PHENIX HV Workshop January 11, 2002 Summary

Draft 1.0 January 14, 2002

John Haggerty, BNL

  1. Upgrades to the present high voltage control hardware and software should be done in such a way as to not prevent operation of high voltage during the coming shutdown. Specifically, it should remain possible to operate the present EPICS system, and operation of the high voltage for detector development and testing shall have priority over high voltage software or hardware development.
  2. It seems unlikely that the present EPICS driver can be repaired in any reasonable amount of time, given that significant effort has already been devoted to the problem (which has improved the reliability significantly, it should be noted, due mainly to Ed Desmond's efforts). If we were to retain the EPICS framework, it seems clear that one would develop a new driver from first principles. However, the expense of the present hardware, and the fact that it is reaching the end of its lifetime, suggest that new work, beyond repairs necessary to make the above possible, should be devoted to a new server and control and monitoring framework.
  3. We should determine whether the mainframes can reliably communicate at 38.4 kbaud, and how serial communication with a mainframe interacts with communication via ARCNET.
  4. Any new server should cleanly separate the communication technology adopted (serial, ARCNET, Ethernet) from higher level control and monitoring operation.
  5. Most of the wire chamber detectors emphasized the importance of rapid updates of measured voltage and currents for channels that are being displayed or manipulated. Updates once every 1-2 seconds for the number of channels that can be assimilated by the human intellect are desirable, during periods of ramping or resetting trips, for example.
  6. Most of the wire chamber detectors would like an easier way to accomplish "stepwise ramping," in which voltages are raised to a maximum in a series of steps, perhaps with pauses at the end of each sub-ramp. This is so that the current trip circuit in the supplies has the opportunity to trip (the modules' current trip is not active during ramping).
  7. Trips of most of the LeCroy supplies are quite easy to reset, but the 1469's are somewhat more complicated due to the eight channels ganged to one "bulk supply." A new server should deal directly with subtleties of this sort.
  8. The only channels in any system that are tightly coupled to other channels (i.e., where a set of channels should be manipulated instead of individual channels) is in the Drift Chamber. In this case, all channels in a set are in the same 1469 module. It would be prohibitively expensive to replace these 1469's with 1471's, and the space is not really available.
  9. The SAVE and RESTORE functions of the mainframe may be useful for implementing "stepwise ramping" and/or quickly changing configurations from injection to operating, for example.
  10. A new server could have significant functionality not provided by the high voltage hardware, or simulate more desirable behavior, like resetting trips of 1469's.
  11. It is desirable to be able to monitor all the high voltage in the experiment in a single place. This includes the current monitor system developed for the PbSc calorimeter high voltage, the PbGl calorimeter high voltage, and the TOF high voltage, which are all provided by systems which differ significantly from the LeCroy 1450 system although not all systems share this enthusiasm: "The PbGl viewpoint is that we're very happy with our "stealth" system and would not change it unless requested (at gunpoint maybe!)."
  12. The case of TOF high voltage control and monitoring is particularly thorny, since the TOF group developed, at the demand of PHENIX management (OK, I'll take responsibility for it) an EPICS driver for the CAEN system (which works much more reliably than the EPICS drivers we use for the LeCroy system). If it becomes desirable to replace that software as well, to bring it into a single, uniform system for PHENIX high voltage control and monitoring, the ONCS group will develop the software in consultation with the TOF group.
  13. It is essential that any future control and monitoring framework have functional and reliable: