Controls Meeting 20-Oct-2011
- can order stuff again
- new EE Giles Brown
- Minutes from previous meeting 22-Sep-2011
- order status - Mark, Hovanes
- 435NBX and Structured Text - Elliott
- Relation between conditions database and archiver
- Solenoid_Controls_Redesign - Josh, Elliott
- rework connections to coil taps
- talk to controls engineers at JLab?
- need specs for fast daq, slow daq and QD
- order relays and test SOE scheme
- shunt resistor value?
- rework connections to coil taps
- Projects for Nerses - all
- test alarm system, archiver, backup/restore?
- develop EPICS device drivers?
- Controls test stand update - Hovanes
- PLC programming, Viewpoint, licenses, etc. - Josh
- Test Lab status - Mark, Scot
- Networking - Hovanes, Elliott
- order status
- ELog? - all
- /group/halld-online available to halldonline group
- for scripts, kernels, programs, etc, but not for voluminous data
- PyIRMIS - Elliott, Hovanes
- Mantis task tracker
9:30 AM Thurs 20-Oct-2011 CC L207
Giles Brown, the new JLab Physics Department power engineer, joined us. He will be working mainly on Hall D projects for about a year. Amongst many other projects, he will help us to run UPS power to the hall.
Thurs 3-Nov 9:30am CC L207
Present: Elliott W, Mark S, Cody D, Giles B, Nerses G, Josh B, Scot S.
UPS power to the hall
Giles will work on this, and a discussion ensued on what devices need power. We will follow Mark's guideline concerning potential recovery time if power is lost to determine which devices need UPS power. Cryo systems are obvious candidates, as well as places where loss of vacuum is a concern (the UPS could close valves, but likely could not run pumps for any length of time).
Budgets and equipment
We can order stuff again. Just about everyone has something they need to order.
Relation between controls archive and offline
Elliott described how the offline will need to access controls data long after it was recorded. Nerses mentioned that controls data can be stored in the data stream. Elliott noted that after the fact you always need some data you didn't think would be needed, so that the offline will always need to access the EPICS archive. A project for the next few months is to choose an EPICS archiver, and ease of programmatic access is one concern.
Solenoid controls redesign
Elliott showed the long list of items being considered by the redesign subgroup. Some items were discussed, but there were too many to go over all of them. He noted that we must address the concerns from the Director's review last year. Josh pointed out that many had to do with avoiding fast dumps, but our experience in the Test Lab showed that fast dumps are no problem (that's all we had!). We also briefly discussed new schemes for connecting to the coil taps...Josh and Elliott will generate requirements for the fast daq, slow daq and quench detector. We also realized that all taps in the high-current circuit must be considered, including the shunt. Mark noted that we probably cannot increase the shunt resistance, but that other technologies may work...he will look into these and report back later.
Projects for Nerses
Nerses is currerntly working on an EPICS driver for the JLab discriminator. When this is done he will work on other drivers and on evaluation of high-level EPICS utilities. Elliott noted that the choice of archiver affects how we implement a conditions database.
PLC and A-B software
Josh reported that the license server is running and that so far no other group has parked their licenses there. He has not tried Viewpoint yet.
Everything is out of the Test Lab except the computers and related equipment. Tom will move the remaining stuff out soon.
Elliott and Hovanes will order the networking equipment soon, now that the budgets are unfrozen. It is not clear what will go into the trailer, probably, a fiber run on poles from the counting house to the trailer.
User ratings of the Accelerator elog were: one ok but not great, two rasberries, and one double-rasberry. I.e. no one at the meeting was particularly happy with it. Complaints included difficulty searching, no way to create drafts and later modify them, and no way to extend the fields to include e.g. run number. One can enhance this elog via extensive scripting, with the scripts making the actual entries, but this is akin to writing an elog almost from scratch.
Nerses suggested looking at the Hall B elog. Elliott noted that he designed that elog (gui written by Greg Riccardi), and that he designed a follow-on version that was never implemented. Elliott further noted that there are almost a dozen elog-like applications needed in the Hall D online system. Also, he is about to test the SNS elog, and a company sells an elog development/deployment kit that looks quite interesting (based on Python and MySql/Postgres).
Elliott asked that we begin moving controls related stuff to this new area on the M drive. It is for programs, documents, etc, not for data. If we need a large area to store data Elliott will get one.
We have begun using the Mantis sytem for task tracking but need to expand our use. Elliott will begin putting more tasks into Mantis, and everyone should add notes and comments to the tasks as needed.