Physical Aspects not Included in the Simulations
Effect of Increasing Closure Force (adding masses)
Vibration of the headphone body. Manual restraint of the headphone cup during acoustic measurement resulted in very high variations in SPL at low frequency and therefore masses were added to the body of the headphone to force closure against the measurement baffle. In the simulations, the headphone body is assumed to be ideally restrained.
Under the headphone body's own mass, a separate resonance was evident at 100Hz. Bearing in mind that the idealised scenario (solid clamping) would produce a flat response to DC, adding mass to increase the amount of restraint effectively lowered this rogue resonance effect to ~35Hz and further increases in mass produced increases in SPL <35Hz. It is possible that the noisier response below 35Hz is contaminated with noise from air leaks however it was noted that very small vibrations of the headphone body produced large variations in SPL - restraining the headphone by hand produced noisy SPL data..
The additional "closure force" crushed the cushions, increasing the SPL at low frequency in the reduced front cavity and producing a more sparse modal response between 5kHz and 20kHz due to the diminished depth dimension. 500g of additional mass was considered optimal using this approach however more realistically the closure force would be provided by the headband. Based on these results, it is likely that the very lowest frequencies would be strongly affected by full body vibration.
Measurement of Headphone Body Velocity (Laser Vibrometer)
Laser measures the mobility of the headphone body. Initially, it is interesting to see if the masses added to increase the closure force affect the resonance in the mechanical domain as was observed in the acoustic results above. Further to this, casing velocity at higher frequency suggest a general excitation of the mechanical parts of the headphone body. This movement was not accounted for in the simulation models (but could be) and represents a mechanism for loss and therefore an explanation for general differences between simulated and experimental results.
Headphone Module Sampled at Centre of Rear Enclosure
Module with Added Mass (heavy motor parts applying closure force)
Results:
Graphing the laser vibrometer velocities with the acoustic measurements of the headphone on baffle confirms that the resonance (100Hz shifted down to 35Hz by adding mass) has a mechanical cause. Theoretically, if the headphone was completely stationary the SPL would remain constant to DC as indicated in simulations. In reality, the closure force is supplied by the headband - an additional stiffness rather than an added mass and it is likely this situation will change where the mechanical resonance occurs but in either case the result is a loss of low bass.
Effect of Cushion Leakage
Another practical issue affecting the bass response is sealing of the cushion to the ear. The effect of increasing leakages between the cushion and the measurement baffle for 2 driver variant was tested.
This effect was previously examined using simulations, the general effect was that the low bass was diminished.
Measurement - Driver A with Increasing Cushion Leakage
Similar to the prediction, small leaks rapidly curtail the bass at the lowest frequencies. As this leak increases, a resonance emerges that climbs in frequency with increasing leakage.
Measurement - Driver B with Increasing Cushion Leakages to Large Gaps
As observed with driver A, increasing the leakage drops the bass content significantly. As the headphone body is spaced from the baffle, the peak in the response approaches 1kHz and the upper cup resonance (5kHz to 20kHz) drops in magnitude.