In March of 2013 the results of ASHRAE Research Project Report 1360-RP was made available for distribution. The 415 page Report is entitled “How Do Pressure Drop, Efficiency, Weight Gain, and Loaded Dust Composition Change throughout Filter Lifetime. The Contractor was RTI International and the Principal Investigator was Kathleen Owen. Ms. Owen presented the findings from the report at the September 2013 Annual Meeting of the National Air Filtration Association (NAFA). Questions were raised about the data in the report at the meeting. After further analysis of the data there are still more questions about filter selection that need to be answered and it appears, at least to his author, that some of the conclusions cannot be supported.
This Project had a number of objectives. The ones that we are focusing on in this article are:
“* to investigate the performance of in-situ filters over their lifetimes for pressure drop, dust weight gain and changes in filtration efficiency
* Compare standard ASHRAE 52.2 testing, with and without Appendix J conditioning, to the in-situ filter performance.”
In order to meet these objectives the researchers used a “Four City” study where they supplied filters to volunteers in four cities. The filters were selected and obtained by the researchers. The volunteers installed the filters in both residential and commercial HVAC systems and returned them at prescribed time intervals for laboratory testing to determine how efficiency, weight gain and pressure drop change with use. Each volunteer received two filters. After the first filter was removed and sent to RTI for testing, the second filter was installed and run in the HVAC system for the prescribed time. It was then removed and sent to RTI for testing. Filters were not returned to the volunteers for additional usage time. Standard laboratory tests were conducted on each type of filter to allow comparisons of performance in laboratory tests to performance in real life. ASHRAE 52.2-2007 tests with and without the optional Appendix J were run on each filter type.
The optional Appendix J is described in the Report as “a test method for conditioning filters intended to better reflect the potential real-life minimum efficiency” of air filters. As we shall see this wording is very important to understand the interpretation of the data in the conclusions of the report. In Appendix J tests, KCl nano-particles are used to “age” the filter and reduce or eliminate the effect of the electret charge on the fibers of electret filters.
There were three filters chosen for this research – a 2″ – MERV 7 mechanical filter, a 1″ – MERV 11 electret filter and a 12″ – MERV 15 mechanical filter. This article will focus on the data and the analysis of that data for the MERV 7 mechanical filters and the MERV 11 electret filters.
Our first concern with this research report is the selection of these filters to represent the whole classes of mechanical pleated filters and electret pleated filters. The mechanical MERV 7 actually tested out to be a MERV 5. It is not clear why this filter would be a 2″ since this is relatively rare in residential applications and is different from the 1″ MERV 11. The MERV 11 filter is also a questionable choice. The MERV 11 filters have a one-piece frame. In addition, they look like they were purchased in a home improvement store at the lowest possible price. Oftentimes, these types of filters are made with the lowest cost materials and achieve their efficiency by using highly charged, relatively large fibers. This type of electret filter has the least mechanical properties and is not representative of other filters on the market that combine electret and mechanical efficiency to obtain better results as they are used. In addition, the MERV 11 actually tested to be a MERV 12.
Two of the conclusions stated in the Report were: 1. the mechanical MERV 7 filter increased in efficiency “for almost every particle size for almost every exposure.” And 2. “For the MERV 11 charged filter, all of the filters decreased in efficiency with use.” By using a “MERV 7” filter that was actually a MERV 5 and a “MERV 11” filter that was actually a MERV 12, the researchers over-stated these conclusions in all of their tables and charts depicting filter performance efficiency.
For example, the figures in the table below are taken from “Table 3-8 Comparison of Computed Values for MERV 7 Pleat” in the Report. We have added the MERV numbers derived from the tests on the in-situ filters.
| MERV | E1 | E2 | E3 | dP in wg | |
| Standard Tests | |||||
| Pleated MERV 7 52.2 wo/App J | 5 | 4 | 20 | 33 | 0.11″ |
| Pleated MERV 7 52.2 w/App J | 5-A | 1 | 20 | 30 | 0.12″ |
| In-situ Filters | |||||
| MN – 1 month | 7 | 3 | 31 | 57 | 0.14″ |
| MN – 2 month | 6 | 2 | 26 | 48 | 0.14″ |
| NC – 1 month | 7 | 11 | 55 | 66 | 0.23″ |
| NC – 2 month | 11 | 19 | 70 | 90 | 0.37″ |
| GA – 4 month | 7 | 4 | 25 | 65 | 0.12″ |
| GA – 13 month | 8 | 6 | 32 | 78 | 0.14″ |
| CA – 1 month | 5 | 2 | 20 | 33 | 0.10″ |
| CA – 2 month | 6 | 3 | 26 | 37 | 0.12″ |
These mechanical “MERV 7” filters when tested by the ASHRAE 52.2 test method received a MERV 5. After Appendix J testing the “MERV 7” filters tested as a MERV 5-A. Of the 8 in-situ filters tested only two showed an increase from the stated MERV 7 of the test filters. One of these with a tested efficiency of MERV 11 had a pressure drop of over 3 times the initial pressure drop and had a higher efficiency because of significant filter cake. The other with a tested efficiency of MERV 8 had been left in the HVAC system for 13 months. Three of the filters reached MERV 7 when tested. However, 3 of the filters did not reach the stated MERV 7 after use. Two of these filters reached MERV 6 and the third remained at MERV 5. If we were forced to make conclusions based on this data alone, we could state that mechanical filters do not meet stated MERV values and even after extended use do not perform as advertised. Of course, this would not be a true reflection of reality.
A second conclusion from the Report is that for the MERV 11 filter type, “the ASHRAE 52.2 initial efficiency and the post-Appendix J conditioning efficiency curves did a good job of bracketing the efficiencies of the filters after actual use.” The data in the table below is from “Table 3-9 Comparison of Computed Values for MERV 11 Pleat” in the Report. Again, MERV values have been calculated and added for the in-situ filters.
| MERV | E1 | E2 | E3 | dP in wg | |
| Standard Tests | |||||
| Pleated Electret MERV 11 – 52.2 wo/App J | 12 | 37 | 80 | 93 | 0.22″ |
| Pleated Electret MeRV 11 – 52.2 w/App J | 8-A | 13 | 51 | 74 | 0.22″ |
| In-situ Filters | |||||
| MN – 2 months | 9 | 14 | 55 | 72 | 0.35″ |
| MN – 4 months | 11 | 21 | 74 | 95 | 0.27″ |
| NC – 1.5 months | 11 | 15 | 66 | 86 | 0.39″ |
| NC – 3.5 months | 10 | 12 | 61 | 91 | 0.26″ |
| GA – 1.5 months | 8 | 13 | 52 | 71 | 0.23″ |
| GA – 5 months | 11 | 35 | 74 | 87 | 0.25″ |
The initial ASHRAE 52.2 test on the “MERV 11” filter yielded a MERV 12 while the ASHRAE 52.2 test after Appendix J conditioning came out with a MERV 8-A. For every in-situ filter except for one, the tested MERV exceeded the MERV 8-A of the filter after Appendix J conditioning. In fact, three of the six filters that were in use from between 1.5 and 5 months showed a MERV 11 when tested. It is clear from this data that the Appendix J conditioning overstates the decrease in efficiency of electret filters and does not reflect “real-life minimum efficiency.” What the data shows is that while electret filters decrease in efficiency with use they continue to perform at a high level for an extended period of time.
Electret filters provide some notable advantages for filter users. Generally, they are much more efficient on smaller particles – particularly those in the E1 and E2 size ranges. They also provide lower resistance and pressure drop. This improves both indoor air quality and HVAC system efficiency. Currently, there is a movement to make the optional Appendix J a mandatory part of the ASHRAE 52.2 test protocol. As we have seen in this re-analysis of the data in the ASHRAE Research Project Report 1360-RP this would be a costly and unnecessary mistake.