieved but not too confident. I conclude by advising something along the lines of: “Your system is not too small, but we need to tighten up the house some, and it would be great if we can test for duct leakage specifically.”

“It’s like eating soup with a fork!” says my friend, the installer, in an outburst of newfound confidence. “You wouldn’t eat soup with a fork, would you? Not if you are hungry, anyways.”

Genry Garcia – Comfort Dynamics, Inc.

Steven Rogers – The Energy Conservatory.

Russ King – Coded Energy, Inc.



To continue you need to agree to our terms.

The HVAC School site, podcast and tech tips
made possible by generous support from