Apr 2001

   John Sender quits..Tony is temporary assigned to AO (Oh No!).
   Meet with John Sender / Kevin Jim to learn about AO - Here are some notes:meeting.txt.

   Before leaving, John was able to statup the LP, CP and GUI.

   Studing AO code.
   When to manoa and brought the 2nd AO PC to hilo
   Installed slackware + RTlinux on Aohilo. Slackware has problems. Need RedHat 6.2?
   Started new version of driver and dio_test as this is needed to test hardware.

MAY 2001

 1-4:
   - Used modfied dio_test program. Able to read from data from MF board in manoa lab.
 7-11 & 12(sat):
   - Prepared graphs for AO Design review.
   - Installed RH 6.2 and RTlinux and DIO32 board on Aohilo pc. - Works OK.
   - Able to recompile driver and dio_test program.
   - Started structure for AO LP process.
   - Backup to basking.
   - copied the aopcm:/home/ao account to aohilo:/home/zao.
     Fixed things up to it runs on aohilo.
14-18:
   - worked on way to get commmnds into lp process. 1st try using
     shmeme failed because sem_() function are not supported across
     linux & RTL on mbuff memory. So will try using fifos.
   - Support lab work - try to read in sensor data error free. - no working so good.
21-25:
   - implemented ao commands to lp using fifos. Got first few commands written.
   - started on IC code.
   - convert xshow to be regular xui. Got macros working.
   - Meet Wolfgang Gaessle (wgaessle2naoj.org). Rm 225. He is working on subrau AO system.
     Offer to shared some source code from hist system.
28- 1:
   - worked on feedback state on new ao program.
   - peter testing writes to HVAMP in dio_test.
   - limited work week: hoiday, and 2 day spend on system admin duties.

June 2001

04-08:
   - continued work on state_imat in the lp command.
   - Created a copy of the aohilo disk to install in aopcm.
   - create the lp command dictionary web page.

12-15: (4 day work week)

   - Wed,Thur,Fri - work at ifa manoa.
   - install RAM and new harddisk in aopcm (manoa system). This will replace
     the JS configured RTlinux with tony's installation.
   - Got the dio32 driver to work like the old system.
   - Tried to improve read speed, but everything we write just make
     it worst. Estimated time for LP loop (microseconds):
      75 (read) + 50 (calcuations) + 75 (writes) = 200.
   - Starting Adding data display to XUI.

18-22:
   - Taking Vaction Tues, Fri, Mon, Tues (Jun: 19, 22, 25, 26).
   - Finished 1st cut at XUI's sensor and actuator display diagram.
   - Started adding drivers hooks to AO software: Needs 1 or 2 more
     days with aopcm to finish and test. Then we can read sensor, display data,
     write actuators using the AO software.
   - Started looking for code to do the singular value decomposition. This is needed
     to transform the interaction matrix into the inverse matrix. I wrote this
     IDL example to illustrate the problem and provide some test data for the C code.

        This is the idl program: svdex.pro.
        This is program's output: svdex.txt.

25-29:
   - Vacation on Mon/Tues
   - Improved svdex.pro (now imat2inv). Program reads in imat.txt and produces
     the U, V, W, and inverse matrix as text files.
   - started developing C matrix routines.
   - The XUI/IC/LP runs on aopcm, reading and displaying counter board values.
     Problems exist:
       - PC clock doesn't work when dio32 is enabled. Known problem on National Instruments
         web site. it claims dio32 interrupts denial the CPU from updating the system clock.
         It suggest using DMA, not interrupt transfers. System clock not updating causes
         serious problem with software running in the 'normal' linux.
   - Need to resolve the dio32 problem. (2 steps backwards):
      - try the 6534?
      - Re do the dio32 driver.
      - Another solution.

July 2001

02-06:
   - short week due to meetings (tues) and holiday (wed).
   - Found C code to do the SVD from slalib. Wrote some basic matrix routines.
     Duplicated IDL results in C, but stop as I'm unsure were to use these functions.
   - AO group does not wish address the dio32/linux timer problem . Will defer the situation
     until ????. This will compromise the software design/performance.
   - Continued with the LP imat state code  used to collect the imat data.

09-13:
   - Able to collect Imat data and save to a file.
   - aohilo was hacked into. Required to remove OS and restore
     from previous backup ( using current aopcm image).
   - taking vacation days on 13/14.
   - Got contact for malcolm: Cell # (808) 783-4879.

16-20:
   - spend 2 day on spex documentation - done.
   - aohilo OS reinstalled. Additional security added by raja.
   - I strongly recommend we replace the DIO32 (the IO between the PC & MFB )
     due to the PCI/system clock problems.  AO Mangement doesn't want to fixed it now
     and says to just live with deficiency.  I disagee as it will handicap
     the software's ability and efficency. Plus ignoring the problem doesn't
     make it go away. Software issues:

        - fundanmentally changes the design of the software's communication structure
          And eliminating the ONE BIG of the advantages of using RTL.
        - Problem of how to handle non-RTL OS periodic stuff still unsolved.
          And I'm not sure I can completely program around this problem.
        - 2nd system required to run AO.
        - Work arounds will handicap the resulting software:
        - need to decouple XUI from shared memory; develop extensive means of
          transfering RTL shared memory data over network; XUI commands need
          a network layer included; slow data update rates.
        - upto 2 weeks of extra software work and an additional layers of
          software complexity just to get around hardware problems.

   - Dont_fix_the_dio32 work arounds: 35% completed.

23-27:
   - Split computer so there is an AOIC and AOXU computer. Split programs
     so there is a suite of both IC and XUI applications.
      - Implement network based commands/replay between IC & XUI.
      - Coping IC shared memory to XI at ~ 5Hz.
      - Require network disk sharing on between IC/UI.
   - Misc improvment to XUI.
   - Purchasing better video card for AOUIM.
   - Updated Documentation of AO IC commands.

August 2001

July 30 - Aug 03:

   - installed version 01.07 on 7/30.
       - added display of phase data on 8/1 (version 01.07a).
   - improved AO Figures (Data diagrams).
       - more colors.
       - no autoscale when data goes out of range & no floating scale.
       - can selection more data sources (ac, dc, dc.mean, signal, signal.mean, delta.drive,
         primary.out, phaseA, phaseB).
       - Should work on 24/16 bit true color, and 8 bit pseudo color.
   - Wrote up tt_loop.doc - simple description of the tt loop POA.
   - Setup AOUIH computer to enable development of motor code.

06 - 10:
   - tested IMAT data capture routines. Found and fixed divide by cycles bug
     in zyoptics algorthm.
   - added capture feature to AO software.
   - installed and tested version 01.08 on 8/9

13 - 16:
   - Wrote some IDL routines to graph the capture data. These
     are available in the AO user's account. Documented on the Web.
   - Converted 1st IMAT to inverse data. Simplified to do this and
     added FITS output along with the text output. Document on
     web site in software branch.
     The matrix caculation for the imat seems correct. However, don't
     know how do judge how well the data is.
   + AO Meeting.

20 - 24:
   - did some work compairing Mark Chun's Imat Data/procedures vs. ours.
     Good New: reduction procedures checked out OK; Bad News: Our Imat Data
     isn't good.
   - Started on version 01.09:
       + added commands to exercise TT actuators.
       + Added figure graphics for tt_out and tt_delta_drive to XUI.
       + re-structured code to allow adding TT imat stuff.
   - Low progress as I spend 2.8 days on spex related work.

27 - 31:
   - Installed version 01.09 as /dev.
   - Added with TT Imat data collection / TT_inverse_loop code. (But not tested).
   - Added Data2DV feature to XUI - can transmit most data items to DV as fits.
   - Implemented histogram style graphics in XUI. See fig.type command.

September 2001

03 - 07:
   - Added control for TT actuators.
   - Improved figure for DM and Actuators.
   - Worked on X color map problem. It is better understood, but not
     completely fixed. Mixing server/clients betweens Sun/PCs is
     still a problem.

10 - 11:
   - Version 01.09 was able to perform feedback mode on primary actuators.
     Result unclear but electronic seem to check out. Driver/RT loop consumes
     about 160 usec in 500 usec period (32%), or 16% overall.
   - new figure Diagrams in XUI based on drawing given by Dan.
   - Able to reading inverse data into lp (RT loop process) module.
   - Made Config dialog box in XUI.
   - Gave Manoa group graphics card for AOUIM - Needed as current graphic
     display is bad.
   - Need to change feedback code to accumulate correction rather that
     apply gain on desired correction.

17 - 21:
   - Data2Dv improvements: Able to transmit AO data to DV.
   - Improved text2fits, and wrote fit2text (Data2Dv files to text).
   - Worked with DigiPort server and animatic motor:
       + Got computer to communication to motor using digiport.
       + We only have 1 cable (3 animatic and power supply )
   - Started developing code for telnet based motor control.

24 - 28:
   - Motor control and feedback is integrated in XUI. Only FCS/POM is done.
   - Low level code for FIB (Fiber Calibration Source) is done (initial version).
	- Minor tweek to software: 
		+ added averaging for Delta.drive.

October 2001

01 - 05:
   - Unable to get a working system using RetHat 7.1 and RTLinux 3.0.
	  After 3 days of trying I gave up and put every back to old
	  configuration: IC: RH6.2/RTL3.0, UI: slackware 8.0.
   - Prep hilo computers/system to be used when AO hardware arrives.
 
08 - 12:
   - Added shutter code in RT.lp, IC and XUI. On the actually control
	  of device is missing.
   - Removed shadow mode from the RT.lp software.
	- XUI fixes: Motor icons

15 - 19:
	- More XUI fixes: work on Motor icons status updates & control widgets.
		 + added mfocus code (simulation mode only).

22 - 26:
   - Setup AO ui and ic computer in hilo ao lab.
	- summitted PO for real AO computer.


Nov 2001

oct 30 - nov 2
   - Looked a sixnetio devices - I have no objection to these products.

   * Low activity on AO as I was concentrating on smokey development.
	  As the AO system was being transfer to Hilo (from Manoa) - which
	  results in rebuilding the optic bench and its components. The
	  control hardware for the mechanical device (utility box) 
	  are being designed and purchased by the manoa group - ETA is January.

Dec 2001

03 - 07
 - AOUI computer (purchased from ABS has arrived).
   I installed slackware 8.0 on the machine. Hostname is aouih.
	Replace borrowed rackmount with new aouih - everything copied to 
	new computer.
	
 * 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).

 - Tony is on vacation Dec 18 to Dec 31.

POA (Plan of Action)

   - If shutter is SET to on, the goto ldle loop?
   - After primary loop work, finish developing TT imat and correction.
   - Misc improvements:
      + Default inverse & tt.inverse at startup.

LongTerm: