In the abstract – adopting ISO 16890 by the US sounds like a great idea. We would have one international air filter test standard replacing EN779 and ASHRAE/ANSI 52.2. This would reduce confusion, increase the size of the market and make it easier for professionals to specify air filters for a variety of applications.
Unfortunately, ISO 16890 is fatally flawed on a number of levels and simply is not the vehicle that should be used for a worldwide standard. Everyone seriously considering adopting ISO 16890 needs to carefully read the Standard. Instead of an air filter test standard what you will find is a document that tries to be “more than it is.” It attempts to link filtration with health effects, calculate particle mass based on incomplete and inaccurate data and negate any benefits of a potentially useful filtration media technology.
We have discussed some of the deficiencies of the Standard in prior articles:
- The fact that the Standard appears to address issues relating to filters used for ventilation and not for HVAC as we know it in the USA with our dependence on air conditioning: http://www.texairfilters.com/the-usa-should-not-adopt-iso-16890-for-air-filter-testing-heres-why/
2. The fact that ISO 16890 makes an erroneous and unsubstantiated connection between the negative health effects of exposure to the particles in outdoor air pollution in the PM1, PM2.5 and PM10 ranges (especially PM1) and the ePM1, ePM2.5 and ePM10 ranges in the air filter test standard: http://www.texairfilters.com/is-iso-16890-adoption-related-to-pm1-removal-in-indoor-air/
3. The fact that ISO 16890 labels filters either “urban” or “rural” based on a questionable “bi-modal” outdoor air distribution observation: http://www.texairfilters.com/iso-16890-uses-questionable-air-filter-classification-system/
Let’s take a closer look at the way that PM levels are calculated in ISO16890. This calculation of PM levels in ISO16890 is based on 100% outdoor air (in many cases called “ambient air”). There is no question that we need to consider proper filtration levels when we bring in outside air. However, outdoor air is only part of the equation when it comes to filtration – particularly with 98% of the systems in the US that recirculate indoor air and do not depend on 100% outdoor air. Particles in an indoor environment are a function of outdoor to indoor particle transfer, activity in the space, particle deposition, system run time and particle removal mechanisms (filters). Typically, particles of outdoor origin constitute from 30-50% of indoor particle counts. The generally accepted use of the word “ambient” is the outdoor environment. But “ambient” can also mean “surrounding air” and in the case of filters this would include all particles exposed to the filter – including those created indoors.
There is a good Review Article in the Journal “Indoor Air” (12/2013) on the subject of indoor particles entitled “Indoor Aerosols: from personal exposure to risk assessment.” In it the authors summarize findings on I/O ratios (Indoor/Outdoor) and state that in “naturally ventilated buildings” (ie. open windows) I/O ratios in the absence of indoor sources for PM10 and PM2.5 range from 0.5 to 0.98 while when indoor sources are present the PM10 and PM2.5 I/O ratios range from 1.14 to 3.91.
This illustrates two points: 1. particles generated by indoor sources constitute a large percentage of total indoor particles and 2. particles entering a building through an open window do not pass through a filter. As stated above, the United States is unique in the world in the numbers of buildings that have both heating and air conditioning.
Let’s move on to some other negatives of ISO 16890. Electret media, especially when it is used for pleated filters, has some distinct advantages over mechanical only media. The most important of these is lower resistance and higher efficiency of sub-micron particles. Tests (ours and others) have shown that these advantages are maintained through the life of the filters. (even though there is a reduction in efficiency) In ISO 16890 all filters are artificially conditioned to eliminate all charge by exposing them to IPA vapor. They are then tested both with the electret charge and without and the two results are averaged. All three numbers are on the test report. Maybe I lack experience, but I have never seen or heard of IPA vapors in an HVAC system. The test is used to eliminate “electretcraft” from the filter world. Why use the word “electretcraft”?
Here’s why: The mandatory IPA vapor conditioning reminds me of the Salem Witch Trials and the 17th Century drive to eliminate witchcraft. During the trials all accused “witches” were subjected to the “ducking stool.” The suspected witches were tied to a stool on a long poll and “ducked” or submerged in water. If the person died by drowning – then they weren’t a witch. But if they survived the “ducking” they must have been infested with evil spirits and they were killed. IPA vapor discharging of filters does not happen in real life. The averaging is a lame attempt to be “fair.” What do you think sales and marketing people will emphasize in their presentations about competitive products? The average or the totally discharged efficiencies? The bottom line is the discharging is biased and inaccurate. It is no more in the “right” than the Salem Witch Trials.
One of my favorite books is “How to Lie with Statistics” by Darrell Huff. It has sold over 1.5 million copies. I read it in high school and college and still have a copy. In it the author reviews many of the ways people misuse statistics to prove their points. An example from ISO 16890 could be used for almost every chapter. A case in point is sample size. Always be leary of small sample sizes. ISO 16890 has a sample size of one. (as do other filter tests) What this means is that you must look at test results with a comfortable margin of error. In the case of 52.2 I recommend that users look at a MERV number for a filter plus or minus one. (In other words, simplify.) ISO 16890 goes the opposite direction with four efficiency groups – ISO Coarse, ISO ePM10, ISO ePM2.5 and ISO ePM1. Each of these groups has numerous sub-divisions. There is simply not enough good data from a single pass filter test to legitimately and truthfully classify filters with this complicated matrix.
Another statistical fallacy to watch out for is false precision. This is where the data presented has the appearance of accuracy because it looks precise due to its level of specificity. A sure sign of this is numbers with lots of digits behind the decimal point. ISO 16890 fails on this count also. In the Standard there are examples of how to calculate PM. This consists of multiplying the particle distribution in a size channel (based on inaccurate outdoor air only numbers) by the efficiency of the filter in the size channel (without standard margin of error) which results in downstream PM numbers with 5 digits behind the decimal point. Yes, 5 digits behind the decimal point! Garbage in – garbage out. It seems to me the only reason to do this would be to give the illusion of accuracy. This is not the type of thing that should be in an international standard.
Two more points. First, as mentioned above all members of the ASHRAE 52.2 committee need to read ISO 16890 very carefully and completely. In an earlier discussion I was admonished (correctly so) for my cursory reading of the standard. My copy is now well-worn. I think if one reads ISO 16890 carefully, it will become apparent that at least some (if not all) of the deficiencies I have pointed out are valid.
Secondly, voting to adopt ISO 16890 as a US standard is a serious and essentially irrevocable step. There are some real benefits of international agreements. They can save time and money for all. Unfortunately, if ISO 16890 is adopted, the US (with it’s one vote) will be at the mercy of other countries (even to make necessary changes or improvements) with a filter market less than the size of the state in which I live. This is not a very good idea for the future of the filter industry.