Kevin Li: Simulations of the 4 MeV OBLA Test Stand Room: WBGB 020 Selected questions and discussion: AO: GPT cannot handle a velocity difference, but in the cavity you have an explicit velocity difference...? KL: This is only for the space-charge computation, the energy chirp should be correct. TS: Are the RAISTLIN and CARAMON codes stand-alone? KL: Yes. TS: Can they be used by others? KL: In principle yes. FLP: Why did you write new codes instead of just using GPT? KL: My simple codes are much faster! YK: What about non-linear effects during velocity bunching? Are they treated in your codes? KL: No, the purpose of these codes is just the evaluation of the effect of external fields on the emittance. TS: In the 2D diode simulation with MAFIA, what is the transverse boundary condition? KL: It's a perfect conductor put at a very large distance. FLP: The cathode is not flat? KL: No, I simulated the FEA configuration (but the difference is small). YK: What about thermal emittance in your simulations? KL: Not considered - I start with a perfectly parallel beam. JZR: (on p. 27) What is the sensitivty of the peak current to the RF phase? KL: I observe a rather high sensitivity; one or two degrees difference can easily give a factor of two in peak current! YK points out deceleration from 500 keV down to 300 keV in the simulation with the third harmonic. YK (on p. 29) points out strong nonlinearity in longitudinal phase space for the solution with the third harmonic (f+h), probably due to space charge. The simulation does not consider wake fields, but they can probably not be neglected. FLP: Are the 48 MV/m gradient (fundamental mode) realistic? JYR: Yes. As for the third harmonic, it is limited by the surface field on the irises. YK commments that 30 A peak current is more amibitious than actually needed. The bunch compressor can also handle 20 A peak current, which would relax considerably the requiremenents on the gun. JYR: Have you tried to form a longitudinally rectangular bunch after the velocity bunching? KL: The bunch always ends up in a Gaussian shape, probably due to space charge effects. TS: At the MAFIA/GPT interface, you use a space dump of all particles? KL: Yes. In fact, I can do either space or time dump and there is very little difference for a drift. However in the presence of external fields, it is much more consistent to have a boundary in space rather than time when going from one code to the next.