The Center for the Study of Wireless Electromagnetic Compatibility was established at the University of Oklahoma on the Norman Campus in Fall 1994. Its charter is to work with industry and government to investigate and resolve interaction issues between wireless phones and other electronic devices. The Center began its initial examination of the interaction between wireless phones and hearing aids in May 1995. This report documents further efforts in hearing aid immunity measurement following the Center’s Phase I Clinical Study (Ravindran et al., 1996). This Phase II-A study was primarily concerned with the development of repeatable laboratory acoustic measurements to characterize the RF interference generated by digital wireless phones within the audible spectrum of hearing aids. The report also summarizes the effects of important experi mental and technological factors on the interference.

Intermediate Phase II-A results have been presented on an ongoing basis to the ANSI ASC C63.19 working group, representatives of hearing aid user groups, wireless service providers and manufacturers, hearing aid manufacturers, the Federal Communications Commission, and at the 1996 EMC Forum #3. This report presents the total accumulation of test results using equipment configured at the C enter for determining hearing aid interference. A major contribution of this study was the development of the Input Referenced Interference Spectrum (IRIS) for quantifying the immunity levels of hearing aids over the audio frequency range.

The Phase II-A project involved the development of a repeatable laboratory protocol for testing the interference between digital wireless phones and hearing aids. Since the RF emissions from hearing aids do not present any appreciable interference to other electronic equipment, only hearing aid immunity was considered. The test procedure was developed for standard, commer cially manufactured phones programmed for continuous transmission at maximum output power, and for linear gain hearing aids providing directly measurable acoustic output. Therefore, testing was conducted under "worst-case" conditions, as oppose d to testing at power levels typical of normal use. Input-rectified hearing aids were also not tested, although some aids incorporated specialized shielding and input filters.

The main focus of this study was to investigate near-field effects or user issues, rather than far-field effects that represent bystander interference. However, many studies have pointed out the importance of developing an interference detection threshold. Hence, relative orientation and distance of separation between the phone and the hearing aid were varied to determine their effects on the interference.

Since this study was concerned with the development of a quick, repeatable setup for hearing aid testing, advanced environments such as RF and acoustic anechoic chambers, waveguides, and TEM/GTEM cells were avoided. No RF survey of the test environment was conducted, nor was any significant RF grounding i ncorporated. However, standard RF isolation practices were judiciously applied.

Hearing aids were tested in the microphone mode only, and the effects using telecoil (T-coil) input were not investigated. Preliminary tests in the T-coil mode indicated that some interference existed with the phone in a standby mode (not transmitting). This confirms the existence of magnetic audio-frequency or baseband interference. Only linear gain aids were tested.

An experimental setup was constructed to measure the acoustic output of a hearing aid. The Brüel & Kjær instrumentation consisted of an acoustic coupler (B&K DB0375), a one-inch microphone (B&K 4144) and a pre-amplifier (B&K 2669) connected to an octave-band frequency analyzer (B&K 2144). An alternate setup using the IEC 711 ear simulator (B&K 4157) was employed in some tests. The hearing aid receiver output was directed to the coupler via a 50 cm length of #13 (OD = 3 mm; ID = 2 mm) TygonÒ tubing. The purpose of the tubing was to distance the metallic coupler from the RF field generated by the phone to minimize undesirable effects. All tests were conducted in a sound isolation chamber constructed at the OU EMC Center to reduce the ambient noise floor. Fixtures were constructed to encourage consistent positio ning and orientation of the hearing aid and phone and to improve the accuracy of establishing the relative separation distances and orientations.

The interference spectrum obtained by systematically introducing a phone in the vicinity of the hearing aid was measured using the frequency analyzer. Prior to testing the hearing aid, the Reference Test Gain (RTG) was set using the Fonix 6500-C hearing aid test system (per ANSI S3.22 - 1987) without the attached TygonÒ tubing. The gain curve for the aid was t hen determined with the TygonÒ tubing attached. The gain curve was subtracted from the hearing aid output interference spectrum to yield the Input Referenced Interference Spectrum (IRIS). The IRIS was used to quantify the interference and to pro vide an objective measure of the immunity level of the aid. The higher the sound level at a particular frequency, the lower the immunity of the aid at that frequency.

The Phase II-A research followed the premise that the perception of the interference or "buzz" by the hearing aid wearer is a function of the combined effect of the acoustic characteristics of the IRIS and the frequency response of the particular hearing aid. Additionally, the pitch of the "buzz" depends on the fundamental and harmonic frequencies present in the IRIS, and the intensity of the "buzz" is determined by the amplitudes of these frequencies. It may be noted that infinite combinations of variations in frequency and amplitude exist, resulting in as many "buzz" patterns. Combined with the user’s own acoustic loss characteristics, a vast range of annoyance and intelligibility effects is possible. Hence, relating a purely objective (laboratory) measurement of acoustic characteristics to speech intelligibility, annoyance and detection thresholds can be a very complex task. Frequency-weighting indices such as the articulation index (AI) a nd the speech interference level (SIL) can, however, provide some objective measures for modeling human performance and perceptions. This study concentrated on evaluating the IRIS as an objective measure for characterization of the audio interference. T hrough a systematic manipulation of experimental variables, an attempt was made to observe the differences in the IRIS in the gestalt, rather than through any statistical analysis conducted on parameterized characteristics of the spectra.

Three phone technologies were studied during the Phase II-A testing: (1) TDMA-50 Hz at 800 MHz (NADC; TIA/EIA 627), (2) TDMA-217 Hz at 1900 MHz (PCS 1900; J-STD-007), and (3) CDMA at 800 MHz (IS-95), where the numbers in parentheses refer to the common name and the specific industry standard for the modulation scheme used. Twelve different hearing aids were tested, representing two main hearing aid technologies: behind-the-ear (BTE) and in-the-ear (ITE). For determining the influence of acoustic termination, comparisons of various lengths of tubing and of two different couplers (2 cc vs. IEC 711 ear simulator) were conducted. Repeatability tests were conducted to determine the importance of time of day, measurement settling time, microphone calibration, phone placement, and hearin g aid placement on measurement of the interference. Three distances of separation (4 cm, 16 cm and 60 cm), multiple orientations (0°, 45°, 90°, 135°, 180°, 225°, 270°, 315°, and 360°) about an axis perpendicular to the base plane, and several orientations (0°, 30°, 45°, 60°, and 90°) about an axis joining the center of the phone speaker to the center of the hearing aid microphone were examined. The primary dependent meas ure used to make inferences regarding these tests was the IRIS.

• phone technology (TDMA-50 Hz, TDMA-217 Hz, CDMA),
• hearing aid type (BTE, ITE),
• separation distance, and
• orientation of the phone with respect to the hearing aid.

2. The specific acoustic termination (tubing and coupler) modified the IRIS pattern, providing attenuation or amplification through interaction with other experimental variables. While in many cases the acoustic levels developed in the IEC 711 ear simulator were larger than those generated using the 2 cc coupler over a broad range of frequencies and test conditions, a reverse trend was found in a few cases.

3. Changes in the Tygonâ tubing length created significant differences in the measured output response of the hearing aid. Longer tubing lengths exhibited lower fundamental resonant frequencies, but a greater number of harmonics. Conversely, very short tubing lengths increased the fundamental resonance frequency, but gener ated few harmonics in the measured audible range. For tubing lengths from 25 cm to 75 cm, no significant difference in attenuation was experienced, although differences existed in the location of the resonant frequencies.

4. Repeatability tests indicated that interference levels did not change appreciably as a function of measurement settling time, time of day, phone removal and replacement in the fixture, hearing aid removal and immediate replacement, and microphone recalibration. However, significant differences were found in the repeatability measurements involving phone placement.

5. The IRIS for each of the three phone technologies exhibited a unique signature characterized by significant power at distinct frequency components based on the RF pulse repetition rate or equivalent characteristic of the phone technology. Therefore, a unique "sound" was generated by each of the phone technologies. Differences in the intensity of the sound pattern were apparent as a function of the output power of the phone and the immunit y level of the hearing aid. Shielding and filtering often led to a reduction in the magnitude of the interference. Hence, the effectiveness of a particular solution applied to a hearing aid could be evaluated from the broad-band IRIS curves. Further, t he relative immunity of each aid to the different phone technologies could be determined.

6. The TDMA-217 Hz phone produced significant, sharp peaks in the interference spectrum, especially at the lower frequencies. This may be due to the fact that the harmonic recurrence of the fundamental frequency is spaced at 217 Hz instead of the 50 Hz rate of TDMA-50 Hz. At higher frequencies (above 1.3 kHz), some s ignal ‘muddying’ was experienced as several harmonics fell within the same measurement band. This trend was especially evident when the separation distance between the phone and the aid was very small (4 cm). In most cases, the TDMA-217 Hz phone produce d greater interference with BTE aids than with ITE aids.

7. Interference effects were not consistent within a particular hearing aid type. With respect to ITE aids, several aids exhibited greater interference when tested with TDMA-217 Hz phones, while others showed greater interference with CDMA and TDMA-50 Hz phones.

8. The pattern of interference produced by the TDMA-50 Hz and CDMA phones was spread across the entire frequency range, with very little dominance at any particular frequency. The peak-to-valley differences in the IRIS for the TDMA-50 Hz and CDMA phones were also smaller than those experienced with TDMA-217 Hz phones. The CDMA phones produced sharper peaks than the TDMA-50 Hz phones. These peaks corresponded to the puncture rate dependent "bumps" for CDMA and to harmonics of the 50 Hz repetition rate for TDMA-50 Hz. Interestingly, the CDMA phone interference occurred at consistent frequencies across the various aids, although CDMA does not adhere to a definite frame rate. The interference obtained with t he TDMA-217 Hz phone was greater in amplitude in many cases than that obtained with the TDMA-50 Hz and CDMA phones.

9. Tests examining the separation distance for different aids and phone technologies produced mixed results, with some combinations showing a smaller decrease in interference from 4 cm to 16 cm and other combinations showing a smaller decrease when going from 16 cm to 60 cm. As expected, in all cases, with all phone technologies, at a specific orientation, the 4 cm separation distance produced the greatest interference.

10. The relative orientation of the phone to the hearing aid was shown to affect the IRIS. However, the effect was inconsistent with respect to hearing aid type, hence, no rule-of-thumb could be formulated.

Phase II of the Hearing Aid - Wireless Phone Interaction Study involved both laboratory and clinical testing with an emphasis on identifying the mechanisms of interaction and leading to the development of standards for hearing aid immunity and phone emissions. Phase II also focused on relating the speech intelligibility and subjective annoyance ratings of hearing aid users to acoustic measurements.

While the Phase II-A study is complete with regard to its objectives, caution should be exercised in applying the results since they are not necessarily typical of actual phone use. The research focused on evaluating the worst-case interaction between the hearing aids and phone technologies. Additional investigation is required to test these three phone technologies and others, including DECT, DCS, DITB, Digital TV, TDMA-11 Hz, and TDMA-22 Hz under "normal" operating conditions. Additionally, the role of the human head in modifying the RF field must be investigated.

This project was significantly influenced by the Phase I clinical tests. Similarly, the results of the Phase II-A study led to several developments in the Phase II-B and Phase II-C studies. Preliminary phone power measurement tests were conducted by the EMC Center (Kelley and Schlegel, 1997) to ensure the consistency of phone power levels across time for all phone technol ogies. To obtain a more comprehensive understanding of wireless phone-hearing aid interactions, this report should be read in conjunction with the reports of the Phase I Clinical Study (Ravindran et al., 1996), the Phase II-B Clinical Study (Schlegel et al., 1997), the Phase II-C Waveguide Testing (Schlegel et al., 1997), and Kelley and Schlegel (1997).