Acoustic Test Station


Suite.png (618,752 bytes)

The Acoustic Test Station (ATS) shown above was developed at Poiesis Research, Inc. to facilitate the development of new acoustic materials and the application of these new materials to hearing protection devices. The ATS is modular, and can be set up for testing sheet materials, or hearing protection earcups, or hearing protection earplugs. The general properties of the ATS will be discussed first, as these properties are applicable to all of the above configurations.

General Properties


Test stimulus:

The ATS incorporates a dedicated electronic interface to store and produce the test stimulus downloaded from the controlling computer, which in most cases is a digitized "noise" file of roughly 3 seconds duration. Output of the stimulus file is continuous once initiated by the controlling computer, which is to say that the interface "rolls back" to the beginning of the file and keeps reading at the beginning after it reaches the end. Stimulus production begins one second before data collection begins to allow the system to equalize, and ends with the end of data collection. If care is taken in producing the stimulus file by arranging the length so that the end of the file is matched to the beginning, "roll back" is seamless and the length of stimulus production could theoretically be endless. Output of data is clocked at 50 KHz with a precision of 16 bits. The use of precisely the same "random noise" across tests insures that stimulus variation is not a confounding variable.
Rack.png (447,786 bytes)
Stimulus and microphone interfaces

Analog data from the integral 16-bit Digital to Analog converter is routed from the dedicated interface through a Mitsubishi DA-P10 pre-amplifier; then through a Teac EQA-20 Equalizer; then through two Fender SPL-9000 Power Amps (900 watts, each); and finally to the speakers in the test chamber. The equalizer and amplifiers allow analog control of the magnitude and composition of the stimulus without resorting to additional digital stimulus files. The two Fender amplifiers can provide a total of 1800 watts of power, easily resulting in 136 dB SPL test noise within the test chamber, or more if done briefly.

Data recording:
Chamber.png (555,639 bytes)
Test Chamber

The test chamber has hard internal surfaces (stainless steel) to facilitate the build up of internal acoustic reflections, thus magnifying the effect of the acoustic stimulus. Two microphones are used for data collection, one reference and one for data. Both microphones are Knowles BL-1785 Instrument Microphones. These microphones have a very flat response out to 4 KHz, built-in preamplifiers, and are very sensitive while also being rugged. Both microphones are simultaneously sampled, and are clocked at 50 KHz. Analog to Digital conversion from the reference and data microphones is done with full 16-bit precision (two independent Analog Devices AD976BN A/D converters). Simultaneous data collection insures that sound pressure levels (which are a function of instantaneous wave phase) are comparable across samples. Data collected from both microphones are held in a dedicated FIFO memory until they can be asynchronously downloaded to the controlling computer for processing.

Data processing:

A single data sample is comprised of two 16,384 word 16-bit precision arrays, one from each microphone. So, for example, if eight samples are collected for a particular test, then 8*2*16,384 16 bit numbers will be produced for analysis at a clock rate of 50 KHz. Two of these arrays will be collected simultaneously (one from each microphone), so the total collection time will be 8*16,384*(1/50,000) which equals 2.62144 seconds.
First page.png (20,261 bytes)
Main screen

After the samples are collected, selecting "Apply window to data, FFT & average sets" and pressing [enter] will initiate the windowing selected on a sample by sample basis, will perform a Fast Fourier Transform of each sample, and will do a bin-wise average of each sample. If a reference file has already been loaded (more about this below) the values from the current reference microphone trial will be compared bin-wise to the reference file values for the same microphone and the magnitude of the largest discrepancy will be displayed. This provides a convenient "sanity check" and shows whether or not the stimulus has changed from trial to trial. Values less than a single dB are usually insignificant.

At this point in an experiment the results can be displayed in either linear or logarithmic mode, saved to disk, or printed.

Testing procedure in detail:

First, the test chamber is configured for the tests that are contemplated (more below), so, for example, if a flat acoustic material sample is to be tested everything is assembled for that test EXCEPT that the sample is left outside the chamber!

The ATS is powered up and the software is loaded.

The stimulus file is loaded. For example, you might select the default "Noise.noi" and press [page up].

The windowing desired is selected (if you do not know what this means, you might use the default "Hamming") and the number of samples to be collected is entered. Experience has shown that the default number of samples "8" is quite sufficient for most purposes.

Most important: YOU MUST CLOSE THE CHAMBER DOOR! The author has forgotten this detail many times, in spite of the jarring and very memorable experience of a 136 dB blast of sudden noise.

Make sure everyone present is wearing hearing protection earcups.

Select "Collect Data" on the screen and press [enter]. If this is the first data collection you will hear the computer downloading the stimulus file, the tone will change as the stimulus begins being outputted, and one second after the tone changes actual data collection will commence. The stimulus data is retained in memory, so subsequent trials do not require a stimulus download. If 8 samples are being collected, then roughly 3 seconds later the stimulus will abruptly cease.

If everything has gone well to this point, you would select "Apply window to data, FFT & average sets" and press [enter]. When the status window shows ***DONE***, you would select "Make file REF" and press [enter]. These data will then be the data against which all subsequent trials are compared.

Now, without changing anything else, you would place your sample to be tested in the test chamber as shown below. It is important to seal the sample well to prevent acoustic leaks around the sample, so in our lab we use either a cheap KY Jelly substitute or a denture adhesive for this purpose. The former has the advantage of sealing fairly well while remaining water soluble, the latter actually works slightly better but is more difficult to clean up.

Select "Collect Data" on the screen and press [enter].

When data collection completes, select "Apply window to data, FFT & average sets" and press [enter]. After the status window shows ***DONE***, the data just collected will have been analyzed as above and then subtracted, bin-wise, from the reference data to produce the attenuation data for the sample just tested. If you had not placed a sample between the data collection microphone and the stimulus source all conditions would have been the same and the result would have been zero attenuation. The only thing that is different is your sample, so any difference in the result is the effect of that sample. By testing in this way all irrelevant variables cancel out if they have been precisely controlled in the first place. Which is why exactly the same stimulus is used for reference and test, and why our two microphones are sampled simultaneously.

Displaying and saving the data:

The ATS has been specifically designed to measure low-frequency attenuation, and has an effective frequency range from roughly 10 Hz to a practical high of roughly 1,200 KHz where the stimulus energy available falls off rapidly. That is, using the "pink" noise file that we supply which is limited to roughly this range. In all but very specialized cases this range will effectively test any acoustic material. This is because there does not appear to be any acoustic material that is an effective low-frequency attenuator that is not also a very much better high-frequency attenuator. Which is to say that, if your material will attenuate low frequencies, it will also do a much better job of attenuating high frequencies whether you want it to or not!
Log DumDat.png (18,219 bytes)
Log Data Display

The above comment is made to encourage you to use the logarithmic data display, because the low frequencies are much more finely depicted with this display and that is where your real information is located. So, select "Log display" and press [enter].

A graph of the attenuation data is displayed, with attenuation in dB on the ordinate and log frequency on the abscissa. Attenuation and both relevant axis are in GREEN with greater attenuation being downward in direction.

Superposed upon this graph are the processed microphone outputs from the most recent test, with the reference microphone data in RED and the data microphone data in ORANGE. Both the reference and data microphone data increase in magnitude upward.

Note that in the screen-capture of logarithmic data shown here: a) The attenuation of the sample increases up to roughly 1 KHz, then falls off. This is because there is insufficient energy in the stimulus to probe higher frequencies, and this can be seen by looking at the energy distribution from the reference microphone (RED). b) There is a lot of noise in all traces. This is because of resonances and the like and is quite normal. c) There is a useful stimulus range of roughly 60 dB from 10 - 1000 Hz.

Once the data have been viewed, pressing the [spacebar] will return you to the main screen. Once there, select the "Data & Summary filenames ==>" box and enter the name of the files you wish to save, then press [page down] to save them. This will save the FFT data, (***.ftd) and the 1/3 octave summary data (***.txt). Both files are ASCII text files and are compatible with almost any word processor or spreadsheet. Data can also be uploaded from disk for viewing by entering the appropriate name and using the [page up] key.


Specific Testing Configurations


Sheet Material and Hearing Protection Earplugs:
MT open screw.png (639,082 bytes)
Double wall test head


The ATS has two data collection "heads." One, used primarily for testing sheet material, is comprised of a heavy double-walled chamber containing a sound absorbing material between the walls and with a microphone embedded in a heavy stainless steel "torpedo" suspended within the chamber from a three-point rubber suspension. The double wall chamber itself is suspended on a circular stainless steel gantry which is, in turn, suspended on a three point spring/damper suspension. The top of this test head is threaded to accept various fixtures.

MT open flat.png (643,300 bytes)
Flat Fixture Affixed


The next two photos show a flat fixture screwed onto the double wall chamber, a sample placed upon this flat fixture, and a sealing fixture placed over the sample. The sealing fixture mates tightly with an "O" ring on the flat fixture, but the sample itself should be placed using KY Jelly or denture adhesive as mentioned above to absolutely assure that no leaks are present.

MT sample covered.png (649,305 bytes)
Test sample and sealing fixture in place


The "data" microphone is within the double wall chamber, and the "reference" microphone is within the fixture on the end of the "gooseneck" shown in the pictures.

MT earplug.png (641,772 bytes)
Earplug test fixture


Earplugs may be tested by removing the flat fixture and replacing it with the fixture shown at right. This is NOT a good test, as the internal volume of the double wall chamber is much larger than any human ear. It may be used for rough comparative studies however. If more accurate testing is required it is strongly suggested that a sleeve be made to fit the cavity of the two-cc flat plate coupler described below. This sleeve should be sized to reduce the diameter of the cavity to one suitable for testing the earplugs of interest.


Hearing Protection Earcups:

ET open full.png (340,785 bytes)
Two-cc Flat-Plate Coupler


Hearing protection earcups should ALWAYS ultimately be tested on live human heads. Nevertheless, much preliminary data can be obtained with the use of a two-cc flat-plate coupler as shown here. For those not familiar with the terminology just used, the "two-cc" refers to a cavity in the center of the coupler with an internal volume of two cubic centimeters. This internal volume is close to the volume of the external auditory meatus of a human ear. The "flat plate" refers to the fact that the coupler is flat, as opposed to the contours on the side of a human head.

The coupler shown at left masses over 20 Kg and is almost entirely solid in construction, thus making it quite unresponsive to a mere 136 dB stimulus. It may be mounted in the test chamber by suspending it from the circular stainless steel suspension ring after removing the double wall chamber shown above.

ET close empty.png (511,681 bytes)
Two-cc Flat-Plate Coupler


A close up view of the flat plate portion of the coupler shown above, with the two-cc cavity clearly visible

ET close with earcup.png (643,133 bytes)
Two-cc Flat-Plate Coupler


The same coupler with an experimental (military helmet) earcup in place.



Terms and Conditions


We just reduced the asking price by 25% and are now asking $37,750 for the Acoustic Test Station shown here, plus $1/mile from Pensacola Florida for delivery (Continental US only, please).

Should this price be acceptable, upon payment the entire Acoustic Test Station shown above (minus the desk) will be delivered to your facility and set up by a scientist familiar with its operation at a time that is mutually convenient.

In addition to the obvious components shown above, all control software, with source code ("C" and assembly), and interface circuit diagrams will be provided. Moreover, as much as a full work day of instruction with hands-on testing will also be provided to make sure that everything works as described and your personnel understand and are comfortable with its use. Operation is actually very simple, and should not require more than 15 or 20 minutes for someone familiar with acoustic testing to learn.


Seal 2.png (57,716 bytes)

Home
Contact
Golf Driver
Alternate Power
Surplus Equipment