Fig.1: data acquisition using a Doppler motion at the monochromator.

The backscattering
monochromator is moved with a velocity **v**_{D} parallel to the
reciprocal lattice vector **τ** . Thus the energy of the reflected neutrons
is modified via a longitudinal *Doppler effect*:
(in first order and for v_{D} << v_{i}; the neutrons "see"
a different lattice constants in case of a moving lattice). One registers the
scattered neutrons as a function of the Doppler velocity v_{D}. The
maximum achievable Doppler speed v_{D} determines the maximum energy
transfer that can be reached. The Doppler velocity is varied periodically (sinusoidal
function in v(t) diagram of Fig.1) and for identical analyser- and monochromator-crystals
(same temperature and orientation) the corresponding energy transfer is centred
around E=0. If we count a neutron at time t_{D} then we have to know
which speed v(t_{M}) the monochromator had earlier at time t_{M}
= t_{D}.-T_{MD} (T_{MD} = neutron flight time from the
monochromator to the detector) when the neutron was at the monochromator. Neutrons
are then accumulated into the corresponding equidistant velocity channels v(t_{M})
(Fig.1; rhs for detector and lhs for monitor). The neutron flight times between
the monochromator and the detector or monitor, T_{MD} and T_{MM,}
are assumed to be constant, because **k _{f} **is fixed and we neglect the variation of the
monochromator position and the small change of neutron speed (valid for IN10,
IN16 and BSS; HFBS and RSSM have 3 times larger monochromator amplitudes and
higher speed, thus T is no longer a constant). Data are acquired with constant
time bins of typically several 10 µs and thus, for a non-linear Doppler velocity,
a constant rate signal (e.g. background) shows up as the inverted function of
the periodic Doppler motion (see monitor spectrum in the left vertical graph
of Fig.5a). The monitor spectrum indicates how much time is spent measuring
each velocity channel, g(t

October 20, 2005