Step V - Time-dependent Simulation
To enable time-dependence simply set ITDP to one. In addition IALL should be zero, SHOTNOISE 1.0 and PRAD0 zero for SASE simulations. The start-up regime has typically modes with strong diffraction. Therefore it is recommended to increase the grid size RMAX0 or DGRID by roughly 50-100%. To compensate the coarser grid NCAR can be increased as well.
The desired time-window requires some minor calculations. For easier calculation 'time' is scaled by the speed of light (s=ct). Assuming a time-window, which starts at the tail s0 and ends at the head s1 the required number of slices is NSLICE = (s1-s0)/(ZSEP*XLAMDS). The first slice has the position s0. Therefore NTAIL is s0/(ZSEP*XLAMDS). Note that in the frame of GENESIS 1.3 s0 has to be smaller than s1.
The parameter IOTAIL is an indication for GENESIS 1.3 whether radiation is generated outside the time-window but might slip at s0 into the time-window while the beam propagates through the undulator. This is the case if only a subsection of the electron beam is simulated. Because GENESIS 1.3 can do only a wild guess about the radiation field which slips into the time-window the tail part of the time-window is physically incorrect. To exclude these slices in the output IOTAIL should be set to zero. The number of suppressed slices is simply the number of undulator periods of the entire undulator divided by ZSEP.
As an example we assume a Gaussian beam profile with an rms length of 1 mm (CURLEN = 1.e-3) and a radiation wavelength of 5 m for an undulator with 200 periods. The spacing between the slices is ZSEP=2.0. The output should be between -2 mm and 2 mm which cuts out a part of the bunch (IOTAIL=0). The total number of slices is 400 slice of output plus 100 suppressed slices (=500). The first simulated slices is NTAIL=-300 which is equivalent to s0=-3mm.
In this example the time-window is actually very close to cover the entire beam. With s0=-3 mm the local beam current is only 1% of the peak current. It is a good assumption that the tail s<s0 will not significantly amplify any radiation. With only 100 more slices the entire bunch can be simulated. The parameters would be now: NSLICE=600, NTAIL=-300 and IOTAIL=1.