ChangeLog 2002

Jan 2002

 - Repeat from December entry (And it still applies)

 * Shifted focus to new guider (smokey), so AO work temporary halted.
   Doug is still putting together the AO optics. Due to ordering and
   machining factors, I don't see this being completed until Jan or Feb.
   (although Doug has estimated  completion date in the staff meeting).
   The utility box components are being ordered and assembled by Greg in manao.
   I also don't see this happening until Jan - And when the utility box arrives,
   I still expect delays getting the final hardware assembled due to cabling
   and the detailed worked (which to me isn't being planned out).


 - work on device communication for:
    ao_computer -> digione -> animatics -> FCS linear slide
 - Greg & Peter delivers the AO WFS utility box. Tony's
   feedback & questions:
    + configured digiOne using hostname ao_mfocus_digione.
      Able to talk to fcs animatics using ao_mfocus_digione, but not the mfocus animatics.
      Noticed NEAT limit are feed into animatics encoder input (not limit inputs).
      Also need to put memory module on motor.
    + Confused on shutter control with EtherTrak. Thought
      the plan was to control WFS shutter from AO electronic/RTLinux.
      Using etherTrak will move shutter control to XUI computer & apps.
    + What is the plan for the spool mechanism?
       Item: Lamp Power Supply, Animatic x 2 (POM, CFS slide), digi_something?
    + Look at sixnet stuff. The unit support network protocols:
        modbus TCP: A standard industry protocol using TCP/IP.
        Universial Protocol: A sixnet protocol using UDP/IP.
      UDP is a connection-less, more efficient protocol. I will try to
      do a Unix implementation of this. They have a DOS/RS-232 sample
      on their web site. This code is poorly written (protocol, serial,
      and DOS all interconnected). Will need to develop new code (almost)
      from scratch.

 - Make Progress on sixnet Universial Protocol Unix/UDP code.
   UP can only do simple commands: Read & write analog and Digitial values.
   All configuration and setup must be done using windows/IO toolkit.
   - For steering mirror, as AO goes 0->pos, AI goes 0->neg. Plus
        there is an offset than changes with voltage:

          Est V.    AO    AI[0]    AI1]   offset 0/1
          ------  ------  ------  ------  ------------
          -5.0    -16383  16244    16325    120/58
          -2.5     -8191   8048     8049    143/122
           0.0         0   -167     -174    167/174
           2.5      8191  -8342     8383    172/172
           5.0     16383 -16577   -16577    194/255

 - Finished a simple Unix UDP implementation of the Universial Protocol.
   Have a test_sixnet that can read, write analog & digitial values.
   Ran under solaris & linux.
 - Wrote low level application to control position of steering Mirror
   and Filter wheel. Commands are:
      smirror_init, smirror_pos, filter_init, filter_pos.
 - Tested membrane focus motor. Found problem with mechanism travel:
   Neat stage mount will bang into steering mirror hardware before
   reaching the positive limit.
   Found -limit, and make these measurements using step counts as units.
   ( assume 0 to at -limit).

            0 = minus limit
       100000 = neat stage is centered on mount.
       174224 = point where the membrane focus neat stage bangs
                into the steering mirror hardware.
       200000 = Where we estimate the + limit is located.

       200000 steps = ~1.8 inches (eye ball guess)

 - Wrote XUI code to control motor, and display status for:
      filter, pom, smirror.

Feb 2002


   - Wrote code for mfocus.
   - took calibration data on smirror binary to volts & got functions for v=f(binary)
     Results are here: sixnet_volt_vs_binary.txt

   - Put MFocus sixnet AO & AI calibaration function into the AO software.
   - Received the IPC (Power Controller) for the FCS laser. Tested on unit
     with linux and the port server -> It functions well. Gave the one unit
     to Greg for AO Relay Utility box.
   - Asked Greg to make a Animatic cable so we can setup the FCS and POM
     in the AO lab using spare while waiting for the utility box.

   - Spend 1 man-day on TCS justification memo for Alan.
   - Setup AO Realy Utility box clone using spares.
       + recieved animatics cable from greg. Now both POM and FCS motor
         are online. Increased the speed of POM x 2 (40000).
       + Added FCS Laser P/W control to ic and xui. IPC works but
         not connected to laser power supply.

   - IPC to laser power is connected (thanks to cable by greg).
   - Assisted in getting system running for the
     1st_Time_We_Run_AO_in_the_Hilo_Lab_Day (2/25)
     There are some jpeg of the 1st 38 imats...02.02/1st_imat
   - Lots of misc fixes:
      + aoui and aoic's /home/ao & /home/aort are served by aoui computer.
      + imat/, tt_imat/, capture/, public_html/ are map to duke:/home/ao
      + Checked TT Actuator diagram - seems OK now.
      + Made configuring View->diagram easiler (however suggest using macros to
        configure). Increased the number of diagram from 6 to 9.
      + Made the data columns View->text (was View->Sensors) user selectable.
        Added widgets to allow easy configuration (but I say use the macros).
      + Added MFocus.Init button to Mfocus dialog.
      + Re-did the smirror dialog box - still some operational problems,
        but it's should be better that the old one.
      + Added 'AMPS' current limit to animatics init procedures.
   - Installed new version 0202 as default.

Mar 2002

   - Worked on tcs1_status program (1.5 days) and attended tech group meeting (0.5 days)
   - supported lab testing.
   - Peter took spare IC computer to manoa.
   - Started modifications to DV to display AO data.
      + Allow ImageRange type to be float (was long).
      + New DypType - AO figure.
         + Data display.
         + Parameters.
         + Mouse feedback.
         + Printing.

   - Completed DV modifications.
   - AO RT/IC/XUI changes:
      + fixed bug in configure diagram dialog box.
      + IMAT restores old membrane stoke when done. (It was setting it to 0 when done).
      + Including membrane freq, and phase in imat FITS header.
   - recompiled new version 0202 on 3/15.
   - Tried to download new GUI library (GTK2.0), but could compile on messier
     so I gave up.
   - setup windows computer in AO Lab to be duel screen XTerminal for AO.

Apr 2002
   - new GTK 2.0.1 came out - got it to compile on messier.
   - Slow for AO due to lab testing.
   - Did some spex, DV, and xgfit work.

   - Re-organized the directory structure. /home/ao on all computer sun, ui, ic
     is the same physical disk. Served by duke.
   - Installed IDL on UI computer.
   - Other that being the file server & web server the sun isn't needed for AO.

May 2002
   Changes to dev version:
      + PR & TT IMAT:
        Changed integration.cycles to coadds.
        Added cycles which is a sweep thought the actuators.
      + PR & TT Feedback loop: added pr.sign & tt.sign to set sign
         for delta_drive = sign * signal * inverse.

   + Install new graphics card on the aoui computer and
     configured for dual-head X Windows. Don't need
     winPC/Exceed & Sun for dual-head.
   + Made new color map on DV for ao. Called
   + Using nis for host table on ui and ic computers.
   + Move widget to handle Imat data: Imat2DV & read Pr.Inverse.File.
   + Move Pr.Inverse file from IC to XUI: workaround for IC NFS update problem.
   + Made text2fits, fits2text part of the dv package.

   1st Engineeing run May 21-24:
      + See Doug's plan toomey_email_020508.txt.

   Sucessfully first run. You can see the data in /home/ao/data/imat/020524.

   Stuff for the next run:
      - AOIC brought to hilo for testing super capture.
      - fixed super.capture using SIM=1.

June 2002
 1- 7
   - Tested super.capture on AOIC with SIM=1. Works.
   - Tested taking and converting tt_imats on AOIC with SIM=1.
     Needed many fixes. Updated "Tony's Notes" documentations.

   2nd Engineering Run.
     + fixed super.capture mode.

July 2002

   + AO system was returned to manoa for 'counter board termination' problems.
     Came backed a bit improved, but now fixed.
   + engineering run (9-12): - used AO software basicly unchanged since last run
     Found and Fixed TT imat bug - TT now works better.

   + Doug email AO to-do list for observing:
      + See Doug's plan 020712_toomey_email.txt.

   + changes in new version 0207:
      + setup hilo development system on wobbegong.
      + make imat WAIT_COUNT a variable.
      + remove /aux partition from AOIC. (/usr/local -> /aux/local)) -
      + use libir1
      + small pr tt gain controls improvement
      + smirror support arcseconds coordinates
      + redesigned smirror dialog box
      + new data item: primary.out.mean
      + default file to load primary.flat at start up.
      + 1/2 way done with APD power control.

   + engineering run (22-24):

Aug/Sept 2002
   + Updated AOIC to RedHat 7.3/RTL3.1. - Works but, the stop_ao script
     hangs the computer. RedHat is questionable distribution, or is kgcc
   + installed hexed: replaced TT's ttconx/hexe with multiplex hexed.
   + Add IDL code to produce a primary flat input files from capture data.
   + converted input for inverse, tt_inverse, primary_flat from text to FITS.
   + mount /benchmark directory on ic and xui.
   + took out shutter code in LP module.

   Used new code for sept engineering. See Doug's email
   020918_toomey_email.txt on the sept 16/17th engineering.

Oct 2002

   + worked on setting up the IFA AO computer 'ehu'. A clone of the aoic
     (RedHat 7.3/RTL3.1).
   + New version for 10/28/2002 engineering incorporates many changes:
      + wrote Focus Offloading code.
      + wrote Tilt Offloading code.
      + Tilt has 2 methods:
           ttdm - old method - leading DM corrects to TT.
           ttsa - Tip-Tilt using Sensor Array information.
           Use selects the method using: tt.servo { off | dm | sa }
      + tt-imat & loop using 2 actuators.
         - removed tt-normalize from TT correction in lp.
         - have doug move TT[2] to TT[1].
      + save imat don't save .txt, .info.txt files. - only FITS.
      + Freq, Phase, Stroke commands added ('membrane' command broken down).
      + Display more Driver performance data: isa_usec.
      + Installed all updates for RH7.3 as of 02.09.26.
           Installed on aoic, ehu, wobbegong.
      + Installed the  kgcc for compiler on aoic, ehu, and wobbegong.
           recompiled kernel: /usr/src/linux
           recompiled rtlinux: /usr/src/rtlinux
        kgcc works for wobbegong.
      + config dialog for Opr Diagrams.
      + Add H85 stuff to the  code.
         #define AO_H85          0
         #define AO_IRTF36       1
      + TT.offload.cmd { error | offset } make the
        xui send TCS guiding or error correction.

   + during the 10/28 engineering run, we tried out 2 ideals on
     taking out tiptilt information from the primary servo loop.

     1. Modify the signal[s] (with tt information removed) before
        caculating delta drive. details are:

       Tried just once, but the loop was unstable. The primary
       and TT actuators slow walk off in different directory.

     2. Try to take out tt information from the primary_out[].
        Details are: 0210_rm_tt_2.txt.

       This method was stable. Took some data to see measure performance.
       It is unclear weather these methods improved the performance of the

    Two versions of AO develop from the 0210 version. It seem we
    will continue development using the method #2 version.

    All other major changes for the 0210 version worked.
    ie: we got the focus offload, and tilt offload to tcs working.

    This major improve overall seem to be the increased bandwidth
    due to the DM voltage fix (voltage POD were limit to about
    +/- 135 voltage - fixed to give full range +/- almost 400 volts)

    Installed an ran with the new tcs deamon.

Nov 2002

new tcsd is installed on stefan.
   + need logging function on tcsd.
   + also log any buffer clears.

Changes for v0211:
	+ fixed up header and make files so compile
	  options, such as AO_IRTF36, AO_H85, and LP_SIM,
	  are located in
	+ Also ~/dev does not have to be the default
	  directory before compile. Difference source copies
	  can be compiled independently.
	+ Removed pr_out_limit_min, pr_out_limit_max.
	+ added /2 for to all imat collection procedures.
	+ focus/tt_offload average over 5 secs (was 1 second)
	  And the XUI off_load period is a parameter.
	+ Updated xui setup display.
	+ Now the IC will start the sock_server process. The User
	  doesn't need to do this.
	+ Added the aoio command to command the IC over the network.
	+ Made {pr|ttsa|ttdm} {imat|inverse} {read|write} dialog boxes.
	+ Made a dialog box.

Dec 2002

Changes for v0212: (used on 12/2 to 12/4 engineering run).

	+ Need to restructure AO UI and IC processes and communications.
	+ added log_message() facilities.
	+ remove capture. Rename super capture to capture.
	+ need new method for super capture save.
	+ swapped smirror X & Y on sixnet so that X,Y matches ra,dec.
	  Got new numbers for smirror translation, rotation, and translation.

Mark's input on changing the PRTT algorithms: 0212_IRTF_AO_tiptilt.txt.

12/6 - Added apt programs to wobbegong, aoic, and ehu.
       And applied all RedHat Software updates.

12/11 - Started new source version: v0301.

12/30 - Spent 1 day in manoa tring the new NI DIO board. New model is the 
	 NI PCI-6534. Was said to be software compatiable with the PCI-DIO 32HS.
	 Tried switching boards in the aoicm, but was not able to output any command
	 via the IO interface.