New MAJOR release 06.00.00 Note: There has been a big version jump (major from 2 to 6) to avoid confusion with the versions of the old splitup and pure-DOOCS servers llrfctrlsrf, llrfctrlnrf etc. The version is now bigger then the biggest version of the other servers, so when they get replaced by this server there is no jumping back in version. Changes since 02.05.05: Many changes break the interface to previous versions of this server but make it more compatible to the servers used at XFEL etc. - Changes to the variable structure: - /Controller/FeedForward/scale was renamed to .../flattopRatio and is now multiplied with 2. .../fillScale is removed and now hardcoded to 1. This now matches the XFEL servers etc. - /Controller/FeedBack/MaximumOutput now is a single variable only instead of seperate variables for I and Q. This is an interface change! - Moved many configuration/calibration variables into /Configuration to improve the structure. This affects e.g. all samplig scheme parameters but also other calibration parameters like the bitScaling factor. - Moved /Controller/Calibration/X -> /Controller/X/Calibration, where X is VectorSum or Output - Moved /Controller/FeedForward/X -> /Controller/FeedForward/DAQ/X, where X is I, Q, amplitude or phase - Controller status word is now decoded - Added mimo parameters - Added gain table delay - Added ADC delay - Introduce a config file LLRFBaseConfig.xml to change fundamental parameters of the setup - Added some more missing table paramters - Allow change of the number of samples per pulse (needed e.g. for regae). - Added some information about the sampling scheme to the DAQ - Changed how attenuators are written. Now the status register is used to determine how long one has to wait before writing the next value. Also only the changed value is actually written to the hardware. - Gain computation is now identical to the llrfctrlsrf server (factor of sqrt(2) was missing but it was scaled with the ff scaling factor, which is wrong) - The raw data from the primary DAQ of the controller must be treated as double, since the gain has a 10 bit fixed point conversion - Removed table limits (now done by DeviceAccess which obeys the proper map file settings) and write tables to hardware as double, so fixed point converter is properly used - Use WORD_IIR_COEF_VALID to mark the filter coefficients as invalid while writing them - Added server version as process variables - Output server version to console on startup - Allow disabling the MicroDAQ system and the automation module - Allow configuration of channel names - Preliminary support for multi-board setups - Removed prefix from timer device module, since this can be included in the dmap file and thus gives use more flexibility - Removed support for integral controller - Some bug fixes and code improvements