I analyzed the data of the acoustic step-wise swept-sine at WEB, yesterday.

• Fig.1: The spectrograms and the ASDs for Hrec and MIC during the test.
• Fig.2: The spectrograms relative to their median.
  • Since there are so many glitches in Hrec found in the 1st take, I used the 2nd take for the forrowing analysis.
  • At the 111 Hz injection, the glitched are found in the microphones signal for both 1st and 2nd takes.
• Fig.3: Picking up each injected frequency and corresponding time.
• Fig.4: The GIF animation of the ASDs for each injection.
  • The 81 Hz and 85 Hz injection data were glitchy and removed from a projection.
• Fig.5: The response function (frequency conversion coupling function) from the sound to the Hrec.
• Fig.6: The projection of the acoustic noise in the Hrec.
  • The rainbow colors correspond to the sound frequency.
  • The green line is the quadratic sum of them.

As far as I remember, the noise visible in both acoustic injections when the signal passes near 110 Hz is a metal structure (scaffolding? clothes closet?) inside the WEB room that starts to vibrate, thus creating a loud and wide noise (TBC).
The highest signal seems to be detected by the ENV_WE_ALS_MIC and ENV_WE_ALS_ACC_Z sensors.

Also:
a question arose about the linearity of the amplifier+speaker transducer (the total harmonic distortion of the DAC is very low, about 4 orders of magnitude below): in the Hall microphone the harmonics are seen a factor of 56 (second harmonic) and 68 (third harmonic) below the fundamental.
This also takes into account the linearity of the microphone, but the latter should be much better. We can roughly say that the acoustic injection system, when used at these levels, has a harmonic distortion around 1 % (-40 dB)