August 11, 1999 All That Fits is News to Print Vol. 12, No. 1

Contents of Vol. 12, No.1

  1. Index Panel 2000
  2. New Feedback Loops in PEP-II
  3. Scheduler for LER Orbit and IP Luminosity Feedbacks
  4. PEP2 Bunch-Toggle and Bucket-Entry Buttons for Fast Feedback
  5. PEP2 Bucket Numbers used by Fast Feedback and Beam Abort
  6. EPICS StripTool
  7. Beam Abort Trigger System
  8. Correlation Plots Modifications in Support of LER
  9. Orbit Fitting from Correlation Plots
  10. PEP PIC Almost Automatic Configuration
  11. Faster BPM buffered acquisitions
  12. Multiknob Enhancements

Page contact and owner at end of this issue.

Index Panel 2000

August 11, 1999

Author: Hamid Shoaee Subsystem: Index Panel User Impact: Hopefully Good
Panel Changes: None Documentation: None Help File: Yes

This issue marks the end of publishing the Index Panel newsletter in paper format. Since its inception in 1987, each issue has had a run of about 200 copies requiring substantial effort to reproduce and distribute the newsletter to the members of the Operations, Controls and Accelerator Departments.

Starting with the next issue both the submittal of articles and the publication of Index Panel will be done entirely via the web. Future Web-based releases of the newsletter will be announced via email messages to the current list of hardcopy recipients.

The URL for the Index Panel home page is: {http://www-group.slac.stanford.edu/cdsoft/indexpanel.html}. Previous issues from the past five years have also been converted to HTML and placed on this page. A link to this page will be added to the Accelerator Department home page.

New Feedback Loops in PEP-II

January 22, 1999

Author: Grossberg, Hendrickson Subsystem: PEPII User Impact: Some
Panel Changes: Few Documentation: No Help File: No

There are now several SLC-style beam-based feedback loops running in the PEP-II rings. These feedbacks are accessible from the feedback panel on the SCP. Feedback loops which control the beam are in the PEP-II group. There are also several compute-only loops which provide diagnostic history information, found in the DIAG PEP-II group.

For transverse orbit control, the feedback loops ``PR02 HER ORBIT'' and ``PR02 LER ORBIT'' control the effective kick at a corrector at the IP. The incoming position and angle changes are fitted out of the calculation. Incoming and outgoing ARC BPMs are used.

When the beams are colliding, optimization feedbacks, ``PR02 IP_X LUMIN'' and ``PR02 IP_Y LUMIN'' keep the beams in collision at the interaction point by dithering the IP position with a closed bump of the LER IP position and maximizing the luminosity monitor signal. Since the optimization feedbacks use correctors in common with the ``PR02 LER ORBIT'' feedback, none of these loops may be on at the same time. Furthermore, these feedbacks must be turned off when users need to tune or knob the feedback correctors. Currently the optimization feedbacks use devices and coefficients in the ``LERIP_X'' and ``LERIP_Y'' multiknobs. The devices and coefficients are stored in feedback database. In order to change the devices, dbedits are required. Users may review or change the coefficients from the ``PR02 KNOB COEFFS'' panel, found off of the feedback magnet panel.

To check which devices are included in a feedback system, the DISPLY VECTOR button on the feedback panel may be used, after selecting the correct group and loop from the panel. To turn a loop in and out of feedback control, users may select the ``Change loop HSTA'' button and enter ``F'' for feedback, or ``C'' for compute. Note, however, that if the scheduler is controlling a particular feedback loop, the user must turn off the scheduler in order for manual control to work. To check which feedback loops are on, from the feedback panel the user may select the PEP-II group and request a SUMMRY DISPLY, or select the loop and look at the CHANGE LOOP HSTA button.

Scheduler for LER Orbit and IP Luminosity Feedbacks

January 26, 1999

Author: Grossberg, Hendrickson Subsystem: PEP-II User Impact: Small
Panel Changes: Few Documentation: No Help File: Coming

An SLC-style feedback scheduler has been implemented for the PR02 LER Orbit feedback and for the IP X and Y luminosity loops. The reason for this scheduler is that they each need to control the same correctors (XCOR,PR02,3027 and YCOR,PR02,3024), so only one of the loops can be on at a time. The functionality of the loops is orthogonal since the luminosity feedbacks use closed bumps which shouldn't affect the range of the BPMs of the orbit feedback.

In order to manage turning on and off the fast feedback loops, a slow feedback scheduler has been added, called PEP-II LUM WATCH. The user may access this either from the slow feedback panels, or from a new panel specifically for the PR02 feedbacks. This new panel is found from the fast feedback panel, to the MAGNET panels, by selecting the PEP2 PR02 PANEL.

On this new panel, one can turn on or off the scheduler by using the GRPALL HSTA STATE button and turning it off (O) or to normal (N). Turning it off takes the fast loops out of scheduler control, and then the user may go to the fast feedback panel and turn on or off one of the loops. Turning the scheduler to NORMAL mode (N) gives the scheduler control of the fast loops if the normal mode is feedback. Ordinally, the NORMAL HSTA for each feedback loop is set to the currently recommended value, so that they can be safely reinstated by setting the GRPALL HSTA to N. We do not recommend directly setting the GRPALL HSTA to FEEDBACK, because some feedbacks in the group may not be intended to be turned on. Currently the loops are set up with a one minute timer, so usually one of the loops will be on. When the fast loop comes within a set tolerance or when a timeout is passed (currently 3 minutes), it may be turned off. The tolerances for each fast loop are accessible from the STATE panel off of the main fast feedback panel, after selecting the loop of interest.

Note that if the user wishes to manually move any of the correctors in either feedback loop, the user should turn off the scheduler using GRPALL HSTA STATE (to O) before assigning or moving a knob. Also it is important that users not manually turn on both the luminosity feedbacks and the LER ORBIT feedback at the same time.

In the future, this scheme may be modified if the luminosity feedbacks are changed to control HER correctors instead of LER.

PEP2 Bunch-Toggle and Bucket-Entry Buttons for Fast Feedback

April 16, 1999

Author: Phyllis Grossberg Subsystem: Fast Feedback User Impact: Small
Panel Changes: One Documentation: No Help File: Yes

Four panel buttons have been added to the SCP for toggling between single bunch and bunch train for the HER and LER feedbacks in the PR micros, and for entering the HER and LER bucket numbers which Feedback (and Beam Abort) should use. These new buttons are at the top of the Fast Feedback PR02 Params panel reachable from the feedback Magnet Index. In addition, because the bucket numbers can be changed automatically from other software, a fifth button has been added for updating the button texts with current database values.

The two HER/LER bunch mode buttons toggle between single bunch and bunch train for the HER or LER fast feedback loops in the PR micros. Button text indicates the currently selected bunch mode as well as the associated BPM definition used by feedback. Toggling one these buttons causes a database write to FBCK BPMD to redefine the HER or LER BPM definition for Feedback, and the re-creation of BPM bufrings in each PR micro containing feedback loops. The BPM bufrings are used for communication between the BPM job and feedback and their creation takes a few seconds.

The two HER/LER bucket-entry buttons allow the user to change the HER or LER bucket number to be monitored by PEP2 feedbacks in the PR micros. (The same bucket number is used by Beam Abort.) Button text indicates the bucket number currently being used as well as the activity responsible for enacting that bucket number. A new value entered via this button is written to BPMD PRMS(3) and the responsible activity is written to BPMD PRMS(4) for all BPMD unit numbers used by the selected ring (currently 14 and 38 for HER and 24 and 39 for LER). Responsible activities are

                                                            0: DIA: diag list execution
                                                            1: BIC: BIC injection
                                                            2: OPS: Operator entry

New bucket numbers also cause the micros' BPM bufrings to be re-created.

The refresh button updates the HER/LER bunch-toggle and bucket-entry buttons with the current database values. Since some of these values are changeable by other software, unless the SCP panel is newly entered, the values displayed on the buttons may be out-of-date. Prior to using any of the four buttons, it is a good idea to refresh them to make sure the latest values are being displayed.

Additional information about these buttons are included in the following article.

PEP2 Bucket Numbers used by Fast Feedback and Beam Abort

April 7, 1999

Author: Grossberg,Hendrickson Subsystem: Fast Feedback User Impact: Large
Panel Changes: One Documentation: No Help File: Yes

Fast Feedback and Ring Abort use three separate functions to decide which ring bucket should be monitored:

  1. A new button on the Fast Feedback PR02 Params panel allows an operator to enter any desired bucket number for HER or LER.
  2. Code behind the 'Execute Diag List' button on the PEP2 Inject Timing Diags panel still automatically sets up the first bucket from the Diag list for Feedback's use if that bucket is different from the bucket Feedback is currently using.
  3. A slow feedback watchdog monitors the EPICS database for database writes made by BIC during injection to indicate the first bucket it will fill. If the watchdog is on after BIC injection, if the slow feedback finds a bucket number in the EPICS database that is different from the bucket Feedback is currently using, then Feedback will be re-initialized to use the BIC bucket number.

These are independent and conflicting actions and, until someone requests a change, the rule is: ``Last one to act gets his way." This means that an operator can use the bucket-entry button to override any selection made by the Diag List code or the slow feedback BIC watchdog. It also means that if someone subsequently executes a diag list, the operator's override might be overridden; and it means that if the slow feedback watchdog is running after BIC injection, it might do some overriding of its own.

Operational procedures must be created to manage these activities so they do not conflict with each other. For example, when BIC and the diag list activity are being used alternately, it would be advantageous if the first bucket to be filled were the same on both their lists. When you want an operator-entered bucket number to prevail, the BIC watchdog could be turned off and diag list execution could be avoided.

To turn the slow feedback BIC watchdog off (or on): From the Fast Feedback main panel, go to the Slow FDBK panel and select SLOW WATCH DOGS; go to the Loop Select panel and select the BUCKWTCH loop; select Loop HSTA State button (bottom right) and respond in the dialog box with O for Off (or S for Sample).

EPICS StripTool

December 9, 1998

Author: R. Chestnut, S. Allison Subsystem: EPICS User Impact: Small
Panel Changes: Few Documentation: No Help File: No

Since the SLC database is now visible to the EPICS subsystem, the EPICS StripTool now adds a flexible tool which can replace the ``realtime" correlation plot solution for monitoring interesting data. With this tool, up to 10 data channels can be connected to a strip chart. The scan rate, appearance, colors, and scales can be modified to meet your requirements, and that configuration saved and restored.

The StripTool has been long used by the PEP-II RF subsystem, and has recently been added to the PEP-II BIC, to provide access to the newly implemented Normalized Luminosity data (PB60:DCCT:LUM3 and PB60:DCCT:LUM4). The Normalized Luminosity canned plot is invoked from the BIC EXPAND button, by pushing the right mouse button and selecting the Luminosity menu item. When these ``canned" plots are started, a specific configuration is used to start the display.

The StripTool can also be accessed from the Special Displays SCP panel (right edge, third row from the bottom). This invocation will invoke a control panel for the StripTool. Channels to be monitored are entered in the upper left, general characteristics are changed on the lower left, and individual channels customized on the right. Configurations can be saved or loaded via the FILE menu. When any valid channels are entered, the strip chart itself appears. To rid yourself of the control panel, push DISMISS.

On the strip chart, there are buttons at the bottom of the chart to go forward/backward, zoom in/out (time x-axis only), and auto-scroll. A pulldown menu appears with the right mouse button key and provides options to bring the control panel back, print, dismiss the strip chart, and quit the StripTool.

For PEP-II RF and BIC EPICS data, to determine the name of a channel, the user can point to the data display on any EPICS DM display and click the middle mouse button key. The channel name is then highlighted and the user can paste it into the ``Plot New Signal" box on the StripTool control panel using the middle mouse button key. For data from the SLC VMS database, the user will have to type in the channel name. Currently, the SLC data is updated at the async rate of the micros so depending on the micro job, the data may be up to a minute old. The following analog channels are currently supported. Note that other digital and discrete channels are also supported but not listed here since it's unlikely they'd be used in a strip chart.

                                Analog Status         - MICR:ASTS:CHANNEL.VAL, .RMS, .REF, .CON

                                 Magnet-Like Data - MICR:PRIM:UNIT:QNTY.VAL, .RMS, .REF, .CON, .MON
                                                                       QNTY = B, V, I, or A (VAL is same as ACT,
                                                                                                            REF is same as DES)

                                 Feedback Element - MICR:LOOPNAME:ELEMNAME.VAL, .RMS, .REF
                                 BPM-Like Data     - MICR:PRIM:UNIT:BPMD_QNTY.VAL, .RMS
                                                                       BPMD = 4-char BPMD NAME, QNTY = X, Y, TMIT, or PHSB
                                                                       Note: BPM acq done with standard option settings.

Beam Abort Trigger System

February 22, 1999

Author: Kenneth Underwood Subsystem: PEP-II User Impact: none
Panel Changes: none Documentation: none Help File: none

The PEP-II Beam Abort Trigger (BAT) system protects the beam line components from damage caused by an errant high power stored beam by extracting the beam within one turn of a detected fault condition. Faults may come from a variety of sources such as a tripped klystron, a Beam Loss Monitor (BLM), or even an explicit operator request.

When the BAT system is armed, a clock is sent from the Beam Abort Trigger Clock (BATC) module through a series of Beam Abort Trigger Processor (BATP) modules and then back to the BATC module which controls the beam abort kicker. The BATP modules are essentially an 8 input AND gate which will interrupt the clock when a fault occurs on any of its unmasked inputs. The BATC module stops sending the clock when it detects that the clock has been interrupted and causes the abort kicker to fire during the next ion clearing gap. The instant that each BATP module detects that the clock has stopped, the current state of its 8 inputs is latched so that the initiating fault can be identified.

Since a sudden loss of any significant amount of current will at the very least perturb the RF systems with reflected power faults, collateral faults would make diagnosing the instigating fault difficult. The hardware is designed such that only those inputs that are faulted at the time the missing clock is detected will be latched. This should only be the originating faulted input although occasionally an unrelated or collateral fault will show up due to the clock propagation time.

When a beam abort occurs for other than an operator initiated request, do not reset the BAT system until you have determined the source of the fault. Subsequent failed reset attempts of the BAT system will reveal all currently faulted BATP inputs which will help resolve any problems associated with rearming the BAT system.

Correlation Plots Modifications in Support of LER

June 25, 1998

Author: Greg White Subsystem: Correlation Plots User Impact: Small
Panel Changes: None Documentation: Yes Help File: No

A few additions have been made to Correlation Plots to help with the LER commissioning and operation. Firstly, the PEP-II tune analyzer for LER is now supported through PHYS variables. Additionally, two new parameters can now be retrieved from the tune analyzers. These are POWER_X and POWER_Y, the peak spectrum power in db.

As before, the format for tune variable specification is ``PHYS ring variable'', where ring may be LER or HER and variable may now be one of NU_X NU_Y NU_MIN NU_MAX NU_DIFF POWER_X POWER_Y.

Also, a new variable named ATIM has been added to Correlation Plots. The value of an ATIM variable is the number of seconds since (the previous) midnight. Variable ATIM is similar to TIME, but it is not stepable.

Orbit Fitting from Correlation Plots

26th May 1998

Author: Greg White Subsystem: Correlation Plots User Impact: Small
Panel Changes: None Documentation: None Help File: None

An interface to the BPM orbit fitting has been added to Correlation Plots. A new PHYS secondary ORBT has been added, and the variables available under it; XPOS, XANG, YPOS, YANG, ENERGY, XKICK, YKICK and FITRMS mirror the fit parameters that would be displayed on a BPM Orbit Fits plot.

All of the set-up for the data acquisition and fitting is done from the BPM panels. In particular the left-most two columns of the BPM DIAGNOSTICS panel govern the arguments of an orbit fit. All of these buttons take defaults though, so just entering the PHYS variables is enough to perform an orbit fit. Note that unlike other PHYS variables, this one does not specify an area of the machine in its name, like PHYS FB31 etc - all of the context information is taken from the values on the buttons in BPM panels.

Please see the Index Panel of 24th Feb 1994 article title ``BPM Orbit Fitting'' for more details.

PEP PIC Almost Automatic Configuration

June 24, 1998

Author: Karey Krauter Subsystem: MPS User Impact: Some
Panel Changes: A couple Documentation: This is it Help File: In Progress

User friendly PIC configuration and parameter entry has been added to the SCP for PEP-II ion chambers eliminating the need for DBEDITs and dialog entries on the MPS device panel. No more manual computations of trip thresholds, wondering whether to enter DAC counts or microamps. No more individual entries for every threshold in every pic anytime the max beam rate changes.

Go to the PEP-II index touch panel, then the PEP injection index touch panel, then either the HE or the LE MPS touch panel. On the HE MPS panel, look under the HEINJ MPSPIC PANEL or the HENIT MATCH MPSPIC PANEL. On the LE MPS panel, look under the LEINJ MPSPIC PANEL or the LESIT MATCH MPSPIC PANEL.

The new PIC configuration capability is on the above four PIC panels, each providing access to the PICs in each of the four PEP PIC dgrps. On each panel, all of the PICs associated with that dgrp are given a pair of buttons: one showing each PIC's trip status and another showing cable status. There is also an ALL* button to allow working on all the PICs in this dgrp at once.



button displays the parameters programmed in all the PICs in this dgrp; the


button allows choosing engineering units in which to display and enter PIC parameters. You may choose microamps, rads per hour, DAC counts, and ADC counts.

Choosing one of the PICs on one of these touch panels, by selecting its button, you are delivered to a panel that focuses on that particular PIC. (The ALL* button delivers you to a slightly different panel, discussed in a minute.) The PIC trip status and cable status are again shown here, along with the PIC CALIB DISPLAY button and the TOGGLE DISPLAY UNITS button. There is also an ENTER PASSWD button which is required before performing any PIC configuration function on this panel (standard choices of operator, eoic, expert, software privilege levels).

The following buttons are available to perform configuration functions on a single PIC. Note all queries and echos are shown in the engineering units selected on the TOGGLE DISPLAY UNITS button.

The following buttons are available to perform configuration functions on all (ALL*) of the PICs in a dgrp at one time.

Faster BPM buffered acquisitions

April 30, 1998

Author: Linda Hendrickson Subsystem: BPM User Impact: Small
Panel Changes: Few Documentation: No Help File: Yes

For some experiments, it is helpful to have BPM buffered acquisition run as fast as possible. An option has been added to allow faster scans at the expense of micro memory. This is only helpful for repeated scans with exactly the same units; the first one is not faster. When doing repeated scans, the user may enable faster acquisitions before the first scan. When finished, users are STRONGLY URGED to disable the faster scans to release micro memory. Otherwise, the wasted micro memory may cause subsequent scans to fail. The new buttons to enable and disable faster buffered acquisition scans are available on the More Buffer Units panel, found off of the BPM buffered acquisition panel.

Multiknob Enhancements

March 30, 1999

Author: Ron Chestnut Subsystem: Multiknobs User Impact: Small
Panel Changes: None Documentation: No Help File: None

The following changes have been recently made to the multiknob facility.

RAMP verb

In tuning the PEP-II ring, many ``bumps" are created and driven by the Multiknob facility. These bump multiknobs are driven both by hand and assigned as correlation plot driving functions. A problem was noted when devices in the multiknob file respond at different speeds. In this case the slow devices lag when taking large steps and cause some degree of beam disruption. To address this problem, a multiknob file can now have a /RAMP specifier:  `` SET /RAMP=(maximum_step_size, wait_time_after_step) ". The maximum step size is in knob units and wait time is in seconds. Often a zero wait time will be appropriate, since the system lag time for updating the devices allows the slow devices to catch up. If zero wait time is desired, an alternative form of the command is: `` SET /RAMP=maximum_step_size ". This feature is implemented as follows:  calculate the number of steps to take (desired_step/max_step) Minimum=1; loop over the number required; if the loop is not on the last step, take step of max_step; otherwise step to the desired maximum; if the loop is not on the last step, wait the specified time.

Multiknob Slider Limits

In response to caters concerning the large default knob range for timing devices (-1.E10 to +1.E10), code has been implemented which scans the devices to be attached via a multiknob and uses reasonable limits for the multiknob. This is visible only on Calfs, which do not have hardware knobs.

Limits for magnet-like devices, with BDES, VDES, or PDES secondaries, are taken from the ..MIN and ..MAX secondaries in the database. Phases are given the range -360 to +360, the KLYS and SBST DRVR secondary runs from 0 to 100, and timing devices are given the limits current_value +/- 800 (about 100 ticks). If you should need to move a timing value further with a Multiknob, deassign and reassign the knob when you are at the limit. The 200 tick range should be plenty though (according to Jim Turner).

Multiknobs and Feedback

The multiknob facility has been modified to check for feedback control of magnet-like devices. This checking is performed before every attempt to move a device. Thus the knob can be assigned, but will complain upon being turned, and NOT move any of the devices on the multiknob, until feedback relinquishes control of every device on the knob. This blocking also affects any RESTORE attempts.

Back to top of this issue

August 11, 1999 Index Panel Vol.12, No.1

Translated from original PlainTeX by index2html.pl.

*Links followed by an asterisk are limited to SLAC clients only.
  Last modified on Wednesday, July 26, 2006 Webmaster