Field System Topics Ed Himwich, John Gipson, Jonathan Quick, and Brian Corey

1. Field System Topics Ed Himwich, John Gipson, Jonathan Quick, and Brian Corey

2. Current Status • FS 9.7.7 • Some relatively small bugs known • Tape labels • Ground pick-up table in RXG files • S2 Help files not selected properly • Station help files not selected properly • disk2file=;in2net=;disk_record= • Time tag display information errors in data_check, scan_check • …

3. FS 9.8.2 (Summer 2006) • Many minor bugs fixed • Mark IV formatter bit slip detection (1 Ms/s) • Improved GNPLT version by Andrea Orlati • Nutation offset support in VEX parser • Mark5 support for “continuous” tape schedules • RDV geodetic schedules • VLBA correlation of FS recorded Mark5 disks • Field labels designed for disk packs • Automated ftp transfer • fs_check.tcl • Other small items generated by this meeting

4. FS 9.8.3 (Fall 2006) • S2 DAS support in DRUDG • Eliminate use of some mk5=… commands for normal operations • Phase-cal extraction support for Mark IV decoders? • CHEKR monitoring of Mark5 • Update Mark 5 “Remaining Capacity” display while recording

5. Automated disk2file ftp transfer • Two options • Daemon running on Mark5 computer detects new file and manages transfer • Detects problems, can manage multiple files • May be implemented by message sent after disk2file=abort • FS initiates transfer after disk2file=abort • Use a SNAP script to start a process on the Mark5

6. Operating System Issues • FS Linux 6 Distribution • Based on Debian “Sarge” • Clean Install or Upgrade from FS Linux 5 (woody) • 3rd Quarter 2006 (Jon) • Firewalls • Inexpensive, <US$100 • Should be used with older, no longer supported distributions • Good idea for current distribution

7. Longer term development items I • Documentation Update • Improve prediction of disk pack change times • Pointing software clean-up • Eliminate redundancies in pointing configuration information by introducing a source coordinate database file and reorganizing point.prc and ctlpo.ctl (aquir control file). • Documentation clean-up to reflect new procedures and utilities • Improved Tsys • Most items completed • Post processing program to generate AIPS (ANTAB) format TSYS files from Cormac Reynolds • Periodic firing of calibration diode with flagging needed • Convert from fort77/f2c to g77 • Will allow use of source level debugger • Must maintain compatibiltiy with f2c for older distributions

8. Longer term development items II • TNX (“Thanks”) command expansion • This command currently provides a syntax to eliminate the display of errors that the operator knows can’t be fixed and should be ignored • Will be extended to allow different error messages with the same number and two letter code to be distinguished • Will not allow an error that has not occurred to be masked off. • FSVUE • Integrated X client (Tcl/Tk) FS Operator Interface for FS • Second iteration now, uses client-server model • Allows multiple “Operator Input/Log Display” consoles for remote operation • Initially add-on, eventually main interface • Includes log display (scrolling doesn’t jump on new input) and operator input windows • Buttons to allow you to examine logs, help, and run various utilities: monit, logpl, msg, and others.

9. Longer term development items III • Band changes • Band configuration procedures added to set-up by DRUDG. • The DRUDG control file will be expanded to include a table of station defined procedures that can be used to set-up local station equipment for a band. These procedures can also be used manually by the operator as needed. Note that use of the existing SAVE_FILE command can be used in these procedures and INITI to recover the receiver set-up between FS terminations and restarts. • Add a command LOCONFIG. • This will be used after the LO= commands in the IFD set-up procedure. This can be used as trigger to start configuring the Los, if doing so in each individual LO= command might cause problems. By default, LOCONFIG will be a NO-OP, but it can be defined locally and do whatever is necessary for LO configuration in one step. • CALON and CALOFF SNAP variables. • This intended to deal with stations that have different cal control methods for different bands. The idea is that variables will be introduced into SNAP, specifically two: CALON and CALOFF. These can be defined by the band set-up procedures described above and used as $CALON and $CALOFF in procedures when the noise diode needs to be controlled.

10. Additional Future Items I • IF patching automation for Mark IV racks • EVN has hardware design, but not implemented in field yet. We will need one relatively simple SNAP command to support it: • A special version of PATCH and a way to control which version is used. • Mark IV decoder support • This is beyond the phase-cal monitoring mentioned above, mainly a few SNAP commands to control the decoder manually. Most of the effort here is actually divining what is needed and developing documentation

11. Additional Future Items II • Phase-cal control monitoring from VEX schedules • Mark IV decoder support already planned • Mark 5B support expected when device available • AIPS format file generated from post-processing • Support for VLBA digital switch board

12. Issues with making antenna calibration reliable • The biggest single issue is that antcn must be reliable about returning the onsource status • Bad onsource status means that the collected data is bad • Most common issues: • After a new source/offset is commanded, antcn returns the status from before the new command • Onsource test does not require close tracking • Antenna hasn’t settled yet when onsource is first claimed

13. Calibration problems with onsource • These problems can me ameliorated but not solved by: • Driving farther than is necessary off-source (won’t help problems with onsource points) • Increasing integration time • Correct solution is to make antcn reliable, this can usually be done, by coding antcn to “outsmart” antenna control computer • Tracking tolerance must be tight enough, typically 0.1*FWHM, first LO frequencies from shared memory can be used to calculate beamwidth.

14. Simple workarounds • If control computer returns tracking information from old source/offset, add logic to antcn to remember when the source/offset was changed and the not allow an “onsource” state to be returned until enough time has passed to be sure information is accurate, usually one or more iterations of tracking loop. • If antenna overshoots target or has trouble settling, require antenna to report good tracking for two or more times consecutively (or for some period of time) before reporting “onsource”. • Other methods are possible depending on the details of the antenna

15. Issues with making antenna calibration reliable + • Second biggest issue is that cal noise level must be stable by the end of the calon and caloff procedures • The system must be set-up this way for caltsys to work • Easy solution is to add extra delay, for example, !+2s at end of calon and caloff procedure • Only add extra delay if there is a real problem with cal stabilization, otherwise everything is slowed down unnecessarily

16. Diagnostic Software • Developed to automate station check-out for CONT05, but useful for other stations • Most tests require a Mark IV decoder • Small amount of customization for each station • General scheme for each test: • Station set-ups equipment in mode to test • Run test SNAP procedure • Analyze/plot using Perl/Pgperl scripts • Can easily be adapted for Mark 5B • Expected to greatly improved sensitivity

17. Diagnostics Tests • Baseband Bandpass shapes • Baseband L/USB image rejection • Phase-cal power • Phase-cal spurious signals • Sampler statistics • Stability tests • Orientation-dependent • Diurnal • Short-term • Need to add Mark5 tests

18. Test Procedures • Example test procedure • proc=rd0501gc • setupsx Set desired mode • proc=systests • log=test01.log Open a new log • pcalamp Run test • pcalamp test01.log Outside FS, plot data • Repeat last three steps for updates and other tests • Tests