"Stop Westeners": That's what they said...

Sorry for the typo: Make that 90 microseconds instead of 90ms. The remainder is correct, though.

I cannot access the PDF, so that I have to assume the 3 nm/s^2 for 2 ns are correct. That's small, but not hopeless.

The experiment is probably involving a resonating mass, which is tuned to the packet frequency of the LHC. With 2800 packets making one cycle per 90ms, this gives a frequency of about 3 * 10^7 Hz. If the resonator is of sufficient quality, we might get (I am guessing) 10^9 impulses before the oscillation is damped significantly. This amounts to a maximum velocity of the oscillator of 3 * 10^-9 m/s^2 * 2 * 10^-9 s * 10^9 = 6 * 10^-9 m/s. Assuming a harmonic oscillator, that's a displacement of 6 * 10^-9 m/s /(2 * pi * 3 * 10^7 Hz) ~ 3 * 10^-17 m.

This effect is minuscle, but it might be just at the limit of dectability. A significant problem will be the shielding of the detector, because the LHC should generate a lof of 3 * 10^7 Hz noise, which would influence the experiment. Maybe it would pay to use a higher frequency resonator, despite the smaller displacement.