AAMA ratings for U-Value vs. NFRC ratings

NFRC is mostly a residential/light commercial window standard. AAMA is more commercial. Notice the big residential window makers like Anderson and Pella are not AAMA. There is some overlap. Either thermal test would be acceptable for LEED purposes. The most important thing that project teams miss when it comes to the window performance of the Proposed is producing a reasonable value. The AAMA test would be preferable to trying to model the frames and the glazing within the modeling software as that often tends to produce U-value much lower than the typical test results. If you do use the frame modeling method make sure your results are reasonable and check them against the manufacturers testing data.

I was looking into this issue myself and discussed it with the technical team at a manufacturer. They said that the AAMA 1503 testing method is actually based on the NFRC 100 test. Here's the process as I understand it:

1. A manufacturer physically tests their assembly using a glazing product with a known CoG U-factor. This gives them the thermal performance of their frame at a specific CoG U-factor 

2. Using a computer model, they model the assembly with different CoG U-factors at standard increments to generate a "U-factor Matrix" of assembly U-factors using the results from #1 above as a baseline. The Matrix is small table with 2 columns showing CoG U-factors and their corresponding assembly U-factors using that specific product. 

This essentially generates the NFRC 100 test results. Important to note here is that the tested assembly has a fixed size (you'll see it in the fine print, something like ~70in x 70in) and a fixed ratio of frame vs. glazing area (I think ~89% for storefront and ~87% for storefront or so). All these figures are prescribed in the test procedure. 

The AAMA 1503 takes the results from the test above and adds a further step. They use a computer model to plot the performance of an assembly at different ratios of frame vs. glazing. This yields the "U-factor Thermal Chart" that we use for determining assembly U-factor for storefront and curtainwall systems. So, if your design uses the same frame-to-glazing ratio as the NFRC 100 test, theoretically the U-factor will be identical between the NFRC Matrix and the AAMA Thermal Chart. If not, the difference can be quite substantial. 

We had a project recently where the CoG U-factor was U-0.29 and the NFRC 100 Matrix for the frame product showed the assembly to be U-0.39. However, the same product cutsheets listed both NFRC and AAMA tests, and per the AAMA Thermal Chart, given our design's glazing-to-frame ratio, the assembly U-factor was U-0.48. That's quite a difference. Our design had a 81% ratio vs. 87-89% per NFRC. 

Another thing to note is, as far I understand it, that the AAMA test is a "Voluntary" test, and I don't believe it is typically required by AHJs. Even on the LEED template for Minimum Energy Performance it lists several methods to choose from for establishing how fenestration U-factor was calculated, NFRC is one of the options, but AAMA is not! 

Moreover, ASHRAE 90.1 2010 Section 5.8.2.4 explicitly requires the NFRC 100 test to establish the U-factor for glazing, with no mention of AAMA. 

This caused an issue for us during bidding, as we came to realize that the goals we had set for assembly U-factor -using the specific glazing type we wanted- was impossible using the AAMA method (which we listed in the spec language), even though it was possible with the NFRC test. We're still trying to work through the issue. 

The technical team at the manufacturer didn't have specific guidance as to which method should be used when. They said the AAMA (the more accurate one) is recommended for HVAC calcs and establishing performance, but it not required. And that the NFRC figure is what will be required for permitting/code compliance.

You could enter the physicall attributes of your system into a program like WINDOW from LBNL which will simulate the NFRC 100 test to give your the values for that test proceedure. LEED reviewers in my experience will use the simulation results as a stand in for manufacturer spec sheets with missing information like NFRC rating values. In Europe this is the only way.