From: SLCW47::LUCHINI 16-NOV-1998 13:54:18.48 To: TMH CC: luchini Subj: bunch injection efficiency Hi Tom, I'm gathering information for Ron about the PEP-II Bunch Injection Efficiency requirements. This is what I have so far from what Ron has described. I would appreciated any additional information about what is needed and what would be helpful to the physist and operators. The functional requirements thus far is to obtain toroid and pid data during PEP-II injection on a pulse to pulse basis over a specified period of time. A minimum of least 10% of the bunches injected is required to determine Bunch Injection Efficiency. This data results of the bunch efficiency will be displayed graphically on a EPICS display. The method that will be used to determine Bunch Injection Efficiency is not discussed herin. Possible Solution: -------------------- 1. The simplest and less expensive solution is to utilize as much existing software and hardware as possible. By taking this approach, a callable interface to the SLC BPM Buffered Acquisition Utility is available and provides the desired functionality to acquire Toroid and pid data. The use of the BPM Buffered Acquisition will require a standalone routine or scp utility to be written on the Alpha, which calls bufacq_takedata() interface to sample the data. This routine would then ship this data to the BIC ioc (bic06pep00) via channel access, depositing it into an EPICS waveform channel(s). In addition, the MPG job on micro MP01 will need to be modified such that the PEP-II Bunch Injection Control data shared between MP01 and the BIC ioc via bit3 shared memory (DPR), includes pid data. The change to the existing data structure shared between the two micros will need tobe decided between Tony G. and Ron C. Software on the BIC ioc will also need tobe written to analyze the Toroid data. Depending upon the complexity of the analysis, an EPICS device support routine or sequence would be written proform this task. Currently, it has not been specified how the Bunch Injection Efficiency will be engaged. For example, will the operator push a button from an EPICS scp, SLC scp or should this happen automatially during injection? In addition, should there be an "Abort" button which stops Injection Efficiency process once it has begun? Will the user want to enter the duration of Bunch Injection Efficiency evaluation? Will data be taken continuously or will the user only need to grab data for a certain period and then wait before grabing data again? If so, should these parameters that the user change from the SCP before starting injection efficiency? Can these parameters change during injection efficiency evaluation? Should the bunch injection efficiency data be history buffered? Additional Information: --------------------------- The maximum lenght of time that data can be acquired from bpm acquisition is 22 seconds. The micro will ignore a request for more time. The maximum length of data is 64K that can be acquired from a call to bufacq_takedata(). The data acquired for a Toroid is: scaled data, TMIT and STAT. Concerns that need tobe addressed: ---------------------------------- 1. How often and over how long a period of time can this type of data be acquired from a micro? (From Linda): 2. BPM bandwidth is of concern, and the implications of BPM applications from automated processes on the functionality of the rest of the BPM system. 3. There's an issue with injection (which Tony understands much better than I do) and BPM/TORO acquisition where we need an inclusion bit for the software to know which pulses are really being injected. Except for when the beam is really being injected, the use of this injection bit really ties up the rest of the BPM system. 4. What amount of data you are going to grab and which micros it comes from? 5. Which micro and toroid is being monitored? TORO:PI01,9404 TORO:PI11,6404 6. Will fast feedback in that micro will be able to get any data? New Requirements as of 11/13/98: ---------------------------------- The pulse to pulse data is to be taken throught the entire injection period, up to 5 minutes (1Hz or 60Hz). The operator will want to view the Injection Efficiency data continuously during injection so that they can try to improve the efficiency.