ims Record R2-1-0 Release Notice
================================================================================

The ims dedicated record software in this release is a re-write/simplification
based on motor record R6-4-4, without the general abstraction layers of device
support and driver support.  It conforms to the original motor record interface,
and is compatible with EPICS base R3-14-12.

NOTE: Re R2.2.0
Scott found there are problems w/ the R2.2.0 release so it's been deprecated
and removed from the release area.

The problematic commit is:
6d38eae Force to "Re-initialize" when controller comes online after iocInit
or goes offline and comes back more than a few minutes later when the automatic
"ping" kicks in.

As of 1/30/2023, this commit has been removed but the rest of the current history
has been restored.

Modification Log from R2-0-0 to R2-1-0
======================================

New features:
 1) Implemented a triggered relative move mode.  Once activated, the motor will
    move the TWV amount for every trigger signal.  The TWV can be negative.

Bug fixes:
 1) Only check S9 for smart motors with internal encoder.

 2) Do not start jogging when the motor is already outside the relevant limit;
    Stop jogging when the motor reaches the relevant limit.

Modification Log from R1-0-1 to R2-0-0
======================================

New features:
 1) Implemented a trigger mode.  Once activated, the motor will wait for a
    trigger signal on IO-13 (TTL, active high) after receiving a move command.

 2) Implemented a mode to monitor the encoder position even when the encoder
    is disabled.

 3) Support caput callback.

 4) Added S9 setting for smart motors.  This is to deal with brakes.

Bug fixes:
 1) When BDST was negative, it might do an unnecessary move thus using more
    time.  However the final position is correct.

 2) Allow backlash correction even when there is an encoder and the encoder
    is activated.  Even though the backlash correction, in this configuration,
    may not be reliable, it is now up to the user to decide whether to use
    it, instead of being forbidden by the software.

 3) The saved position (:SAVED_P) value was not correct.  This is only for
    display purpose.

 4) When multiple control applications (screens, camonitor's or caput's) were
    involved, the motor would respond to commands sent from anywhere, but the
    applications which did not send the command would not show the new VAL or
    DVAL.

Modification Log from R1-0-0 to R1-0-1
======================================

 1) About 45 seconds after a move, the controller would like to save the
    position, it askes the IOC for permission.  If this request or the reply
    is lost, the controller would not save the position.  Added checking in
    the MCode program so that about 15 seconds after it sends the request, it
    will save the position anyway even if it has not received permission.


Initial Release R1-0-0
======================================

Major improvements are as follows:

 1) Reduced the lag on the end-of-motion detection from ~100 ms to ~10 ms.
    This is achieved by changing the status checking from polling to quasi-
    interrupt driven, with the help of an extensively enhanced MCode program
    running on the controllers.

    A further reduction of ~70 ms on the overhead could be gained by changing
    the RS485 baud rate from the default 9600 to 115200.

 2) A "SCAN" mode is implemented to further reduce the overhead, by ~20 - 30 ms.
    This is achieved by only checking the most important status variables
    instead of the complete state, while moving.  For non time critical moves,
    it is advisable to use the "NORMAL" mode.

 3) Near instaneous (within 100 ms) notification of any position drift or
    status change.  This was impossible with the old status polling -- would
    cost too much CPU time and network traffic.

 4) Ensured that heavy weight work is done in separate threads for each motor
    so that the response of working motors will not be impacted by motors in
    the same IOC which have been disconnected or are having problems.

 5) Greatly reduced network traffic by combining status messages for individual
    varibles to a single message pushed by the MCode from the controller.

 6) Introduced a new field FREV -- full steps per revolution, to clearly
    reflect the actual hardware.  Previously the SREV field was too much
    overloaded.

 7) Introduced an EL field to reflect the encoder line setting.  It is a float
    number instead of an integer as one would expect, to accommodate any
    arbitrary relationship between a motor revolution and the number of encoder
    lines to be detected.

 8) No more confusion/hacking of the motor intrinsic resolution (MRES,
    = UREV / (FREV * MS)) and the encoder resolution (ERES, = UREV / (4 * EL)),
    and seamless switching between using encorder or not.

 9) Easy closed loop motion against an external PV (absolute encoders, LVDTs,
    or any other EPICS PV which does not have to be position related).

10) Implemented an automatic procedure to home to end-of-travel limit switches
    for stages without built-in home switches or encoder reference marks.

11) Implemented an automatic procedure to assist in debugging the encoder
    wiring and finding the encoder lines.

12) Implemented human-readable pull-up menus wherever possible to assist
    in parameter setting.  Previously these were mostly done with numbers or
    combination of numbers which were very hard to remember or decipher.

13) Implemented a procedure to copy parameters from a similar motor to speed
    up deployment.

14) Present log/warning messages on the edm.  Previously the log messages
    from all motors in the same IOC were all mixed in the IOC log file, and
    it was very hard to trace the messages from a particular motor.  And
    the warnings only showed up as bits in the status bit map (MSTA), and
    therefore very hard to decipher.
    
15) Support plug-and-play by enabling reinitialization of a single motor after
    hardware work, without needing to reboot the IOC.

16) Added an auxiliary record to remember the last (saved or loaded) config
    file.