These measurements describe NFS's HPM-100s, which he has modified as described above. While they are
not definitive of stock units, they do illustrate known HPM-100 performance characteristics. Note that NFS's attenuators have been bypassed; he normally uses them EQ'd.
Here are the published FR curves as we know them to date, courtesy of AK member Tripqzon, for comparison:
We kept the color coding constant across the measurements, i.e., Red is unit #1 full range in all plots, Black is unit #2 full range, Green is #1 at 30° off-axis, etc.. The mic is aimed midway between the three HF drivers.
1) RTA, 1/6 octave. The two units are nearly identical on axis full range, Red and Black. HF is up ~8 dB above 3 kHz where the midrange crosses in. There is a hole in the response just below that, down to 1 kHz, which I believe Ray observed in his listening evaluation. The supertweeter generates a peak at 18 kHz.
Additional measurements were taken on Unit #1. Green is 30° off axis, a typical listening alignment, where the high frequencies are somewhat attenuated, as is also shown in the published directional frequency response curves, above, though not so much. The supertweeter peak is gone. Violet is the HF alone, without the woofer running; more regarding that follows. Ignore the dips in low-frequency response below 500 Hz; those are room nodes. See 3) and 4), below, for the LF response, measured nearfield:
2) MLS, valid only above 200 Hz, but quasi-anechoic, i.e., room reflections have been largely eliminated via time windowing, providing a higher-resolution view of the same measurements. Off-axis (Green,) the notch between 1 and 2 kHz is substantially deepened (-10 dB.) This may derive from woofer phase interference with an anomalous extension of mid driver response in this region (Violet.) The factory curves show similar between 2 and 3 kHz, perhaps from a different measurment axis. A high-Q anomaly appears between 8 and 10 kHz, also shown, to a lesser degree, in the factory 30° response curve. Here, again, the supertweeter peak is gone, suggesting that it is ineffective off-axis:
3) SINusoidal, gated and delayed, also minimizes the effects of room reflections at higher frequencies, confirming the above findings. Additionally shown are 1/4" nearfield measurements of woofer and port responses, Blue and Cyan, respectively, reflecting elevated bass response in the region of 100 Hz. There is a substantial (~6 dB) peak at 900 Hz in all of the measurements warranting further determination.
4) SINusoidal again, with the HF response, no woofer (Violet) and Unit #2 response, Black.
We were unable to measure the response of the individual drivers, as the connections to each were soldered to the crossover leads. NFS is bringing me crossovers alone for voltage drive measurements next week....
]
