unity over the whole frequency range, while from the
sensitivity to measurement noise point of view the gain
should be as low as possible at all frequencies – a fundamental
trade-off. The analyses for different controllers
in different schemes are given in Figures 12–15.
According to the preceding analysis, the methods
can be put into two groups based on their performance.
The IHC, FSCF and SFSCF methods have very
low sensitivity (–80 dB) to the measurement noise outside
the disturbance tones. However, the phenomenon
already addressed in the performance analysis is
clearly visible again. The controllers are very sensitive
to measurement noise between the disturbance
frequencies in the cases where multiple tones are
being controlled. The IDC and DOFC methods are
somewhat more sensitive to the measurement noise,
yet still being well below the 40 dB threshold.
The biggest difference between the IDC and DOFC
is the sensitivity at the frequency bands below and
above the disturbances. At the low frequencies; the
former converges to a certain value, while the latter
approaches the negative infinity (being completely
insensitive to static measurement noise). At the
higher frequencies, the IDC method provides lower
sensitivity. The drawback of the good tolerance to
measurement noise is the loss of the reference tracking
if the same actuator is being used and no precompensation
is applied.