Monday, June 20, 2011

Accurate Sound Data Matters

For years, sound has been specified in terms of NC, or noise criterion. ASHRAE has published tables of recommended NC levels for many years, which are often the basis if sound specifications. A few years back some acousticians were promoting RC, or Room Criteria, and for a while, ASHRAE replaced the NC table with RC, which had the same numerical ratings, but added a letter (always an “N”) which described the shape of the sound curve. RC never took off in engineer’s specifications; it was flawed in that it didn’t really properly describe the sound shape of devices which contained fans (but worked ok for air outlets).

Fast forward to today where ASHRAE lists both NC and RC; however, most specifications still require NC. The problem with NC is that it is both a single number rating and an estimate based on assumptions. Manufacturer’s don’t actually measure NC. We measure octave band Sound Power in a reverb room and then use standard assumptions (outlined in AHRI Standard 885, 2008, Appendix E) to estimate NC (or RC or dBA or whatever is specified), as required by AHRI when data is presented in a catalog that also has certified data.

We can only guarantee Octave Band sound power as measured in a reverb room. In fact, VAV box sound levels are certified by octave band at a certification point and checked annually. At present, only VAV boxes are certified for sound generation through AHRI. Fan Coils, WSHP’s, and Chilled Beams are proposing to do so in the future, also in a reverb room.

Specifications, however, seldom require meeting stated maximum sound power levels, as would seem logical, since they could be both validated and verified. Instead, we typically see schedules based on manufacturer’s calculated NC values, and sometimes the required assumptions are listed, but more often not. This, of course, allows the suppliers to use whatever assumptions they wish, with the potential to underestimate the sound that will result in the space.

Worse, we see schedules generated by software that list NC for a single manufacturer’s device (the “basis of design”), often at very low NC levels for small units at low flows. Requiring an NC=21 in an office, because that is what was spit out of a manufacturer’s computer program, is pretty silly, as sound levels less than an NC=30 can’t likely be heard in most environments. There is often a note requiring the addition of a silencer if “NC falls below scheduled values”. Silencers are likely ineffective at the low frequencies that set the NC, add cost and pressure drop, and simply don’t make sense.

Finally, with fan powered terminals, which have an induction port open to the plenum, radiated sound levels will be significantly affected if the insulation inside the units is covered up by foil, or worse, sheet metal. We have seen increases as great as 12 NC in some reverb room tests. When the specification calls for the insulation to be covered up, and someone submits data with exposed insulation, there is likely to be a big surprise when the job is commissioned.

And the surprises will be discovered far more readily these days, as one can now download a real time sound analyzer, often for free, as an iPhone app! In the past, in order to validate a space sound level, it was necessary to hire someone with the real time sound analyzer equipment, which often costs at least $4,000. That also meant hiring the services of the guy who owned the equipment. Today, however, it seems everyone occupying the space can do their own sound analysis.

So in short, we need for engineers to produce meaningful and verifiable sound specifications, based on real needs (not some manufacturer’s computer output) and based on real product options, such as interior linings and expected inlet pressures. The AHRI 885 spreadsheet is available on the Krueger website ( to get realistic sound estimations, as well as a reverse spreadsheet ( to determine realistic octave band sound power limitations.

Authored by: Dan Int-Hout, Chief Engineer Krueger