By Rick Andrew

When I mention to folks that I’m involved in testing and certification of water treatment products, they are immediately interested and have questions on the topic. I can’t help but keep a mental tally regarding the most frequently asked. Nine times out of ten, it’s: Do those carafes with the filters in them really work?

I love being asked this question, because it provides an opening for me to explain just how much science goes into the design and construction of those products and how much rigorous laboratory evaluation is required, all so that I can confidently reply: As long as they have been independently tested and certified, you can rest assured that they really do work as advertised!

In terms of their fit into NSF/ANSI 42 and 53, these products are categorized as: nonplumbed pour-through-type batch treatment systems. There are some unique aspects of procedures for contaminant reduction testing for these products that are not connected to the water supply. As such, NSF/ANSI Standards 42 and 53 recognize this uniqueness and address testing of these products in separate sections. These sections are brief, however, and leave out details that can be helpful to those interested in understanding how the testing is performed. This column will describe some of the uniqueness in testing these systems and compare and contrast it to testing of plumbed-in systems.

NSF/ANSI 42 and 53 requirements for contaminant reduction testing
NSF/ANSI 42 and 53 Contaminant Reduction Tests are separated into two broad categories: chemical reduction and mechanical reduction. Chemical reduction involves adsorption or absorption of contaminants from water by media within the tested system. Mechanical reduction is a physical separation of contaminants from the water by the system. The following discussion is based on chemical reduction testing, but the procedures for mechanical reduction testing are very similar.

NSF/ANSI 42 and 53 require chemical reduction testing to be conducted to an end point derived from the manufacturer’s rated capacity for the system. This rated capacity is based on a treated volume. These requirements are in place regardless of whether the system is connected to the plumbing supply or not.

When testing systems are connected to plumbing, introduction of the contaminated challenge water is cycled on and off. The allowable cycle times vary by test. Testing is conducted 16 hours per day, with an eight-hour overnight stagnation period. The flow rate for testing is either controlled at the manufacturer’s rated service flow if testing to NSF/ANSI 42 or at the maximum flow rate attainable through the system with 60 psig dynamic pressure if testing to NSF/ANSI 53.

NSF/ANSI 42 and 53 are less specific with respect to systems that are not connected to the plumbing. NSF/ANSI 53-2006 specifies the following for chemical reduction testing of non-plumbed-in systems:

“Two systems shall be tested using the appropriate challenge and influent water after establishment of the manufacturer’s recommended use pattern, with automatic cycling. If there is not a recommended use pattern, the systems shall be operated on the basis of four times the bed volume per batch. The cycle shall include a rest period of 15 to 60 seconds between batches, timed from the cessation of streamed flow.”

The key to the above-referenced section of NSF/ANSI 53 is “manufacturer’s recommended use pattern”. Manufacturers typically advise consumers to process a maximum amount of water per day through their system. For a carafe-type system, this maximum recommended daily throughput is typically two to five gallons.

Taking into account the standard requirement and the manufacturer’s recommended use pattern, the daily processing of water through the system during contaminant reduction testing follows. The number of batches processed per day is determined by dividing the daily throughput by the size of the untreated water reservoir or hopper as many manufacturers call it. For example, a carafe system with a two-gallon per day use pattern and a one-quart hopper would have eight batches processed per day during laboratory testing.

The standard directs that a rest period of 15 to 60 seconds shall be included between batches. This language makes it clear that complete batches must be processed. This requirement rules out testing with a continuous influent challenge feed system that maintains a certain level in the hopper and lets the water drip through constantly. This is not a normal consumer usage pattern and it may result in either better or poorer contaminant reduction test results than would be obtained by testing with complete batches.

It is much more likely that in actual usage, batches would be processed at random intervals. Sometimes, a batch might be processed immediately after the previous batch has finished. At other times, several hours may pass between batches. This causes some contradiction in the standard’s requirements: recommended use pattern versus 15 to 60 seconds between batches. NSF has interpreted that testing with a typical usage pattern is more appropriate than testing strictly with 15 to 60 seconds between batches. Therefore, when testing at NSF, the correct number of batches are processed each day, but at varying intervals. These intervals are dictated in the laboratory by practical considerations. Another batch is processed when required by the use pattern and as the technicians have time. The result is varying batch processing patterns each day of testing, which is representative of actual consumer use patterns.

Carafe systems testing challenges
One result of testing based on rated capacity and recommended use patterns is that testing takes a relatively long time, with relatively small volumes of influent challenge being used each day. For example, a system with a 40-gallon rated capacity and a two-gallon-per-day use pattern being tested to 200 percent of the manufacturer’s rated capacity under NSF/ANSI 53 would be challenged for 40 days.

It would seem to follow that a batch of 80 gallons of influent challenge could be prepared when starting the test, the influent challenge being analyzed in the laboratory for conformance to the standard’s requirements and the batch of challenge used for the entire 40-day test. If this approach were possible, testing laboratories would be faced with minimal daily labor requirements once the test was up and running.

Unfortunately, this simplified approach does not work in most cases because the influent challenges are not stable enough. Contaminants in the influent challenge must remain within a specific concentration range in order to meet the standard’s requirements. If those contaminants tend to adhere or plate onto the container’s surface or photodecompose or volatilize and evaporate out or react with dissolved oxygen, then the contaminant challenge will not remain stable for the entire test. Depending on the specific contaminant, the influent challenge must be made frequently enough to account for any of the possibilities that can lead to contaminant concentrations drifting outside the acceptance ranges specified in the standard. And with each batch of influent challenge and at every sample point, the contaminant concentration in the influent must be analyzed to verify that the challenge has the correct concentration.

These same issues of influent stability exist when testing systems that are connected to plumbing, but the shorter duration of the test reduces their impact. Because of the 16-hour per day on/off cycling, typically hundreds of gallons per day are being processed. Tests of plumbed-in systems usually last no more than about 10 days and sometimes as little as one day. Even when the test takes several days, the higher throughput means that 500-gallon batches are processed in a short period of time, minimizing stability issues.

Carafe systems enjoy popularity
Carafe systems tend to be relatively low cost, easy to operate and available at many popular retail outlets. Even non-scientists are usually quite interested in the capabilities of the systems and the details of how the systems are tested. And again, the answer to their question is: As long as they have been independently tested and certified, you can rest assured that they really do work as advertised!

About the author
Rick Andrew is the Operations Manager of the NSF Drinking Water Treatment Units Program. Prior to joining NSF, his previous experience was in the area of analytical and environmental chemistry consulting. Andrew has a bachelor’s degree in chemistry and an MBA from the University of Michigan. He can be reached at 1-800-NSF-MARK or email:


Comments are closed.